LAPAR: Linearly-Assembled Pixel-Adaptive Regression Network for Single Image Super-Resolution and Beyond
This addresses the efficiency-quality trade-off in image super-resolution for practical applications, though it appears incremental as it builds on existing deep learning approaches.
The paper tackles the challenge of balancing model complexity and quality in single image super-resolution by proposing LAPAR, a linearly-assembled pixel-adaptive regression network that achieves state-of-the-art results on benchmarks while being lightweight and easy to optimize.
Single image super-resolution (SISR) deals with a fundamental problem of upsampling a low-resolution (LR) image to its high-resolution (HR) version. Last few years have witnessed impressive progress propelled by deep learning methods. However, one critical challenge faced by existing methods is to strike a sweet spot of deep model complexity and resulting SISR quality. This paper addresses this pain point by proposing a linearly-assembled pixel-adaptive regression network (LAPAR), which casts the direct LR to HR mapping learning into a linear coefficient regression task over a dictionary of multiple predefined filter bases. Such a parametric representation renders our model highly lightweight and easy to optimize while achieving state-of-the-art results on SISR benchmarks. Moreover, based on the same idea, LAPAR is extended to tackle other restoration tasks, e.g., image denoising and JPEG image deblocking, and again, yields strong performance. The code is available at https://github.com/dvlab-research/Simple-SR.