Towards Compact Single Image Super-Resolution via Contrastive Self-distillation
arXiv:2105.11683
Abstract
Convolutional neural networks (CNNs) are highly successful for super-resolution (SR) but often require sophisticated architectures with heavy memory cost and computational overhead, significantly restricts their practical deployments on resource-limited devices. In this paper, we proposed a novel contrastive self-distillation (CSD) framework to simultaneously compress and accelerate various off-the-shelf SR models. In particular, a channel-splitting super-resolution network can first be constructed from a target teacher network as a compact student network. Then, we propose a novel contrastive loss to improve the quality of SR images and PSNR/SSIM via explicit knowledge transfer. Extensive experiments demonstrate that the proposed CSD scheme effectively compresses and accelerates several standard SR models such as EDSR, RCAN and CARN. Code is available at https://github.com/Booooooooooo/CSD.
arXiv commentAccepted by IJCAI-21
References in corpus (8)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Distilling the Knowledge in a Neural Network
- A Simple Framework for Contrastive Learning of Visual Representations
- ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks
- PACT: Parameterized Clipping Activation for Quantized Neural Networks
- DRFN: Deep Recurrent Fusion Network for Single-Image Super-Resolution with Large Factors
- GhostSR: Learning Ghost Features for Efficient Image Super-Resolution
- PAMS: Quantized Super-Resolution via Parameterized Max Scale