Apparatuses, systems, and techniques are presented to predict annotations for objects in images. In at least one embodiment, one or more neural networks are used to help generate one or more segmentation boundaries of one or more objects within one or more digital images, wherein the one or more neural networks are to transform one or more representations of one or more portions of the one or more objects into one or more lower-dimensional representations of the one or more portions of the one or more objects.

Medical imaging methods for processing a three-dimensional (3D) image data set with two-dimensional X-ray images from an X-ray machine using a target function. Methods can include providing a 3D image data set of at least one examination zone in which anatomical structures are present, segmenting the image data set to provide a 3D vascular structure model and a 3D bone structure model, recording a first two-dimensional (2D) X-ray image containing at least a portion of the vascular structure and at least a portion of the bone structure, recording a second 2D X-ray image of the examination zone at a different contrast agent concentration, and subtracting the first and second 2D X-ray images to generate a subtraction image. An optimum projective geometry may then be determined using a three-part target function based on the 3D image data and the 2D X-ray images.

An image processing apparatus includes an acquisition unit configured to acquire first medical image data and second medical image data obtained by imaging a subject, an intermediate deformation information acquisition unit configured to acquire intermediate deformation information obtained by applying registration processing up to a predetermined stage in first registration processing including a plurality of stages to the acquired first medical image data and second medical image data, a determination unit configured to perform determination of a deformation abnormality with respect to the acquired intermediate deformation information, and a deformation unit configured to perform, in a case where the determination unit determines that there is the deformation abnormality, second registration processing different from the first registration processing with respect to the first medical image data and the second medical image data and calculate deformation information.

A method of generating a correction plan for a knee of a patient includes obtaining a ratio of reference bone density to reference ligament tension in a reference population. A bone of the knee of the patient may be imaged. From the image of the bone, a first dataset may be determined including at least one site of ligament attachment and existing dwell points of a medial femoral condyle and lateral femoral condyle of the patient on a tibia of the patient. Desired positions of contact in three dimensions of the femoral condyles of the patient with the tibia of the patient may be obtained by determining a relationship in which a ratio of bone density to ligament tension of the patient is substantially equal to the ratio of reference bone density to reference ligament tension.

An ultrasound diagnostic apparatus includes an ultrasound probe, a reference image holding unit that holds an ultrasound image acquired by fixing a position of the ultrasound probe as a reference image, a movement vector calculation unit that calculates a movement vector between two ultrasound images, a movement vector integration unit that integrates the movement vector from a time when the reference image is held to a current time, a deformed image generation unit that generates a deformed image in which the current ultrasound image is moved and changed to a time when the reference image is held based on an integration result, a tomographic plane determination unit that compares the deformed image with the reference image to determine whether tomographic planes of the current ultrasound image and the reference image are the same as each other, and a determination result notification unit that notifies a user of a determination result.

The invention provides methods and apparatus for image processing that perform image segmentation on data sets in two- and/or three-dimensions so as to resolve structures that have the same or similar grey values (and that would otherwise render with the same or similar intensity values) and that, thereby, facilitate visualization and processing of those data sets.

A learning model provided in a segmentation device is a learning model which is generated using training data such that segmentation data of a biologically important region is output when data of a constituent maxillofacial region is input.

A system of deep learning neural network in prostate cancer bone metastasis identification based on whole body bone scan images includes a pre-processing module for receiving input whole body bone scan images, and a neural network module for detecting whether there is a prostate cancer bone metastasis. The neural network module includes: a chest portion network module for establishing first stage faster R-CNN and segmenting training images of chest portion according to the input whole body bone scan images, and using the training images to train second stage faster R-CNN and categorizing the lesions of cancerous bone metastasis; and a pelvis portion network module for establishing first stage faster R-CNN and segmenting training images of pelvis portion according to the input whole body bone scan images, and using the training images to train the convolutional neural network to categorize whether it is a bone metastasis image.

A method and apparatus for planning an orthopedic procedure is disclosed. The method comprises retrieving medical imaging data, identifying a plurality of landmarks including at least a first landmark at a portion of the first bone and at least a second landmark at a portion of the second bone comprised in the medical imaging data. The portion of the first bone and the portion of the second bone may be segmented such that the portion of the first bone is moveable relative the portion of the second bone. At least one of a first implant component and a second implant component can be selected from among a plurality of implant components in a database based on information obtained from the first landmark and the second landmark The first implant component and/or the second implant component can be fitted in a space at least partially defined by the first landmark and the second landmark

According to a method for correcting a 2D measurement value is described, 2D image data of an examination object is received. Landmarks in the 2D image data are detected, and 2D positions of the landmarks are calculated. A corrected measurement value of the examination object is predicted, using a trained model, which depends on the received 2D image data, the estimated 2D positions of the landmarks and a reference parameter of a reference 3D orientation of the examination object.