Exploration of Lightweight Single Image Denoising with Transformers and Truly Fair Training
This work addresses the need for efficient denoising models for real-world applications, though it is incremental as it builds on existing Transformer methods.
The paper tackles the problem of lightweight single image denoising with Transformers by providing seven baseline models and highlighting the impact of randomly cropped patches on performance, achieving competitive results on benchmarks like SIDD and DND with PSNR gains of up to 0.5 dB.
As multimedia content often contains noise from intrinsic defects of digital devices, image denoising is an important step for high-level vision recognition tasks. Although several studies have developed the denoising field employing advanced Transformers, these networks are too momory-intensive for real-world applications. Additionally, there is a lack of research on lightweight denosing (LWDN) with Transformers. To handle this, this work provides seven comparative baseline Transformers for LWDN, serving as a foundation for future research. We also demonstrate the parts of randomly cropped patches significantly affect the denoising performances during training. While previous studies have overlooked this aspect, we aim to train our baseline Transformers in a truly fair manner. Furthermore, we conduct empirical analyses of various components to determine the key considerations for constructing LWDN Transformers. Codes are available at https://github.com/rami0205/LWDN.