Abstract
Since the initial weight matrix between the last hidden layer of the network and the classification layer is usually generated randomly, the weight matrix does not have the discrimination ability to accurately classify the facial expression recognition, which results in that the features obtained by the weight matrix mapping cannot be guaranteed to be suitable for classification tasks. To solve this problem, a novel linear discriminant deep belief network is proposed in this paper. Firstly, the traditional linear discriminant analysis method is improved, and a new type of inter-class dispersion matrix is designed to solve the rank limitation problem in the traditional Linear Discriminant Analysis Method (LDA). Then, the weight matrix between the last hidden layer and the classification layer of the deep belief network is initialized by the improved linear discriminant analysis method, so that the network is more suitable for the classification task. In the experiments, our proposed deep network obtains respectively the recognition rates of 78.26% and 94.48% on the JAFFE database and the Extended Cohn-Kanade database. In addition, using our proposed algorithm for aggregating linear discriminant analysis into a deep belief network, we were able to produce an accuracy of 81.03% on the challenge test set.
Similar content being viewed by others
References
Kotsia I, Pitas I (2007) Facial expression recognition in image sequences using geometric deformation features and support vector machines[J]. IEEE Trans Image Process 16(1):172–187
Calder AJ (2003) Facial expression recognition across the life span[J]. Neuropsychologia 41(2):195–202
Calder AJ, Young AW (2005) Understanding the recognition of facial identity and facial expression[J]. Nat Rev Neurosci 6(8):641–651
Sprengelmeyer R, Young AW et al (2003) Facial expression recognition in people with medicated and unmedicated Parkinson's disease[J]. Neuropsychologia 41(8):1047–1057
Gu W, Xiang C, Venkatesh YV et al (2012) Facial expression recognition using radial encoding of local Gabor features and classifier synthesis[J]. Pattern Recogn 45(1):80–91
Kim BK, Roh J, Dong SY et al (2016) Hierarchical committee of deep convolutional neural networks for robust facial expression recognition[J]. Journal on Multimodal User Interfaces 10(2):173–189
Wood A, Rychlowska M, Korb S et al (2016) Fashioning the face: sensorimotor simulation contributes to facial expression recognition[J]. Trends Cogn Sci 20(3):227–240
Lopes AT, Aguiar ED, Souza AFD et al (2016) Facial expression recognition with convolutional neural networks: coping with few data and the training sample order[J]. Pattern Recogn 61:610–628
Zhao K, Chu WS, Fernando DLT, et al. (2016) Joint patch and multi-label learning for facial action unit and holistic expression recognition[J]. IEEE Trans Image Process, 1–1
Zhang T, Zheng W, Cui Z et al (2016) A deep neural network-driven feature learning method for multi-view facial expression recognition[J]. IEEE T Multimedia 18(12):2528–2536
Uçar A, Demir Y, Güzeliş C (2016) A new facial expression recognition based on curvelet transform and online sequential extreme learning machine initialized with spherical clustering[J]. Neural Comput & Applic 27(1):131–142
Zhen Q, Huang D, Wang Y et al (2016) Muscular movement model-based automatic 3D/4D facial expression recognition[J]. IEEE T Multimedia 18(7):1438–1450
Sun B, Li L, Zhou G et al (2016) Facial expression recognition in the wild based on multimodal texture features[J]. J Electron Imaging 25(6):061407
Yanpeng L, Yibin L, Xin M et al (2017) Facial expression recognition with fusion features extracted from salient facial areas[J]. Sensors 17(4):712
Wang Z, Ruan Q, An G (2016) Facial expression recognition using sparse local fisher discriminant analysis[J]. Neurocomputing 174(174):756–766
Guo Y, Zhao G, Pietikainen M (2016) Dynamic facial expression recognition with atlas construction and sparse representation.[J]. IEEE Trans Image Process 25(5):1977–1992
Zheng H, Geng X, Tao D, et al. (2015) A multi-task model for simultaneous face identification and facial expression recognition[J]. Neurocomputing, S0925231215010097
Kamarol SKA, Jaward MH, Parkkinen J et al (2016) Spatiotemporal feature extraction for facial expression recognition.[J]. IET Image Process 10(7):534–541
Zen G, Porzi L, Sangineto E, et al. (2016) Learning personalized models for facial expression analysis and gesture recognition[J]. IEEE T Multimedia, 1–1
Li H, Sun J, Xu Z, et al. (2017) Multimodal 2D+3D facial expression recognition with deep fusion convolutional neural network[J]. IEEE T Multimedia, 1–1
Yan J, Zheng W, Xu Q, et al. (2016) Sparse kernel reduced-rank regression for bimodal emotion recognition from facial expression and speech[J]. IEEE T Multimedia, 1–1
Chakraborty BK, Kumar S, Bhuyan MK (2016) Extraction of informative regions of a face for facial expression recognition[J]. IET Comput Vis 10(6):567–576
Uddin M Z, Hassan M M, Almogren A, et al. (2017) A facial expression recognition system using robust face features from depth videos and deep learning ☆[J]. Comput Electr Eng, 63
Siddiqi MH, Alam MGR, Hong CS et al (2016) A novel maximum entropy Markov model for human facial expression recognition[J]. PLoS One 11(9):e0162702
Chen J, Chen Z, Chi Z, et al. (2018) Facial expression recognition in video with multiple feature fusion[J]. IEEE Trans Affect Comput, PP(99):1–1
Lee SH, Ro YM (2017) Partial matching of facial expression sequence using over-complete transition dictionary for emotion recognition[J]. IEEE Trans Affect Comput 7(4):389–408
Wen G, Hou Z, Li H et al (2017) Ensemble of deep neural networks with probability-based fusion for facial expression recognition[J]. Cogn Comput 9(5):597–610
Wen G et al (2017) Ensemble of deep neural networks with probability-based fusion for facial expression recognition. Cogn Comput 9(5):597–610
Tsai HH, Chang YC (2017) "Facial expression recognition using a combination of multiple facial features and support vector machine." Soft Comput
Kamarol SKA et al (2017) Joint facial expression recognition and intensity estimation based on weighted votes of image sequences. Pattern Recogn Lett 92:25–32
Yan X, Young AW, Andrews TJ (2016) "Differences in holistic processing do not explain cultural differences in the recognition of facial expression." Q J Exp Psychol :1–15
Liu M et al (2015) Learning expressionlets on spatio-temporal manifold for dynamic facial expression recognition. IEEE Trans Image Process 25(12):5920–5932
Yan H (2016) "Transfer subspace learning for cross-dataset facial expression recognition." Neurocomputing :S0925231216304623
Danelakis A, Theoharis T, Pratikakis I (2016) A spatio-temporal wavelet-based descriptor for dynamic 3D facial expression retrieval and recognition. Vis Comput 32(6–8):1001–1011
Jaina DK, Zhanga Z, Huanga K (2017) Multi angle optimal pattern-based deep learning for automatic facial expression recognition. Pattern Recogn Lett
Mlakar U et al (2017) Multi-objective differential evolution for feature selection in facial expression recognition systems. Expert Syst Appl 89:129–137
Funding
This work was financially supported by the Excellent Specialties Program Development of Jiangsu Higher Education Institutions (PPZY2015C240).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Xiao, Y., Wang, D. & Hou, L. Unsupervised emotion recognition algorithm based on improved deep belief model in combination with probabilistic linear discriminant analysis. Pers Ubiquit Comput 23, 553–562 (2019). https://doi.org/10.1007/s00779-019-01235-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00779-019-01235-y