TAO-Amodal
收藏资源简介:
TAO-Amodal数据集由卡内基梅隆大学创建,包含332,000个边界框,覆盖了2,907个视频中的多种遮挡场景,并针对833个对象类别进行了标注。数据集内容丰富,包括完全遮挡和部分遮挡的对象,以及图像帧内和帧外的对象。创建过程涉及详细的标注协议和质量控制,确保数据的高质量。该数据集主要用于评估和改进对象跟踪算法在处理遮挡情况下的性能,特别是在自动驾驶等应用中对完全遮挡对象的理解至关重要。
The TAO-Amodal dataset was created by Carnegie Mellon University. It contains 332,000 bounding boxes, spanning diverse occlusion scenarios across 2,907 videos, with annotations for 833 object categories. The dataset encompasses a wide range of content, including fully and partially occluded objects, as well as objects within or outside image frames. Its development follows detailed annotation protocols and strict quality control procedures to ensure high data quality. This dataset is primarily used to evaluate and improve the performance of object tracking algorithms when handling occlusion situations, particularly in applications such as autonomous driving where understanding fully occluded objects is critically important.




