Exploiting Quadtree Template Matching for Enhanced Hybrid Video Coding
摘要
In this study, we propose the Quadtree Template Matching (QTM) technique, a sophisticated hierarchical evolution from conventional template matching approaches, crafted to optimize coding efficiency. Seamlessly blending with established research, our method amplifies traditional template matching tenets, leading to superior compression outcomes without compromising image fidelity. At the heart of our approach lies the integration of QTM with the merge mode and advanced motion vector predictions, fundamental tools in today’s video coding technologies. This synergy is realized through the crafting of two unique representative vector designs. Our comprehensive analysis delves into the compression performance and intricacy of our method, further examining QTM’s interoperability with algorithms like low-rank matrix completion and overlapped block motion compensation. Crucially, our simulations offer compelling evidence of our approach’s effectiveness. The empirical data consistently affirms the marked improvements in compression and efficiency brought about by our proposed method.