Efficient disentangled representation learning for multi-modal finger biometrics

Weili Yang, Junduan Huang, Dacan Luo, Wenxiong Kang

Abstract: Highlights•FDRL-Net is the first disentangled representation learning method on finger traits.•A direct Boat-Trackers-based multi-task distillation method is proposed.•Extensive experiments shows the superiority of our proposed method.