Research And Application Of Rate Control Technology For Screen Content Coding

With the rapid development of multimedia and Internet technology,the application of screen content video such as remote online office and remote video education is becoming more and more widespread.The difference between screen content video and traditional video content is that the screen content can include text,charts,graphics directly generated by a computer,and it can also include natural image areas.Therefore,screen content videos usually include large flat areas,limited colors,sharp edges,etc.Due to the unique characteristics of the screen content video make the traditional rate control method ineffective,such as the frequent page turning of the presentation document in the remote video,and the rapid switching of tabs when browsing the web,etc.Due to the small correlation with the coded frame,the abrupt frame is often encoded with intra-frame prediction method,which consumes a lot of bits and cause large fluctuations of the bit rate.Therefore,when encoding screen content video,the encoding performance will drop significantly.In addition,the rate distortion relationship is also very different because screen content video is different from classical natural content video in terms of color range,energy distribution,noise,etc.The efficiency of the two types of video content will be significantly reduced by using the same set of R-λ models for coding.These characteristics of screen content video pose a great challenge to traditional rate control methods.In response to the above problems,this paper has studied the rate control algorithm of screen content video,and the main contributions and innovations are as follows:For the case of online real-time video transmission,this paper proposes an improved rate control algorithm for screen content in low delay scenarios.A pre-analysis mechanism is introduced for the screen content characteristics,which enables the encoder to analyze in advance the types of several frames to be encoded and adjust the rate control and the target buffer level of the buffer by analyzing the frames in advance.The algorithm optimizes the bit control part of the system according to the complexity of the frames,the frame type and the buffer status,and improves the accuracy and efficiency of the rate control without introducing additional delay.With the low delay coding structure,the algorithm proposed in this chapter can reduce the rate control error by an average of 2.08%,obtain a 2.84 dB improvement in reconstructed video quality and the video quality fluctuation is more stable compared to the rate control algorithm in the VVC reference software VTM 10.0.For the case of offline local video storage,this paper presents an improved rate control algorithm for screen content based on content characteristics.According to the content and texture characteristics of the video,the video content is divided into two major categories,natural image and screen content.On this basis,the screen content is further divided into simple screen content and complex screen content,and rate control is performed according to the type of video content to improve rate control efficiency and video quality,and the pre-analyzer mechanism proposed in Chapter 3 is integrated into the algorithm to reduce the effect of abrupt frame.With the random access coding structure,the algorithm proposed in this chapter can reduce the rate control error by 3.13%on average and the BD-rate by 14.33%on average compared with the code rate control algorithm in VVC reference software VTM 10.0.