Joint Source Channel Coding For Image Transmission With JPEG2000

With the rapid development of communications, multimedia and computer technology, the image transmission has become a key component of multimedia communications. Traditional communication system is based on Shannon coding theory, in other words, source coding and channel coding are considered separately. This method often fails to achieve efficient and reliable communication, so Joint Source Channel Coding (JSCC) has become a hot field. JSCC system takes into account of source coding and channel coding, so this method can effectively reduce the distortion of the entire communication system. In the end, we can get better quality of reconstructed image.First, a JSCC scheme was proposed in this paper which was based on the improvement of LDPC codes for unequal error protection (UEP). Different parts of the stream have different effect on the quality of image reconstruction, so we can use different channel coding to protect the important stream. A novel irregular LDPC with weight increasing parity check matrix (WIPC-LDPC) was first put forward, which can simply realize UEP characteristic. Then the different significant JPEG2000code streams were coded by WIPC-LDPC with dissimilar degree bit nodes to realize UEP scheme. Simulation results show that the proposed JSCC scheme can improve the Peak Signal to Noise Ratio (PSNR) and improve the subjective visual effect of reconstructed image, compared with the other two schemes mentioned in the literature.Second, a joint source channel coding based on end-to-end rate distortion function was performed to optimize the Source-channel bit rate allocation. Firstly, channel signal to noise ratio (SNR) was estimated in real-time based on estimated subspace SNR, thus channel error rate was obtained, so as to solve the adaptive problem of channel. An end-to-end rate distortion function was established considering the distortion caused by quantization and channel noise. Secondly, an objective function was established based on taking full advantage of the characteristics of the JPEG2000code steam and adaptive source-channel bit rate to optimize the allocation was achieved by low complexity of computing. Simulation results show that JSCC based on end-to-end rate distortion function can adapt to changes better in wireless channel. Besides the JSCC scheme can achieve better images than the fixed rate schemes.