Branch-structured detector for fast face detection using asymmetric LBP features
- PDF / 1,467,787 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 51 Downloads / 183 Views
ORIGINAL PAPER
Branch-structured detector for fast face detection using asymmetric LBP features Jie-chun Chen1
· Jun Wang1 · Li-ping Zhao2 · Jin He1
Received: 8 December 2019 / Revised: 21 March 2020 / Accepted: 8 May 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Face detection has been widely used to solve many important problems. In this paper, we present a fast face detector that is more suitable to be used by a low-cost embedded system. It can obtain high performance but use less CPU and memory resources. We first present a kind of pixel-level image feature named asymmetric LBP feature (ALBP). Furthermore, we find four kinds of four-bit ALBP features that are suitable to be used to construct cascade classifiers, since the four-bit ALBP features are discriminative features and large feature pools can be generated even if a small image patch is given. Second, we propose a branch-structured face detector that is composed of three ALBP-based cascade classifiers, i.e., one deep ALBP cascade and two shallow ALBP cascades. The detection speed and performance on FDDB dataset of the branch-structured face detector have both been evaluated. Experimental results show that it runs much faster than existing face detectors run on similar platforms, and its performance is close to that of well-known non-CNN face detectors which can achieve high performance. Keywords Face detection · Fast · Branch-structured · Asymmetric LBP
1 Introduction Face detection is one of the research focuses in computer vision. It has been applied to solve many important problems such as face recognition and verification, face tracking for surveillance, and camera’s autofocus. As stated by Zafeiriou et al., the research on automatic face detection can be traced back to more than 50 years [1, 34]. The face detector proposed by Viola and Jones [7] is a line of demarcation between practical face detectors and non-practical face detectors. The main contribution of Viola and Jones is that they constructed a cascade classifier. The success of the cascade classifier lies in its satisfactory efficiency and performance. Since the Viola–Jones face detector was presented, many researchers have tried to further improve its efficiency and performance in different ways: by improving on classical Haar-like features [8–12] or finding new features [2, 5,
B
Jie-chun Chen [email protected]
1
College of Automation Engineering, Northeast Electric Power University, Jilin 132012, China
2
College of Science, Northeast Electric Power University, Jilin 132012, China
6, 13–15], by changing the face detector’s structure [13, 16–19], and by preprocessing image prior to being used in the face detection stage [20]. For instance, Lienhart et al. [8] have extended the set of classical Haar-like features through adding a set of 45° rotated features. Li et al. [5] have proposed three types of features one after another, i.e., MB-LBP feature, aggregated channel feature (ACF) [13], and NPD feature [6]. To improve the performance of a Viola–
Data Loading...
