Extreme Learning Machine-Enabled Coding Unit Partitioning Algorithm for Versatile Video Coding

The versatile video coding (VVC) standard offers improved efficiency compared to the high (HEVC) in multimedia signal coding. However, this increased comes at cost of complexity. This work proposes an efficient unit partitioning algorithm based on extreme learning machine (ELM), which can reduce complexity while ensuring efficiency. Firstly, size decision is modeled as a classification problem. Secondly, ELM classifier trained predict size. In experiment, proposed approach verified VVC reference model. results show that method significantly, and good image quality be obtained.