A New Video Codec Based on Compressed Sensing

In this paper, we propose a new video codec which combines a new compressed sensing theory and some critical ideas of traditional video codecs. Samples needed in the new codec are fewer than in traditional codecs, and the number of samples can be changed according to the encoding mode. The encoder works as a projector, which projects a high dimensional signal to a low dimensional domain, while the decoder works for the solution of underdetermined equations. The decoder is more complex than the encoder, and the process of decoding is no longer the inverse process of encoding. Compared with traditional codecs, the new codec is simpler, but high compression ratio and good quality of reconstructed video are obtained.