Targeting Ultimate Accuracy: Face Recognition via Deep Embedding

TL;DR

This paper introduces a two-stage deep learning approach combining multi-patch CNN and metric learning to improve face recognition accuracy, achieving near-perfect results on benchmark datasets.

Contribution

The paper presents a novel two-stage deep learning framework that enhances face recognition performance by extracting highly discriminative features through multi-patch CNNs and deep metric learning.

Findings

Achieves 99.77% verification accuracy on LFW dataset.

Outperforms state-of-the-art methods on practical protocols.

Highlights the importance of data size and patch number for high performance.

Abstract

Face Recognition has been studied for many decades. As opposed to traditional hand-crafted features such as LBP and HOG, much more sophisticated features can be learned automatically by deep learning methods in a data-driven way. In this paper, we propose a two-stage approach that combines a multi-patch deep CNN and deep metric learning, which extracts low dimensional but very discriminative features for face verification and recognition. Experiments show that this method outperforms other state-of-the-art methods on LFW dataset, achieving 99.77% pair-wise verification accuracy and significantly better accuracy under other two more practical protocols. This paper also discusses the importance of data size and the number of patches, showing a clear path to practical high-performance face recognition systems in real world.