摘要
针对字典学习和分类过程所采用的表示系数l1范数稀疏约束求解过程代价过高的问题,同时为获取更有效的表情相关特征来进行字典学习,提出一种结合分块LBP特征与投影字典对学习的表情识别方法。提取图像的分块LBP特征替代原始数据样本,用来训练和测试。学习一个分析字典和一个综合字典,分析字典可求得表示系数,综合字典具备重构能力。利用综合字典和分析字典求出各类别的重构误差进行分类,从而实现人脸表情识别。在JAFFE和CK+数据库上的实验结果表明,与其他方法相比,所提出的方法不仅可以大大降低训练和测试阶段的时间复杂度,而且可以在分类任务中达到更高的识别率。
Aiming at the excessive costs of the solving process of the l1-norm sparsity constraint on the representation coefficients, which is used in dictionary learning and classification, a new method is proposed, combining blocking LBP feature and projective dictionary pair learning to recognize the facial expression, at the meantime, to obtain more efficient facial expression-related features for dictionary learning. The blocking LBP feature of the image is extracted to replace the original data sample for training and testing. Then, this paper learns an analysis dictionary and a synthesis dictionary, and the former can obtain the representation coefficient while the latter possesses refactoring capability. Both synthesis dictionary and analysis dictionary are used to calculate the reconstruction error of each category and classify them, and facial expression recognition is achieved consequently. According to the experimental results in the JAFFE and CK+databases,compared with other methods, this proposed method not only can greatly reduce the time complexity of training and testing phases, but also can achieve higher recognition rate in classification tasks.
引文
[1]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.
[2]Lee S H,Ro Y M.Partial matching of facial expression sequence using over-complete transition dictionary for emotion recognition[J].IEEE Transactions on Affective Computing,2016,7(4):389-408.
[3]Sun Zhe,Hu Zhengping,Wang Meng,et al.Discriminative feature learning-based pixel difference representation for facial expression recognition[J].IET Computer Vision,2017,11(8):675-682.
[4]Wang Hui,Gao Jing,Tong Lifeng,et al.Facial expression recognition based on PHOG feature and sparse representation[C]//2016 35th Chinese Control Conference,2016:3869-3874.
[5]Jiang Xingguo,Feng Bin,Jin Liangnian.Facial expression recognition via sparse representation using positive and reverse templates[J].IET Image Processing,2016,10(8):616-623.
[6]Guo Yimo,Zhao Guoying,Pietik?inen M.Dynamic facial expression recognition with atlas construction and sparse representation[J].IEEE Journals&Magazines,2016,25(5):1977-1992.
[7]Jiang Zhuolin,Lin Zhe,Davis L S.Label consistent K-SVD:learning a discriminative dictionary for recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(11):2651-2664.
[8]Aharon M,Elad M,Bruckstein A.rmK-SVD:an algorithm for designing overcomplete dictionaries for sparse representation[J].IEEE Transactions on Signal Processing,2006,54(11):4311-4322.
[9]Mairal J,Bach F,Ponce J,et al.Supervised dictionary learning[C]//Conference and Workshop on Neural Information Processing Systems,2008:1-8.
[10]Yang Meng,Zhang Lei,Feng Xiangchu,et al.Fisher discrimination dictionary learning for sparse representation[C]//2011 International Conference on Computer Vision,2011:543-550.
[11]Wang Zhaowen,Yang Jianchao,Nasrabadi N,et al.A maxmargin perspective on sparse representation-based classification[C]//2013 IEEE International Conference on Computer Vision,2013:1217-1224.
[12]Sch?lkopf B,Platt J,Hofmann T.Efficient sparse coding algorithms[C]//Proceedings of the 2006 Conference on Advances in Neural Information Processing Systems,2007:801-808.
[13]Hale E T,Yin W,Zhang Y.Fixed-point continuation for l1-minimization methodology and conver-gence[J].SLAMJournal on Optimization,2008,19(3):1107-1130.
[14]Soltanolkotabi B M,Elhamifar E,Cand`es E J.Robust subspace clustering[J].The Annals of Statistics,2014,42(2):669-699.
[15]Gu Shuhang,Zhang Lei,Zuo Wangmeng,et al.Projective dictionary pair learning for pattern classification[C]//Advances in Neural Information Processing Systems,2014:793-801.