M1-V Deep Learning for Future Video Compression

PI: Zhu Li, Sejun Song

The goal of this project is to develop new Deep Learning based high dimensional signal models and prediction tools for immersive visual signal coding. Shade new insight on deep learning’s potential in signal modeling and compression (vs. classification, enhancement).Achieve significant complexity reduction with deep learning tools to beat info-theoretical limits on R-D performance. Deliver a practical next gen immersive visual content coding tool that achieve significant coding efficiency gains with 10x reduction in complexity.