Optimization Of Screen Codec In Distance Education Scenario

In distance education,we normally need to broadcast teacher's screen to students' learning terminal.Because of high screen resolution and great complexity of content,we need to compress the screen,before we send it out.Hence,video compression becomes a key component of distance education.We have found,through experiment,compared to natural image,screen image has its own temporal and spatial characteristics.As with spatial,screen image owns a narrower color spectrum,less colors,more dramatic changes at some points,because of sharp edges of fonts and controls.As with temporal,without change of scene,sequence of screen image tends to have less changes compared to adjacent frames,meanwhile these changes tend to emerge within a local area.Compared to natural image,human eyes are more sensitive to image quality of screen image.Under the framework of traditional video codec,the computation of motion estimation and motion compensation takes up most of whole computation of encoding a frame.And we have found out,compared to natural image,screen image has more temporal and spatial redundancy.Based on this discovery,this paper presents an improved motion estimation algorithm.Through caching previous motion estimated macro-blocks,we can speed up the process of motion estimation of current macro-block by utilizing the result of previous motion estimation,replacing the search algorithm by index algorithm.We implement this algorithm based on x264,an open implementation of H.264.Through experiment,we find the average time of compress an image is reduced by15%,without change of image quality.On the basis of new algorithm,the paper presents the architecture and implementation of a teaching client of online education system.The second part of this paper has discussed issues concerning implementation of this teaching client,for example image source,concurrency,etc.