隨著人機互動的發展，用各種辨識來做為與機器操作互動的依據，而表情是除了言語之外最能表達人類情緒的方法，因此表情辨識成為一個相當重要的議題。
這篇論文提出使用兩個卷積神經網路應用於表情辨識並將其結合，第一個為原圖資料訓練，第二個為光流資料訓練，合併後做最後的辨識使用，加上擴增資料來提升整體辨識率。資料庫則是使用cohn-kanade公用表情影像資料庫進行實驗，並證明我們的方法可以正確的辨識表情。

With the development of human-computer interaction, all kinds of recognition are used as the basis for interaction with machine operations, and facial expression is the most effective way to express human emotions in addition to speech. Therefore, facial expression recognition becomes a very important issue.
This paper proposes the use of two convolutional neural networks for face recognition and combining them. The first is the training of original data, and the second is the training of optical flow data. After combining, it is used for the final recognition and increasing data to increase overall recognition rate. We use the cohn-kanade expression database to perform experiments and proves our method can correctly identify expressions.

第1章 緒論 1
1.1 研究動機與目的 1
1.2 系統流程 2
1.3 章節架構 3
第2章　相關文獻探討 5
2.1 卷積神經網絡 5
2.1 表情辨識 6
第3章　雙串流卷積神經網路用於表情辨識 9
3.1.1 光流法(optical flow) 10
3.1.2 場景變化偵測(Scenechange detection) 11
3.3　3D卷積神經網路網路與雙串流架構 14
第4章　實驗結果 17
4.1　實驗環境與資料庫 17
第5章　結論與未來研究 23

[1] Yann LeCun, Leon Bottou, Yoshua Bengio, et al. Gradient-based learning applied to document recognition. Proceedings of the IEEE,1998,86(11): 2278-2324
[2] Ji S, Xu W, Yang M, et al. 3D convolutional neural networks for human action recognition[J]. IEEE transactions on pattern analysis and machine intelligence, 2013, 35(1): 221-231.
[3] Du Tran, Lubomir Bourdev, Rob Fergus, Lorenzo Torresani, Manohar Paluri, “LearningSpatiotemporal Features with 3D Convolutional Networks,” arXiv preprint arXiv:1412.0767v3, 2015.
[4] K. Simonyan and A. Zisserman. Two-stream convolutional networks for action recognition in videos. In NIPS, 2014.
[5] Paul Ekman, Wallace V. Friesen. " Constants across cultures in the face and emotion". Journal of Personality and Social Psychology, Vol. 17, No. 2. (1971),
[6] M. Lyons, S. Akamatsu, M. Kamachi, andJ. Gyoba,“Coding facial expressions with gabor wavelets,”Automatic Face and Gesture Recognition, pp. 200-205, April 1998.
[7] C.Shan,S.Gong,and P.McOwan,”Facial expression recognition based on local binary patterns: a comprehensive study,” Image and Vision Computing, vol.27, p.803-816, May2009.
[8] T. Cootes, G. Edwards, and C. Taylor, “Active Appearance Models,”IEEE Trans. Pattern Analysis and Machine Intelligence,vol. 23, no. 6, pp. 681-685, June 2001.
[9] Black, M. J. and Yacoob, Y.”Recognizing facial expressions in image sequences using local parameterized models of image motion,”Int. Journal of Computer Vision, 25(1), pp. 23-48, 1997.
[10] E. Osuna, R. Freund, and F. Girosi, “Training Support Vector Machines: An Application to Face Detection,” Computer Vision and Pattern Recognition,pp.130-136,June1997.
[11] P. Viola and M. J. Jones, “Rapid Object Detection using a Boosted Cascade of Simple Features,”in Proceedings of the IEEE Computer Society International Conference on Computer Vision and Pattern Recognition, vol. 1,pp. 511-518, December2001.
[12] M. Liu, S. Li, S. Shan, R. Wang, and X. Chen. Deeply learning deformable facial action parts model for dynamic expression analysis. In ACCV, 2014, pages 1749–1756. IEEE, 2014
[13] A. Sanin, C. Sanderson, M. T. Harandi, and B. C. Lovell. Spatiotemporal covariance descriptors for action and gesture recognition. In WACV, 2013, pages 103–110. IEEE, 2013
[14] Berthold K. P. Horn, Brian G. Schunck. Determining Optical Flow[J]. Artificial Intelligence, 1981: 185-203.
[15] LUCAS B D， KANADE T. An iterative image registration technique with an application to stereo vision[J]. International Joint Conference on Artificial Intelligence， 1981（81）：674-679.
[16] Lucey, P., Cohn, J., Kanade, T., Saragih, J., Ambadar, Z., Matthews, I.: The extended cohn-kanade dataset (ck+): A complete dataset for action unit and emotionspecified expression. In: CVPRW. (2010)