Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks

TL;DR

This paper introduces a deep Laplacian Pyramid network for fast, high-quality image super-resolution that reduces computational load and parameters by directly extracting features and using recursive layers.

Contribution

The proposed network reconstructs images progressively at multiple pyramid levels, directly from low-resolution inputs, with parameter sharing and deep supervision for improved efficiency and accuracy.

Findings

Outperforms state-of-the-art in speed and quality

Uses fewer parameters due to recursive layers

Achieves high-quality results on benchmark datasets

Abstract

Convolutional neural networks have recently demonstrated high-quality reconstruction for single image super-resolution. However, existing methods often require a large number of network parameters and entail heavy computational loads at runtime for generating high-accuracy super-resolution results. In this paper, we propose the deep Laplacian Pyramid Super-Resolution Network for fast and accurate image super-resolution. The proposed network progressively reconstructs the sub-band residuals of high-resolution images at multiple pyramid levels. In contrast to existing methods that involve the bicubic interpolation for pre-processing (which results in large feature maps), the proposed method directly extracts features from the low-resolution input space and thereby entails low computational loads. We train the proposed network with deep supervision using the robust Charbonnier loss functions and achieve high-quality image reconstruction. Furthermore, we utilize the recursive layers to share parameters across as well as within pyramid levels, and thus drastically reduce the number of parameters. Extensive quantitative and qualitative evaluations on benchmark datasets show that the proposed algorithm performs favorably against the state-of-the-art methods in terms of run-time and image quality.