Wavelet-domain distributed video coding based on continuous-valued syndromes

In current video codecs, the temporal correlation between successive frames of a video sequence is exploited by means of predictive coding. Successful utilization of this paradigm in a wireless and mobile scenario is impaired due to both the computationally-demanding encoder-side motion-compensation procedure and the severe degradation of the decoded video quality when losses of encoder-decoder synchronization occur. The application of distributed source coding principles has been recently proposed as solution to both problems. In this paper, a wavelet-domain video coder is investigated that uses continuous-valued syndromes to efficiently code the coefficients at the encoder. Motion compensation, to form the side-information that allows for reconstruction, takes instead place at the decoder. The proposed video coder, whose performance is similar to other distributed video coders found in literature, does not require any feedback channel and produces a spatial-quality scalable stream.