UVT-VOD2024
收藏资源简介:
UVT-VOD2024是由中国人工智能协会创建的一个用于未对齐RGBT视频对象检测的大型数据集。该数据集包含174个视频,总计30,494对未对齐的RGBT图像,这些图像直接从多光谱相机中捕获。数据集的创建过程包括数据收集、清洗、标注和验证,旨在真实反映多光谱传感器的捕获数据。UVT-VOD2024主要应用于视频对象检测领域,旨在解决现有方法在处理未对齐RGBT图像时的局限性,提升检测的鲁棒性和效率。
UVT-VOD2024 is a large-scale dataset developed by the Chinese Association for Artificial Intelligence for unaligned RGBT video object detection. This dataset comprises 174 videos, with a total of 30,494 pairs of unaligned RGBT images directly captured by multispectral cameras. The dataset creation pipeline encompasses data collection, cleaning, annotation and validation, aiming to faithfully represent the raw data captured by multispectral sensors. UVT-VOD2024 is primarily utilized in the field of video object detection, with the goal of addressing the limitations of existing state-of-the-art methods when processing unaligned RGBT images and improving the robustness and efficiency of object detection.




