Low-Resolution Face Recognition in the Wild via Selective Knowledge Distillation

By Shiming Ge, Shengwei Zhao, Chenyu Li, Jia Li. 2018

209 citations recorded by the source corpus.

Typically, the deployment of face recognition models in the wild needs to identify low-resolution faces with extremely low computational cost. To address this problem, a feasible solution is compressing a complex face model to achieve higher speed and lower memory at the cost of minimal performance drop. Inspired by that, this paper proposes a learning approach to recognize low-resolution faces via selective knowledge distillation. In this approach, a two-stream convolutional neural network (CNN) is first initialized to recognize high-resolution faces and resolution-degraded faces with a teacher stream and a student stream, respectively. The teacher stream is represented by a complex CNN for high-accuracy recognition, and the student stream is represented by a much simpler CNN for low-complexity recognition. To avoid significant performance drop at the student stream, we then selectively distil the most informative facial features from the teacher stream by solving a sparse graph optimization problem, which are then used to regularize the fine-tuning process of the student stream. In this way, the student stream is actually trained by simultaneously handling two tasks with limited computational resources: approximating the most informative facial cues via feature regression, and recovering the missing facial cues via low-resolution face classification. Experimental results show that the student stream performs impressively in recognizing low-resolution faces and costs only 0.15-MB memory and runs at 418 faces per second on CPU and 9433 faces per second on GPU.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Dacheng Tao
2. Jianping Gou
3. S. Maybank
4. Yunhe Wang
5. Baosheng Yu
6. Kai Han
7. Byung Cheol Song
8. Seunghyun Lee

Citations in the corpus

• Knowledge Distillation: A Survey (2020, methodology): “…al. 2019) and then extended to other visual recognition applications, including face recognition (Luo et al. 2016; Kong et al. 2019; Yan et al. 2019; Ge et al. 2018; Wang et al. 2018b, 2019c; Duong et al. 2019; Wu et al. 2020; Wang et al. 2017; Zhang et al. 2020b; Wang et al. 2020b), image/video…”
• Delving Deep Into Label Smoothing (2020, unknown): “Knowledge distillation has been widely used in many tasks [35], [36], [37], [38].”
• Low-resolution Visual Recognition via Deep Feature Distillation (2019, background): “Recently teacher-student learning paradigm is very popular, i.e., directly learning a student network with the indication of a teacher network which consists of more valuable information [17, 18, 19, 20].”
• Self-Distillation: Towards Efficient and Compact Neural Networks (2021, background): “Ge et al. regard neural networks trained on high resolution images as the teachers of neural networks trained on low resolution, achieving both accurate and rapid face recognition [51].”
• Efficient Low-Resolution Face Recognition via Bridge Distillation (2020, methodology): “The most recent SKD [1] can recognize 16 × 16 face images while costing 0.79M parameters.”
• Graph-based Knowledge Distillation by Multi-head Attention Network (2019, methodology): “ This method is simple, so it is good for general purpose, but it is relatively naive and has a small amount of information. Recently, various methods have been introduced to enhance teacher knowledge[6,19]. Multi-connection knowledge. In order to solve the problem that the amount of information of ”