Recent Advances in Complexity Reduction Methods for VVC Inter coding, A Review

Abstract

Video traffic continues to surge, pushing codecs toward higher efficiency at the cost of sharply increased complexity. The H.266/Versatile Video Coding (VVC) standard roughly halves bitrate relative to HEVC at comparable quality but increases encoder complexity substantially. This paper surveys algorithm-level complexity-reduction methods for VVC inter-frame coding, grouping them by decision module—coding unit (CU) partitioning, inter-mode selection, and motion estimation (ME)—and by approach—heuristic/statistical, machine learning (ML), and deep learning (DL). We provide quantitative comparisons of reported complexity–efficiency trade-offs (encoder time vs. BD-rate), and discuss observed trends and potential areas for improvement. Across the literature, CU partitioning is the most heavily targeted module, followed by ME; reported encoder time savings span approximately 10–55% with typically no more than about 3% BD-rate increase, depending on module and method. It is observed that DL methods generally achieve the largest time savings (often around 50% or higher) by predicting partition structures or mode decisions end-to-end, at the expense of training data and inference cost. Heuristic methods remain lightweight and hardware-friendly with small BD-rate impact, and ML methods provide balanced trade-offs. To the best of our knowledge, this is the first comprehensive survey of VVC inter-frame coding, distilling practical lessons and outlining open directions—especially hybrid pipelines that combine inexpensive filters with learned predictors—to guide more efficient VVC implementations and future standards.