APTv2
收藏资源简介:
APTv2是由杭州电子科技大学和悉尼大学合作创建的大规模动物姿态估计和跟踪数据集,包含来自30种不同动物的2749个视频片段,总计41,235帧。数据集通过手动选择和标注,提供了高质量的关键点和跟踪标注,适用于评估动物姿态估计和跟踪方法。APTv2数据集的应用领域包括动物行为理解、野生动物保护等,旨在解决动物姿态估计和跟踪中的挑战。
APTv2 is a large-scale animal pose estimation and tracking dataset jointly developed by Hangzhou Dianzi University and the University of Sydney. It includes 2,749 video clips from 30 distinct animal species, totaling 41,235 frames. The dataset provides high-quality keypoint and tracking annotations obtained through manual selection and annotation, which is suitable for evaluating animal pose estimation and tracking methods. The application fields of the APTv2 dataset cover animal behavior understanding, wildlife conservation and other related areas, with the goal of addressing the challenges in animal pose estimation and tracking.

- 1APTv2: Benchmarking Animal Pose Estimation and Tracking with a Large-scale Dataset and Beyond杭州电子科技大学电子信息学院 · 2023年



