Joint Demosaicing and Denoising with Double Deep Image Priors
This addresses the challenge of processing RAW images in digital cameras without needing large training datasets, though it is incremental as it builds on prior deep learning approaches.
The paper tackled the problem of joint demosaicing and denoising for RAW images by proposing JDD-DoubleDIP, a method that operates on a single image without training data, and it outperformed other methods on Kodak and McMaster datasets in terms of PSNR, SSIM, and visual quality.
Demosaicing and denoising of RAW images are crucial steps in the processing pipeline of modern digital cameras. As only a third of the color information required to produce a digital image is captured by the camera sensor, the process of demosaicing is inherently ill-posed. The presence of noise further exacerbates this problem. Performing these two steps sequentially may distort the content of the captured RAW images and accumulate errors from one step to another. Recent deep neural-network-based approaches have shown the effectiveness of joint demosaicing and denoising to mitigate such challenges. However, these methods typically require a large number of training samples and do not generalize well to different types and intensities of noise. In this paper, we propose a novel joint demosaicing and denoising method, dubbed JDD-DoubleDIP, which operates directly on a single RAW image without requiring any training data. We validate the effectiveness of our method on two popular datasets -- Kodak and McMaster -- with various noises and noise intensities. The experimental results show that our method consistently outperforms other compared methods in terms of PSNR, SSIM, and qualitative visual perception.