The invention relates to methods and systems for automatically detecting an optic disc in a retinal fundus autofluorescence image. A monochromatic image of a retina is obtained by stimulated autofluorescence. The image is processed using filters and machine learning to identify blood vessel segments and to identify pixels as possible optic disc pixels. The possible optic disc pixels are grouped into clusters, and a best fit circle (or portion thereof lying within the image) is fitted to each cluster. The circle may be enlarged to improve image contrast at the circumference of the circle. Blood vessel segments that intersect the circle are identified, and metrics are derived from such intersecting blood vessel segments. These metrics are assessed by machine learning processes to determine the probability that each cluster contains the optic disc, and the contours of the optic disc are further determined by analysis of the possible optic disc pixels.

An information processing apparatus comprising: an extraction unit configured to extract a predetermined region of an anatomical part from an examination image of a subject; a first feature amount acquisition unit configured to acquire a first feature amount of the predetermined region related to a movement of the anatomical part; a second feature amount acquisition unit configured to acquire a second feature amount of the predetermined region related to the movement; a division unit configured to divide the predetermined region; and an integration unit configured to integrate the first feature amount and the second feature amount based on a result of division by the division unit.

Techniques and systems are described for automatic artifact removal in a digital image. A segmentation map is generated that describes a magnitude of difference among pixels in a digital image. Contours may be generated that describe boundaries of objects described in the segmentation map. The contours may be filtered according to two-dimensional and three-dimensional cues to identify contours corresponding to artifacts in the digital image. For each contour corresponding to an artifact, an object mask and a sampling mask may be generated. The object mask and the sampling mask may be utilized as part of a content filling operation upon the digital image to remove the artifact, and a corrected digital image is generated that does not include the artifact.

A method for determining respiratory induced blood mass change from a four-dimensional computed tomography (4D CT) includes receiving a 4D CT image set which contains a first three-dimensional computed tomographic image (3D CT) and a second 3D CT image. The method includes executing a deformable image registration (DIR) function on the received 4D CT image set, and determining a displacement vector field indicative of the lung motion induced by patient respiration. The method further includes segmenting the received 3D CT images into a first segmented image and a second segmented. The method includes determining the change in blood mass between the first 3D CT image and the second 3D CT image from the DIR solution, the segmented images, and measured CT densities.

A method of identifying fluid retention in a body part of the patient by directly or indirectly measuring a first parameter relating to a size of a body part of the patient to obtain an actual measurement of the body part, obtaining an estimated measurement of said first parameter relating to the size of the body part of the patient by measuring alternative predefined parameters of the patient, wherein said estimated measurement is calculated based on a mathematical relationship between said alternative parameters and the size of the body part; and correlating the actual and estimated measurements of the body part of the patient to assess any fluid retention in the body part.

The disclosure provides a multi-target 3D ultrasound image segmentation method based on simulated and measured data. The method includes: presetting conventional acoustic parameters; collecting raw 3D data; employing an initial segmentation algorithm to segment the raw 3D data; substituting with the conventional acoustic parameters according to probability in order to form a transitional image model; performing a simulation operation; performing transformation to obtain simulated data; performing a comparison operation; adjusting corresponding magnitude of the probability in each probability variable, and returning to the step of substituting with the conventional acoustic parameters. According to the probability, the conventional acoustic parameters are substituted into a model of an incomplete target tissue, the simulation operation is then performed, the probability is adjusted, and repeatedly, corrections are performed continuously in an iterative convergence manner till each incomplete target tissue is completely substituted by a certain normal tissue or lesion tissue.

A method, system and computer readable storage media for segmenting individual intra-oral measurements and registering said individual intraoral measurements to eliminate or reduce registration errors. An operator may use a dental camera to scan teeth and a trained deep neural network may automatically detect portions of the input images that can cause registration errors and reduce or eliminate the effect of these sources of registration errors.

Systems and methods are disclosed for performing operations comprising: receiving first and second images depicting an anatomy of a subject; obtaining a segmentation associated with the first image; applying a trained neural network to estimate the adapted segmentation corresponding to the anatomy depicted in the second image, the trained network consisting of three sub-networks: a registration sub-network, generating an initial segmentation estimate representing a deformation of the segmentation associated with the first image to fit the anatomy depicted in the second image, a segmentation sub-network, generating a second initial segmentation estimate for the second image, and a third refinement sub-network, combining the two initial segmentations and generating a refined segmentation for the second image.

A method for adaptive treatment planning is provided. The method may include obtaining a planning image volume of a subject, a treatment image volume of the subject, and a first treatment plan related to the planning image volume of the subject, each of the planning image volume and the treatment image volume including an ROI of the subject. The method may also include registering the planning image volume and the treatment image volume, and determining a first contour of the ROI in the registered planning image volume and a second contour of the ROI in the registered treatment image volume. The method may also include evaluating whether an error exists in at least one of the registration or the contour determination based on the first contour and the second contour, and determining a second treatment plan with respect to the treatment image volume based on the evaluation result.

Provided herein are systems and methods for automatically segmenting a 3D model of a patient's teeth. A patient's dentition may be scanned with a 3D scanning system, such as CT, CBCT, or MRI. The 3D scan data may be automatically segmented with one or more neural networks. The segmented 3D scan can be incorporated into a dental treatment plan.