Embodiments of the present invention relates to an X-ray contrast agent. The X-ray contrast agent has an X-ray absorption the change of which between at least two different X-ray photon energy levels differs from the change in X-ray absorption of calcium between the at least two different X-ray photon energy level. Embodiments of the present invention also relates to an X-ray imaging method. Embodiments of the present invention additionally relates to an image reconstruction device. Embodiments of the present invention further relates to an X-ray imaging system.

Among other things, techniques are described for cross-modality active learning for object detection. In an example, a first set of predicted bounding boxes and a second set of predicted bounding boxes is generated. The first set of predicted bounding boxes and the second set of predicted bounding boxes are projected into a same representation. The projections are filtered, wherein predicted bounding boxes satisfying a maximum confidence score are selected for inconsistency calculations. Inconsistencies are calculated across the projected bounding boxes based on filtering the projections. An informative scene is extracted based on the calculated inconsistencies. A first object detection neural network or a second object detection neural network is trained using the informative scenes.

A method and system for facilitating identification and marking of a target in a displayed Fluoroscopic Three-Dimensional Reconstruction (F3DR) of a body region of a patient. The system includes a display and a storage device storing instructions for receiving an initial selection of the target in the F3DR, fining the F3DR based on the initial selection of the target, displaying the fined F3DR on the display, and receiving a final selection of the target in the fined F3DR via a user selection. The system further includes at least one hardware processor configured to execute said instructions. The method and instructions may also include receiving a selection of a medical device in two two-dimensional fluoroscopic images, where the medical device is located in an area of the target, and initially fining the F3DR based on the selection of the medical device.

Systems and methods for structure footprint detection from oblique imagery are disclosed, including a computer system configured to receive geo-referenced oblique images; analyze pixels of the images to: identify pixels representing a structure with walls; determine ground locations for the walls, geographic locations and orientations of pixels representing vertical edges of the walls, and relative lengths of the walls to produce horizontal line segments representing the base of the walls and having a relative length and an orientation, the horizontal line segment(s) determined from horizontal edge(s) extending a length between vertical edges above the bottoms of the vertical edges such that the horizontal edge is above the base of the structure; and assemble the horizontal line segments based on their relative lengths and orientations to form a footprint of the structure.

An image processing apparatus includes at least one processor, and the processor derives three-dimensional coordinate information that defines a position of a structure in a tomographic plane from a tomographic image including the structure, and that defines a position of an end part of the structure outside the tomographic plane in a direction intersecting the tomographic image.

A system and method for generating reports on perfusion blood volume from computed tomography (CT) data acquired from a subject. The method includes receiving multi-faceted CT data acquired from the subject using one of a multi-energy or polychromatic CT acquisition and deriving an iodine concentration in an artery feeding a volume of interest (VOI) in the multi-faceted CT data. The method further includes determining an effective atomic number of a spatial distribution in the VOL calculating a perfused blood volume of the VOI using the iodine concentration and the effective atomic number, and generating a report of the perfused blood volume of the VOI.

Data processing systems (DPS) and related methods for nuclear medicine imaging. At an input interface (IN), first projection data (λ), or a first image (V) reconstructable from the first projection data, is received. The first projection data is associated with a first waiting period (ΔT*). The first waiting period indicates the time period from administration of a tracer agent to a start of acquisition by a nuclear medicine imaging apparatus (IA) of the projection date. A trained machine learning module (MLM) estimates, based on the first projection data (λ) or on the first image (V), a second projection data (λ′) or a second image (V′) associable with a second waiting period (ΔT), longer than the first waiting period (ΔT*). Nuclear imaging can thus be conducted quicker. Similar machine learning based data processing systems and related methods are also envisaged to reduce acquisition time periods or the time it takes to reconstruct imagery.

An X-ray imaging system using multiple puked X-ray sources to perform highly efficient and ultrafast 3D radiography is presented. There are multiple puked X-ray sources mounted on a structure in motion to form an array of sources. The multiple X-ray sources move simultaneously relative to an object on a pre-defined arc track at a constant speed as a group. Electron beam inside each individual X-ray tube is deflected by magnetic or electrical field to move focal spot a small distance. When focal spot of an X-ray tube beam has a speed that is equal to group speed but with opposite moving direction, the X-ray source and X-ray flat panel detector are activated through an external exposure control unit so that source tube stay momentarily standstill equivalently. 3D scan can cover much wider sweep angle in much shorter time and image analysis can also be done in real-time.

A method includes recording a plurality of projection recordings along a linear trajectory. An X-ray source and an X-ray detector move in parallel opposite to one another along the linear trajectory and the examination object is arranged between the X-ray source and the X-ray detector. The method includes reconstructing a tomosynthesis dataset, respective depth information of the examination object is respective determined along an X-ray beam bundle spanned by the motion along the linear trajectory and an X-ray beam fan of the X-ray source perpendicular to the linear trajectory so that different respective depth levels in the object parallel to a detection surface of the X-ray detector are respectively scanned differently. Finally, the method includes determining a first slice image with a first slice thickness in a depth level, among the respective depth levels, substantially parallel to the detection surface of the X-ray detector based on the tomosynthesis dataset.

Techniques are described for domain adaptation of image processing models using post-processing model correction According to an embodiment, a method comprises training, by a system operatively coupled to a processor, a post-processing model to correct an image-based inference output of a source image processing model that results from application of the source image processing model to a target image from a target domain that differs from a source domain, wherein the source image processing model was trained on source images from the source domain. In one or more implementations, the source imaging processing model comprises an organ segmentation model and the post-processing model can comprise a shape-autoencoder. The method further comprises applying, by the system, the source image processing model and the post-processing model to target images from the target domain to generate optimized image-based inference outputs for the target images.