TL;DR
This paper introduces NormFace, a face verification method that employs L2 hypersphere embedding and normalized features during training, improving accuracy by addressing normalization challenges in deep face recognition.
Contribution
It presents a mathematical analysis of normalization issues, proposes two training strategies with normalized features, and demonstrates consistent performance improvements on face verification benchmarks.
Findings
Improved face verification accuracy by 0.2% to 0.4% on LFW.
Identified key issues in feature normalization during training.
Proposed normalization-based training strategies outperform baseline methods.
Abstract
Thanks to the recent developments of Convolutional Neural Networks, the performance of face verification methods has increased rapidly. In a typical face verification method, feature normalization is a critical step for boosting performance. This motivates us to introduce and study the effect of normalization during training. But we find this is non-trivial, despite normalization being differentiable. We identify and study four issues related to normalization through mathematical analysis, which yields understanding and helps with parameter settings. Based on this analysis we propose two strategies for training using normalized features. The first is a modification of softmax loss, which optimizes cosine similarity instead of inner-product. The second is a reformulation of metric learning by introducing an agent vector for each class. We show that both strategies, and small variants,âŠ
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
NormFace: Hypersphere Embedding for Face Verification
(2017) â â journalyear: 2017â â copyright: acmcopyrightâ â conference: MM â17; ; October 23â27, 2017, Mountain View, CA, USA.â â price: 15.00â â doi: https://doi.org/10.1145/3123266.3123359â â isbn: ISBN 978-1-4503-4906-2/17/10
