The Research On Error Control Technology In Video Communication

Now Error Control Technology (ECT) had became a key technique which was wildly used in video communication filed. This technique could effectively improve the visual quality of decoded video information output from receiver. Based on analysis the video codec system such as H.26 serial standard proposed by ITU-T, this paper studied on several problems of error control technique in video transmission by improving error correction performance and elevating the efficiency of error control arithmetic. The following works are addressed in this paper:1. Studying on video codec system such as MPEG and H.26 serial standards, especially the technical characteristic of new generation video coding standard - H.264 recommendation. By simulation tests, the affection on coding performance of chief improved arithmetic in H.264 was analyzed.2. Based on analyzing H.263 recommendation, a new scene adaptive video coding scheme which used two layer filter mechanism was proposed. Combined the efficiency of histogram based examination scheme with the veracity of content based scheme, proposed scheme made use of two layer filter mechanism to solve the quality decline in some parts of video frame due to scene changing. In this mechanism, an image similarity examination operator was introduced to reexamine the result of histogram examination scheme in order to get rid of the fake abrupt frame. Simulation has shown that the proposed scheme can achieve more accurately in similarity examination than scheme which proposed by [102]. So the bit number of coded video stream was decreased about 22.95 percent and thus the compression ratio was improved. Although the coding cost was increased about 8.25 percent by using proposed scheme.3. Based on analyzing the affection on coding performance of two operating modes in H.264 NAL layer, such as data partition mode (DPM) and signal slice mode (SSM), an Adaptive Channel UEP scheme(ACUEP) for H.264 video stream was proposed. This scheme firstly determined the operating mode of coded video stream in NAL layer according to the current channel state before they were packed. The coded video stream operated under SSM in error-free channel so as to reduce the packing tradeoff and increase the coding efficiency. While in error-prone channel, they operated under DPM combined with improved unequal error protection (UEP) scheme based on human eyes characteristic in order to elevate their robustness to channel error. In improved UEP scheme, the syntax elements belonged to luminance parts and chrominance parts was reassigned to different data partitions. Those packets containing syntax elements belonged to luminance parts were got error protection at high level so as to guarantee their correctness during transmitting in error-prone network. Simulation has shown that using proposed scheme, the packing tradeoff is decreased 8240 bytes in error-free channel and also the output bits and bit rate of coded video stream are decreased 2.70kbits and 0.33 kbps respectively. The visual quality of decoded video information is improved at some extent and PSNR value of luminance parts is increased nearly 1 dB in error-prone channel.4. An adaptively RD-optimal coding algorithm was proposed to be aimed at the computing redundancy in H.264 recommendation algorithm. A relation between packet loss rate and times of simulating decoding in encoder was setup in proposed algorithm for the first time. Usingthis relation, encoder could adaptively determine the times of simulating decoding according packet loss rate and the number of simulating channel states. So the computing complexity and coding time in encoder was decreased effectively. Simulation has shown that the proposed scheme can effectively reduce the computing complexity and can save up to nearly 60 percent of coding time by decreasing the times of simulating decoding in encoder. Under the same coding time as H.264 algorithm, the average standard deviation of decoding video information estimated distortion can be decreased about 90 percent by...