RENOIR stands for Real Low-Light Image Noise Reduction. It is a color image dataset under low-light conditions, containing 500 images from 120 scenes. These images were collected by three cameras in low-light environments (Cannon T3i, Cannon S90, and Xiaomi MI3 mobile phone). The dataset is divided into natural noise images and low-noise images, each of which is about 10 million pixels.

This dataset can be used to understand the noise formation and noise statistics in images, and can also be used to train and test image denoising algorithms or for other purposes.