Natural Image Noise Dataset (NIND)
收藏资源简介:
自然图像噪声数据集(NIND)是由天主教鲁汶大学的研究人员创建的一个大型数据集,专注于DSLR类图像中ISO噪声的处理。该数据集包含616张图像,涵盖多种ISO噪声水平,适用于训练盲去噪模型。数据集内容丰富,包括室内外场景,旨在解决高质量图像的通用去噪问题。创建过程中,每个场景均捕捉多次,确保数据多样性和质量。NIND数据集的应用领域广泛,主要用于图像去噪算法的训练和评估,特别是在处理不同相机和光线条件下的图像噪声问题。
The Natural Image Noise Dataset (NIND) is a large-scale dataset developed by researchers from KU Leuven (Catholic University of Leuven), focusing on the mitigation of ISO noise in DSLR-style images. This dataset contains 616 images covering various ISO noise levels, and is suitable for training blind denoising models. With diverse content including both indoor and outdoor scenes, the dataset aims to address the general denoising problem for high-quality images. During the creation process, each scene was captured multiple times to ensure data diversity and quality. The NIND dataset has broad application scenarios, and is primarily used for the training and evaluation of image denoising algorithms, particularly for handling image noise issues under different camera and lighting conditions.

- 1Natural Image Noise Dataset天主教鲁汶大学 · 2019年



