Abstract: The present disclosure provides systems and methods for diagnosing disease. In some aspects, an imaging system is provided that includes a light source configured to illuminate a retina of the eye with light, one or more imaging devices configured to receive light returned from the retina to generate one or more spatial-spectral images of the retina, and a computing device configured to receive the one or more spatial-spectral images of the retina, evaluate the one or more spatial-spectral images, and identify one or more biomarkers indicative of a neurogenerative pathology.

Abstract: Embodiments of the present systems and methods may provide techniques that are applicable to a variety of imaging modalities and that utilize prior knowledge of the dynamics of a physiological system to analytically generate augmented samples for machine learning. For example, in an embodiment, a method implemented in a computer comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the method may comprise generating a transform for an image of tissue based on deformation of the tissue under compression, obtaining an image of tissue using an imaging modality, and generating an output image by transforming the image of the tissue using the transform.

Abstract: A method involves performing a mathematical estimation operation identifying a risk score threshold. The operation identifies the risk score threshold as a point on a curve rather than a value of a particular risk score. Such a curve approximates the distribution of risk score values output over a time interval and represents a function embodied by a plot of risk score percentile vs. risk score value. The risk engine, rather than selecting a particular risk score, selects a curve from a family of curves that is known to accurately represent such risk score distributions. For example, the risk engine may choose the curve that provides the best fit to the previous week's risk scores over the family of curves. The risk engine identifies the risk score threshold by finding a risk score value such that the function evaluated at that risk score value produces a specified risk score percentile.

Abstract: Embodiments of the present systems and methods may provide techniques that are applicable to a variety of imaging modalities and that utilize prior knowledge of the dynamics of a physiological system to analytically generate augmented samples for machine learning. For example, in an embodiment, a method implemented in a computer comprising a processor, memory accessible by the processor, and computer program instructions stored in the memory and executable by the processor, the method may comprise generating a transform for an image of tissue based on deformation of the tissue under compression, obtaining an image of tissue using an imaging modality, and generating an output image by transforming the image of the tissue using the transform.

Abstract: A method comprising using at least one hardware processor for automatically receiving, using a user interface associated with the hardware processor(s), an annotation for a medical image of a patient, the medical image associated with a suspected disease of the patient and metadata of the patient. The hardware processor(s) are configured for performing a search query in real time on a database for the annotation. The hardware processor(s) are configured for receiving at least one result of the search query. The hardware processor(s) are configured for retrieving at least one other medical image comprising at least one other annotation, wherein the at least one other medical image is associated with the at least one result. The hardware processor(s) are configured for displaying, on the user interface, the at least one other medical image and the at least one other annotation.

Abstract: There is provided, a method of selecting patients for treatment, comprising: feeding anatomical image(s) of a patient depicting a target tissue, and non-imaging clinical parameters of the patient into neural network component(s) of a model, outputting by the neural network component(s), an intermediate vector storing a plurality of embedding values computed for the anatomical image(s), a plurality of values outputted by a dense layer of the neural network component(s) in response to an input of at least some of the non-imaging clinical parameters, and an intermediate value indicative of likelihood of malignancy for the target tissue, feeding into a classifier component of the model, a feature vector created from the intermediate vector and the plurality of non-imaging clinical parameters, and selecting patients for treatment according to an indication of likelihood of malignancy in the target tissue outputted by the model.

Abstract: In some examples, a system for training a neural network can include a processor to detect a trained neural network application. The processor can also detect a set of images, wherein the neural network application is not trained with the set of images. Additionally, the processor can train an adapter network based on the trained neural network application and the set of images, wherein the adapter network is to be trained by freezing weights of the trained neural network and modifying weights of the adapter network. Furthermore, the processor can use the trained adapter network to process at least one additional image, the processed additional image to be transmitted to the trained neural network to generate an output signal.

Abstract: In some examples, a system for training a neural network can include a processor to detect a trained neural network application. The processor can also detect a set of images, wherein the neural network application is not trained with the set of images. Additionally, the processor can train an adapter network based on the trained neural network application and the set of images, wherein the adapter network is to be trained by freezing weights of the trained neural network and modifying weights of the adapter network. Furthermore, the processor can use the trained adapter network to process at least one additional image, the processed additional image to be transmitted to the trained neural network to generate an output signal.

Abstract: Systems and techniques are disclosed for improvement of machine learning systems based on enhanced training data. An example method includes generating an interactive classification user interface concurrently displaying a first group of medical images and a second group of medical images, each group depicting objects associated with a respective classification. User input indicating movement of medical images from the first group to the second group is detected. The moved medical images are classified according to the second group. The re-classified medical images are provided to a machine learning system, with the machine learning system updating based on analysis of object characteristics of the re-classified medical images to increase accuracies associated with automated assignment of classifications.

Abstract: A method comprising using at least one hardware processor for: receiving a digital medical image and a binary image mask, wherein the binary image mask depicts a segmentation of a lesion in the digital medical image; computing a plurality of layers of the lesion; for each of the plurality of layers of the lesion, extracting layer features; and sending the extracted layer features to a lesion classifier.

Abstract: A method comprising using at least one hardware processor for automatically receiving, using a user interface associated with the hardware processor(s), an annotation for a medical image of a patient, the medical image associated with a suspected disease of the patient and metadata of the patient. The hardware processor(s) are configured for performing a search query in real time on a database for the annotation. The hardware processor(s) are configured for receiving at least one result of the search query. The hardware processor(s) are configured for retrieving at least one other medical image comprising at least one other annotation, wherein the at least one other medical image is associated with the at least one result. The hardware processor(s) are configured for displaying, on the user interface, the at least one other medical image and the at least one other annotation.

Abstract: A method comprising using at least one hardware processor for: receiving a digital medical image and a binary image mask, wherein the binary image mask depicts a segmentation of a lesion in the digital medical image; computing a plurality of layers of the lesion; for each of the plurality of layers of the lesion, extracting layer features; and sending the extracted layer features to a lesion classifier.

Abstract: Systems and techniques are disclosed for improvement of machine learning systems based on enhanced training data. An example method includes generating an interactive classification user interface concurrently displaying a first group of medical images and a second group of medical images, each group depicting objects associated with a respective classification. User input indicating movement of medical images from the first group to the second group is detected. The moved medical images are classified according to the second group. The re-classified medical images are provided to a machine learning system, with the machine learning system updating based on analysis of object characteristics of the re-classified medical images to increase accuracies associated with automated assignment of classifications.

Abstract: Techniques of performing authentication involve comparing current user authentication factors with previous authentication factors selected from multiple users during a single authentication session. Along these lines, suppose that an authentication server receives current browser characteristics from a user computer during a current authentication session. Based on the current browser characteristics, the authentication server selects previous browser characteristics received from devices used by multiple users during previous authentication sessions. For example, the authentication server may select previous browser characteristics based on the whether any of the results of a modified, locally sensitive hashing (LSH) of the previous browser characteristics match any of the results of a modified LSH of the current browser characteristics.