A Survey on Video Compression using Deep Neural Networks
Rishab J Shivalli, Shrinkhla Sinha, Siddharth Subramanian, Vijayalakshmi Suresh, P. Dayananda · CiiT international journal of networking and communication engineering · 2021
The rise in video content shared over the internet demands the need for state-of-the-art video compression. A quality compression can help enable the efficient transmission of video data over limited bandwidth. Particularly during the COVID-19 pandemic, we saw that the increasing internet traffic used for video conferencing, gaming, and online learning forced Netflix, YouTube, and other streaming providers to limit their video quality in Europe and other countries as well. Standard video compression algorithms represent them as a sequence of reference frames following residual frames, and these methods lack versatility. The advent and recent advances in deep learning can solve such issues. This paper summarizes some of the Deep Video Codec (DVC) frameworks and their characteristics.