AwA Pose
收藏资源简介:
AwA Pose是由普雷里维尤农工大学的研究人员创建的一个大型动物关键点检测数据集,专门用于四足动物的关键点检测。该数据集包含35种不同的四足动物物种,共计10064张图像,每种动物标注了多达39个关键点。数据集的创建过程涉及从AwA数据集中选择图像,并使用LabelMe工具进行详细的关键点和边界框标注。AwA Pose数据集的应用领域广泛,包括动物行为理解、自动化识别与跟踪、部分分割等,旨在解决四足动物关键点检测的复杂性和多样性问题。
AwA Pose is a large-scale animal keypoint detection dataset created by researchers from Prairie View A&M University, specifically designed for quadruped animal keypoint detection. This dataset includes 35 distinct quadruped animal species, with a total of 10,064 images, and up to 39 keypoints are annotated for each animal. The development of the AwA Pose dataset involved selecting images from the AwA dataset and conducting detailed annotations of keypoints and bounding boxes using the LabelMe tool. The AwA Pose dataset has broad application scenarios, including animal behavior understanding, automated recognition and tracking, and part segmentation, aiming to address the complexity and diversity challenges in quadruped animal keypoint detection.

- 1A Novel Dataset for Keypoint Detection of quadruped Animals from Images普雷里维尤农工大学 · 2021年



