RESIDE
收藏资源简介:
RESIDE是一个用于评估单图像去雾算法的大型数据集,由中国科学院信息工程研究所等机构创建。数据集包含14,490张图像,分为五个子集,分别用于不同的训练或评估目的。RESIDE强调多样化的数据来源和图像内容,并提供了从全参考度量到无参考度量,再到主观评价和新颖的任务驱动评价的丰富评价标准。该数据集旨在解决户外场景中由于雾霾导致的图像质量下降问题,适用于视频监控和自动驾驶等多种应用领域。
RESIDE is a large-scale dataset for evaluating single-image dehazing algorithms, created by institutions including the Institute of Information Engineering, Chinese Academy of Sciences, and others. The dataset contains 14,490 images, divided into five subsets tailored for distinct training or evaluation purposes. RESIDE emphasizes diverse data sources and image content, and provides a comprehensive set of evaluation metrics ranging from full-reference metrics, no-reference metrics, to subjective evaluation and novel task-driven evaluation criteria. This dataset aims to address the issue of image quality degradation caused by haze in outdoor scenes, and is applicable to multiple application fields such as video surveillance and autonomous driving.




