Video Super Resolution Using Non-Local Means with Adaptive Decaying Factor and Searching Window

Power consumption, transmission bandwidth, and spatial-temporal resolution tradeoff are among the most important factors that affect video processing on mobile devices. Scalable video coding (SVC) provides a possible solution to overcome these problems. Every video in SVC format consists of high-resolution (HR) frames and low-resolution (LR) frames. For a better viewing experience, super resolution (SR) is introduced to refine the LR frames. Non-local means (NLM) based SR has a promising prospect. Since NLM originates from image denoising, the fixed parameters of NLM is not fit for SR tasks. A fixed decaying factor tends to blur the details in flat regions and a fixed searching window results in mismatches among pixels. Thus, we propose two adaptive parameters to address these problems. We generalize key features that affect SR methods' applicability of implementation on hardware and show NLM is fit for hardware implementation. The experimental results validate the proposed algorithm.