NATIONAL ENGINEERING LABORATORY FOR VIDEO TECHNOLOGY
The National Engineering Laboratory for Digital Video Codec (NELVT) was approved by the National Development and Reform Commission of the People's Republic of China in February 2009 and passed the acceptance in 2013. Under the leadership of the Ministry of Education, the laboratory is operated based on Peking University, and cooperating for research development and integrated application with partners including Huawei Technologies Co., Ltd., National Radio and Television Administration, and the Institute of Computing Technology of the Chinese Academy of Sciences.
The laboratory is the leading unit for the Audio Video Coding Standard (AVS). In 2002, the AVS working group was established, with hundreds partners from academia and industry, the AVS with independent national intellectual property rights was established and became an IEEE international standard. In March 2014, AVS is selected as the only requirement for digital TV terminals. By the end of 2015, 14 HD channels of CCTV and more than two-thirds of provincial satellite TVs nationwide have adopted AVS+ for broadcasting. AVS not only saved tens of billions of patent fees for China, but also effectively promoted the digital video of China from “big to strong” and became a major example of the successful implementation of China's intellectual property strategy and standard strategy. In 2016, the AVS2, an ultra-high-definition video coding standard developed for over five years, was successively promulgated as the national broadcasting industry standard and national standard. The test showed that the AVS2 coding efficiency is higher than the latest international standard HEVC/H.265. AVS2 can double the coding efficiency of surveillance video to HEVC, which has great application value for smart cities.
Besides video codec as a key research direction, NELVT also has strong research strength in image recognition, video analysis, computer vision, and bionic vision. From 2009 to 2012, the laboratory was among the best in the surveillance video event recognition competition organized by the American Institute of Standards and Technology for three consecutive years. In 2015, the laboratory contributed more than half of the technology (based on the number of patents and standard pages) for the international standard “Compact Descriptors for Visual Search” ISO/IEC 15938-13 (CDVS), and was adopted by Baidu, Tencent, etc. in large-scale intelligent video applications.