A Survey on Video Coding Optimizations using Machine Learning

The most common type of data used globally is presently video data. The volume of video data has been rising explosively around the globe as a result of the quick development of video applications and the rising demand for higher-quality video services, giving the biggest challenge to multimedia processing, transmission, and storage. Video coding by compression has become somewhat saturated while the compression ratio has grown in the last three decades. Deep Learning algorithms offer new possibilities for improving video coding technologies since they can make data-driven predictions and learn from vast amounts of unstructured data. We explore machine learning-based video encoding optimization in this research, which lays a solid groundwork for further advancements in video coding. The video service's designer must choose a suitable video coding scheme to satisfy criteria like efficiency, complexity, rate distortion, flexibility, etc. This article also presents challenges associated with machine learning video coding optimization. The survey is mainly presented from two key aspects, first is low complexity optimization with the help of advanced learning tools, such as feed-forward CNN, deep RL, and deep NN, and second is learning-based visual quality assessment (VQA).

Keywords:- Video Coding, Deep Learning, Machine Learning, High-Efficiency Video Coding Standard (HEVC), Versatile Video Coding (VVC), Visual Quality Assessment. (VQA).