Image segmentation review: Theoretical background and recent advances

Image segmentation is a significant topic in image refining and automated image analysis with relevance for instance object recognition, diagnostic imaging scanning, mechanized perception, monitoring cameras, satellite imaging, and image compression, and so on. This technology has become an essential component of image assessment as it facilitates the depiction, taxonomy, and conception of the subject matter in the representation. The latest advances in computer vision procedures and the progressive attainability of substantial databases have made it absolutely typical in the computer vision domain. Lately, because of the progression of deep learning techniques, it is observed that a considerable number of tasks are directed at establishing image segmentation strategies operating deep learning models. As an evolving biomedical image refining mechanization, medical image segmentation has computed significant improvements to sustainable health maintenance. Presently it has evolved into a predominant experimentation direction in the domain of computer vision. With the rapid evolution of deep learning, diagnostic image scanning characterized by deep convolutional neural networks has become a research epicentre. This review covers a survey on existing image segmentation approaches into extensive categorization of their algorithms. Additionally, this review outlines the therapeutic and nontherapeutic image databases deployed in the literature for implementing the experimentation. Apart from this, numerous evaluation metrics are discussed for evaluation comparing the results of different segmentation techniques. Further, a detailed discussion on the distinct domains of applications in image segmentation is provided. In conclusion, a discussion on several issues, especially in therapeutic domain and scope in the domain of image segmentation for implementation in the diverse disciplines is provided