Data Upcycling Knowledge Distillation for Image Super-Resolution

TL;DR

This paper introduces Data Upcycling Knowledge Distillation (DUKD), a novel method for image super-resolution that leverages in-domain data augmentation and regularization to enhance student model performance beyond traditional limits.

Contribution

The paper proposes DUKD, a new knowledge distillation approach that utilizes upcycled in-domain data and label consistency regularization to improve super-resolution models.

Findings

DUKD outperforms previous methods on several SR benchmarks.

The use of in-domain data upcycling enhances knowledge transfer.

Regularization improves model robustness and performance.

Abstract

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.