Recent Trends in Computer Vision Using Large Data Models

Vision dominates the sensory capabilities of the human brain. In fact, most of the perception of the world is by vision. It is a distinction like no other, carried by most of life. Time has transferred some human abilities like vision to machines. Now, there are many vision based tasks that machines do better than humans. Today computer vision stands at a position where it can challenge humans. Computer vision encompasses multiple domains like healthcare, robotics, autonomous driving, defence, gaming and others. Large data models like those used in deep learning extensively use data resources in order to train for subsequent perfromance. Some examples of these models, which are vital for modern day computer vision tasks, are Convolutional neural networks (CNNs), Vision based transformers (ViTs), etc.

Research in the area of computer vision is rich with ever-increasing time and resources being poured into harnessing the powerful and fast computational resources offered by today's computers. Understanding computer vision from the author's perspective may prove highly beneficial to the interested readers. This book is meant for anyone interested in computer vision. The book is divided into five chapters each bringing with it unique information related to the state of the art in computer vision. The book comprises of recent developments in the area of computer vision in the form of popular reviews and high impact research articles. The works are part of the author's research and have been published in top scientific journals in Springer Nature like Pattern Analysis and Applications (PAAA) and Multimedia Tools and Applications (MTAP). These works have been selected here carefully to cater to the reader's interest. Since the book includes state of the art reviews, anyone desiring to get the desired knowledge about computer vision, will find the book delightful. The book also features original research articles in state of the art computer vision areas like deep learning based ensembles, and image to three dimensional model conversion.