Description

Image segmentation is an interesting as well as important concept within computer vision and in image processing. It finds applications in medical diagnostics, Robotic vision and surveillance, augmented reality and scene understanding. There are various algorithms for image segmentation in literature. Recently, due to the huge success of deep learning algorithms, researches have exploited the possibility of utilising deep learning models for image segmentation as well. Literature presents two classes of deep learning based segmentation models. This talk presents deep learning models for both semantic and instance segmentation problems. It will discuss these model architectures and design approaches in detail with the help of used cases.