The entry “The cloud service platform of face super-resolution enhancement based on mobile terminal” is awarded the first prize of “The first of national graduate mobile terminal application design contest”
“Maimang Cup” of the first of national graduate mobile terminal application design contest was hold in Beijing University of Posts and Telecommunications from 2015 August 7th to August 9th. The competition was hold by The Ministry of Education Degree and Postgraduate Education Development Center and Youth Science and Technology Center of China Association for Science and Technology, jointly organized by The National Engineering Professional Degree Graduate Education Steering Committee, undertaken by Beijing University of Posts and Telecommunications.
The “Maimang Cup” is not only the one of topic events of “The national graduate student innovation practice series of activities” but also the one of the highest level of the graduate student five events of the ministry of education. Competition around the current situation and development trend of mobile terminals and mobile Internet, Aiming to improve the creative and practical ability of students ,cultivating innovative talents for China's mobile Internet industry, enhance training quality of graduate education.617 teams which come from 123 units of graduate student training took part in this competition. After the online trial, The final selection of 150 works from 60 colleges and universities entered the finals.
The students Zhong Chuan, Liu Peiwen, Liu Bo, Ding Xin who come from National Engineering Research Center for Multimedia Software took part in this competition with the work “The Cloud service platform of face super-resolution enhancement based on mobile terminal”. With the strong competition of works show and oral examination, our team won the first prize of the national creative category in the preliminary. And won the first prize of national creative category in finals.
In the finals six teams won the first prize of creative category .And two of them come from National Engineering Research Center for Multimedia Software which account for a third of a winning team.
In recent years, China has invested 320 billion in building the world’s largest city security monitoring network, more than 20 million video camera has been put in use. Video monitoring plays a more and more important role in the maintenance of social security. Face is the most important key in solving crimes, but because the distance between the camera and the suspect is usually too far, the resolution of the camera image is too low. So that, the face is very difficult to distinguish. Super-resolution(SR) in based on using low-resolution image to rebuild clear high resolution image which can restore facial feature detail efficiently to increase the accuracy of face recognition in order to improve the detection rate of the police.
However, actually using the SR technology is faced with these challenges 1)The technology is difficult for criminal investigators to master and The response of scene image processing is so slow that it’s very easy to miss opportunities of solving crimes. 2)Although there are many cameras, most of them aren’t connected. And a large number of video didn't get effective use, the collection and transmission efficiency is very low. 3)The original Super-resolution Technology do very poor in Human face image enhancement which in actual noise, so that it is difficult to identify faces image.
To solve the problem above, the entry references the thinking of “Internet +”,proposing the cloud service platform of face super-resolution enhancement based on mobile terminal by joining the mobile internet and the traditional security image processing. Take the advantage of mobile internet, criminal investigators can upload the image of criminal place though the mobile terminals anytime and anywhere, and the professional and technical personnel of the cloud service platform can process instant and reply the results of super-resolution enhancement immediately, increasing the speed of response dramatically. In order to avoid high costs of Image processing apparatus, the platform is paid by service.
The entry which referenced by National Engineering Research Center for Multimedia Software uses the robustness of face super-resolution algorithm based on the local constraints to promote the quality of face rebuild in noisy environment significantly and improve distinguishing degree. Our entry has four advantages “wider coverage, quicker response, lower cost and better efficient”, significantly improve the processing efficiency and the identification value of surveillance images, increasing the rate of solving crimes. The applications of this entry also can be extended to the field of digital entertainment, and can provide mobile users with life photo enhancement and other diversified image processing services in order to build a more beautiful digital life.