Distributed video codec with complexity scalability based on mixed resolution coding

Recently, new requirements in video coding have emerged. These requirements include bandwidth fluctuation, quality of service and energy constraints. Typically, the encoder has a higher complexity, requiring more energy consumption than the decoder. A new paradigm in video coding, called distributed video coding, enables a reversed complexity coding mode, where the decoder requires more computational effort than the encoder. Distributed video coding is based on the Wyner-Ziv theorem for separately coding but jointly decoding correlated sources. A distributed video codec can be used in power constrained scenarios, like in hand-held mobile devices. We propose a mixed resolution framework based on full resolution reference frames and spatial-reduction-based coding of the non-reference frames. Improved rate-distortion performance is achieved by enabling better side-information generation at the decoder and better rate-allocation at the encoder. The framework enables reduced encoding complexity by low-resolution encoding of the non-reference frames, followed by Wyner-Ziv coding of the residue. The quantized transform coefficients of the residual frame are mapped to cosets without the use of a feedback channel. For that purpose, a study to select optimal coding parameters in the creation of the memoryless cosets is made. Furthermore, a correlation estimation mechanism that guides the parameter choice process is proposed. At the decoder, coset decoding is carried using the sideinformation to obtain a higher quality version of the decoded frame. Results for the coding complexity and rate-distortion performance are presented using the H.264/AVC codec. It is shown that the proposed Wyner-Ziv coding mode is competitive. It allows a scalable complexity reduction and supports a low-decoding-complexity mode. Keywords-Distributed video coding; Wyner-Ziv; H.264; scalable complexity.