Neural Network based Video Coding: Opportunities & Challenges

Up to now, video coding standards have been developed as a combinatory set of a large number of coding tools. Each tool was primarily based on signal-processing algorithms or experimental optimizations. Recently, a new paradigm called Learned Video Coding (LVC), a.k.a. Neural Network-based Video Coding (NNVC), has emerged. By exploiting the end-to-end training with a large number of video datasets, some representative models already outperform the latest video coding standard, VVC (Versatile Video Coding). In this talk, the LVC technology will be briefly outlined and then the opportunities and remaining challenges will be discussed.
Veranstalter
Vortragende(r)
Prof. Hui Yong Kim
Dept. of Computer Science and Engineering
and Dept. of Artificial Intelligence
Kyung Hee University
Kontakt
Dr. Gerhard Leitner (Gerhard [dot] Leitner [at] aau [dot] at)












