Compressing Deep Image Super-resolution Models
arXiv:2401.00523 · doi:10.1109/PCS60826.2024.10566374
Abstract
Deep learning techniques have been applied in the context of image super-resolution (SR), achieving remarkable advances in terms of reconstruction performance. Existing techniques typically employ highly complex model structures which result in large model sizes and slow inference speeds. This often leads to high energy consumption and restricts their adoption for practical applications. To address this issue, this work employs a three-stage workflow for compressing deep SR models which significantly reduces their memory requirement. Restoration performance has been maintained through teacher-student knowledge distillation using a newly designed distillation loss. We have applied this approach to two popular image super-resolution networks, SwinIR and EDSR, to demonstrate its effectiveness. The resulting compact models, SwinIRmini and EDSRmini, attain an 89% and 96% reduction in both model size and floating-point operations (FLOPs) respectively, compared to their original versions. They also retain competitive super-resolution performance compared to their original models and other commonly used SR approaches. The source code and pre-trained models for these two lightweight SR approaches are released at https://pikapi22.github.io/CDISM/.
References in corpus (6)
- Distilling the Knowledge in a Neural Network
- Lightweight Image Super-Resolution with Information Multi-distillation Network
- To prune, or not to prune: exploring the efficacy of pruning for model compression
- MFRNet: A New CNN Architecture for Post-Processing and In-loop Filtering
- ViSTRA2: Video Coding using Spatial Resolution and Effective Bit Depth Adaptation
- CVEGAN: A Perceptually-inspired GAN for Compressed Video Enhancement