基于人脸图像的二阶段性别分类算法

doi:10.3778/j.issn.1673-9418.2006071

摘要/Abstract

摘要：

许多现实场景要求准确的脸部性别识别。深度卷积神经网络在正常状况下取得好的准确率，适用于大规模分类任务，但存在模型可解释性差、易丢失细节信息等问题，并且光照、姿势、表情等因素带来的不确定性会导致分类准确率较低。提出一种基于阴影集的二级分类模型。采用深度卷积神经网络对大规模图像集进行一阶段分类；结合阴影集理论，将图像分类结果划分为接收域、拒绝域和不确定域，得到不确定的脸部图像集，用传统方法进行二阶段分类。在LFW数据集和Adience数据集下，与现有先进算法相比，所提方法能有效地提高总体分类的准确率。

关键词: 性别识别, 卷积神经网络（CNN）, 阴影集, 不确定域

Abstract:

Many scenes need facial gender identification with good accuracy. Deep convolutional neural networks (CNN) with large set of training data normally give good accuracy, however, to achieve good accuracy with uncertain training data is a difficult task due to their lower explanation and potential information loss. Moreover, the uncertain-ties resulted from illumination, postures and facial expressions can lead to low accuracy of the classification. In this paper, a shadowed sets based two-level model for gender classification is proposed to address the problem. Deep convolutional neural networks are used as one-level classifier. Combining the concept of shadowed sets, one-level classification results are divided into three parts: accept domain, reject domain and uncertain domain. Samples in the uncertain domain are selected as uncertain facial images for two-level reclassification. Results show that the proposed method can further improve the accuracy compared with several existing state-of-the-art methods on the LFW dataset and Adience dataset.

Key words: gender identification, convolutional neural networks (CNN), shadowed sets, uncertain domain

杨晨旭, 蔡克参, 张红云, 苗夺谦. 基于人脸图像的二阶段性别分类算法[J]. 计算机科学与探索, 2021, 15(3): 524-532.

YANG Chenxu, CAI Kecan, ZHANG Hongyun, MIAO Duoqian. Facial Image Based Two-Level Model for Gender Classification[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(3): 524-532.

参考文献

[1] WANG Q W, YING Z L. A face detection algorithm based on Haar-Like T features[J]. Pattern Recognition & Artificial Intelligence, 2015, 28(1): 35-41.
王庆伟, 应自炉. 一种基于Haar-Like T特征的人脸检测算法[J]. 模式识别与人工智能, 2015, 28(1): 35-41.
[2] BHATTACHARYYA A, SAINI R, ROY P P, et al. Recogni-zing gender from human facial regions using genetic algorithm[J]. Soft Computing, 2019, 23(17): 8085-8100.
[3] ZHOU Z, GAN Y, SUN F J. Research of face recognition method based on Gabor and SIFT features[J]. Electronic Tech-nology, 2019, 32(4): 5-9.
周柱, 甘屹, 孙福佳. 一种融合Gabor+SIFT特征的人脸识别算法[J]. 电子科技, 2019, 32(4): 5-9.
[4] YE F, YE X Y, LUO X H, et al. Face recognition based on SA-CRC of multi-layer AR-LBP and WLD feature fusion[J]. Computer Engineering and Applications, 2019, 55(14): 134-141.
叶枫, 叶学义, 罗宵晗, 等. 多层AR-LBP与WLD特征融合的SA-CRC人脸识别[J]. 计算机工程与应用, 2019, 55(14): 134-141.
[5] TOEWS M, ARBEL T. Detection, localization, and sex cla-ssification of faces from arbitrary viewpoints and under occ-lusion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(9): 1567-1581.
[6] CORTES C, VLADIMIR V. Support-vector networks[J]. Mac-hine Learning, 1995, 20(3): 273-297.
[7] MOGHADDAM B, YANG M H. Learning gender with sup-port faces[J]. IEEE Transactions on Pattern Analysis and Mac-hine Intelligence, 2002, 24(5): 707-711.
[8] KEKRE H B, SARODE T K, SAVE J K. Performance com-parison of wavelet transforms, generated from orthogonal transforms, in classification of image database[J]. International Journal of Advanced Research in Computer Science and Software Engineering, 2012, 2(12): 10-18.
[9] TIAN Q, TAL A, JAMES J C. Fisher pruning of deep nets for facial trait classification[J]. arXiv:1803.08134, 2018.
[10] MATHIVANAN P, POORNIMA K. Biometric authentication for gender classification techniques: a review[J]. Journal of the Institution of Engineers, 2018, 99(1): 79-85.
[11] LU Y, LAI Z, XU Y, et al. Low-rank preserving projections[J]. IEEE Transactions on Cybernetics, 2016, 46(8): 1900-1913.
[12] LU Y, LAI Z, LI X, et al. Low-rank 2-D neighborhood pre-serving projection for enhanced robust image representation[J]. IEEE Transactions on Cybernetics, 2019, 49(5): 1859-1872.
[13] LUY W, WONG K W, LAI Z H, et al. Robust flexible pre-serving embedding[J]. IEEE Transactions on Cybernetics, 2020, 50(10): 4495-4507.
[14] BALCI K, ATALAY V. PCA for gender estimation: which eigenvectors contribute?[C]//Proceedings of the 16th Inter-national Conference on Pattern Recognition, Quebec City, Aug 11-15, 2002. Washington: IEEE Computer Society, 2002: 30363.
[15] ZHANG T, TANG Y Y, FANG B, et al. Face recognition under varying illumination using gradientfaces[J]. IEEE Tran-sactions on Image Processing, 2009, 18(11): 2599-2606.
[16] RAI P, KHANNA P. A gender classification system robust to occlusion using Gabor features based(2D)2 PCA[J]. Journal of Visual Communication and Image Representation, 2014, 25(5): 1118-1129.
[17] WANG D, MAO Z Y, WU M D. Outlier detection algorithm on shadowed sets clustering[J]. Journal of Frontiers of Com-puter Science and Technology, 2012, 6(11): 985-993.
王丹, 毛紫阳, 吴孟达. 融合Shadowed Sets聚类的离群点检测算法[J]. 计算机科学与探索, 2012, 6(11): 985-993.
[18] ZHOU Y, QIAN X, WANG Z Q. Performance improvement of extension neural network using data selection method based on shadowed sets[J]. Journal of Beijing University of Technology, 2013, 39(3): 430-437.
周玉, 钱旭, 王自强. 基于阴影集数据选择的可拓神经网络性能改进[J]. 北京工业大学学报, 2013, 39(3): 430-437.
[19] SU X H, ZHAO L L, XIE L, et al. Shadowed sets-based sample selection method for fuzzy support vector machine[J]. Journal of Harbin University of Technology, 2012, 44(9): 78-84.
苏小红, 赵玲玲, 谢琳, 等. 阴影集的模糊支持向量机样本选择方法[J]. 哈尔滨工业大学学报, 2012, 44(9): 78-84.
[20] MITRA S, KUNDU P P. Satellite image segmentation with shadowed C-Means[J]. Information Sciences, 2011, 181(17): 3601-3613.
[21] ZHANG H Y, ZHANG T, PEDRYCZ W, et al. Improved adaptive image retrieval with the use of shadowed sets[J]. Pattern Recognition, 2019, 90: 390-403.
[22] GOEL A, VISHWAKARMA V P. Gender classification using KPCA and SVM[C]//Proceedings of the 2016 IEEE International Conference on Recent Trends in Electronics, Information & Communication Technology, Bangalore, May 20-21, 2016. Piscataway: IEEE, 2016: 291-295.
[23] KEKRE H B, SARODE T K, SAVE J K. Gender classi-fication of human faces using class based PCA[J]. Internati-onal Journal of Scientific and Research Publications, 2014, 4(2): 1-9.
[24] DHOMNE A, KUMAR R, BHAN V. Gender recognition through face using deep learning[J]. Procedia Computer Science, 2018, 132: 2-10.
[25] NG C B, TAY Y H, GOI B M. Training strategy for convo-lutional neural networks in pedestrian gender classification[C]//Proceedings of the 2nd International Workshop on Pattern Recognition, Singapore, May 1-3, 2017. Bellingham: SPIE, 2017: 1-5.
[26] LAPUSCHKIN S, BINDER A, MüLLER K R, et al. Und-erstanding and comparing deep neural networks for age and gender classification[C]//Proceedings of the 2017 IEEE Inter-national Conference on Computer Vision Workshops, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 1629-1638.
[27] PEDRYCZ W. Shadowed sets: representing and processing fuzzy sets[J]. IEEE Transactions on Systems, Man and Cyber-netics, Part B, 1998, 28(1): 103-109.
[28] PEDRYCZ W. From fuzzy sets to shadowed sets: interpre-tation and computing[J]. International Journal of Intelligent Systems, 2009, 24(1): 48-61.
[29] QIAN Y H, LIANG J Y, DANG C Y. Knowledge structure, knowledge granulation and knowledge distance in a know-ledge base[J]. International Journal of Approximate Reason-ing, 2009, 50(1): 174-188.
[30] CATTANEO G, CIUCCI D. An algebraic approach to shad-owed sets[J]. Electronic Notes in Theoretical Computer Science, 2003, 82(4): 64-75.
[31] PEDRYCZ W. Interpretation of clusters in the framework of shadowed sets[J]. Pattern Recognition Letters, 2005, 26(15): 2439-2449.
[32] ZHOU J, PEDRYCZ W, MIAO D. Shadowed sets in the characterization of rough-fuzzy clustering[J]. Pattern Recog-nition, 2011, 44(8): 1738-1749.
[33] ZHOU Y, ZHU A F, ZHOU L, et al. Sample data selection method for neural network classifiers[J]. Journal of Huazhong University of Science and Technology (Natural Science Edition), 2012, 40(6): 39-43.
周玉，朱安福，周林, 等.一种神经网络分类器样本数据选择方法[J]. 华中科技大学学报(自然科学版), 2012, 40(6): 39-43.
[34] ZHANG T, LI Y J, HU H H, et al. A gender classification model based on cross-connected convolutional neural net-works[J]. Journal of Automation, 2016, 42(6): 858-865.
张婷, 李玉鑑, 胡海鹤, 等. 基于跨连卷积神经网络的性别分类模型[J]. 自动化学报, 2016, 42(6): 858-865.
[35] TIAN Q, ARBEL T, CLARK J. Deep LDA-pruned nets for efficient facial gender classification[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 512-521.
[36] SHI X C, ZHOU Y T, CHI Y. Face gender recognition based on multi-layer feature fusion convolution neural network with adjustable supervisory function[J]. Application Research of Computers, 2019, 36(3): 940-944.
石学超, 周亚同, 池越. 基于多层特征融合可调监督函数卷积神经网络的人脸性别识别[J]. 计算机应用研究, 2019, 36(3): 940-944.
[37] ASTHANA A, ZAFEIRIOU S, CHENG S Y, et al. Increme-ntal face alignment in the wild[C]//Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recog-nition, Columbus, Jun 23-28, 2014. Washington: IEEE Com-puter Society, 2014: 1859-1866.
[38] GONZALEZ-SOSA E, FIéRREZ J, VERA-RODRíGUEZ R, et al. Facial soft biometrics for recognition in the wild: recent works, annotation and COTS evaluation[J]. IEEE Trans-actions on Information Forensics and Security, 2018, 13(8): 2001-2014.
[39] GAJJAR V. 2B3C: 2 box 3 crop of facial image for gender classification with convolutional networks[J]. arXiv:1803.02181, 2018.
[40] CHEN J N, LI S B, GAO Z, et al. Age and gender classi-fication using improved convolutional neural networks[J]. Computer Engineering and Applications, 2018, 54(16): 135-139.
陈济楠, 李少波, 高宗, 等. 基于改进CNN的年龄和性别识别[J]. 计算机工程与应用, 2018, 54(16): 135-139.
[41] MITTAL S. Gender recognition from facial images using convolutional neural network[C]//Proceedings of the 2019 5th International Conference on Image Information Processing, Solan, Nov 15-17, 2019. Piscataway: IEEE, 2019: 347-352.
[42] AFIFI M, ABDELHAMED A. AFIF4: deep gender classifi-cation based on AdaBoost-based fusion of isolated facial features and foggy faces[J]. Journal of Visual Communica-tion and Image Representation, 2019, 62: 77-86.
[43] ZHOU Y, NI H, REN F, et al. Face and gender recognition system based on convolutional neural networks[C]//Procee-dings of the 2019 IEEE International Conference on Mecha-tronics and Automation, Tianjin, Aug 4-8, 2019. Piscataway: IEEE, 2019.
[44] ZHOU Y Y, QIN K. Improved VGG-Net for increasing pre-cision of age and gender prediction[J]. Computer Engineer-ing and Applications, 2019, 55(18): 173-179.
周玉阳，秦科. 改进的VGG网络可提升年龄与性别预测准确率[J]. 计算机工程与应用, 2019, 55(18): 173-179.
[45] LEVI G, HASSNCER T. Age and gender classification using convolutional neural networks[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition, Boston, Jun 7-12, 2015. Washington: IEEE Computer Society, 2015: 34-42.
[46] WOLFSHAAR J V D, KARAABA M F, WIERING M A. Deep convolutional neural networks and support vector mac-hines for gender recognition[C]//Proceedings of the 2015 IEEE Symposium Series on Computational Intelligence, Cape Town, Dec 7-10, 2015. Piscataway: IEEE, 2015: 188-195.
[47] ÖZBULAK G, AYTAR Y, EKENEL H K. How transferable are CNN-based features for age and gender classification?[C]//Proceedings of the 2016 International Conference of the Bio-metrics Special Interest Group, Damrstadt, Sep 21-23, 2016. Piscataway: IEEE, 2016: 39-50.