(Informative)
Filtering of video frames is required in several application. A major example is provided by video compression where the data rate of the input video sequence undergoes the following processing: down-sampled, encoded, transported, decoded, up-sampled, and rendered.
Typical up-sampling technologies used to up-sample a video sequence are Bi-cubic and Lanczos filters. They have a major limitation because they perform purely mathematical operations on the decoded video sequence.
Up-sampling filters resulting from a process of training on a large number of video sequences display a highly non-linear performance and are better capable at reproducing the original-resolution video sequences beyond the simple interpolation of pixels in the sub-sampled video frames.
Technical Specification: AI-Enhanced Video Coding (MPAI-EVC) – Up-sampling Filter for Video applications (EVC-UFV) V1.0 specifies technologies enabling the implementation of an up-sampling filter for video applications based on Super Resolution.
