Data Upcycling Knowledge Distillation for Image Super-Resolution
This work addresses the problem of inefficient knowledge transfer in super-resolution networks for researchers and practitioners, representing an incremental advancement in distillation techniques.
The paper tackles the limited effectiveness of knowledge distillation in image super-resolution by proposing Data Upcycling Knowledge Distillation (DUKD), which uses upcycled in-domain data and label consistency regularization, resulting in significant performance improvements over previous methods on several SR tasks.
Knowledge distillation (KD) compresses deep neural networks by transferring task-related knowledge from cumbersome pre-trained teacher models to compact student models. However, current KD methods for super-resolution (SR) networks overlook the nature of SR task that the outputs of the teacher model are noisy approximations to the ground-truth distribution of high-quality images (GT), which shades the teacher model's knowledge to result in limited KD effects. To utilize the teacher model beyond the GT upper-bound, we present the Data Upcycling Knowledge Distillation (DUKD), to transfer the teacher model's knowledge to the student model through the upcycled in-domain data derived from training data. Besides, we impose label consistency regularization to KD for SR by the paired invertible augmentations to improve the student model's performance and robustness. Comprehensive experiments demonstrate that the DUKD method significantly outperforms previous arts on several SR tasks.