In this paper, we propose a convolutional neural network (CNN) filter for super-resolution (SR) with the RPR functionality in VVC.
ABSTRACT. High resolution (HR) videos are easy to go beyond the band- width limit during transmission after encoding. Downsam-.
Their filter ranks first among NN-based in-loop filters at the 20th JVET meeting. In the current JVET conference, as a main research direction in the field ...
Sep 11, 2023 · In this paper, we introduce the RPR technique into the SR method, to effectively restore the details lost in the down-sampling and encoding ...
Missing: Functionality | Show results with:Functionality
Ming Li's 9 research works with 15 citations and 212 reads, including: RTNN: A Neural Network-Based In-loop Filter in VVC Using Resblock and Transformer.
People also ask
What is super resolution CNN?
How do I choose a filter value in CNN?
The proposed Efficient Sub-Pixel Convolutional Neural Network is used for the conversion of low-resolution images into a high-resolution format for onward ...
[18] Shimin Huang, Cheolkon Jung, Yang Liu, and Ming Li, “AHG11: CNN Filter for Super-Resolution with RPR functionality in VVC,” JVET-AA0065, Joint Video ...
Herein, we integrate a CNN-based super resolution method which leverages the side information to perceive compression distortion during the encoding or ...
Missing: Functionality | Show results with:Functionality
Abstract—In Versatile Video Coding (VVC), reference picture resampling (RPR) is supported to change resolution without inserting an instantaneous decoder ...
[M1] Shimin Huang, Cheolkon Jung, Yang Liu, and Ming Li, “AHG11: CNN Filter for Super-Resolution with RPR functionality in VVC,” JVET-AA0065, Joint Video ...