Titlebook: Biometric Recognition; 11th Chinese Confere Zhisheng You,Jie Zhou,Qijun Zhao Conference proceedings 2016 Springer International Publishing

xanthelasma

enfeeble

Monolithic

泰然自若

978-3-319-46653-8Springer International Publishing AG 2016

全國性

希望

Occlusion-Robust Face Detection Using Shallow and Deep Proposal Based Faster R-CNNe detectors cannot fulfill the requirements in real-world scenarios especially in the presence of severe occlusions. This paper proposes a novel and effective approach to occlusion-robust face detection. It combines two major phases, . proposal generation and classification. In the former, we combin

Bravado

CORD

吸引力

Binary Classifiers and Radial Symmetry Transform for Fast and Accurate Eye Localizationon. The algorithm is based on binary classifiers and fast radial symmetry transform. First, eye candidates can be detected by the fast radial symmetry transform in infrared image. Then three-stage binary classifiers are used to eliminate most unreliable eye candidates. Finally, the mean eye template

JAMB

Robust Multi-view Face Alignment Based on Cascaded 2D/3D Face Shape Regressionand estimates 2D and 3D facial feature points simultaneously. In stage one, 2D and 3D facial feature points are roughly detected on the input face image, and head pose analysis is applied based on the 3D facial feature points to estimate its head pose. The face is then classified into one of three c