Abstract: In one embodiment, a method includes indexing a whole slide image dataset to generate one or more dataset embeddings corresponding to one or more respective regions of one or more whole slide images. Each dataset embedding includes a feature vector mapping the respective region to a feature embedding space. The method includes accessing a query image and generating an embedding for the query image that includes a feature vector mapping the query image to the feature embedding space. The method includes identifying result tiles by comparing the embedding for the query image to one or more of the dataset embeddings. The comparison is based on one or more distances between the embedding for the query image and the one or more of the dataset embeddings in the feature embedding space. The method includes generating a user interface including a display of the result tiles.

Abstract: Systems and methods relate to predicting disease progression by processing digital pathology images using neural networks. A digital pathology image that depicts a specimen stained with one or more stains is accessed. The specimen may have been collected from a subject. A set of patches are defined for the digital pathology image. Each patch of the set of patches depicts a portion of the digital pathology image. For each patch of the set of patches and using an attention-score neural network, an attention score is generated. The attention-score neural network may have been trained using a loss function that penalized attention-score variability across patches in training digital pathology images labeled to indicate no or low subsequent disease progression. Using a result-prediction neural network and the attention scores, a result is generated that represents a prediction of whether or an extent to which a disease of the subject will progress.

Abstract: The method includes receiving digital images of tissue samples of patients, the images having assigned a label indicating a patient-related attribute value; splitting each received image into a set of image tiles; computing a feature vector for each tile; training a Multiple-Instance-Learning program on all the tiles and respective feature vectors for computing for each of the tiles a numerical value being indicative of the predictive power of the feature vector associated with the tile in respect to the label of the tile's respective image; and outputting a report gallery including tiles sorted in accordance with their respectively computed numerical value and/or including a graphical representation of the numerical value.

Abstract: Apparatus and methods are described including a processor (20) configured to receive first and second sets of extraluminal images of a lumen, the second set being acquired while an endoluminal device is being moved through the lumen. The processor designates at least one of the first set of images as a roadmap image, and designates, within the lumen in the roadmap image, a roadmap pathway (42). A plurality of features (44, 46, 48, 50) that are visible within at least some of the second set of extraluminal images are identified, and an arrangement of three or more of the features within the image are compared to a shape of at least a portion of the roadmap pathway. Based upon the comparing, the identified features are mapped to locations along the roadmap pathway within the roadmap image. Other applications are also described.

Abstract: The method includes generating, for each of a plurality of original images, a first artificially degraded image by applying a first image-artifact-generation logic on each of the original images; and generating the program logic by training an untrained version of a first machine-learning logic that encodes a first artifacts-removal logic on the original images and their respectively generated first degraded images; and returning the trained first machine-learning logic as the program logic or as a component thereof. The first image-artifact-generation logic is A) an image-acquisition-system-specific image-artifact-generation logic or B) a tissue-staining-artifact-generation logic.

Abstract: Apparatus and methods are described including, using a computer processor, automatically identifying whether a given pixel within an image corresponds to a portion of an object. A set of concentric circles that are disposed around the pixel are sampled, and a first function is applied to each of the circles such that the circles are defined by a first set of rotationally invariant descriptors. A second function is applied to the set of circles to generate a second set of descriptors, each of which represents a difference between respective pairs of the circles. A third function is applied such that the second set of descriptors becomes rotationally invariant. The processor identifies whether the given pixel corresponds to the portion of the object, based upon the first and second sets of rotationally invariant descriptors. Other applications are also described.

Abstract: The method includes generating, for each of a plurality of original images, a first artificially degraded image by applying a first image-artifact-generation logic on each of the original images; and generating the program logic by training an untrained version of a first machine-learning logic that encodes a first artifacts-removal logic on the original images and their respectively generated first degraded images; and returning the trained first machine-learning logic as the program logic or as a component thereof. The first image-artifact-generation logic is A) an image-acquisition-system-specific image-artifact-generation logic or B) a tissue-staining-artifact-generation logic.

Abstract: The method includes generating, for each of a plurality of original images, a first artificially degraded image by applying a first image-artifact-generation logic on each of the original images; and generating the program logic by training an untrained version of a first machine-learning logic that encodes a first artifacts-removal logic on the original images and their respectively generated first degraded images; and returning the trained first machine-learning logic as the program logic or as a component thereof. The first image-artifact-generation logic is A) an image-acquisition-system-specific image-artifact-generation logic or B) a tissue-staining-artifact-generation logic.

Abstract: The invention relates to a method of identifying a biomarker in a tissue sample. The method comprises receiving an acquired image depicting a tissue sample, the pixel intensity values of the acquired image correlating with an autofluorescence signal or of an X-ray induced signal or a signal of a non-biomarker specific stain or a signal of a first biomarker specific stain adapted to selectively stain a first biomarker. The acquired image is input into a trained machine learning logic—MLL which automatically transforms the acquired image into an output image highlighting tissue regions predicted to comprise a second biomarker.

Abstract: The method includes, for each of a plurality of tiles of an image, extracting a feature vector from the tile; providing a Multiple-Instance-Learning program configured to use a model for classifying any input image as a member of one out of at least two different classes based on feature vectors extracted from the tiles; for each of the tiles, computing a certainty value indicating the certainty of the model regarding the contribution of the tile's feature vector on the classification of the image; for each of the images, using, by the MIL-program, a certainty-value-based pooling function for aggregating the feature vectors of the image or predictive values computed from the feature vectors of the image into an aggregated predictive value as a function of the certainty values of the tiles; and classifying each of the images as a member of one of the classes based on the aggregated predictive value.

Abstract: The method includes receiving a digital image of a slide, the slide including the tissue sample, a target dot and a fluoro dot. The digital image includes intensity values of the tissue sample, the target dot and the fluoro dot. The method further includes receiving pixel intensities of the fluoro dot depicted in the image; pixel intensities of the target dot depicted in the image; pixel intensities of an area of the tissue sample depicted in the image; fluorescence-inducing molecule density information indicative of the known density of the fluorescence-inducing molecules in the fluoro dot; target molecule density information indicative of the known density of the target molecules in the target dot; and computing the density of target molecules in the area of the tissue sample as a function of the received pixel intensities and the received molecule density information.

Abstract: The method includes receiving a plurality of digital images each depicting a tissue sample; splitting each of the received images into a plurality of tiles; automatically generating tile pairs, each tile pair having assigned a label being indicative of the degree of similarity of two tissue patterns depicted in the two tiles of the pair, wherein the degree of similarity is computed as a function of the spatial proximity of the two tiles in the pair, wherein the distance positively correlates with dissimilarity; and training a machine learning module—MLM—using the labeled tile pairs as training data to generate a trained MLM, the trained MLM being configured for performing an image analysis of digital histopathology images.

Abstract: At least one embodiment relates to an image analysis system for tumor classification. The system is configured for receiving at least one digital image of a tissue sample; analyzing the at least one received image for identifying immune cells and tumor cells in the at least one received image; for each of the identified tumor cells, determining the distance of the tumor cell to the nearest immune cell; computing a proximity measure as a function of the determined distances; in dependence on the proximity measure, classifying the identified tumor cells into tumor cells of an inflammatory tumor or as tumor cells of a non-inflammatory tumor; and storing the classification result on a storage medium and/or displaying the classification result on a display device.

Abstract: Apparatus and methods are described for use with a tool that is inserted into a portion of a body of a subject that undergoes cyclic motion. The apparatus and methods include using an imaging device, acquiring a plurality of images of the tool inside the portion of the body, at respective phases of the cyclic motion. Using at least one computer processor, an extent of movement of the tool with respect to the portion of the body that is due to the cyclic motion of the portion of the body is determined. In response thereto, an output is generated that is indicative of the determined extent of the movement of the tool with respect to the portion of the body. Other applications are also described.

Abstract: Apparatus and methods are described for use with a tool that is inserted into a portion of a body of a subject that undergoes cyclic motion. The apparatus and methods include using an imaging device, acquiring a plurality of images of the tool inside the portion of the body, at respective phases of the cyclic motion. Using at least one computer processor, an extent of movement of the tool with respect to the portion of the body that is due to the cyclic motion of the portion of the body is determined. In response thereto, an output is generated that is indicative of the determined extent of the movement of the tool with respect to the portion of the body. Other applications are also described.

Abstract: Apparatus and methods are described for imaging a tool inside a portion of a subject's body that undergoes motion. A plurality of image frames are acquired of the portion of the subject's body. The image frames are image tracked by (a) automatically identifying at least a feature of the tool in at least a portion of the image frames, and (b) aligning the tool in image frames of the portion of the image frames, based on the automatic identifying. The image-tracked image frames of the portion of the subject's body are displayed as an image stream. Other embodiments are also described.

Abstract: The method includes receiving digital images of tissue samples of patients, the images having assigned a label indicating a patient-related attribute value; splitting each received image into a set of image tiles; computing a feature vector for each tile; training a Multiple-Instance-Learning program on all the tiles and respective feature vectors for computing for each of the tiles a numerical value being indicative of the predictive power of the feature vector associated with the tile in respect to the label of the tile's respective image; and outputting a report gallery including tiles sorted in accordance with their respectively computed numerical value and/or including a graphical representation of the numerical value.

Abstract: Apparatus and methods are described for use with an imaging device (12) configured to acquire a set of angiographic images of a lumen. At least one processor (10) includes blood-velocity-determination functionality (16) that determines blood velocity within the lumen, via image processing. Current-flow-related-parameter-determination functionality (18) determines a value of a flow-related parameter at the location based upon the determined blood velocity. Flow-related-parameter-receiving functionality (19) receives an indication of a value of a second flow-related parameter of the subject, and index-determination functionality (21) determines a value of a luminal-flow-related index of the subject at the location, by determining a relationship between the value of the current flow-related parameter and the value of the second flow-related parameter. Other applications are also described.

Abstract: Apparatus and methods are described including a processor (20) configured to receive first and second sets of extraluminal images of a lumen, the second set being acquired while an endoluminal device is being moved through the lumen. The processor designates at least one of the first set of images as a roadmap image, and designates, within the lumen in the roadmap image, a roadmap pathway (42). A plurality of features (44, 46, 48, 50) that are visible within at least some of the second set of extraluminal images are identified, and an arrangement of three or more of the features within the image are compared to a shape of at least a portion of the roadmap pathway. Based upon the comparing, the identified features are mapped to locations along the roadmap pathway within the roadmap image. Other applications are also described.

Abstract: Apparatus and methods are described for use with a plurality of angiographic image frames of a moving lumen of a subject, including aligning the image frames with each other. Using the aligned image frames, a time it takes a contrast agent to travel a known distance through the lumen is determined. At least partially in response thereto, a characteristic of the lumen is determined, and, in response to the determined characteristic, an output is generated on a display. Other applications are also described.