【作者单位】1College of Computer Science, Sichuan University, Chengdu 610065, China 2School of Information Science and Technology, Tibet University, Lhasa 850000, China
【摘要】 A part-based convolutional network is proposed for yak identification. • An auxiliary task based on self-supervised learning is devised. • Horizontal mirroring as new entities is investigated for model training. • Our method improves Rank-1 accuracy ...