The Design And Implementation Of Content-based Transmission Control Model Towards Media Stream

In the age of big data, the data storage, transmission, retrieval and analysis processing capacity need to become more enormous along with the sharp growth of the data size of information. In terms of network transmission, the applications such as video conferencing, video surveillance, etc. make multimedia transmission be widely used. To transport the media streams of those applications require more and more bandwidth, the bottleneck of network bandwidth will have a serious impact on the streams’ integrity and instantaneity.For decreasing the transmission bandwidth, a content-based transmission control model has been designed and implemented. The model consists of audio stream transmission control and video stream transmission control module. In the audio stream transmission control module, the packets of audio streams will be multiplexed. Because of the image information redundancy existed in video streams and in consideration of the H.264code’s characteristics, the video stream transmission control module is designed to eliminate redundant data from the streams. The video stream transmission control module consist of four sub-modules, content extraction, content feature contrast, redundant content judgment, transmission control and synchronization sub-module. In content extraction sub-module, the key frames are extracted to represent the video content. Then by contrasting the DCT coefficients of two adjacent key frames, the similarity between these two frames is obtained in content feature contrast sub-module. According to the obtained similarity, redundant content judgment sub-module looks out the redundant frames between the two adjacent key frames. Finally in transmission control and synchronization sub-module, a packet contained the number of discarded frames is sent to client. When the decoder detects the packet, the cached image should be play repeat according to the analyzed number to implement the synchronization.The experimental results proved the model is effective to save transmission bandwidth towards the video streams consist of static or slowly changing scenes and eliminate redundant data from the streams. And the video could be retrieved to satisfy the play time and synchronization in decoder. So this model can be widely used in video conferencing, video surveillance, distance education, etc.