Deep Video Lab is in Wangxuan Institute of Computer Technology, Peking University. We have interest in these research topics: video codec technology, the Internet plus education, deep learning, and data mining.
Research Directions
Video codec technology
Deep Video Lab has been committed to video codec research. The laboratory has developed the HEVC / H.265 codec based on the latest video codec standard HEVC, called Lentoid (Download Here). The Lentoid has the characteristics of higher compression efficiency, faster compression speed, higher concurrency, etc. Lentoid has been widely used in the industry. With the support of the National High-tech R&D Program of China (863 Program), the members of the laboratory have applied for 22 patents, and published numbers articles in journals such as TCSVT, TMM, and conferences such as ICME, ICASSP, etc.
Deep learning and wisdom education
Since 2018, the Deep Video Lab has been committed to combining Internet to promote education development, and studying research topics related to deep learning and intelligent approach for education. Based on the original real education data collected by our lab, we carry out the research about automated assessment, text correction with image, exercise recommendation, student behavior modeling, and so on.
Admissions Information
The laboratory recruits undergraduate interns, masters and doctors, etc. The research direction is about deep learning, data mining and smart education, video coding and decoding. Welcome to contact us by this Email: jsun@pku.edu.cn.
News and Activities
[2021 MM] Congratulations to Xiaopeng Guo for his paper, “Enhancing knowledge tracing via adversarial training” accepted by ACM Multimedia (MM)!
[2021 TCSVT] Congratulations to Zhijie Huang for his paper, “One-for-all: An Efficient Variable Convolution Neural Network for In-loop Filter of VVC” accepted by IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)!!
[Activity] at AXiYa Restaurant
[2021 TIP] Congratulations to Zhijie Huang for his paper, “Adaptive Deep Reinforcement Learning-Based In-Loop Filter for VVC” accepted by IEEE Transactions on Image Processing (TIP)!
More>>>
Partners
