Low Complexity Image And Video Coding At Low Bit-rate

As the core technology of multimedia data storage and communication, digital image and video compression has been a research hotspot in the field of digital media processing. With the development of portable equipment and mobile terminal, applications of image and video coding under low bit-rate and complexity become more and more popular. From sea video surveillance, unmanned plane detection and wild information collection to mobile video, on-board television and videophone, all these applications will encode and transmit image and video under low bit-rate because of the limitation of environment which accesses to unstable and limited bandwidth. At the same time, the low process capability and power limitation call for the low complexity of coding algorithm. As a result, this paper investigated the image and video coding scheme under low bit-rate and complexity.In image coding, wavelet transform and zerotree coding are widely used in coding standard such as JPEG/JPEG2000. However, implementation complexity of wavelet transform and long coding process of zerotree structure may limit the application in some cases. In this paper, an idea of multilevel DCT-based zerotree image coding is proposed, in which Wavelet Transform(WT) is replaced by Discrete Cosine Transform(DCT) to make transform easy to implement. A conventional zerotree with a parent-child relation of 2×2 offspring is expanded to the one of 4×4 offspring to reduce the complexity of zerotree coding process. Simulation results show the proposed multi-level DCT based zerotree image coding algorithm can significantly reduce the complexity of coding process without sacrificing image subjective quality.In video coding, High Efficiency Video Coding(HEVC) improves encoding performance a lot compared with the previous standards. However HEVC is designed for high definition video and the high compression rate calls for high complexity, which limiting the application in some low bandwidth and low complexity conditions. In this paper, an idea of wavelet transform-based downsampling scheme for intra mode is used before HEVC coding to ensure the bandwidth for low frequency content of video in low bit-rate condition. Furthermore the downsampled size of video can reduce the complexity of HEVC encoding procedure. To make the performance better, we also allocate bandwidth for high frequency content with bit-plane encoding and frame partition to encode the high frequency coefficients efficiently. Simulation results show proposed wavelet transform-based downsampling scheme reduces about 70% complexity with similar coding performance as HEVC...