Era of deep neural networks: A review

Abstract

Deep learning has achieved remarkable success in various machine learning and computer vision applications. The learning allows multiple processing layers to learn features by themselves opposite to conventional machine learning approaches which were not able to process the data in their natural form. Deep convolution networks have shown great performance in processing images and videos, whereas recurrent nets have shown great success for sequential data. This paper reviews all the aspects and researches done till now in this area along with their future possibilities.