Rethinking Noise Synthesis and Modeling in Raw Denoising
This addresses the challenge of accurate noise modeling for raw image denoising, which is crucial for improving image quality in photography and computer vision applications, though it is incremental in refining noise synthesis techniques.
The paper tackles the problem of synthesizing realistic raw image noise for training denoising models by directly sampling from sensor noise, outperforming existing methods and showing wide generalization across sensors and lighting conditions.
The lack of large-scale real raw image denoising dataset gives rise to challenges on synthesizing realistic raw image noise for training denoising models. However, the real raw image noise is contributed by many noise sources and varies greatly among different sensors. Existing methods are unable to model all noise sources accurately, and building a noise model for each sensor is also laborious. In this paper, we introduce a new perspective to synthesize noise by directly sampling from the sensor's real noise. It inherently generates accurate raw image noise for different camera sensors. Two efficient and generic techniques: pattern-aligned patch sampling and high-bit reconstruction help accurate synthesis of spatial-correlated noise and high-bit noise respectively. We conduct systematic experiments on SIDD and ELD datasets. The results show that (1) our method outperforms existing methods and demonstrates wide generalization on different sensors and lighting conditions. (2) Recent conclusions derived from DNN-based noise modeling methods are actually based on inaccurate noise parameters. The DNN-based methods still cannot outperform physics-based statistical methods.