Cow Face Detection and Recognition Based on Automatic Feature Extraction Algorithm
Automatic farm livestock detection and recognition have high importance in the management of livestock due to the increasing potentials in dairy cow welfare as well as production efficiency.In contrast to the general object(e.g.,person,car and bird),the recognition of farm livestock still remains challenging due to the open complex scenarios,similar appearance,shape deformation,occlusion and insufficient annotated data and needs to be solved.In this paper,we discuss the problem of cow face detection and recognition by releasing a new large-scale cow dataset which containing about 50,000 annotated cow face detection data and probably 18,000 cow recognition data.Moreover,a cow face recognition framework is proposed which hybrids the detection and recognition model to improve the recognition performance.Experimental results show the superiority of the proposed method.The accuracy of the detection is 98.3%,and the accuracy of the cow face recognition is up to 94.1%.
Cow face detection Cow face recognition Deep neural network
Liyao Yao Zexi Hu Caixing Liu Hanxing Liu Yingjie Kuang Yuefang Gao
College of Mathematics and Informatics South China Agricultural University,China School of Computer Science University of Sydney,Sydney
国际会议
2019国图灵大会(ACM Turing Celebration conference-China 2019 )
成都
英文
195-199
2019-05-17（万方平台首次上网日期，不代表论文的发表时间）