SYSTEM AND METHOD FOR SYNTHETIC BREAST TISSUE IMAGE GENERATION BY HIGH DENSITY ELEMENT SUPPRESSION

Abstract

A method and breast imaging system for processing breast tissue image data includes feeding image data of breast images to an image processor, identifying image portions depicting breast tissue and high density elements and executing different processing methods on input images. A first image processing method involves breast tissue enhancement and high density element suppression, whereas the second image processing method involves enhancing high density elements. Respective three-dimensional sets of image slices may be generated by respective image processing methods, and respective two-dimensional synthesized images are generated and combined to form a two-dimensional composite synthesized image which is presented through a display of the breast imaging system. First and second image processing may be executed on generated three-dimensional image sets or two-dimensional projection images acquired by an image acquisition component at respective angles relative to the patient's breast.

Claims

1. (canceled)

2. A computer-implemented method for breast tissue image data processing, the computer-implemented method comprising: reconstructing a stack of image slices collectively depicting breast tissue based on image data acquired from a breast imaging system; processing the reconstructed stack of image slices in different ways to generate a first processed stack of image slices in which a high density object is suppressed and breast tissue is enhanced, and a second processed stack of image slices in which the high density object is enhanced; generating a first two-dimensional (2D) synthesized image based at least in part upon the first processed stack of image slices in which the high density object is suppressed, and generating a second 2D synthesized image based at least in part upon the second processed stack of image slices in which the high density object is enhanced; generating a 2D composite synthesized image that is free of the high density object and incorporates portions of the first 2D synthesized image and the second 2D synthesized image; and presenting the 2D composite synthesized image through a display of a computing system.

3. The computer-implemented method of claim 2, further comprising acquiring a plurality of projection images of the breast tissue at the breast imaging system, wherein the image data includes the plurality of projection images.

4. The computer-implemented method of claim 2, wherein generating the first 2D synthesized image includes emphasizing breast tissue image portions by use of a plurality of target object recognition modules, each configured to recognize and emphasize a predetermined type of object.

5. The computer-implemented method of claim 2, wherein processing the reconstructed stack of image slices comprises: segmenting image portions identified as depicting the high density object to determine respective pixel values of segmented image portions; and generating a high density mask based on the respective pixel values, wherein the high density mask is utilized to determine which portions of the first 2D synthesized image or the second 2D synthesized image to include in the 2D composite synthesized image.

6. The computer-implemented method of claim 5, wherein the high density mask includes a binary pixel-level mask.

7. The computer-implemented method of claim 5, wherein generating the 2D composite synthesized image includes executing a modulated combination of the first 2D synthesized image and the second 2D synthesized image with the high density mask.

8. A computer-implemented method for processing images of breast tissue, the computer-implemented method comprising: feeding a reconstructed stack of image slices depicting the breast tissue and a high density object as an input into an image processor of an image generation and display system; processing, by the image processor, the reconstructed stack of image slices to suppress the high density object and generate a first processed stack of image slices; processing, by the image processor, the reconstructed stack of image slices to enhance the high density object and generate a second processed stack of image slices; generating, by the image processor, a first two-dimensional (2D) synthesized image based at least partially on the first processed stack of image slices; generating, by the image processor, a second 2D synthesized image based at least partially on the second processed stack of image slices; combining, by the image processor, the first 2D synthesized image and the second 2D synthesized image to generate a 2D composite synthesized image; and displaying the 2D composite synthesized image through a display of the image generation and display system.

9. The computer-implemented method of claim 8, wherein processing the reconstructed stack of image slices to suppress the high density object also includes enhancing breast tissue portions.

10. The computer-implemented method of claim 8, wherein the first 2D synthesized image is based on a subset of the first processed stack of image slices and the second 2D synthesized image is based on a subset of the second processed stack of image slices.

11. The computer-implemented method of claim 8, wherein the high density object depicted in the reconstructed stack of image slices includes at least one of a radiopaque object, a metallic object, a shadow cast by the high density object, or a calcification in the breast tissue.

12. The computer-implemented method of claim 8, wherein processing the reconstructed stack of image slices to suppress the high density object or processing the reconstructed stack of image slices to enhance the high density object comprises: segmenting image portions identified as depicting the high density object to determine respective pixel values of segmented image portions; and generating a high density mask based on the respective pixel values, wherein the high density mask is utilized to determine which portions of the first 2D synthesized image or the second 2D synthesized image to include in the 2D composite synthesized image.

13. The computer-implemented method of claim 8, wherein generating the second 2D synthesized image includes executing a morphological operation to dilate or erode image edges of enhanced image portions depicting at least a portion of the high density object.

14. The computer-implemented method of claim 8, further comprising: acquiring, by an x-ray image acquisition component, a plurality of projection images of the breast tissue with the high density object; and reconstructing at least some of the plurality of projection images to generate the reconstructed stack of image slices.

15. A system comprising: an x-ray image acquisition system having an x-ray source and an x-ray detector, the x-ray image acquisition system configured to generate a plurality of images of breast tissue; a display coupled in communication to the x-ray image acquisition system; an image processor; and memory storing instructions that, when executed by the image processor cause the system to perform a set of operations, the set of operations comprising: feeding a reconstructed stack of image slices depicting the breast tissue and a high density object as an input into the image processor; processing, by the image processor, the reconstructed stack of image slices to suppress the high density object and generate a first processed stack of image slices; processing, by the image processor, the reconstructed stack of image slices to enhance the high density object and generate a second processed stack of image slices; generating, by the image processor, a first two-dimensional (2D) synthesized image based at least partially on the first processed stack of image slices; generating, by the image processor, a second 2D synthesized image based at least partially on the second processed stack of image slices; combining, by the image processor, the first 2D synthesized image and the second 2D synthesized image to generate a 2D composite synthesized image; and displaying the 2D composite synthesized image through the display.

16. The system of claim 15, wherein within the set of operations, processing the reconstructed stack of image slices to suppress the high density object also includes enhancing breast tissue portions.

17. The system of claim 15, wherein within the set of operations, the first 2D synthesized image is based on a subset of the first processed stack of image slices and the second 2D synthesized image is based on a subset of the second processed stack of image slices.

18. The system of claim 15, wherein the high density object includes a radiopaque object, a metallic object, a shadow cast by the high density object, or a calcification in the breast tissue.

19. The system of claim 15, wherein within the set of operations, processing the reconstructed stack of image slices to suppress the high density object or processing the reconstructed stack of image slices to enhance the high density object comprises: segmenting image portions identified as depicting the high density object to determine respective pixel values of segmented image portions; and generating a high density mask based on the respective pixel values, wherein the high density mask is utilized to determine which portions of the first 2D synthesized image or the second 2D synthesized image to include in the 2D composite synthesized image.

20. The system of claim 15, wherein within the set of operations, generating the second 2D synthesized image includes executing a morphological operation to dilate or erode image edges of enhanced image portions depicting at least a portion of the high density object.

21. The system of claim 15, wherein the set of operations further comprises: acquiring, by the x-ray image acquisition system, a plurality of projection images of the breast tissue with the high density object; and reconstructing at least some of the plurality of projection images to generate the reconstructed stack of image slices.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0027] The drawings illustrate the design and utility of embodiments of the disclosed inventions, in which similar elements are referred to by common reference numerals. These drawings are not necessarily drawn to scale. In order to better appreciate how the above-recited and other advantages and objects are obtained, a more particular description of the embodiments will be rendered, which are illustrated in the accompanying drawings. These drawings depict only typical embodiments of the disclosed inventions and are not therefore to be considered limiting of its scope.

[0028] FIG. 1 is a block diagram illustrating the flow of data through an exemplary breast image acquisition and processing system in accordance with embodiments of the disclosed inventions;

[0029] FIG. 2 is a block diagram illustrating the flow of data through a 2D synthesizer that utilizes multiple target object recognition/enhancement modules to identify respective objects in an image stack in accordance with embodiments of the disclosed inventions;

[0030] FIG. 3 illustrates one embodiment of applying target object recognition/enhancement modules on an image stack to recognize respective objects and reduce the objects onto the 2D synthesized image;

[0031] FIG. 4 illustrates a flow of data when applying a single target object recognition/enhancement module on an image stack;

[0032] FIG. 5 illustrates a flow of data when applying multiple target object recognition/enhancement modules on an image stack;

[0033] FIGS. 6A and 6B illustrate a sequential combination technique of combining data from multiple target object synthesis modules;

[0034] FIGS. 7A and 7B illustrate a parallel combination technique of combining data from multiple target object synthesis modules; and

[0035] FIGS. 8A and 8B illustrate two example flow diagrams of generating 2D synthesized images using the sequential combination and parallel combination techniques respectively.

[0036] FIG. 9 depicts a 3D stack of image slices and how a high density element such as shadows generated by imaging metallic objects in breast tissue can obscure breast tissue and clinically important information;

[0037] FIG. 10 is a block diagram illustrating the flow of data through an exemplary breast image acquisition and processing system in accordance with embodiments of the disclosed inventions involving selective suppression and enhancement of high density elements in breast images;

[0038] FIG. 11 illustrates a flow diagram of multi-flow or differential image processing executed to generate a composite 2D synthesized image according to one embodiment;

[0039] FIG. 12 illustrates a flow diagram of breast tissue enhancement and high density suppression of a first image processing flow executed on input images;

[0040] FIG. 13 illustrates a flow diagram of high density element enhancement of a second image processing flow executed on input images;

[0041] FIG. 14 illustrates a flow diagram of multi-flow or differential image processing that is executed directly on an input 3D stack of image slices;

[0042] FIG. 15 illustrates an image-flow diagram illustrating an embodiment of multi-flow or differential image processing of FIG. 14 in which a multi-flow image processor receives inputs of 3D image slices;

[0043] FIG. 16 illustrates an image-flow diagram of breast tissue enhancement and high density suppression of a first image processing flow in which high density elements are suppressed and breast tissue elements are enhanced;

[0044] FIG. 17 illustrates an image-flow diagram of high density element enhancement of a second image processing flow in which high density elements are enhanced;

[0045] FIG. 18 illustrates an image-flow diagram of a 2D image synthesizer utilized in a second image processing flow in which high density elements are enhanced;

[0046] FIG. 19 illustrates an image-flow diagram of combining multiple 2D synthesized images to generate a 2D composite synthesized image;

[0047] FIG. 20A depicts a 2D synthesized image showing how shadows generated by a metallic biopsy marker obscure breast tissue and clinically important information, and FIG. 20B depicts a 2D synthesized composite image generated according to embodiments and showing how shadows of FIG. 20A are suppressed or eliminated to provide for a clearer and unobstructed view of breast tissue and clinically important information;

[0048] FIG. 21 illustrates a flow diagram of multi-flow or differential image processing that is executed directly on an input 2D images;

[0049] FIG. 22 illustrates an image-flow diagram illustrating an embodiment of multi-flow or differential image processing of FIG. 21 in which a multi-flow image processor receives inputs of 2D images;

[0050] FIG. 23 illustrates an image-flow diagram of breast tissue enhancement and high density suppression of a first image processing flow executed by a multi-flow image processor that executes on 2D image inputs and during which high density elements are suppressed and breast tissue elements are enhanced; and

[0051] FIG. 24 illustrates an image-flow diagram of high density element enhancement of a second image processing flow of a multi-stage image processor that executes on 2D image inputs and during which high density elements are enhanced.

DETAILED DESCRIPTION OF THE ILLUSTRATED EMBODIMENTS

[0052] All numeric values are herein assumed to be modified by the terms about or approximately, whether or not explicitly indicated, wherein the terms about and approximately generally refer to a range of numbers that one of skill in the art would consider equivalent to the recited value (i.e., having the same function or result). In some instances, the terms about and approximately may include numbers that are rounded to the nearest significant figure. The recitation of numerical ranges by endpoints includes all numbers within that range (e.g., 1 to 5 includes 1, 1.5, 2, 2.75, 3, 3.80, 4, and 5).

[0053] As used in this specification and the appended claims, the singular forms a, an, and the include plural referents unless the content clearly dictates otherwise. As used in this specification and the appended claims, the term or is generally employed in its sense including and/or unless the content clearly dictates otherwise. In describing the depicted embodiments of the disclosed inventions illustrated in the accompanying figures, specific terminology is employed for the sake of clarity and ease of description. However, the disclosure of this patent specification is not intended to be limited to the specific terminology so selected, and it is to be understood that each specific element includes all technical equivalents that operate in a similar manner. It is to be further understood that the various elements and/or features of different illustrative embodiments may be combined with each other and/or substituted for each other wherever possible within the scope of this disclosure and the appended claims.

[0054] Various embodiments of the disclosed inventions are described hereinafter with reference to the figures. It should be noted that the figures are not drawn to scale and that elements of similar structures or functions are represented by like reference numerals throughout the figures. It should also be noted that the figures are only intended to facilitate the description of the embodiments. They are not intended as an exhaustive description of the invention or as a limitation on the scope of the disclosed inventions, which is defined only by the appended claims and their equivalents. In addition, an illustrated embodiment of the disclosed inventions needs not have all the aspects or advantages shown. For example, an aspect or an advantage described in conjunction with a particular embodiment of the disclosed inventions is not necessarily limited to that embodiment and can be practiced in any other embodiments even if not so illustrated.

[0055] For the following defined terms and abbreviations, these definitions shall be applied throughout this patent specification and the accompanying claims, unless a different definition is given in the claims or elsewhere in this specification:

[0056] An acquired image refers to an image generated while visualizing a patient's tissue. Acquired images can be generated by radiation from a radiation source impacting on a radiation detector disposed on opposite sides of a patient's tissue, as in a conventional mammogram.

[0057] A reconstructed image refers to an image generated from data derived from a plurality of acquired images. A reconstructed image simulates an acquired image not included in the plurality of acquired images.

[0058] A synthesized image refers to an artificial image generated from data derived from a plurality of acquired and/or reconstructed images. A synthesized image includes elements (e.g., objects and regions) from the acquired and/or reconstructed images, but does not necessarily correspond to an image that can be acquired during visualization. Synthesized images are constructed analysis tools.

[0059] An Mp image is a conventional mammogram or contrast enhanced mammogram, which are two-dimensional (2D) projection images of a breast, and encompasses both a digital image as acquired by a flat panel detector or another imaging device, and the image after conventional processing to prepare it for display (e.g., to a health professional), storage (e.g., in the PACS system of a hospital), and/or other use.

[0060] A Tp image is an image that is similarly two-dimensional (2D), but is acquired at a respective tomosynthesis angle between the breast and the origin of the imaging x rays (typically the focal spot of an x-ray tube), and encompasses the image as acquired, as well as the image data after being processed for display, storage, and/or other use.

[0061] A Tr image is a type (or subset) of a reconstructed image that is reconstructed from tomosynthesis projection images Tp, for example, in the manner described in one or more of U.S. Pat. Nos. 7,577,282, 7,606,801, 7,760,924, and 8,571,289, the disclosures of which are fully incorporated by reference herein in their entirety, wherein a Tr image represents a slice of the breast as it would appear in a projection x ray image of that slice at any desired angle, not only at an angle used for acquiring Tp or Mp images.

[0062] An Ms image is a type (or subset) of a synthesized image, in particular, a synthesized 2D projection image that simulates mammography images, such as a craniocaudal (CC) or mediolateral oblique (MLO) images, and is constructed using tomosynthesis projection images Tp, tomosynthesis reconstructed images Tr, or a combination thereof. Ms images may be provided for display to a health professional or for storage in the PACS system of a hospital or another institution. Examples of methods that may be used to generate Ms images are described in the above-incorporated U.S. Pat. Nos. 7,760,924 and 8,571,289 and also U.S. application Ser. No. 15/120,911, published as U.S. Publication No. 2016/0367120 on Dec. 22, 2016 and entitled System and Method for Generating and Displaying Tomosynthesis Image Slabs, PCT Application No. PCT/US2018/024911, filed Mar. 28, 2018 and entitled System and Method for Hierarchical Multi-Level Feature Image Synthesis and Representation, PCT Application No. PCT/US2018/024912, filed Mar. 28, 2018, and entitled System and Method for Synthesizing Low-Dimensional Image Data From High-Dimensional Image Data Using an Object Grid Enhancement, and PCT Application No. PCT/US018/0249132, filed Mar. 28, 2018, and entitled System and Method for Targeted Object Enhancement to Generate Synthetic Breast Tissue Images, the contents of all of which are incorporated herein by reference as thought set forth in full.

[0063] It should be appreciated that Tp, Tr, Ms and Mp image data encompasses information, in whatever form, that is sufficient to describe the respective image for display, further processing, or storage. The respective Mp, Ms. Tp and Tr images, including those subjected to high density element suppression and enhancement, are typically provided in digital form prior to being displayed, with each image being defined by information that identifies the properties of each pixel in a two-dimensional array of pixels. The pixel values typically relate to respective measured, estimated, or computed responses to X-rays of corresponding volumes in the breast, i.e., voxels or columns of tissue. In a preferred embodiment, the geometry of the tomosynthesis images (Tr and Tp) and mammography images (Ms and Mp) are matched to a common coordinate system, as described in U.S. Pat. No. 7,702,142. Unless otherwise specified, such coordinate system matching is assumed to be implemented with respect to the embodiments described in the ensuing detailed description of this patent specification.

[0064] The terms generating an image and transmitting an image respectively refer to generating and transmitting information that is sufficient to describe the image for display. The generated and transmitted information is typically digital information.

[0065] The term high density element is defined as an element, when imaged with breast tissue, partially or completely obscures imaged breast tissue or clinically important information of breast tissue such as malignant breast mass, tumors, etc. A high density element may be detected based on pre-determined criteria or filters involving one or more of contrast, brightness, radiopacity or other attribute. A high density element may be a foreign object or naturally occurring within breast tissue and may be partially or completely radiopaque. For example, one type of high density element is a metallic object such as a metallic biopsy marker inserted into breast tissue. Such markers are designed to be radiopaque such that they are clearly visible when using x-rays. Another example of a high density element is a calcification within the breast tissue. A high density element may also be a non-metallic or non-calcified element such as a shadow artifact generated by imaging a metallic marker, and which may not be considered to be radiopaque. Accordingly, a high density element is defined to include metallic objects such as a biopsy marker or a skin marker, radiopaque materials or objects, and shadows or shadow artifacts generated by imaging of same.

[0066] The terms differential or multi-flow image processing are defined to refer to the input images being processed in different ways to generate different image results and is defined to include one flow involving suppression of an imaged high density element and involving enhancement of an imaged high density element. Different image processing flows can be executed in parallel and simultaneously, and images input to image processors of embodiments may be of different dimensional formats.

[0067] In order to ensure that a synthesized 2D image displayed to a reviewer or end-user (e.g., an Ms image) includes the most clinically relevant information, it is necessary to detect and identify 3D objects, such as malignant breast mass, tumors, etc., within the breast tissue. Towards this end, in accordance with embodiments of the presently disclosed inventions, 3D objects may be identified using multiple target object recognition/synthesis modules, wherein each target recognition/synthesis module may be configured to identify and reconstruct a particular type of object. These multiple target synthesis modules may work together in combining information pertaining to respective objects during the reconstruction process of generating one or more synthesized 2D images, ensuring that each object is represented accurately, and preserving clinically significant information on the 2D synthesized images that are the displayed to the end-user.

[0068] The synthesized 2D image that is displayed to an end-user should also be clear such that clinically relevant information and objects are not obscured by undesirable image elements or artifacts, which may include a high density element such as a biopsy marker and/or a shadow generated by imaging of same during breast imaging. Towards this end, in accordance with embodiments of the presently disclosed inventions, a multi-flow image processor is utilized to generate a 2D synthesized image by suppressing high density elements in one image processing method and enhancing high density elements in another image processing method such that when different 2D synthesized images generated by different image processing flows are combined, high density elements such as shadows are reduced or eliminated resulting in a composite 2D synthesized image that is clearer and more accurately depicts breast tissue and breast tissue objects while providing for more accurate and efficient radiologist review.

[0069] Embodiments designed to generate a 2D synthesized image that maintains and enhances clinically interesting characteristics are described with reference to FIGS. 1-8B, and embodiments that utilize a multi-flow image processing method for reducing of high density elements such as shadows and generate a clearer 2D composite synthesized image are described with reference to FIGS. 9-24.

[0070] FIG. 1 illustrates the flow of data in an exemplary image generation and display system 100, which incorporates each of synthesized image generation, object identification, and display technology. It should be understood that, while FIG. 1 illustrates a particular embodiment of a flow diagram with certain processes taking place in a particular serial order or in parallel, the claims and various other embodiments described herein are not limited to the performance of the image processing steps in any particular order, unless so specified.

[0071] More particularly, the image generation and display system 100 includes an image acquisition system 101 that acquires tomosynthesis image data for generating Tp images of a patient's breasts, optionally using the respective 3D and/or tomosynthesis acquisition methods of any of the currently available systems. If the acquisition system is a combined tomosynthesis/mammography system, Mp images may also be generated. Some dedicated tomosynthesis systems or combined tomosynthesis/mammography systems may be adapted to accept and store legacy mammogram images, (indicated by a dashed line and legend Mp.sub.legacy in FIG. 1) in a storage device 102, which is preferably a DICOM-compliant Picture Archiving and Communication System (PACS) storage device. Following acquisition, the tomosynthesis projection images Tp may also be transmitted to the storage device 102 (as shown in FIG. 1). The storage device 102 may further store a library of known 3D objects that may be used to identify significant 3D image patterns to the end-user. In other embodiments, a separate dedicated storage device (not shown) may be used to store the library of known 3D objects with which to identify 3D image patterns or objects.

[0072] The Tp images are transmitted from either the acquisition system 101, or from the storage device 102, or both, to a computer system configured as a reconstruction engine 103 that reconstructs the Tp images into reconstructed image slices Tr, representing breast slices of selected thickness and at selected orientations, as disclosed in the above-incorporated patents and applications.

[0073] Mode filters 107 are disposed between image acquisition and image display. The filters 107 may additionally include customized filters for each type of image (i.e., Tp, Mp, and Tr images) arranged to identify and highlight or enhance certain aspects of the respective image types. In this manner, each imaging mode can be tuned or configured in an optimal way for a specific purpose. For example, filters programmed for recognizing objects across various 2D image slices may be applied in order to detect image patterns that may belong to a particular high-dimensional objects. The tuning or configuration may be automatic, based on the type of the image, or may be defined by manual input, for example through a user interface coupled to a display. In the illustrated embodiment of FIG. 1, the mode filters 107 are selected to highlight particular characteristics of the images that are best displayed in respective imaging modes, for example, geared towards identifying objects, highlighting masses or calcifications, identifying certain image patterns that may be constructed into a 3D object, or for creating 2D synthesized images (described below). Although FIG. 1 illustrates only one mode filter 107, it should be appreciated that any number of mode filters may be utilized in order to identify structures of interest in the breast tissue.

[0074] The imaging and display system 100 further includes a 2D image synthesizer 104 that operates substantially in parallel with the reconstruction engine 103 for generating 2D synthesized images using a combination of one or more input Tp (tomosynthesis projection), Mp (mammography projection), and/or Tr (tomosynthesis reconstruction) images. The 2D image synthesizer 104 consumes a set of input images, determines a set of most relevant features from each of the input images, and outputs one or more synthesized 2D images. The synthesized 2D image represents a consolidated synthesized image that condenses significant portions of various slices onto one image. This provides an end-user (e.g., medical personnel, radiologist, etc.) with the most clinically-relevant image data in an efficient manner, and reduces time spent on other images that may not have significant data.

[0075] One type of relevant image data to highlight in the synthesized 2D images would be relevant objects found across one or more Mp, Tr and/or Tp images. Rather than simply assessing image patterns of interest in each of the 2D image slices, it may be helpful to determine whether any of the 2D image patterns of interest belong to a larger high-dimensional structure, and if so, to combine the identified 2D image patterns into a higher-dimensional structure. This approach has several advantages, but in particular, by identifying high-dimensional structures across various slices/depths of the breast tissue, the end-user may be better informed as to the presence of a potentially significant structure that may not be easily visible in various 2D slices of the breast.

[0076] Further, instead of identifying similar image patterns in two 2D slices (that are perhaps adjacent to each other), and determining whether or not to highlight image data from one or both of the 2D slices, identifying both image patterns as belonging to the same high-dimensional structure may allow the system to make a more accurate assessment pertaining to the nature of the structure, and consequently provide significantly more valuable information to the end-user. Also, by identifying the high-dimensional structure, the structure can be more accurately depicted on the synthesized 2D image. Yet another advantage of identifying high-dimensional structures within the various captured 2D slices of the breast tissue relates to identifying a possible size/scope of the identified higher-dimensional structure. For example, once a structure has been identified, previously unremarkable image patterns that are somewhat proximate to the high-dimensional structure may now be identified as belonging to the same structure. This may provide the end-user with an indication that the high-dimensional structure is increasing in size/scope.

[0077] To this end, the 2D image synthesizer 104 employs a plurality of target object recognition/enhancement modules (also referred to as target object synthesis modules) that are configured to identify and reconstruct different types of objects. Each target image recognition/synthesis module may be applied (or run) on a stack (e.g., a tomosynthesis image stack) of 2D image slices of a patient's breast tissue, and work to identify particular types of objects that may be in the breast tissue, and ensure that such object(s) are represented in a clinically-significant manner in the resulting 2D synthesized image presented to the end-user. For example, a first target image synthesis module may be configured to identify calcifications in the breast tissue. Another target image synthesis module may be configured to identify and reconstruct spiculated lesions in the breast tissue. Yet another target image synthesis module may be configured to identify and reconstruct spherical masses in the breast tissue. In one or more embodiments, the multiple target image synthesis modules process the image slice data and populate respective objects in a high-dimensional grid (e.g., 3D grid) comprising respective high-dimensional structures (e.g., 3D objects) present in the breast tissue. This high-dimensional grid may then be utilized to accurately depict the various structures in the 2D synthesized image.

[0078] A high-dimensional object may refer to any object that comprises at least three or more dimensions, e.g., 3D or higher object, or a 3D or higher object and time dimension, etc. Examples of such objects or structures include, without limitation, calcifications, spiculated lesions, benign tumors, irregular masses, dense objects, etc. An image object may be defined as a certain type of image pattern that exists in the image data. The object may be a simple round object in a 3D space, and a corresponding flat round object in a 2D space. It can be an object with complex patterns and complex shapes, and it can be of any size or dimension. The concept of an object may extend past a locally bound geometrical object. Rather, the image object may refer to an abstract pattern or structure that can exist in any dimensional shape. It should be appreciated that the inventions disclosed herein are not limited to 3D objects and/or structures, and may include higher-dimensional structures. It should be appreciated that each of the target image synthesis modules is configured for identifying and reconstructing respective types of objects. These objects may refer to 2D shapes, 2D image patterns, 3D objects, or any other high-dimensional object, but in any event will all be referred to as objects or 3D objects herein for simplicity, but this illustrative use should not be otherwise read as limiting the scope of the claims.

[0079] In the illustrated embodiment, the 2D synthesizer 104 comprises a plurality of target object recognition/enhancement modules (e.g., 110a, 110b . . . 110n), each configured for recognizing and enhancing a particular type of object. Each of the target object recognition/enhancement modules 110 may be run on a 2D image stack (e.g., Tr image stack), and is configured to identify the respective object (if any is/are present) therein. By identifying the assigned object in the 2D image stack, each target object recognition/enhancement module 110 works to ensure that the respective object is preserved and depicted accurately in the resulting 2D synthesized image presented to the end-user.

[0080] In some embodiments, a hierarchical model may be utilized in determining which objects to emphasize or de-emphasize in the 2D synthesized image based on a weight or priority assigned to the target object recognition/enhancement module. In other embodiments, all objects may be treated equally, and different objects may be fused together if there is an overlap in the z direction, as will be discussed in further detail below. These reconstruction techniques allow for creation of 2D synthesized images that comprise clinically-significant information, while eliminating or reducing unnecessary or visually confusing information.

[0081] The synthesized 2D images may be viewed at a display system 105. The reconstruction engine 103 and 2D image synthesizer 104 are preferably connected to a display system 105 via a fast transmission link. The display system 105 may be part of a standard acquisition workstation (e.g., of acquisition system 101), or of a standard (multi-display) review station (not shown) that is physically remote from the acquisition system 101. In some embodiments, a display connected via a communication network may be used, for example, a display of a personal computer or of a so-called tablet, smart phone or other hand-held device. In any event, the display 105 of the system is preferably able to display respective Ms, Mp, Tr, and/or Tp images concurrently, e.g., in separate side-by-side monitors of a review workstation, although the invention may still be implemented with a single display monitor, by toggling between images.

[0082] Thus, the imaging and display system 100, which is described as for purposes of illustration and not limitation, is capable of receiving and selectively displaying tomosynthesis projection images Tp, tomosynthesis reconstruction images Tr, synthesized mammogram images Ms, and/or mammogram (including contrast mammogram) images Mp, or any one or sub combination of these image types. The system 100 employs software to convert (i.e., reconstruct) tomosynthesis images Tp into images Tr, software for synthesizing mammogram images Ms, software for decomposing 3D objects, software for creating feature maps and object maps. An object of interest or feature in a source image may be considered a most relevant feature for inclusion in a 2D synthesized image based upon the application of the object maps along with one or more algorithms and/or heuristics, wherein the algorithms assign numerical values, weights or thresholds, to pixels or regions of the respective source images based upon identified/detected objects and features of interest within the respective region or between features. The objects and features of interest may include, for example, spiculated lesions, calcifications, and the like.

[0083] FIG. 2 illustrates the 2D image synthesizer 104 in further detail. As discussed above, various image slices 218 of a 3D tomosynthesis data set or stack 202 (e.g., filtered and/or unfiltered Tr and/or Tp images of a patient's breast tissue) are input into the 2D image synthesizer 104, and then processed to determine portions of the images to highlight in a synthesized 2D image that will be displayed on the display 105. The image slices 218 may be consecutively-captured cross-sections of a patient's breast tissue. Or, the image slices 218 may be cross-sectional images of the patient's breast tissue captured at known intervals. The 3D tomosynthesis stack 202 comprising the image slices 218 may be forwarded to the 2D image synthesizer 104, which evaluates each of the source images in order to (1) identify various types of objects (Tr) for possible inclusion in one or more 2D synthesized images, and/or (2) identify respective pixel regions in the images that contain the identified objects.

[0084] As shown in the illustrated embodiment, the 3D tomosynthesis stack 202 comprises a plurality of images 218 taken at various depths/cross-sections of the patient's breast tissue. Some of the images 218 in the 3D tomosynthesis stack 202 comprise 2D image patterns. Thus, the tomosynthesis stack 202 comprises a large number of input images containing various image patterns within the images of the stack.

[0085] More particularly, as shown in FIG. 2, three target object recognition/enhancement modules 210a, 210b and 210c are configured to run on the 3D tomosynthesis stack 202, wherein each of the target object recognition and enhancement modules 210 corresponds to a respective set of programs/rules and parameters that define a particular object, and how to identify that particular object amongst other objects that may exist in the breast tissue depicted by the 3D tomosynthesis stack 202. For example, filtering/image recognition techniques and various algorithms/heuristics may be run on the 3D tomosynthesis stack 202 in order to identify the object assigned to the particular target object recognition/enhancement module 210. It will be appreciated that there are many ways to recognize objects using a combination of image manipulation/filtration techniques.

[0086] For the purposes of illustration, it will be assumed that the each of the target object recognition/enhancement modules 210 identifies at least one respective object, but it should be appreciated that in many cases no objects will be identified. However, even healthy breast tissue may have one or more suspicious objects or structures, and the target object recognition/enhancement modules may inadvertently identify a breast background object. For example, all breast linear tissue and density tissue structures can be displayed as the breast background object. In other embodiments, healthy objects such as spherical shapes, oval shapes, etc., may simply be identified by one or more of the target object recognition/enhancement modules 210. The identified 3D objects may then be displayed on the 2D synthesized image 206; of course, out of all identified 2D objects, more clinically-significant objects may be prioritized/enhanced when displaying the respective objects on the 2D synthesized image, as will be discussed in further detail below.

[0087] In the illustrated embodiment, a first target object recognition/enhancement module 210a is configured to recognize circular and/or spherical shapes in the images 218 of the 3D tomosynthesis stack 202 (e.g., Tr, Tp, etc.). A second target object synthesis module 210b is configured to recognize lobulated shapes. A third target object synthesis module 210c is configured to recognize calcification patterns. In particular, each of the target object synthesis modules 210a, 210b and 210c is run on the Tr image stack 202, wherein a set of features/objects are recognized by the respective target object synthesis modules.

[0088] For example, target object recognition/enhancement module 210a may recognize one or more circular shapes and store these as recognized objects 220a. It will be appreciated that multiple image slices 218 of the 3D tomosynthesis stack 202 may contain circular shapes, and that these shapes may be associated with the same spherical object, or may belong to different spherical objects. In the illustrated embodiment, at least two distinct circular objects are recognized by the target object recognition/enhancement module 210a.

[0089] Similarly, target object recognition/enhancement module 210b may recognize one or more lobulated shapes and store these as recognized objects 220b. In the illustrated embodiment, one lobulated object has been recognized in the 3D tomosynthesis stack 202 by the target object recognition/enhancement module 210b. As can be seen, two different image slices 218 in the 3D tomosynthesis stack 202 depict portions of the lobulated object, but the respective portions are recognized as belonging to a single lobulated object by the recognition/enhancement module 210b, and stored as a single recognized object 220b.

[0090] Finally, target object recognition/enhancement module 210c may recognize one or more calcification shapes and store these as recognized objects 220c. In the illustrated embodiment, a (single) calcification cluster has been recognized by the target object recognition/enhancement module 210c and stored as a recognized object 220c. The recognized objects 220a, 220b and 220c may be stored at storage facilities corresponding to the respective target object recognition/enhancement modules 210a, 210b and 210c, or alternatively at a separate (i.e., single) storage facility that may be accessed by each of the target object recognition/enhancement modules.

[0091] Referring now to FIG. 3, each of the target object recognition/enhancement modules 210 may be configured to identify and synthesize (e.g., to reduce to 2D) a respective 3D object to be displayed on the one or more 2D synthesized images. In other words, once the 3D objects are recognized by the respective target object recognition/enhancement module 210a, 210b or 210c, the target object recognition/enhancement module thereafter converts the recognized 3D object into a 2D format so that the recognized object may be displayed on the 2D synthesized image. In the illustrated embodiment, the target object recognition/enhancement modules 210a, 21b and 210c recognize respective objects, and convert the recognized objects into respective 2D formats. As part of the conversion process, certain of the recognized objects may be enhanced to a greater or lesser degree for the displayed image, as will be discussed in further detail below. Assuming all three target object recognition/enhancement modules 210a, 210b and 210c are considered equally important to the 2D image synthesizer 104, the respective 2D formats of all recognized objects (e.g., two spherical objects, one lobular object, and one calcification mass) depicted on the 2D synthesized image 302.

[0092] FIG. 4 illustrates how a single target object recognition/enhancement module 210 may be run on a 3D tomosynthesis stack to generate a portion of the 2D synthesized image. In the illustrated embodiment, image slices 402 of the 3D tomosynthesis stack are fed through a single target object recognition/enhancement module 404, which is configured to recognize star shaped objects in the stack of images 402. As a result, the single target object synthesis module reduces information pertaining to the recognized star shape gained from various depths of the image slices onto a single 2D synthesized image 406.

[0093] FIG. 5 illustrates an exemplary embodiment for having multiple target object recognition/enhancement modules work together to produce the 2D synthesized image. In the illustrated embodiment, image slices 502 (of a respective 3D tomosynthesis stack) are fed through a first target object recognition/enhancement module 504a configured to recognize and reconstruct circular and/or spherical shapes, a second target object recognition/enhancement module 504b configured to recognize and reconstruct star-like shapes, and a third target object recognition/enhancement module 504c configured to recognize and reconstruct calcification structures. It should be appreciated that any number of target object recognition/enhancement modules may be programmed for any number of object types.

[0094] Each of the target object recognition/enhancement modules 504a, 504b and 504c corresponds to respective algorithms that are configured with various predetermined rules and attributes that enable these programs to successfully recognize respective objects, and reduce the recognized objects to a 2D format. By applying all three target object recognition/synthesis modules 504a, 504b and 504c to the image slices 502, a 2D synthesized image 506 is generated. In particular, rather than simply displaying a single type of object, the 2D synthesized image 506 comprises all three object types that are recognized and synthesized by the three target object recognition/enhancement modules 504a, 504b and 504c, with each of the recognized objects being equally emphasized. While this may be desirable if all the object types are of equal significance, it may be helpful to enhance/emphasize different object types to varying degrees based on their weight/priority. This technique may be more effective in alerting the end-user to a potentially important object, while de-emphasizing objects of lesser importance.

[0095] Referring now to FIG. 6A, a hierarchical sequential approach to combine data from the multiple target object recognition/enhancement modules is illustrated. In particular, a sequential combination technique may be applied if the various object types have a clearly defined hierarchy associated with them. For example, one type of object (e.g., spiculated lesions) may be deemed to be more clinically significant than another type of object (e.g., a spherical mass in breast tissue). This type of object (and the corresponding target object module) may be assigned a particular high weight/priority. In such a case, if two objects are competing for space on the 2D synthesized image, the object type associated with the higher priority may be emphasized/displayed on the 2D synthesized image, and the other object type may be de-emphasized, or not displayed at all. Similarly, in such an approach, each of the target object recognition/enhancement modules may be assigned respective weights based on respective significance.

[0096] In the illustrated embodiment, the image slices 602 of the 3D tomosynthesis stack are sequentially fed through three different target object recognition/enhancement modules (604, 606 and 608) to generate the 2D synthesized image 610, wherein each of the target object synthesis modules is configured to recognize and reconstruct a particular type of object. The first target object recognition/enhancement module 604 (associated with a square-shaped object) is run first on the reconstruction image slices 602, followed by the second target object recognition/enhancement module 606 (associated with a diamond-shaped object), and then followed by the third target object recognition/enhancement module 608 (associated with a circular-shaped object). It should be appreciated that since the target object recognition/enhancement modules are applied (or run) sequentially, the second target object recognition/enhancement module 606 may be considered a higher priority object as compared with the first target object recognition/enhancement module 604, and the third target object recognition/enhancement module 608 may be considered as having a higher priority as compared to the second target object recognition/enhancement module 606. Thus, the third object type may override (or be emphasized over) the second object type, and the second object type may override (or be emphasized over) the first object type.

[0097] FIG. 6B illustrates this hierarchical approach to combining various object types sequentially. In particular, the 3D tomosynthesis image stack 652 includes objects 656, 658 and 660 that can be recognized in various image slices. As illustrated, objects 658 and 660 somewhat overlap in the z direction, which means that they are likely to compete for representation in the 2D synthesized image 654. When using the sequential approach of FIG. 6A to combine data from the multiple target object recognition/enhancement modules 604, 606 and 608, the programmed hierarchy is preserved. Thus, since target object recognition/enhancement module 608 configured to recognize and reconstruct circular-shaped objects has higher priority as compared to target object recognition/enhancement module 604 configured to recognize and reconstruct square-shaped objects, in a case of overlap between the two objects (as is the case in FIG. 6B), circular-shaped object 658 overrides square-shaped object 660 in the 2D synthesized image 654. Of course, it should be appreciated that since diamond-shaped object 656 does not overlap in the z direction with the other two objects, diamond shaped object 656 is also displayed in the 2D synthesized image 654. In other embodiments, instead of completing overriding the lower-priority object, the object with high-priority may be emphasized relative to the lower-priority object (rather than be omitted from display).

[0098] Another approach to running multiple target object synthesis modules on a set of image slices is illustrated in FIG. 7A. As can be seen, rather than running the multiple target object recognition/enhancement modules sequentially with the last-run target object synthesis module having the highest priority, all the target object recognition/enhancement modules may be applied in parallel. In particular, one or more enhancement or fusion modules 712 may be utilized to ensure that the various objects are combined appropriately on the 2D synthesized image. This approach may not follow a hierarchical approach, and all of the objects may be given equal weight.

[0099] The image slices 702 are fed through three different target object recognition/enhancement modules, 704, 706 and 708, in parallel. The first target object recognition/enhancement module 604 (associated with square-shaped object), the second target object recognition/enhancement module 606 (associated with diamond-shaped object), and the third target object recognition/enhancement module 608 (associated with circular-shaped object) are all run in parallel on the image slices 702. In some embodiments, an enhancement and fusion module 712 may be utilized to ensure that the different objects are fused together appropriately in case of overlap between multiple objects. The target object recognition/enhancement modules 704, 706 and 708, run in parallel may generate the 2D synthesized image 710.

[0100] This approach to combining various object types in parallel is illustrated in FIG. 7B. In particular, the tomosynthesis stack 752 depict the same objects as FIG. 6B (e.g., objects 756, 758 and 760) at various image slices. As illustrated, objects 758 and 760 somewhat overlap in the z direction, which means that they are likely to compete for representation and/or overlap in the 2D synthesized image 754. Here, because the multiple target object recognition/enhancement modules are run in parallel, rather than one object type overriding another object type, as was the case in FIG. 6B, both the square-object 760 and the circular object 758 are fused together in the 2D synthesized image 754. Thus, this approach does not assume an innate priority/hierarchy between objects and all objects may be fused together appropriately in the 2D synthesized image 754.

[0101] FIG. 8A depicts a flow diagram 800 that illustrates exemplary steps that may be performed in an image merge process carried out in accordance with the sequential combination approach outlined above in conjunction with FIGS. 6A and 6B. At step 802, an image data set is acquired. The image data set may be acquired by a tomosynthesis acquisition system, a combination tomosynthesis/mammography system, or by retrieving pre-existing image data from a storage device, whether locally or remotely located relative to an image display device, e.g., through a communication network. At steps 804 and 806, for a range of 2D images (e.g., Tr stack), a first target object recognition/enhancement module is run in order to recognize a first object associated with the first target object recognition/enhancement module. Any recognized objects may be stored in a storage module associated with the first target object recognition/enhancement module. At step 808, a second target object recognition/enhancement module is run in order to recognize a second object associated with the second target object recognition/enhancement module. At step 810, it may be determined whether the first recognize object and the second recognized object overlap each other in the z direction. If it is determined that the two objects overlap, only the second object may be displayed (or otherwise emphasized over the first object) on the 2D synthesized image at step 812. If, on the other hand, it is determined that the two objects do not overlap, both objects are displayed on the 2D synthesized image at step 814.

[0102] FIG. 8B depicts a flow diagram 850 that illustrates exemplary steps that may be performed in an image synthesis process carried out in accordance with the parallel combination approach outlined above in conjunction with FIGS. 7A and 7B. At step 852, an image data set is acquired. The image data set may be acquired by a tomosynthesis acquisition system, a combination tomosynthesis/mammography system, or by retrieving pre-existing image data from a storage device, whether locally or remotely located relative to an image display device. At steps 854 and 856, for a range of 2D images (e.g., Tr stack), all the programmed target object recognition/enhancement modules are run to recognize respective objects in the Tr image stack. At step 858, one or more enhancement modules may also be run to determine whether a fusion process needs to occur. At step 860, it may be determined whether any recognized objects overlap in the z direction. If it is determined that any two (or more) objects overlap, the overlapping objects may be fused together, at step 862. If, on the other hand, it is determined that no objects overlap, all the objects are displayed as is on the 2D synthesized image at step 814.

[0103] Having described how a 3D stack of image slices is generated and processed by a 2D synthesizer comprising target object recognition/enhancement modules in order to ensure that a synthesized 2D image displayed to a reviewer or end-user includes the most clinically relevant information, embodiments related to generating clearer, reduced shadow or shadow-free 2D synthesized images are described with reference to FIGS. 9-24. Embodiments described with reference to FIGS. 9-24 eliminate or reduce high density elements such as image portions depicting metal objects and/or shadows generated by imaging of same within 2D acquired or projection images and/or sets or stacks of 3D slices reconstructed based on 2D projection images. With embodiments, high density elements such as shadows are eliminated or reduced resulting in clearer 2D synthesized image that more accurately depicts breast tissue being analyzed and allows for more accurate and efficient radiologist examination since clinically relevant information is not blocked or obscured by shadows within the 2D synthesized image.

[0104] Referring to FIG. 9, and referring again to FIGS. 1-2, reconstructed images Tr form a 3D tomosynthesis stack 902 of image slices 918. As a non-limiting example, a 3D tomosynthesis stack 902 may include about 30 to about 120 image slices Tr (e.g., 60 image slides Tr) derived from or constructed based on about 15 or more 2D projection images Tp acquired by an x-ray image acquisition component 101 such as an x-ray source and detector collectively rotate around the patient or breast being analyzed. FIG. 9 depicts a 3D tomosynthesis stack 902 including image slices 918, e.g., similar to the stack 202 illustrated in FIG. 2. FIG. 9 further illustrates a high density element 920 in the breast tissue 910 and extending across multiple image slices 918. FIG. 9 illustrates the high density element 920 extending across two slices 918, but it will be understood that the high density element may extend to various depths.

[0105] An example of a high density element 920 is a metallic biopsy marker or clip, which may be made of stainless steel or titanium or other radiopaque or dense material. Another example of a high density element 920 is an external skin marker. A high density element 920 may also be a biological or tissue component within the breast tissue 910 such as a calcification or other dense biological or tissue structure that obscures other clinically relevant information or objects of interest in the breast tissue 910. A high density element 920 is also defined to include image artifacts generated thereby including shadows 922 generated by imaging or radiating a high density element 900 during breast imaging. Thus, a high density element may be a foreign or external object that is inserted into breast tissue 910 or attached to an outer breast surface 910 or be a naturally occurring material or component of breast tissue 910 having sufficient density to obscure other breast tissue that is clinically relevant information of breast tissue 910. For ease of explanation and not limitation, reference is made to a high density element 920, and a specific example of a metallic biopsy marker and a shadow 922 generated by imaging the metallic biopsy marker 920, but it will be understood that embodiments are not so limited.

[0106] The high density element 920 is illustrated as extending across multiple image slices 918. As generally illustrated in FIG. 9, the high density element 920 is denser than breast tissue 910 such that when imaged, a shadow 922 is generated, and the shadow 922 (as well as the metallic biopsy marker 920) obscures underlying and/or adjacent breast tissue 910 and clinically relevant information concerning same.

[0107] In the example generally illustrated in FIG. 9, the shadow 922 generated by imaging the metallic biopsy marker 920 is a complete, circumferential or global shadow since the shadow 922 surrounds the metallic biopsy marker 920. Shadows may be caused from various aspects of image acquisition. For example, the type of shadow 922 generally depicted in FIG. 9 may result from one or more of the limited angle of tomosynthesis acquisition and reconstruction, also known as a reconstruction artifact, and image processing and enhancement, also known as an enhancement artifact. The illustrative shadow 922 depicted in FIG. 9 overlaps or obscures 924 objects of interest or clinically relevant information of breast tissue 910 such as lesions and spiculations. The depth and dimensions of shadows 922 depicted in the 3D tomosynthesis stack 902 (or in 2D projection images Tp) resulting from imaging of the high density element 920 may vary based on one imaging and/or material attributes including the angles of the x-ray source utilized and number of projection images Tp acquired, the metallic biopsy marker 920 material, and the size, shape and orientation of the metallic biopsy marker 920 being imaged. Thus, FIG. 9 is provided for purposes of general illustration, not limitation, to illustrate that a high density element in the form of a metallic biopsy marker 920 and/or shadow 922 depicted in one or more images may obscure clinically relevant information. Moreover, FIG. 9 illustrates a single high density element 920 in a 3D tomosynthesis stack 902, but there may be multiple high density elements 920, each of which may generate their own shadow 922, and which may be distinct and independent of each other or overlap with other shadows 922. Thus, multiple markers 920 and shadows 922 can further complicate generation and review of synthesized images since they may obscure relevant information from at multiple viewpoints.

[0108] Referring to FIG. 10, embodiments of inventions provide breast image acquisition and processing systems 100s and multi-flow image processing methods 1000 that address complications with imaging high density elements 920 within breast images as discussed above with reference to FIG. 9 and provide for clearer and more accurate images that have reduced or are free of shadows for more accurate and efficient radiologist review. FIG. 10 illustrates a breast image generation and display system 100s (s referring to a breast image generation and display system that suppresses high density elements) constructed according to one embodiment and configured to execute a multi-flow or differential image processing method 1000 for selective high density element suppression and high density enhancement in breast images. Details of various system 100s components and interoperability thereof such as an acquisition system 101, storage system 102, reconstruction engine 103, a 2D image synthesizer 104 and a display 105 are discussed above with reference to FIGS. 1-8B and not repeated. Different images generated or processed thereby including acquired images, reconstructed images, synthesized images, Tp images (a 2D image acquired at respective tomosynthesis angles), Tr images (type (or subset) of a reconstructed image that is reconstructed from tomosynthesis projection images Tp) and Ms images (type (or subset) of a synthesized image, in particular, a synthesized 2D projection image that simulates mammography images) are also described above with reference to FIGS. 1-8B. For ease of explanation, embodiments of inventions are described with reference to 2D acquired images such as 2D projection images (e.g., Tp), reconstructed images (e.g., Tr) or a 3D stack 902 of image slices 918 and 2D synthesized images.

[0109] In the illustrated embodiment, the breast image generation and display system 100s includes a multi-flow image processor 1000 that is in communication with the reconstruction engine 103 and display 105. The image processor 1000 receives input images or digital image data 1001 of one or more types of images. The input image data 1001 (generally, input data 1001) may be for images of different dimensional formats such as 2D projection images and/or a 3D tomosynthesis stack 902 of image Tp slices 218. The input data 1001 is processed according to a first image processing flow or method 1010, and the same input data 1001 is processed with a second image processing flow or method 1020 different from the first processing flow or method 1010. The resulting 2D synthesized image is based at least in part upon high density element suppression and based at least in part upon high density element enhancement, and an image fusion or merge element 1030 combines the 2D synthesized images generated by respective image processing flows or methods 1010 and 1020 to generate a new 2D composite image 1032, which is communicated to display 105.

[0110] Thus, with the breast image generation and display system 100s, the same input data 1001 is processed in different ways according to different image processing flows to generate different 2D synthesized images, which are merged to generate a single 2D synthesized composite image 1034.

[0111] In the illustrated embodiment, the multi-flow image processor 1000 processes the same input data 1001 in different ways, which may be done by parallel and simultaneous image processing flows. In one embodiment, the input data 1001 is data of 2D projection images (Tp). In another embodiment, the input data 1001 is data of 3D images of a stack 902 of image slices 908. Different image processing methods executed based on the type of input data received are described in further detail below.

[0112] The input data 1001 received by the image processor is first processed in different ways, beginning with one or more image detectors 1011, 1021. Two image detectors 1011, 1021 are illustrated as the beginning of respective first and second image processing flows 1010, 1020. Image detector 1011 identifies and differentiates high density elements 920 and other elements such as breast tissue/background 910. Image detector 1021 identifies high density elements 920.

[0113] Image detectors 1011, 1021 may operate to distinguish a high density element 920 from breast tissue 910 or other image portions based on pre-determined filters or criteria involving, for example, one or more of image contrast, brightness, and radiopacity attributes. For example, high density element 920 may be associated with high contrast and brightness attributes compared to breast tissue or background 910 and thus be identified as a high density element. Detection criteria may involve a group of pixels or adjacent pixels having common characteristics, e.g., contrast or brightness within a certain range such that the group is identified as being a high density element. Image detectors may also distinguish a high density element 920 from best tissue based on shape, orientation and/or location data. For example, the image processor 1000 may be provide with specifications of known metallic biopsy markers. This data may be used in conjunction with image or pixel data such that image portions having similar properties also form a shape similar to a known shape of a biopsy marker, those pixels are identified as depicting a high density element 920. As another example, another factor that can be utilized to differentiate a high density element 920 is that skin markers are typically attached to an outer surface of the breast rather than being inserted into breast tissue. Thus, pixels having similar properties and being located at an outer surface indicative of an external skin marker are identified as a high density element 920. Location data can also be a factor, e.g., if a certain marker is inserted into a particular breast tissue region. Accordingly, it will be understood that image portions corresponding to high density elements and image portions corresponding to breast tissue or background 910 can be differentiated or detected in various ways using various filters, criteria and/or more sophisticated algorithms such as feature-based machine learning algorithms, or deep convolutional neural network algorithm.

[0114] Image detector 1011 is in communication with a high density element suppression module 1012, and image detector 1012 is in communication with a high density enhancement module 1024 such that respective detection results are provided to respective suppression and enhancement modules 1012, 1022. Respective outputs of respective high density element suppression and enhancement modules 1012, 1022 are provided as inputs to respective 2D image synthesizers 1014, 1024.

[0115] According to one embodiment, 2D image synthesizer 1014 used in the first image processing flow 1010 and that executes on high density element suppressed image portions operates in the same manner as 2D image synthesizer 104 that executes object enhancement and recognition modules 110a-n as discussed above with reference to FIGS. 1-8B, except that 2D image synthesizer 1014 receives high density suppressed image data. The 2D image synthesizer of the first image processing flow 1010 is thus referred to as 2D image synthesizer 104supp (supp referring to high density element suppressed). Thus, 2D image synthesizer 1014 is configured to process high density element suppressed data while providing for breast tissue object enhancement and recognition via modules 110a-n.

[0116] In contrast, 2D image synthesizer 1024 does not involve high density element suppression or high density element suppressed data, and instead processes high density element enhanced image data while not enhancing breast tissue. In this manner, the focus of 2D image synthesizer 1024 is high density element 920 enhanced image data rather than breast tissue 910 enhancement such that 2D image synthesizer 1024 may also be referred to as 2D image synthesizer 104enh (enh referring to high density element enhanced). For this purpose, the 2D image synthesizer 1024 may not include object enhancement and recognition modules 110a-n or these object enhancement and recognition modules 110a-n may be deactivated. Thus, 2D image synthesizer 1024 is configured to process high density element enhanced data while breast tissue is not enhanced.

[0117] The 2D image synthesizer 1014/104 supp outputs a 2D synthesized image 1016 that embodies high density element suppression and breast tissue enhancement data, and 2D image synthesizer 1024/104enh outputs a different 2D synthesized image 1026 that embodies high density element enhancement data. These different 2D synthesized images 1016, 1026 are provided as inputs to an image fusion or merging element 1030, which combines or merges the 2D synthesized images 1014, 1024 to generate a 2D composite synthesized image 1032 that incorporates elements of both of the 2D synthesized images 1014, 1024. Multi-flow image processing methods involving different types of input data 1001 and intermediate image and associated processing involving different dimensional formats and image or slice configurations are described in further detail with reference to FIGS. 11-24.

[0118] Referring to FIG. 11, in a multi-flow image processing method 1100 executed by breast image acquisition and processing system 100s according to one embodiment, at 1102, digital input data 1001 of one or more breast tissue images is fed as an input to a multi-flow or differential image processor 1000 of image generation and display system 100s such as a tomosynthesis system. At 1104, portions of images that depict breast tissue 910 and portions of images that depict high density elements 920 are identified or detected. A first image processing flow 1010 is executed on input data to generate first 2D synthesized image 1016.

[0119] Referring to FIG. 12, the first image processing flow or method 1010, or metal suppression flow, includes enhancing image portions depicting breast tissue 910 at 1202, whereas image portions depicting a high density element 920 such as a metallic biopsy marker 920 and/or shadow 922 are suppressed, replaced or eliminated at 1204. At 1206, a first set or stack of 3D image slices (e.g., Tr slices) embodying enhanced breast tissue and suppressed high density element image portions is constructed, and at 1208, a first 2D synthesized image 1016 is generated based at least in part upon first stack of 3D image slices.

[0120] Referring to FIG. 13, the second image processing method 1020 is different from first image processing method 1010 and is executed on the same input data 1001. The second image processing method 1020 generates a different, second 2D synthesized image 1026. At 1302, image portions depicting high density elements 920 are emphasized (without breast tissue enhancement, or by deemphasizing breast tissue), at 1304, a second set of 3D image slices based at least in part upon enhanced image portions depicting high density elements is generated. At 1306, the second 2D synthesized image 1026 is generated based at least in part upon second set of 3D image slices.

[0121] Referring again to FIG. 11, at 1110, the first and second 2D synthesized images 1016, 1026 are combined or merged to generate composite synthesized image 1032, and at 1112, the composite synthesized image 1032 is presented through display 105 of image generation and display system 100s to a radiologist or end user.

[0122] Referring to FIG. 14 and with further reference to FIG. 15, one embodiment of a method 1400 for processing breast images using the system 100s configuration shown in FIG. 10 and as described with reference to FIGS. 11-13 is described. In method 1400, the multi-flow image processing method 1000 is executed on an input of a reconstructed 3D stack of image slices in which breast tissue 910 and high density elements 920 are both visible. Thus, in this embodiment, a stack of 3D image slices 1506 rather than 2D projection images 1502 are provided as an input 1501 to the image processor 1500 such that the multi-flow image processing method 1500 is not executed on 2D projection images 1502.

[0123] At 1402, image acquisition component 101 (e.g., x-ray device of digital tomosynthesis system)) is activated, and at 1404, a plurality of 2-D images 1502 of patient's breast is acquired. For example, in a tomosynthesis system, approximately 15 2D projection images Tp 1502 may be acquired at respective angles between the breast and the x-ray sourcedetector. It will be understood that 15 2D projection images is provided as an example of how many projection images may be acquired, and other numbers, greater than and less than 15, may also be utilized. At 1406, if needed, the acquired or projection images 1502 are stored by the acquisition component 101 to a data store 102 for subsequent retrieval, which may be from a data store 102 that is remote relative to the image processor 1000 and via a communication network.

[0124] At 1408, 2D projection image reconstruction 1504 is executed to generate a 3D stack 1508 of image slices Tr 1506 (e.g., 60 image slices in the illustrative example). At 1410, the first detector 1511 of the first image processing flow 1510 identifies portions of input 3D image slices 1506 depicting breast tissue 910 and portions of image slices 1506 depicting high density elements 920 (e.g., metallic object or calcification, or shadow) generated by imaging high density element 920 in or on breast. A second detector 1521 identifies a high density element 920. For these purposes, the image processor 1500 may utilize one or more criteria or filters as described above to identify and differentiate breast tissue or background 910 and high density element image portions 920 in the 3D stack 1506.

[0125] Continuing with reference to FIGS. 14-15, at 1412, the first image processing flow 1510 involves high density element suppression 1512 of the input stack 1508, the result of which is generation of a first processed 3D stack 1513 in which a high density element 920 is suppressed or eliminated.

[0126] FIG. 16 illustrates in further detail one manner in which high density element suppression 1512 may be executed on the input 3D stack 1508 and also how an optional mask may be generated for subsequent use in generating a 2D synthesized composite image 1032. In the illustrated embodiment, the first image processing flow 1510 on the input 1501 of 3D image slices 1506 involves detection of portions of image slices 1506 depicting a high density element 920 such as a metallic biopsy marker at 1602, segmentation or pixel identification of the detected high density element portions 920 at 1604, and at 1606, a segmentation mask may be generated based on the segmentation results. The mask may be subsequently utilized when generating a 2D synthesized composite image 1032. At 1608, segmented portions are suppressed or eliminated from image slices of the 3D stack for high density element suppression. This may be done by interpolation or replacing segmented portions with other sampled portions of image slice background. High density element suppression results in the elimination of high density element 920 image portions from the 3D stack 1508 of image slices such that the high density element 920 would not be visually perceptible to a radiologist, or visually perceptible or to a lesser degree. Thus, the end result 1610 of the suppression process is a processed 3D stack 1610 of reconstruction image slices, or metal suppressed breast tissue slices, in which breast tissue image portions 910 are maintained while suppressing or eliminating image portions of high density elements 920, while also generating a separate high density mask.

[0127] FIG. 17 illustrates in further detail one manner in which high density element enhancement 1522 within the input 3D stack 1508 may be executed and also how an optional segmentation or pixel mask may be generated for subsequent use in generating a 2D synthesized composite image 1032. In the illustrated embodiment, the second image processing flow 1520 on the input 3D image slices 1501 involves detecting portions of image slices depicting a high density element 920 such as a metallic biopsy marker at 1702, segmentation of the detected high density element portions at 1704, and at 1706, a segmentation mask may be generated and may be subsequently utilized when generating a 2D synthesized composite image 1032. Metal segmentation 1704 information may be recorded as a metal segmentation mask 1706, and masks from different slices can be combined into a single 2D metal mask, which is a side output from the metal synthesizer module 1524. As an example, in the case of using a binary mask, within this 2D metal mask image, the high density element regions are marked with 1 and the background or breast tissue regions are marked with 0. Different mask configurations or designs can also be utilized for these purposes by utilizing other or multiple labels rather than only binary 0 and 1 labels. At 1708, segmented portions are isolated and emphasized or enhanced in image slices of the 3D stack. High density element enhancement 1708 may be executed using, for example, maximum intensity projection or MIP. The end result 1710 generated by the metal enhancement module is a stack 1523 of 3D reconstruction image slices in which breast tissue image portions 910 are not processed or not enhanced, and high density elements 920 are enhanced or emphasized.

[0128] Referring again to FIGS. 10 and 14-15, at 1414, the multi-flow image processor 1000 executes the first 2D image synthesizer 1514 that receives the processed or metal suppressed stack 1513 of 3D image slices as an input. The first 2D image synthesizer 1514 generates a 2D synthesized image 1515 based on the suppressed high density image portions of the metal suppressed 3D stack while enhancing or emphasizing breast tissue image portions by use of the target object recognition/enhancement modules (e.g., 110a, 110b . . . 110n), each configured for recognizing and enhancing a particular type of object. The first 2D image synthesizer 1514 may operate in the same manner as 2D image synthesizer 104 discussed above, except that the 2D image synthesizer receives image data resulting from high density element suppression 1512. As discussed above with reference to FIGS. 1-8B, target object recognition/enhancement modules 110a-n are configured to identify the respective object (if any is/are present) therein such that the resulting 2D synthesized images includes clinically-significant information.

[0129] With continuing reference to FIGS. 14-15 and with further reference to FIG. 18, the multi-flow image processor 1000 executes the second 2D image synthesizer 1524/1801 that receives the processed or metal enhanced stack 1523 of 3D image slices as an input to generate a 2D synthesized image 1525/1802 based on the enhanced high density image portions of the metal enhanced 3D stack while other background or breast tissue portions are maintained or not enhanced, or even deemphasized, e.g., by reducing brightness thereof. For this purpose, the second 2D image synthesizer 1524 does not include, or deactivates, target object recognition/enhancement modules (e.g., 110a, 110b . . . 110n) such that these breast tissue analyses and enhancements are not performed and are not necessary in view of a high density element 920 structure. For example, a metallic biopsy marker may have a less complex geometric shape (e.g., a cylinder), or is typically less complex than breast tissue. For example, rather than employing more complicated target object recognition/enhancement 110a-n, the second image processing flow in which high density elements 920 are enhanced can deploy simply image processing algorithms such as mean-intensity projection or max-intensity projection as the base method to combine 3D stack of metal object slices into a single metal object 2D synthetic image 1810, which may be stored to a buffer. The result generated by the second 2D synthesizer 1524/1801 is generally illustrated by the high density object appearing as a dot in the 2D synthesized image 1525 1802 in FIG. 18.

[0130] FIG. 18 also illustrates the 2D synthetic image 1802 including various artifacts 1810 resulting from imperfections in metal detection and segmentation processes. Morphological operations 1803 (e.g. pixel dilation and/or erosion) can be executed on the 2D synthetic image 1802 to clean these artifacts 1810 by smoothing the high density object boundary to make the boundary in the resulting 2D image 1525/1804 more accurate and more visually appealing.

[0131] Referring again to FIGS. 14-15, having generated a first 2D synthesized image 1515 based at least in part upon first stack of 3D image slices and a second 2D synthesized image 1525 based at least in part upon second stack of 3D image slices, at 1420, these intermediate first and second 2D synthesized images 1515, 1525 generated by respective first and second image processing flows 1510, 1520 are merged or combined 1530 to generate a 2D final or composite synthesized image 1532, which is presented to a radiologist or end user via display 105 at 1422. According to one embodiment, image combination 1530 may involve selecting the best signals of 2D synthetic image data from each synthetic image buffer and ensuring that the transition between the breast tissue 910 and the high density element 920 is seamless. The 2D composite synthesized image 1532 is visually free of shadow artifacts 920 such that unwanted shadow artifacts do not obscure clinically important information while also including enhanced breast tissue or background and sharp delineations between breast tissue 910 and high density elements 920.

[0132] Referring to FIG. 19, according to one embodiment of combining 1030 the first and second 2D synthesized images 1515, 1525, the 2D metal mask 1900 generated by segmentation as discussed above may be utilized for modulated combination 1902 or maximum intensity projection or MIP combination of the intermediate first and second 2D synthesized images 1515, 1525 to generate the 2D composite synthesized image 1532. This embodiment essentially extracts the signals or image portions from each 2D synthetic image 1515, 1525 buffer for seamless transition between breast tissue 910 and high density elements 920 such that the resulting 2D composite image 1032 is visually sharp, free of high density shadow elements while providing for optimal breast tissue background.

[0133] FIGS. 20A-B illustrate an example of how multi-flow image processing embodiments can be executed to generate a 2D synthesized composite image 1032 that is visually sharp and clear with reduced or eliminated shadow 922 artifacts. FIG. 21B illustrates a 2D synthesized image 1032 that is constructed according to multi-flow image processing of embodiments that eliminates obscuring shadow artifacts 922 compared to FIG. 21A, which includes various shadow artifacts 922 around the metallic biopsy marker 920. The final result of a 2D synthesis composite image 1032 generated according to embodiments is sharp and shadow 922 free while breast tissue or background 910 is also enhanced.

[0134] Certain embodiments described above with reference to FIGS. 10-20B involve the multi-flow image processor 1500 receiving reconstructed or generated images or a 3D stack 1508 of image slices 1506 (e.g., 60 reconstructed slices) as an input 1001 such that multi-flow image processing is executed on the same 3D stack 1508. The 3D stack 1508 is generated based on acquired 2D projection images 1502, which are not provided as an input 1001 to the image processor 1000 in these embodiments. Thus, the multi-flow image processing is not executed on the 2D projection images 1502 in these embodiments. In other words, the multi-flow image processing is executed directly on the 3D stack 1508 of image slices, but not the 2D projection images 1502 upon which the 3D stack 1508 of image slices is based, and the multi-flow image processing is executed after reconstruction 1504. Other embodiments may involve the image processor 1000 receiving inputs of different image types and dimensional formats.

[0135] For example, in other embodiments, the multi-flow image processor receives an input of 2D projection images such that the multi-flow image processing is executed directly on the 2D projection images rather than the 3D stack of image slices that is eventually generated after reconstruction. Different 3D stacks of image slices are provided as respective inputs to respective 2D image synthesizers after suppression and enhancement processing has been executed on 2D projection images. Thus, in certain embodiments, high density element suppression and enhancement occurs after reconstruction 1504 of a 3D stack 1508 of image slices 1506, whereas in other embodiments, high density element suppression and enhancement occur before reconstruction of a 3D stack of image slices. Alternative embodiments of multi-flow image processing involving execution of image processing embodiments using 2D projection images as an input to the image processor are described with reference to FIGS. 21-24. System components and their interoperability described above are not repeated.

[0136] Referring to FIGS. 21-22, in an image processing method 2100 according to another embodiment, at 2102, image acquisition component 101 (e.g., x-ray device) of the image generation and display system 100s is activated, and at 2104, a plurality of 2-D images of patient's breast 2201 (e.g., 15 projection images at respective angles between breast and x-ray source-detector) is acquired. At 2106, 2D projection images 2201 are stored to a data store 102, and at 2108, digital image data of the 2-D projection images 2201 is received from the data store 102 and provided as an input to the multi-flow image processor 2200 of image generation and display system 100s. At 2210, a first detection module 2211 identifies portions of individual 2D projection images 2201 depicting breast tissue 910 and portions of individual 2D projection images 2201 depicting high density elements 920 (e.g., metallic biopsy marker or shadow) generated by imaging the metallic biopsy marker, and a second detection module 2221 identifies portions of individual 2D projection images 2201 depicting high density elements 920 (e.g., metallic object or shadow) generated by imaging high density object in or on breast.

[0137] At 2212, the first image processing method or flow 2210 including high density element suppression 2212 is executed on the input 2D projection images 2201 to generate processed/high density element suppressed 2D projection images 2213, and at 2214, the second image processing method or flow 2220 including high density element enhancement 2222 is executed on the input 2D projection images 2201 to generate processed/high density element enhanced 2D projection images 2223.

[0138] In certain embodiments, all of the input 2D projection images 2201 are suppressed in some way, whereas in other embodiments, only certain input 2D projection images 2201 are subjected to high density suppression 2212, e.g., only those determined to include at least a portion of a high density element 920. Thus, in certain embodiments, high density suppression 2212 and high density enhancement 2222 are both executed before any image reconstruction into a 3D stack of image slices. Further, in one embodiment, each input 2D projection image 2201 is processed such that the set of processed of 2D projection images 2213, 2223 is the same as the number of input 2D projection images 2201, but it will be understood that embodiments are not so limited. For example, the number of input 2D projection images 2201 that are subjected to high density element suppression 2212 and enhancement 2213 may be less than the number of input 2D projection images 2201 if only those input 2D projection images 2201 that are determined to include a high density element 920 are processed. Thus, for example, image acquisition may result in 15 input 2D projection images 2201, only eight of which contain at least a portion of a high density element 920, in which case only those eight input 2D projection images 2201 are processed for high density element suppression 2212 and enhancement 2222. The remaining seven input 2D projection images 2201 may be rejoined with the eight that were processed for a set of 15 projection images prior to reconstruction and generation of a 3D stack.

[0139] Accordingly, high density element suppression 2212 and enhancement 2222 may be executed before any 3D image reconstruction, on all of the 2D projection images 2201 of the input set, or on selected 2D projection images 2201 of the input set, e.g., those determined to contain high density elements by detector 2211, since a metallic object or shadow 920 generated thereby may not be present in certain images depending on the high density element size, location and orientation and position relative to a radiation source and detector used for imaging. Moreover, the number of processed 2D projection images 2213, 2223 following suppression 2212 and enhancement 2222 may be the same as the number of input 2D projection images 2201 even if only some of the input 2D projection images 2201 are processed since unprocessed input 2D projection images 2201 may be added to the processed set.

[0140] Continuing with reference to FIGS. 21-22, having generated the processed set of 2D projection images 2213, 2223, at 2216, a first stack 2214 of 3D image slices (e.g., 60 image slices) is generated based at least in part upon the first set of processed 2D projection images 2213 (e.g., 15 images) involving high density element suppression 2212, and at 2218, a second stack 2224 of 3D image slices is generated based at least in part upon the second set of processed 2D projection images 2223 involving high density element enhancement 2222.

[0141] Having constructed the first and second stacks of 3D images 2214, 2224, these stacks are then processed at 2120, 2122 by respective 2D image synthesizers 2215, 2225 to generate respective first and second 2D synthesized images 2216, 2226 based at least in part upon respective first and second stacks 2214, 2224. At 2124, morphological operations may be executed on the second 2D synthesized image 2226 as necessary to dilate or erode image edges of enhanced image portions depicting high density elements as necessary, and at 2226, the first and second 2D synthesized images 2216, 2226 are merged or combined 2230 to generate a 2D composite image at 2232, which is presented to the radiologist or end user via a display 105.

[0142] FIGS. 23-24 further illustrate how respective suppression 2212 and enhancement 2222 processing are executed, and are similar to the processing described with reference to FIGS. 15-17 above except that the detection, segmentation and suppression (FIG. 16) and suppression (FIG. 17) are based on inputs of individual input 2D projection images 1502 rather than on an input 3D stack of image slices, the resulting masks 2306, 2406 generated by segmentation is a mask for an individual image as shown in FIGS. 23-24 rather than for a stack of 3D image slices as shown in FIGS. 15-17, and the result or output of the suppression and enhancement processing is a suppressed or enhanced processed 2D projection image as shown in FIG. 22 rather than an output of a high density element suppressed 3D stack.

[0143] Having described exemplary embodiments, it can be appreciated that the examples described above and depicted in the accompanying figures are only illustrative, and that other embodiments and examples also are encompassed within the scope of the appended claims. For example, while the flow diagrams provided in the accompanying figures are illustrative of exemplary steps; the overall image merge process may be achieved in a variety of manners using other data merge methods known in the art. The system block diagrams are similarly representative only, illustrating functional delineations that are not to be viewed as limiting requirements of the disclosed inventions. It will also be apparent to those skilled in the art that various changes and modifications may be made to the depicted and/or described embodiments (e.g., the dimensions of various parts), without departing from the scope of the disclosed inventions, which is to be defined only by the following claims and their equivalents. The specification and drawings are, accordingly, to be regarded in an illustrative rather than restrictive sense.