Fast Depth Decision Algorithm with Spatial Homogeneity and Threshold Value Modified on Versatile Video Coding (VVC) Partitioning

The VVC standard has achieved great performance improvements and has outperformed its predecessor HEVC (High Efficiency Video Coding) with more than 30% compression rate by VVC with no significant loss in quality. The VVC standard has a very broad use, which can be used for camera capture results up to immersive applications with HDR (High dynamic range) and SDR (Standard Dynamic Range) quality. However, with an increase in VVC’s excellent compression capability, it comes with an increase in the encoding process complexity, especially in the partitioning process because VVC implements a block partition method that is far more flexible than its predecessor HEVC, namely the Quadtree method with multi-type tree structures (QT+MTT) so it requires a lot of work. In this research, we propose a fast depth decision algorithm with spatial homogeneity and threshold value modified method to save the amount of time and overall process complexity using uniform areas and decision-making of CU (coding unit) split type. The VVC reference program VTM 14.2 implements the suggested fast-depth decision algorithm and tests it under common test conditions (CTC) to determine how effective it is. Fourteen sequences and the All-Intra primary configuration are used in the simulation. The average encoding time savings (T_saving) are evaluated and obtained 36.70% with insignificant quality degradation Bjontegaard delta bit rate (BDBR) of 1.37%.