Abstract: Face detection is the first step in many visual
processing systems like face recognition, emotion recognition, gender
recognition and etc. In this paper, we propose a novel method for fast and
reliable face detection, especially in complex background. Firstly, the
image is divided into small blocks to extract Local Binary Pattern (LBP)
features and Simplified Weber Local Descriptor (SWLD) features. Secondly,
these features compares values of Local Binary Pattern and Simplified Weber
Local Descriptor calculated over two adjacent image subregions. These
subregions are similar to those in the Haar masks. Thirdly, the HLBP and
HSWLD features are combined together to ensure the classification capability
of Haar features, an improvement on the Weber Local Descriptor (WLD)
contains differential excitation component and Local Binary Pattern (LBP)
component. These two components are complementary to each other.
Specifically, differential excitation preserves the local intensity
information but omits the orientations of edges. On the contrary, LBP
describes the orientations of the edges but ignore the intensity
information. Finally, the fused features are boosted using AdaBoost in a
framework similar to that proposed by Viola and Jones. Experimental results
show that our method outperforms the methods that are based on Haar
features, SWLD features or LBP features individually.
Keywords: Face detection, haar feature, weber local descriptor (WLD), local binary pattern (LBP), AdaBoost
