Year: 2024
Type: article
Abstract: To achieve goat face detection, head behavior recognition, and classification, this study developed a dataset of goat and sheep facial images. The dataset was created by capturing images of solid-colored goats in natural settings and sheep in their natural environments. It includes goat and sheep facial images taken from various angles, in different quantities, in a range of scenes (both indoors and outdoors), and with varying levels of occlusions. To ensure the quality of the dataset, 2,412 high-quality, high-resolution goat and sheep facial images were obtained through manual processing and selection. These images were then manually labeled for annotation. To assess the quality and authenticity of this dataset, it was divided into training and testing sets at an 8:2 ratio. This division was used to evaluate the performance of goat and sheep face detection models, including Faster R-CNN, SSD, Yolov3, and Yolov5. Experimental results confirmed the effectiveness of the dataset in goat and sheep face detection models, providing valuable resources for research and applications in this field. In addition, the dataset will be continuously refined and updated to meet the growing demands of research and application scenarios.
Source: China Scientific Data
Authors: Xiaoping HUANG, Zihao DOU, Fei HUANG, Huanyu ZHENG, Xiankun HOU, Chenyang WANG, Tao FENG, Yuan RAO
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
SDG: Zero hunger
Open Access status: diamond
OpenAlex ID: W4408636366