Feature-Domain Adaptive Contrastive Distillation for Efficient Single Image Super-Resolution
Open Access
- 23 November 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 131885-131896
Abstract
Convolutional neural network-based single image super-resolution (SISR) involves numerous parameters and high computational expenses to ensure improved performance, limiting its applicability in resource-constrained devices such as mobile phones. Knowledge distillation (KD), which transfers useful knowledge from a teacher network to a student network, has been investigated as a method to make networks more efficient in terms of performance. To this end, feature distillation (FD) has been utilized in KD to minimize the Euclidean distance-based loss of feature maps between teacher and student networks. However, this technique does not adequately consider the effective and meaningful delivery of knowledge from the teacher to the student network to improve the latter’s performance under given network capacity constraints. In this study, we propose a feature-domain adaptive contrastive distillation (FACD) method to train lightweight student SISR networks efficiently. We highlight the limitations of existing FD methods in terms of Euclidean distance-based loss, and propose a feature-domain contrastive loss, which causes student networks to learn richer information from the teacher’s representation in the feature domain. We also implement adaptive distillation that performs distillation selectively depending on the conditions of the training patches. Experimental results showed that the proposed FACD scheme improves student enhanced deep residual networks and residual channel attention networks not only in terms of the peak signal-to-noise ratio on all benchmark datasets and scales but also in terms of subjective image quality, compared to the conventional FD approaches. Code will be release at https://github.com/hcmoon0613/FACD.Keywords
All Related Versions
Funding Information
- Institute of Information and Communications Technology Planning and Evaluation
- Korean Government [Ministry of Science and Information and Communication Technology (MSIT)] (2021-0-00802)
This publication has 50 references indexed in Scilit:
- NTIRE 2017 Challenge on Single Image Super-Resolution: Methods and Results2017
- Deeply-Recursive Convolutional Network for Image Super-Resolution2016
- Deep Residual Learning for Image Recognition2016
- Accurate Image Super-Resolution Using Very Deep Convolutional Networks2016
- Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification2015
- Single image super-resolution from transformed self-exemplars2015
- Image Super-Resolution Using Deep Convolutional NetworksIEEE Transactions on Pattern Analysis and Machine Intelligence, 2015
- Interpolation-Based Image Super-Resolution Using Multisurface FittingIEEE Transactions on Image Processing, 2012
- Image Super-Resolution Via Sparse RepresentationIEEE Transactions on Image Processing, 2010
- A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics2002