MegaFace: A Million Faces for Recognition at Scale

TL;DR

This paper introduces a large-scale face recognition dataset with a million images, evaluates state-of-the-art algorithms and human performance at scale, revealing significant drops in accuracy and robustness compared to small-scale benchmarks.

Contribution

It provides the MegaFace dataset for large-scale face recognition and analyzes algorithm and human performance at this scale, highlighting challenges and robustness issues.

Findings

Recognition performance drops at scale for most algorithms

Deep learning approaches perform better but still less robust at scale

Human recognition accuracy at scale is evaluated and reported

Abstract

Recent face recognition experiments on the LFW benchmark show that face recognition is performing stunningly well, surpassing human recognition rates. In this paper, we study face recognition at scale. Specifically, we have collected from Flickr a \textbf{Million} faces and evaluated state of the art face recognition algorithms on this dataset. We found that the performance of algorithms varies--while all perform great on LFW, once evaluated at scale recognition rates drop drastically for most algorithms. Interestingly, deep learning based approach by \cite{schroff2015facenet} performs much better, but still gets less robust at scale. We consider both verification and identification problems, and evaluate how pose affects recognition at scale. Moreover, we ran an extensive human study on Mechanical Turk to evaluate human recognition at scale, and report results. All the photos are creative commons photos and is released at \small{\url{http://megaface.cs.washington.edu/}} for research and further experiments.