Reconstructive video coding system

We propose a video coding paradigm that we call a reconstructive video coding system along with complementary constituent functional units we have developed. The proposed system reduces the resolution of the input image to facilitate successive conventional video coding (core coding) in terms of computation and compression complexities. On the reception side, the decoded image frames are up-sampled without losing sharpness with the aid of super-resolution techniques. It is possible to dramatically improve subjective appearance by avoiding degradations due to rough quantization parameters (QPs) in the core coding. We also introduce a hardware implementation of the system with PCI Express cards targeting household use.