-
摘要:
针对大多数特征表示算法在挖掘高维数据内在结构时容易受到噪声的影响，以及特征学习与分类器设计割裂导致分类性能降低的问题，提出了一种新的基于特征表示的人脸识别方法，称为块对角投影表示（BDPR）学习。首先，利用样本信息对每类样本的编码系数施加一个加权矩阵，通过局部约束来加强表示系数之间的相似性，从而降低噪声对系数学习的影响，使所提方法能够更好地保持数据的局部结构。其次，为了实现数据与编码系数相关联，降低表示系数的学习难度，构造了块对角化判别约束项来学习一个判别投影，通过投影从低维数据中提取样本表示系数，使系数包含更多的样本间全局结构信息且具有更低的计算复杂度。最后，将系数学习和分类器学习整合到同一框架下，同时增大不同类别样本间的“标签距离”，采用迭代求解的方式交替更新判别投影和分类器，最终得到最适合当前表示特征的分类器，使得所提方法能自动完成分类。多个公开的人脸数据集上的实验结果表明：较之传统的协作表示分类和多个主流的子空间学习方法，所提方法均取得了更优的识别效果。
Abstract:Most feature representation algorithms are susceptible to noise when mining the internal structure of the high-dimensional data. Meanwhile, their feature learning and classifier design are separated, resulting in the limited classification performance in practice. Aimed at this issue, a new feature representation method, Block-Diagonal Projective Representation (BDPR), is proposed in this paper. First, a weighted matrix is imposed on the coding coefficients of samples over each class. By using such local constraints to enhance the similarity between the coefficients and reduce the impact of noise on coefficient learning, the proposed BDPR can well maintain the internal data structure. Second, to closely correlate the data with their coding coefficients and reduce the difficulty of learning the representation coefficients, we construct a block-diagonal constraint to learn a discriminative projection. In this way, the sample representation coefficients can be obtained in the low-dimensional projected subspace, which contains more global structure information between samples and enjoys lower computational complexity. Finally, the representation learning and classifier learning are integrated into the same framework. By increasing the "label distance" between samples of different classes, BDPR updates the discriminative projection and classifier in an iterative manner. In this way, the most suitable classifier can be found for the current optimal feature representation, making the proposed algorithm automatically realize the classification task. The results of experiments on multiple benchmark face datasets show that BDPR has achieved better recognition performance, compared to traditional collaborative representation based classification and several mainstream subspace learning algorithms.
-
图 1 CRC与BDPR在AR数据集上的样本编码系数对比
Figure 1. Comparison of sample coding coefficients learned by CRC and BDPR on AR dataset
图 2 不同数据集上BDPR的函数收敛曲线
Figure 2. Function convergence curves of BDPR method on different datasets
图 3 AR数据集上BDPR分类正确率随参数变化的三维柱状图
Figure 3. Three-dimensional histogram of classification accuracy of BDPR method changing with different parameters on AR dataset
表 1 实验采用的数据集信息
Table 1. Information of dataset used in experiment
数据集 类别 样本数 数据维度 BANCA 52 520 2 576 AR 50 1 300 2 200 YaleB 38 2 414 1 024 
下载: 导出CSV
表 2 各方法在不同数据集的最优分类正确率
Table 2. Highest classification accuracy of each method on different datasets
% 方法 BANCA AR YaleB 4 train 5 train 6 train 5 train 10 train 15 train 20 train 30 train 40 train CRC 46.03 45.92 47.74 68.79 81.85 87.58 92.67 88.69 96.02 PCA 48.40 52.42 56.63 33.73 48.09 57.60 61.90 68.31 72.19 LPP 41.12 38.78 42.84 38.00 48.44 61.20 83.97 88.69 90.83 MFA 73.01 77.62 80.67 86.49 94.43 97.24 93.97 96.04 97.13 CGDA 67.30 73.73 79.47 90.69 96.66 98.69 93.86 95.00 98.02 RLSDP 68.97 76.38 82.12 88.96 96.69 98.69 91.90 95.49 96.86 RLSL 72.18 77.50 82.13 84.97 95.63 97.87 93.47 96.85 98.15 BDPR 75.80 79.88 84.38 93.01 97.38 98.60 94.33 97.37 98.48
下载: 导出CSV
-
[1] TURK M, PENTLAND A. Eigenfaces for recognition[J]. Journal of Cognitive Neuroscience, 1991, 3(1): 71-86. doi: 10.1162/jocn.1991.3.1.71 [2] FISHER R A. The use of multiple measurements in taxonomic problems[J]. Annals of Human Genetics, 1936, 7(2): 179-188. http://ci.nii.ac.jp/naid/10008962389 [3] 蒋晨琛, 霍宏涛, 冯琦. 一种基于PCA的面向对象多尺度分割优化算法[J]. 北京航空航天大学学报, 2020, 46(6): 1192-1203. doi: 10.13700/j.bh.1001-5965.2019.0398JIANG C C, HUO H T, FENG Q. An object-oriented multi-scale segmentation optimization algorithm based on PCA[J]. Journal of Beijing University of Aeronautics and Astronautics, 2020, 46(6): 1192-1203(in Chinese). doi: 10.13700/j.bh.1001-5965.2019.0398 [4] 李可, 刘祎, 杜少毅, 等. 基于PCA和WPSVM的航天器电特性识别方法[J]. 北京航空航天大学学报, 2015, 41(7): 1177-1182. doi: 10.13700/j.bh.1001-5965.2014.0482LI K, LIU Y, DU S Y, et al. Spacecraft electrical characteristics identification method based on PCA feature extraction and WPSVM[J]. Journal of Beijing University of Aeronautics and Astronautics, 2015, 41(7): 1177-1182(in Chinese). doi: 10.13700/j.bh.1001-5965.2014.0482 [5] HE X F, YAN S C, HU Y X, et al. Face recognition using Laplacianfaces[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, 27(3): 328-340. doi: 10.1109/TPAMI.2005.55 [6] HE X F, CAI D, YAN S C, et al. Neighborhood preserving embedding[C]//Proceedings of the IEEE International Conference on Computer Vision. Piscataway: IEEE Press, 2005: 1208-1213. [7] SUGIYAMA M. Dimensionality reduction of multimodal labeled data by local Fisher discriminant analysis[J]. Journal of Machine Learning Research, 2007, 8(1): 1027-1061. doi: 10.5555/1248659.1248694 [8] YAN S C, XU D, ZHANG B Y, et al. Graph embedding and extensions: A general framework for dimensionality reduction[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, 29(1): 40-51. doi: 10.1109/TPAMI.2007.250598 [9] BENGIO Y, COURVILLE A, VINCENT P, et al. Representation learning: A review and new perspectives[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(8): 1798-1828. doi: 10.1109/TPAMI.2013.50 [10] HUANG K, AVIYENTE S. Sparse representation for signal classification[C]//Advances in Neural Information Processing Systems. Cambridge: MIT Press, 2007: 609-616. [11] WRIGHT J, YANG A Y, GANESH A, et al. Robust face recognition via sparse representation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(2): 210-227. doi: 10.1109/TPAMI.2008.79 [12] QIAO L S, CHEN S C, TAN X Y. Sparsity preserving projections with applications to face recognition[J]. Pattern Recognition, 2010, 43(1): 331-341. doi: 10.1016/j.patcog.2009.05.005 [13] LY N H, DU Q, FOWLER J E. Sparse graph-based discriminant analysis for hyperspectral imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2014, 52(7): 3872-3884. doi: 10.1109/TGRS.2013.2277251 [14] ZHANG L, YANG M, FENG X C, et al. Sparse representation or collaborative representation: Which helps face recognition [C]//Proceedings of the IEEE International Conference on Computer Vision. Piscataway: IEEE Press, 2011: 471-478. [15] LY N H, DU Q, FOWLER J E. Collaborative graph-based discriminant analysis for hyperspectral imagery[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(6): 2688-2696. doi: 10.1109/JSTARS.2014.2315786 [16] YANG W K, SUN C Y, ZHENG W M. A regularized least square based discriminative projections for feature extraction[J]. Neurocomputing, 2016, 175: 198-205. doi: 10.1016/j.neucom.2015.10.049 [17] ZHENG C Y, WANG N N. Collaborative representation with k-nearest classes for classification[J]. Pattern Recognition Letters, 2019, 117: 30-36. doi: 10.1016/j.patrec.2018.11.005 [18] LI W, DU Q. Collaborative representation for hyperspectral anomaly detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015, 53(3): 1463-1474. doi: 10.1109/TGRS.2014.2343955 [19] KAYA M, BILGE H Ş. Deep metric learning: A survey[J]. Symmetry, 2019, 11(9): 1066. doi: 10.3390/sym11091066 [20] ZHONG G Q, WANG L N, LING X, et al. An overview on data representation learning: From traditional feature learning to recent deep learning[J]. The Journal of Finance and Data Science, 2016, 2(4): 265-278. doi: 10.1016/j.jfds.2017.05.001 [21] ZHOU X Z, JIN K, XU M, et al. Learning deep compact similarity metric for kinship verification from face images[J]. Information Fusion, 2019, 48: 84-94. doi: 10.1016/j.inffus.2018.07.011 [22] CHU R H, SUN Y F, LI Y D, et al. Vehicle re-identification with viewpoint-aware metric learning[C]//Proceedings of the IEEE International Conference on Computer Vision. Piscataway: IEEE Press, 2019: 8282-8291. [23] 车畅畅, 王华伟, 倪晓梅, 等. 基于深度学习的航空发动机故障融合诊断[J]. 北京航空航天大学学报, 2018, 44(3): 621-628. doi: 10.13700/j.bh.1001-5965.2017.0197CHE C C, WANG H W, NI X M, et al. Fault fusion diagnosis of aero-engine based on deep learning[J]. Journal of Beijing University of Aeronautics and Astronautics, 2018, 44(3): 621-628(in Chinese). doi: 10.13700/j.bh.1001-5965.2017.0197 [24] WEN J, ZHANG B, XU Y, et al. Adaptive weighted nonnegative low-rank representation[J]. Pattern Recognition, 2018, 81: 326-340. doi: 10.1016/j.patcog.2018.04.004 [25] FANG X Z, TENG S H, LAI Z H, et al. Robust latent subspace learning for image classification[J]. IEEE Transactions on Neural Networks, 2018, 29(6): 2502-2515. doi: 10.1109/TNNLS.2017.2693221 [26] BAILLY-BAILLIÉRE E, BENGIO S, BIMBOT F, et al. The BANCA database and evaluation protocol[C]//International Conference on Audio- and Video-based Biometric Person Authentication. Berlin: Springer, 2003: 625-638. [27] GEORGHIADES A S, BELHUMEUR P N, KRIEGMAN D J. From few to many: Illumination cone models for face recognition under variable lighting and pose[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2001, 23(6): 643-660. doi: 10.1109/34.927464 -
下载:
点击查看大图
计量
- 文章访问数: 579
- HTML全文浏览量: 85
- PDF下载量: 78
- 被引次数: 0
