Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks

TLDR

Face detection and alignment in unconstrained environments are difficult due to varying poses, illumination, and occlusion, but recent deep learning approaches have shown impressive performance on these tasks. This work proposes a deep cascaded multi‑task framework that exploits the inherent correlation between detection and alignment to boost their performance. The framework uses a three‑stage cascaded CNN that predicts face and landmark locations in a coarse‑to‑fine manner and incorporates an online hard‑sample mining strategy to improve performance automatically. The method outperforms state‑of‑the‑art techniques on the FDDB, WIDER FACE, and AFLW benchmarks while maintaining real‑time speed.

Abstract

Face detection and alignment in unconstrained environment are challenging due to various poses, illuminations and occlusions. Recent studies show that deep learning approaches can achieve impressive performance on these two tasks. In this paper, we propose a deep cascaded multi-task framework which exploits the inherent correlation between them to boost up their performance. In particular, our framework adopts a cascaded structure with three stages of carefully designed deep convolutional networks that predict face and landmark location in a coarse-to-fine manner. In addition, in the learning process, we propose a new online hard sample mining strategy that can improve the performance automatically without manual sample selection. Our method achieves superior accuracy over the state-of-the-art techniques on the challenging FDDB and WIDER FACE benchmark for face detection, and AFLW benchmark for face alignment, while keeps real time performance.