Various Approaches in Deep Learning for Medical Modality Image Segmentation

Abstract

In computer vision and image processing the surveillance of Video, Virtual reality, robotic perception, and image compression are just a few of the uses for image segmentation. There are a plethora of segmentation algorithms out there right now. These new approaches to picture segmentation have been spurred on by the extensive victory of Deep Learning (DL). Semantic and instance segmentation literature is thoroughly reviewed, with particular attention paid to groundbreaking work in CNN models, architecture of encoder-decoder, recurrent networks, attention models, and GAN settings. This work is a comprehensive review of the most recent research in these fields. Using a wide range of datasets, we assess the performance of different deep learning-based segmentation representation, examine their strengths and drawbacks, and identify promising future avenues