Image Segmentation Using Deep Learning: A Survey

This paper provides a comprehensive review of deep-learning approaches to image segmentation, motivated by the broad success of deep learning and the importance of segmentation across scene understanding, medical image analysis, robotic perception, video surveillance, augmented reality, and image compression. It surveys the spectrum of pioneering efforts in semantic and instance segmentation, spanning convolutional pixel-labeling networks, encoder-decoder architectures, multiscale and pyramid-based approaches, recurrent networks, visual attention models, and generative models in adversarial settings.

Beyond cataloguing methods, the survey investigates the relationships, strengths, and challenges among these models, examines the datasets that are widely used to train and evaluate them, and compares their reported performances. It mattered as a consolidated reference that maps the fast-moving field, helps researchers situate their work, and identifies promising directions for future segmentation research.