Coding efficiency improvement on distributed video coding utilizing super resolution approach

Distributed Video Coding (DVC) has recently been attracting attention because it encodes frames independently without taking heavy computation at the inter-frame prediction coding. This paper proposes to introduce super resolution technique in conventional wavelet-based DVC utilizing JPEG2000. New key frame determination algorithm exploiting block segmentation with respect to translational motion feature is introduced at encoder. Down-sampled Wyner-Ziv frames are reconstructed by means of super resolution technique in the DVC context. The proposed method is compared with other existing DVC methods through experiments and it demonstrates efficiency in lower bit rate compression cases.