Rate Control Algorithm Based On Subjective Video Quality

In the modem age, with the rapid development of multimedia information teclinology and network communication teclinology, digital video technology, as an important branch of the field of information processing, has being broadened in many application filed. Video coding has become one of the basic core technology of information industry, such as digital TV, internet TV, mobile multimedia and so on. Because the uncompressed video source often associated with mass generated data, so the compressed video image quality, coding efficiency and real-time performance of video coding teclinology are key factors to large-scale application in multimedia communication fields. At present, H.264/AVC is a new video coding standard proposed by ISO/IEC and ITU/T, and it has been widely used in multimedia industry for its more high efficiency and stable network adaptability. Rate control technology is an important part of video coding standards, which will directly affect the quality of video image.JVT-G012is currently the mainstream of the H.264/AVC rate control algorithm, although it can solve the problem of "laying hens paradox", there are a lot of shortcomings, for example, it doesn’t according to the human eye’s subjective feeling to control bit rate, According to the shortcomings of JVT-G012’frame layer and basic unit layer rate control algorithm, this paper presents a corresponding improvement algoritlim as below.1、SSIM and MAD MB Layer Rate Control Algorithm for H.264:In order to control the rate of Macro block better, this paper propose a kind of algorithm which adopts the weighted value of MAD and SSIM instead of a single value by MAD to code-rate allocation. The experimental result shows that, this algorithm reflects human’s subjective feeling better, increased the video continuity, and the SSIM of the video picture is improved.2n Frame Layer Perceptual Rate Control Based on Distortion of Structural Similarity for H.264:The H.264rate control algoritlim allocates frame layer bit rate according to the objective frame complexity and fails to gain an optimal subjective quality. Therefore, a frame layer rate control algorithm based on distortion of structural similarity is put forward in this paper. The algorithm defines the distortion ratio of structural similarity of a frame for target bits allocation. The experimental results find better subjective quality of the proposed algorithm from its accurate rate control and structural similarity improvement of video output sequence.3> SSIM Prediction Model for H.264and Application to Frame Layer Rate Control:In order to make the H.264rate control algorithm to get better subjective video quality, this paper proposes a linear prediction model based on structural similarity. This algorithm works out structural similarity of the current frame through the linear prediction model, and presents the structure similarity factor to represent the frame subjective complexity and then apply it to the current frame bit allocation. The experimental results demonstrate that the proposed algorithm improves the structure similarity of the video sequence effectively, thus gaining better subjective video quality.