Meerkat Behaviour Recognition Dataset

Meerkat Behaviour Recognition Dataset是由NAO研究所和奥克兰大学的研究团队创建，专注于记录和分析动物园中獴科动物的行为。该数据集包含20个视频，总计848,400帧的标注数据，涵盖了獴科动物的多种社会互动行为，如警戒、觅食和相互梳理等。数据集的创建过程涉及使用GoPro® Max摄像机在惠灵顿动物园的特定区域进行视频采集，并通过CVAT工具进行标注。此数据集的应用领域主要集中在动物行为学研究，旨在通过计算机视觉技术提高动物行为监测的效率和准确性，从而更好地理解和评估动物的福利状况。The Meerkat Behaviour Recognition Dataset was developed by research teams from the NAO Institute and the University of Auckland, focusing on recording and analyzing the behaviors of meerkats (family Herpestidae) in zoos. This dataset includes 20 videos totaling 848,400 annotated frames, covering a wide range of social interactive behaviors of meerkats such as vigilance, foraging, and allogrooming. The creation of this dataset involved video collection using GoPro® Max cameras at designated areas within Wellington Zoo, followed by annotation via the CVAT tool. The main application scenarios of this dataset lie in animal behavioral research, aiming to improve the efficiency and accuracy of animal behavior monitoring through computer vision technologies, so as to better understand and assess the welfare status of animals.