Single-image super-resolution using clustering-based global regression and propagation filtering

In this paper, we present a novel single-image superresolution (SR) algorithm that utilizes clustering-based global regression to generate desired high-resolution (HR) patch with its low-resolution (LR) counterpart. Propagation filtering can achieve smoothing over image while preserving image context like edges or textural regions. Furthermore, to preserve the edge structures of super-resolved image and suppress artifacts, a propagation filtering-based constraint is introduced into the SR reconstruction framework. Experimental comparison with state-of-the-art single-image SR algorithms validates the effectiveness of proposed approach.