Research On Operation Command Driven Mobile Video Processing System

With the rapid development of mobile Internet in recent years,mobile video processing applications are concerned by more and more people.However,mobile devices have limitations on computation capability and storage capability as well as having poor cruising ability which make it difficult to effectively support compute-intensive services of video processing.Thus leads to a lot of performance issues,such as lower video processing efficiency,longer processing time,lower scalability and practicality.While the requirements of mobile users on more complicated mobile video processing operations and high processing performance are continually increasing,a research on mobile video processing system with higher performance is of great significance.According to the limitation of mobile devices and the factor of performance issues,the paper defines common operation set as the unity of client and server while editing and processing and presents the research of operation command driven mobile video processing system based on the solution of transferring native compute-intensive service into network transferring and video processing on server.The main contribution of this paper is as follow:(1)Propose a framework of mobile video processing system in cloud environment.According to the limitations brought by native mobile video processing method's computation overhead and energy consumption,this paper proposes the framework of mobile video processing by migrating video processing to cloud.The mobile terminal is in charge of video operation collecting,operation recording and video playing which reduces the workload of the processor of mobile device.(2)Propose a mobile video processing method based on common operation set.Based on the consistency issue in the cloud migration solution,the method defines a common operation set to be the unity of video editing on client and video processing on server which guarantees the consistency.The client captures and records the operations based on the common operation set.Then the client transfers and synchronizes the operation set to server based on the pre-operation feedback way.Next step is the parsing of the operation set on server and proceeding the video processing phrase according to the command set after operation parsing phrase.The last thing is providing the processed video to the client using adaptive encoding under different network conditions.(3)Propose the multi-granularity parallel performance optimization solution.Considering the case of large scale video processing,the paper optimizes the performance of operation parsing phrase,video processing phrase and video posting back phrase.For operation parsing phrase,adopt FastJson technique to optimize the performance.For video processing phrase,improve the processing efficiency with a multi-granularity multi-thread parallel pattern.A buffer feedback mechanism is proposed which utilizes the character of buffer to optimize the performance in video posting back phrase.(4)Design and implement the mobile video processing system in cloud environment and finish the test and analyzing work about the prototype system.This paper utilizes the advantage of cloud environment and proposes the video editing method based on common operation set to reduce the computational overhead and energy consumption thus to implement more mobile video editing functions and improve the video processing efficiency on mobile devices.The definition of common operation set has established the foundation of video editing and processing on different mobile devices which improves the scalability of system.The multi-granularity optimization solution has made positive contribution to the performance of video processing under large-scale situation.