Efficient Low-Resolution Face Recognition via Bridge Distillation

TL;DR

By learning low- resolution face representations and mimicking the adapted high-resolution knowledge, a light-weight student model can be constructed with high efficiency and promising accuracy in recognizing low-resolution faces.

Abstract

Face recognition in the wild is now advancing towards light-weight models, fast inference speed and resolution-adapted capability. In this paper, we propose a bridge distillation approach to turn a complex face model pretrained on private high-resolution faces into a light-weight one for low-resolution face recognition. In our approach, such a cross-dataset resolution-adapted knowledge transfer problem is solved via two-step distillation. In the first step, we conduct cross-dataset distillation to transfer the prior knowledge from private high-resolution faces to public high-resolution faces and generate compact and discriminative features. In the second step, the resolution-adapted distillation is conducted to further transfer the prior knowledge to synthetic low-resolution faces via multi-task learning. By learning low-resolution face representations and mimicking the adapted high-resolution knowledge, a light-weight student model can be constructed with high efficiency and promising accuracy in recognizing low-resolution faces. Experimental results show that the student model performs impressively in recognizing low-resolution faces with only 0.21M parameters and 0.057MB memory. Meanwhile, its speed reaches up to 14,705, ~934 and 763 faces per second on GPU, CPU and mobile phone, respectively.

Figures (6)

• Figure 1: Motivation of the bridge distillation. The direct knowledge transfer from private high-resolution faces to target low-resolution faces may be difficult. Therefore, we use public high-resolution and low-resolution faces as a bridge to step-wisely distil and compress the knowledge via cross-dataset distillation and resolution-adapted distillation. Note that the public low-resolution faces are generated from the public high-resolution faces to simulate the probable distribution of target low-resolution faces.
• Figure 2: The bridge distillation framework, which consists of a teacher stream and a student stream. 1) The teacher stream is first pre-trained on high-resolution private faces, which extracts the learned knowledge about informative facial details. Then, cross-dataset distillation adapts the learned knowledge to the high-resolution public faces so as to preserve compact and discriminative features. 2) The student stream is trained on low-resolution face recognition via resolution-adapted distillation by jointly performing two tasks: feature regression to mimic the adapted high-resolution knowledge, and face classification on low-resolution public faces. Thus, the resulting student models could be deployed to recognize low-resolution target faces.
• Figure 3: The performance of the adapted models with and without cross-dataset distillation on UMDFaces.
• Figure 4: The face verification performance of various teacher and student models on LFW.
• Figure 5: ROC curves of various models on LFW.