# Natural Image Noise Dataset

## TL;DR

This paper introduces the Natural Image Noise Dataset (NIND), a large collection of DSLR-like images with varying ISO noise levels, enabling improved training of blind denoising models that outperform existing methods like BM3D.

## Contribution

The creation of NIND, a comprehensive dataset for training and benchmarking image denoising models on real-world noise, and demonstrating its effectiveness in improving denoising performance.

## Key findings

- Model trained on NIND outperforms BM3D on unseen ISO noise.
- The dataset enables generalization across different camera types.
- Open access on Wikimedia Commons encourages community contributions.

## Abstract

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.