YONET: A Unified Model for Simultaneous Identification and Segmentation of Structures in Medical Images

The proposed novel model, YONET, performs the identification and segmentation of medical images collaboratively. Many types of image-processing software are being developed to extract meaningful information from medical photos to enhance patient diagnosis. Simultaneous segmentation and object detection is a computer vision task that combines object detection and image segmentation, two related functions usually done separately. It involves detecting objects in an image and generating a segmentation mask for each object. This assortment of tasks can enhance the precision and velocity of object detection by providing more accurate object boundaries and, in some cases, eliminating the need for post-processing steps. The image segmentation module in YONET handles intermediate abstract representations and utilizes them as input for object detection. YONET will be trained on bounding boxes that delineate the detected objects and pixel-wise segmentation information. The resultant system is optimized for segmenting an optionally distinct class of structures and detecting a class of objects.