A subpixel-based image downsamping using learned 2D FIR filters

This paper proposes a subpixel-based image down-sampling algorithm using content-adaptive two-dimensional (2D) FIR filters. The proposed algorithm consists of a learning stage and an inference stage. In the learning stage, using a sufficient number of low-resolution (LR) and high-resolution (HR) patch pairs, we compute optimal 2D FIR filters to synthesize LR patches of the highest quality from a specific HR patch and store the patch-adaptive 2D FIR filters. In the inference stage, we explore candidates that best match to each HR input patch and synthesize LR patches by using their corresponding 2D FIR filters on a subpixel basis. The experimental results show that the proposed algorithm produces higher-quality LR images on a patch basis than existing methods and entails no blur and aliasing artifacts.