FaceNet: A unified embedding for face recognition and clustering

TLDR

Despite recent advances in face recognition, scaling efficient verification and recognition remains a serious challenge. This paper presents FaceNet, a system that learns a mapping from face images to a compact Euclidean space where distances directly reflect face similarity. FaceNet employs a deep convolutional network trained with triplet loss and online triplet mining on roughly aligned matching and non‑matching face patches, producing embeddings that enable straightforward recognition, verification, and clustering. Using only 128‑byte embeddings, FaceNet achieves state‑of‑the‑art accuracy—99.63% on LFW and 95.12% on YouTube Faces—reducing error rates by 30% compared to the best published results.

Abstract

Despite significant recent advances in the field of face recognition [10, 14, 15, 17], implementing face verification and recognition efficiently at scale presents serious challenges to current approaches. In this paper we present a system, called FaceNet, that directly learns a mapping from face images to a compact Euclidean space where distances directly correspond to a measure offace similarity. Once this space has been produced, tasks such as face recognition, verification and clustering can be easily implemented using standard techniques with FaceNet embeddings asfeature vectors. Our method uses a deep convolutional network trained to directly optimize the embedding itself, rather than an intermediate bottleneck layer as in previous deep learning approaches. To train, we use triplets of roughly aligned matching / non-matching face patches generated using a novel online triplet mining method. The benefit of our approach is much greater representational efficiency: we achieve state-of-the-artface recognition performance using only 128-bytes perface. On the widely used Labeled Faces in the Wild (LFW) dataset, our system achieves a new record accuracy of 99.63%. On YouTube Faces DB it achieves 95.12%. Our system cuts the error rate in comparison to the best published result [15] by 30% on both datasets.