Natural Image Noise Dataset
This provides a dataset for training blind denoising models, addressing the gap in realistic noise data for image processing applications, though it is incremental as it builds on existing dataset efforts.
The paper tackles the problem of image denoising by introducing the Natural Image Noise Dataset (NIND), a large dataset of DSLR-like images with varying ISO noise, and demonstrates that a model trained on it significantly outperforms BM3D on unseen images, even generalizing to different camera types.
Convolutional neural networks have been the focus of research aiming to solve image denoising problems, but their performance remains unsatisfactory for most applications. These networks are trained with synthetic noise distributions that do not accurately reflect the noise captured by image sensors. Some datasets of clean-noisy image pairs have been introduced but they are usually meant for benchmarking or specific applications. We introduce the Natural Image Noise Dataset (NIND), a dataset of DSLR-like images with varying levels of ISO noise which is large enough to train models for blind denoising over a wide range of noise. We demonstrate a denoising model trained with the NIND and show that it significantly outperforms BM3D on ISO noise from unseen images, even when generalizing to images from a different type of camera. The Natural Image Noise Dataset is published on Wikimedia Commons such that it remains open for curation and contributions. We expect that this dataset will prove useful for future image denoising applications.