A Review of Deep Learning Models for Computer Vision

Abstract

Computer Vision has given a way for computers to see by interpreting the surrounding objects. Deep Learning is often used while training neural networks with image data. Many different models in Deep Learning are used to perform various tasks like classification and segmentation. To solve such tasks to give a better accuracy, the size and depth of modern deep learning models have been increasing. Thus transfer learning holds paramount importance as pre-trained weights can be used for further training and avoid expensive computation. Therefore, this can save a lot of expensive computing power. This paper aims to review such transfer learning models and compare their performances.