唇动视觉信息是说话内容的重要载体。受嘴唇外观、背景信息和说话习惯等影响，即使说话者说相同的内容，唇动视觉信息也会相差很大。为解决唇语视觉信息多样性的问题，提出一种基于long short-term memory（LSTM）的新的唇语识别方法。以往大多数的方法从嘴唇外表信息入手。本方法用嘴唇关键点坐标描述嘴唇形变信息作为唇语视频的特征，它具有类内一致性和类间区分性的特点。然后利用LSTM对特征进行时序编码，它能学习具有区分性和泛化性的空间-时序特征。在公开的唇语数据集GRID、MIRACL-VC和OuluVS上对本方法做了针对分割的单词或短语的说话者独立的唇语识别评估。在GRID和MIRACL-VC上，本方法的准确率比传统方法至少高30%；在OuluVS上，本方法的准确率接近于最优结果。以上实验结果表明，本文提出的基于LSTM的唇语识别方法有效地解决了唇语视觉信息多样性的问题。
Visual speech information is the important carrier of conversation. However, visual speech informations from different speakers are different due to various appearances of lips, various backgrounds, and various talking ways even the content of the conversation is the same. To address the problem of variety of visual speech information, we propose a new approach for lip-reading recognition based on long short-term memory (LSTM). We compute the positions of lip landmarks which describe the dynamic information of the shape as the features of the lip-reading video, and it has the characteristics of within-class consistency and between-class distinctiveness. Then we use LSTM to encode temporal information, and it learns spatio-temporal features which have the ability of discrimination and generalization. Our approach is evaluated on three public databases (GRID, MRIALC, and OuluVS) for lip-reading recognition of isolated words or phrases in speaker independent experiments. On GRID and MRIALC, the accuracy of our approach is more than 30% highter than that of the conventional approach. On OuluVS, the accuracy of our approach is comparable to state of the art. The experiment results indicate that our lip-reading recognition approach solves the problem of variety of visual speech information effectively.
[1] McGurk H, MacDonald J. Hearing lips and seeing voices[J]. Nature, 1976, 264:746-748.
[2] Graves A. Supervised sequence labelling with recurrent neural networks[M]. Berlin:Springer Berlin Heidelberg, 2012.
[3] Zhou Z, Zhao G, Hong X, et al. A review of recent advances in visual speech decoding[J]. Image & Vision Computing, 2014, 32:590-605.
[4] Zhao G, Barnard M, Pietikainen M. Lipreading With Local Spatiotemporal Descriptors[J].IEEE transactions on multimedia, 2009, 11:1254-1265.
[5] Zhou Z, Zhao G, Pietikainen M, Towards a practical lipreading system[C]//Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference on. IEEE, 2011:137-144.
[6] Bakry A, Elgammal A. Mkpls:manifold kernel partial least squares for lipreading and speaker identification[C]//Computer Vision and Pattern Recognition (CVPR), 2013 IEEE Conference on. IEEE, 2013:684-691.
[7] Lan Y, Harvey R, Theobald B J, et al. Comparing visual features for lipreading[C]//International Conference on Auditory-Visual Speech Processing. 2009:102-106.
[8] Pei Y, Kim T K, Zha H. Unsupervised random forest manifold alignment for lipreading[C]//Proceedings of the IEEE International Conference on Computer Vision. 2013:129-136.
[9] Rekik A, Ben-Hamadou A, Mahdi W. A new visual speech recognition approach for RGB-D cameras[C]//International Conference Image Analysis and Recognition. Springer International Publishing, 2014:21-28.
[10] Noda K, Yamaguchi Y, Nakadai K, et al. Lipreading using convolutional neural network[J]. Interspeech, 2014:1149-1153.
[11] Graves A, Jaitly N. Towards end-to-end speech recognition with recurrent neural networks[C]//Proceedings of the 31st International Conference on Machine Learning, 2014:1764-1772.
[12] Sutskever I, Vinyals O, Le Q V. Sequence to sequence learning with neural networks[C]//Advances in Neural Information Processing Systems, 2014:3104-3112.
[13] Donahue J, Hendricks L A, Guadarrama S, et al. Long-term recurrent convolutional networks for visual recognition and description[C]//Computer Vision and Pattern Recognition (CVPR), 2015 IEEE Conference on. IEEE, 2015:2625-2634.
[14] Wand M, Koutn J. Lipreading with long short-term memory[C]//Acoustics, Speech and Signal Processing (ICASSP), 2016 IEEE International Conference on. IEEE, 2016:6115-6119.
[15] Simonyan K, Zisserman A. Two-stream convolutional networks for action recognition in videos[C]//Advances in Neural Information Processing Systems, 2014:568-576.
[16] Cooke M, Barker J, Cunningham S, et al. An audio-visual corpus for speech perception and automatic speech recognition[J]. The Journal of the Acoustical Society of America, 2006, 120:2421-2424.
[17] Xiong X, Torre F D L. Supervised descent method and its applications to face alignment[C]//Computer Vision and Pattern Recognition (CVPR), 2013 IEEE Conference on. IEEE, 2013:532-539.
[18] Hochreiter S, Schmidhuber J. Long short-term memory[J]. Neural computation, 1997, 9:1735-1780.
[19] Yarin G, Ghahramani Z. A theoretically grounded application of dropout in recurrent neural networks[C]//Advances in Neural Information Processing Systems, 2016:1019-1027.
