针对二分支神经网络匹配的人脸检测算法研究

doi:10.3778/j.issn.1673-9418.1911037

摘要/Abstract

摘要：

在尺度、遮挡等因素的影响下，人脸检测算法速度和精度不匹配进而表现出检测性能差等系列缺点，采用融入模板匹配思想的神经网络检测器算法解决这一不足，提出了二分支神经网络匹配的人脸检测算法。以此得到检测速度与精度相匹配的改进方法。首先根据五官位置关系构建人脸模板框架并设置一定的阈值；其次依据构建的模板和锚策略对不同尺度人脸的鲁棒性，在DSFD人脸检测器的特征增强模块层中实现对样本数据中人脸区域的提取及标记；最后依据提取、标记的人脸区域与构建的模板进行相关性匹配。改进算法在Face Detection Data Set and Benchmark、FDDB、WIDER FACE人脸数据集的实验结果表明，在提升检测速度的同时保证算法的精度，相比几类相近算法，该算法的优势更加明显。

关键词: 人脸检测, 模板匹配, 人脸检测器, 特征提取

Abstract:

Under the influence of scale, occlusion and other factors, the speed and precision of the face detection algorithm are not matched at the same time, and the detection performance is poor. The neural network detector algorithm incorporating the idea of template matching is adopted to solve this deficiency. A face detection algorithm for dual-shot neural network matching is proposed. Thus, an improved method for matching the detection speed and precision is obtained. Firstly, constructing a face template framework and setting a certain threshold according to the five-feature position relation. Secondly, according to the robustness of the constructed template and anchor strategy to different scales of face, extracting and labeling the face area in the sample data in the feature enhancement module layer of the DSFD (dual shot face detector). Finally, matching the correction between the extracted and marked face area and the constructed template. The experimental results on face data sets, such as Face Detection Data Set and Benchmark, FDDB and WIDER FACE indicate that the improved algorithm ensures the accuracy while improving the detection speed, and its advantage is more obvious than that of several similar algorithms.

Key words: face detection, template matching, face detector, feature extraction

王茹，贺兴时，杨新社. 针对二分支神经网络匹配的人脸检测算法研究[J]. 计算机科学与探索, 2020, 14(10): 1727-1734.

WANG Ru, HE Xingshi, YANG Xinshe. Research on Face Detection Algorithm for Dual-Shot Neural Network Matching[J]. Journal of Frontiers of Computer Science and Technology, 2020, 14(10): 1727-1734.

参考文献

[1] Shao P, Yang L M, Huang H B, et al. Rapid algorithm of tem-plate matching for face detection[J]. Journal of Computer Applications, 2006, 26(7): 1607-1610.邵平, 杨路明, 黄海滨, 等. 适于人脸检测的模板匹配快速算法[J]. 计算机应用, 2006, 26(7): 1607-1610.
[2] Liang L H, Ai H Z, He K Z, et al. Face detection based on the matching of multiple related templates[J]. Journal of Software, 2001, 12(1): 94-102.梁路宏, 艾海舟, 何克忠, 等. 基于多关联模板匹配的人脸检测[J]. 软件学报, 2001, 12(1): 94-102.
[3] Crosswhite N, Byrne J, Stauffer C, et al. Template adaptation for face verification and identification[J]. arXiv:1603.03958v3, 2016.
[4] Liu H Z, Yang S P, Yuan J Z, et al. Multi-scale face detec-tion based on single neural network[J]. Journal of Electro-nics and Information Technology, 2018, 40(11): 2598-2605.刘宏哲, 杨少鹏, 袁家政, 等. 基于单一神经网络的多尺度人脸检测[J]. 电子与信息学报, 2018, 40(11): 2598-2605.
[5] Tang X, Danial K D, He Z Q, et al. Pyramidbox: a context-assist single shot face detector[J]. arXiv:1803.07737, 2018.
[6] Wang R, He X S. Face detection algorithm based on golden section[J]. Journal of Beihua University (Natural Science), 2019, 7(4): 556-560.王茹, 贺兴时. 基于黄金分割的人脸检测算法[J]. 北华大学学报(自然科学版), 2019, 7（4）: 556-560.
[7] Viola P, Jones M J. Robust real-time face detection[C]//Pro-ceedings of the 8th International Conference on Computer Vision, Vancouver, Jul 7-14, 2001. Washington: IEEE Com-puter Society, 2001: 747.
[8] Yang S, Luo P, Chen C L, et al. Faceness-net: face detection through deep facial part responses[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(8): 1845-1859.
[9] Wu P L, Yang X, Mao B Y, et al. A perspective-independent method behavior recognition in depth video via temporal-spatial correlating[J]. Journal of Electronics and Information Technology, 2019, 41(4): 904-910.吴培良, 杨霄, 毛秉毅, 等. 一种视角无关的时空关联深度视频行为识别方法[J]. 电子与信息学报, 2019, 41(4): 904-910.
[10] Li J, Wang Y B, Wang C A, et al. DSFD: dual shot face detector[J]. arXiv:1810.10220, 2019.
[11] Shen J, Zuo X, Li J, et al. A novel pixel neighborhood diff-erential statistic feature for pedestrian and face detection[J]. Pattern Recognition, 2016, 63: 127-138.
[12] Chi C, Zhang S F, Xiang J L, et al. Selective refinement net-work for high performance face detection[J]. arXiv:1809. 02693v1, 2018.
[13] Ren S, He K, Girshick R B, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelli-gence, 2017, 39(6): 1137-1149.
[14] Xie X, Zhang L Q, Wang J. Application of residual network to infant crying recognition[J]. Journal of Electronics and Information Technology, 2019, 41(1): 233-239.谢湘, 张立强, 王晶. 残差网络在婴幼儿哭声识别中的应用[J]. 电子与信息学报, 2019, 41(1): 233-239.
[15] Ranjan R, Patel V M, Chellappa R. HyperFace: a deep multi-task learning framework for face detection, landmark locali-zation, pose estimation, and gender recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 41(1): 121-135.
[16] Wang L, Peng Y, Fu B F, et al. Face detection network based on multi-stage optimization[J]. Journal of Computer Appli-cations, 2019, 39(S1): 18-20.王丽, 彭垚, 傅冰飞, 等. 多级调优的人脸检测网络[J]. 计算机应用, 2019, 39(S1): 18-20.
[17] Lin T, Goyal P, Girshick R B, et al. Focal loss for dense object detection[J]. arXiv:1708.02002v2, 2017.
[18] Luo W, Li Y, Urtasun R, et al. Understanding the effective receptive field in deep convolutional neural networks[J].arXiv:1701.04128v2, 2017.
[19] Sun G H, Chen S R. An improved RefineDet multi-scale face detection method[J]. Application of Electronic Techni-que, 2019, 45(8): 34-39. 孙贵华, 陈淑荣. 一种改进的RefineDet多尺度人脸检测方法[J]. 电子技术应用, 2019, 45(8): 34-39.
[20] Luo M Z, Xiao Y W. Multi-scale face detection of full con-volution neural network[J]. Computer Engineering and App-lications, 2019, 55(5): 124-128.罗明柱, 肖业伟. 全卷积神经网络的多尺度人脸检测研究[J]. 计算机工程与应用, 2019, 55(5): 124-128.
[21] Wang X Y, Han C L, Hu X H. Densely connected convolu-tional networks face recognition algorithm based on wei-ghted feature fusion[J]. Journal of Frontiers of Computer Science and Technology, 2019, 13(7): 1195-1205.王小玉, 韩昌林, 胡鑫豪. 加权特征融合的密集连接网络人脸识别算法[J]. 计算机科学与探索, 2019, 13(7): 1195-1205.
[22] Jian V, Learned-Miller E G. FDDB: a benchmark for face detection in unconstrained settings: UM-CS-2010-009[R]. Amherst: University of Massachusetts, 2010.
[23] Zhu X, Raman D. Face detection, pose estimation, and land-mark localization in the wild[C]//Proceedings of the 2012 IEEE Conference on Computer Vision and Pattern Recogni-tion. Washington: IEEE Computer Society, 2012: 2879-2886.
[24] Yang S, Luo P, Loy C C, et al. WIDER FACE: a face detec-tion benchmark[C]//Proceedings of the 2016 IEEE Confer-ence on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Society, 2016: 3676-3684.