Darmstadt Noise Dataset (DND)
收藏资源简介:
Darmstadt Noise Dataset (DND) 是由达姆施塔特工业大学计算机科学系开发的,用于评估真实照片去噪技术的基准数据集。该数据集包含50个场景,使用不同传感器大小的消费者相机在各种ISO值下拍摄。DND数据集旨在解决传统基于合成高斯噪声评估去噪技术的不现实性问题,通过捕捉具有不同ISO值的图像对,使用精确的后处理方法来获取真实噪声的基准。数据集的应用领域主要集中在图像处理和计算机视觉,特别是在去噪算法的评估和改进上。
The Darmstadt Noise Dataset (DND) is a benchmark dataset developed by the Department of Computer Science, Technische Universität Darmstadt, for evaluating real-photograph denoising techniques. It contains 50 scenes captured by consumer-grade cameras with varying sensor sizes across a range of ISO values. The DND dataset aims to address the unrealistic nature of traditional denoising evaluation methods based on synthetic Gaussian noise, by capturing paired images across different ISO settings and utilizing precise post-processing workflows to establish a benchmark for real-world noise. Its primary application domains cover image processing and computer vision, particularly for the evaluation and improvement of denoising algorithms.




