The present disclosure relates to computer-implemented systems and methods for training and using generative adversarial networks to detect abnormalities in images of a human organ. In one implementation, a method is provided for training a neural network system, the method may include applying a perception branch of an object detection network to frames of a first subset of a plurality of videos to produce a first plurality of detections of abnormalities. Further, the method may include using the first plurality of detections and frames from a second subset of the plurality of videos to train a generator network to generate a plurality of artificial representations of polyps, and training an adversarial branch of the discriminator network to differentiate between artificial representations of the abnormalities and true representations of abnormalities. Additionally, the method may include retraining the perception branch based on difference indicators between the artificial representations of abnormalities and true representations of abnormalities included in frames of the second subset of plurality of videos and a second plurality of detections.

Computer systems and computer-implemented methods for performing classification, detection, and/or prediction based on processing of ocular images obtained from various imaging modalities are disclosed. Use of delayed near-infrared analysis (DNIRA) as one of the imaging modality is also disclosed.

A method for processing breast tissue image data includes processing the image data to generate a set of image slices collectively depicting the patient's breast; for each image slice, applying one or more filters associated with a plurality of multi-level feature modules, each configured to represent and recognize an assigned characteristic or feature of a high-dimensional object; generating at each multi-level feature module a feature map depicting regions of the image slice having the assigned feature; combining the feature maps generated from the plurality of multi-level feature modules into a combined image object map indicating a probability that the high-dimensional object is present at a particular location of the image slice; and creating a 2D synthesized image identifying one or more high-dimensional objects based at least in part on object maps generated for a plurality of image slices.

Disclosed is a method for analyzing a lesion based on a medical image, which is performed by a computing device. The method may include: obtaining positional information of a suspicious nodule which exists in the medical image; generating a mask for the suspicious nodule based on a patch of the medical image corresponding to the positional information; and determining a class for a state of the suspicious nodule based on the patch of the medical image and the mask for the suspicious nodule.

Disclosed is a system and a method for adapting a report of nodules in computed tomography (CT) scan image. A CT scan image may be resampled into a plurality of slices. A plurality of region of interests may be identified on each slice using an image processing technique. Subsequently, a plurality of nodules may be detected in each region of interest using the deep learning. Further, a plurality of characteristics associated with each nodule may be identified. The plurality of nodules may be classified into AI-confirmed nodules and AI-probable nodules based on a malignancy score. Further, feedback associated with the AI-confirmed nodules and the AI-probable may be received form a radiologist. Furthermore, data may be adapted based on the feedback. Finally, a report comprising adapted data may be generated.

A deep learning-based examination and diagnosis assistance system and apparatus for early digestive tract cancer comprising a feature extraction network, an image classification model, an endoscope classifier, and an early cancer recognition model. The feature extraction network is used for performing initial feature extraction on endoscope images based on a neural network model; the image classification model is used for performing extraction on the initial features to acquire image classification features; the endoscope classifier is used for performing feature extraction on the initial features to acquire endoscope classification features and classify gastroscope/colonoscope images; the early cancer recognition model is used for splicing the initial features, the endoscope classification features, and the image classification features to acquire the probability of early cancer lesions in white light images, electronic dye images or chemical dye images of a corresponding site or acquire a flushing prompt or position recognition prompt for the corresponding site.

An object detection model training method includes: inputting an unannotated first sample image into an initial detection model of a current round, and outputting a first prediction result for a target object, transforming the first sample image and a first prediction position region within the first prediction result to obtain a second sample image and a prediction transformation result in the second sample image; inputting the second sample image into the initial detection model, and outputting a second prediction result for the target object; obtaining a loss value of unsupervised learning according to a difference between the second prediction result and the prediction transformation result; and adjusting model parameters of the initial detection model according to the loss value and returning to the operation of inputting a first sample image into an initial detection model of a current round to perform iterative training, to obtain an object detection model.

An AI-based object classification method and apparatus, a computer-readable storage medium, and a computer device. The method includes: obtaining a target image to be processed, the target image including a target detection object; separating a target detection object image of the target detection object from the target image; inputting the target detection object image into a feature object prediction model to obtain a feature object segmentation image of a feature object in the target detection object image; obtaining quantitative feature information of the target detection object according to the target detection object image and the feature object segmentation image; and classifying the target detection object image according to the quantitative feature information to obtain category information of the target detection object in the target image.

Computer-assisted methods for assisting in anomaly identification relating to a dysmorphology analysis comprise retrieving and/or generating a sequence of progressively changing images that depict morphing of at least a subject's first physical feature, between at least a first state representing an identification of a first anomaly relating to a dysmorphology and a second state representing lack of the identification of the first anomaly. The number of images in the sequence is selected to provide a substantially continuous transition between the first and second states. The retrieved and/or generated images are displayed on a display of a processing device, and a user interface on the processing device is provided for a user to make an image selection, which can then be used to determine identification or lack of identification of the anomaly.

The present system provides methods and systems of detecting lung abnormalities in chest x-ray images using at least two neural networks.