MobileFaceNets: Efficient CNNs for Accurate Real-Time Face Verification on Mobile Devices
This work addresses the need for accurate real-time face verification on mobile and embedded devices, offering a significant improvement in efficiency over previous mobile CNNs.
The paper tackles the problem of efficient face verification on mobile devices by introducing MobileFaceNets, which use less than 1 million parameters and achieve 99.55% accuracy on LFW and 92.59% TAR@FAR1e-6 on MegaFace, with inference times as low as 18 milliseconds on a mobile phone.
We present a class of extremely efficient CNN models, MobileFaceNets, which use less than 1 million parameters and are specifically tailored for high-accuracy real-time face verification on mobile and embedded devices. We first make a simple analysis on the weakness of common mobile networks for face verification. The weakness has been well overcome by our specifically designed MobileFaceNets. Under the same experimental conditions, our MobileFaceNets achieve significantly superior accuracy as well as more than 2 times actual speedup over MobileNetV2. After trained by ArcFace loss on the refined MS-Celeb-1M, our single MobileFaceNet of 4.0MB size achieves 99.55% accuracy on LFW and 92.59% TAR@FAR1e-6 on MegaFace, which is even comparable to state-of-the-art big CNN models of hundreds MB size. The fastest one of MobileFaceNets has an actual inference time of 18 milliseconds on a mobile phone. For face verification, MobileFaceNets achieve significantly improved efficiency over previous state-of-the-art mobile CNNs.