UG2+ Challenge Track 2
收藏资源简介:
UG2+ Challenge Track 2数据集由北京大学王选计算机研究所的研究团队创建,旨在提升恶劣天气和低光环境下计算机视觉任务的理解能力。该数据集包含三个子集,分别针对雾、雨和低光条件下的物体或面部检测。每个子集均包含在真实世界中收集的图像,并带有标注的物体/面部信息,以支持对增强方法的全面检查和公平比较。此数据集的创建是为了激发对低级视觉技术如何惠及各种场景下高级自动视觉识别的全面讨论和探索。
The UG2+ Challenge Track 2 dataset was developed by the research team from the Wangxuan Institute of Computer Technology, Peking University, aiming to enhance the understanding of computer vision tasks under adverse weather and low-light conditions. This dataset comprises three subsets, which target object or face detection under foggy, rainy and low-light conditions respectively. Each subset contains real-world captured images with annotated object/face information, to support comprehensive evaluation and fair comparison of enhancement methods. This dataset was created to inspire comprehensive discussions and explorations on how low-level vision technologies can benefit high-level automatic visual recognition across various scenarios.




