| 摘要(英) |
With the developments of research in brain science and the novel technologies of brain-related applications technology, electroencephalography (EEG) provides an effective tool to unveil the mystery of human brain. Especially, "emotion" is the most attractive topic which can influence decision and be the key to victory or defeat. In this dissertation, emotion stimulation materials are provided by the emotional plan of National Science Council, which is implemented by the National Taiwan University. Eight-channel EEG machine and high-sensitivity dry electrodes are used as the experimental equipment to collect the emotional EEG training data. Human emotions were represented in valence and dominance domains. The relations between EEG temporal-frequency features and emotion markers verified the the correctness of the experiment. Since brain activities are different across individuals owing to their ages and habitual hand, neural network is a suitable choice for our emotion classification study. In the study, fifteen subjects were recruited to view a twenty-minute emotion stimulation material. EEG data measured from the fifteen subjects were used as training data for deep-learning neural network. After network training, real-time recognition of subjects’ emotions were performed on TX2 platform whose GPU module was configured by the pre-trained Deep Learning neural network. In our system, simultaneous EEG data recording and real-time emotion recognition was available with the use of double-threading system. The accuracies of valence and dominance indexes have achieved 81.48% and 74.18%, respectively. The research results of our eight-channel EEG study in this thesis have attained comparable results compared to other previous literatures which utilized 32- or 64-channel EEG system in emotion studies. In addition, we also introduced virtual reality (VR) to integrate with our proposed eight-channel EEG emotion detection system. The VR-EEG emotion detection system has already implemented in 3D environment and combined with social communication software which can achieve a variety of entertainment applications in future studies. |
| 參考文獻 |
[1] J.-S. Lee, A. Yazdani, T. Ebrahimi, T. Pun, A.
Nijholt, I. Patras, S. Koelstra, C. Muehl, M. Soleymani,IEEE Transaction on Affective Computing, under review "DEAP: A Database for Emotion Analysis using Physiological Signals",
[2] Franc,ois Chollet et al. Keras. https://github.com/fchollet/keras, 2015.
[3] A. Paszke, S. Gross, S. Chintala, G. Chanan, E. Yang, Z. DeVito, Z. Lin, A. Desmaison, L. Antiga, and A. Lerer. "Automatic differentiation in pytorch". In NIPS Workshop, 2017
[4] Y. LeCun. B. Boser. J. S. Denker. D. Henderson. R. E. Howard. W. Hubbard. L. D. Jackel. "Backpropagation Applied to Handwritten Zip Code. Recognition. "
[5] M.WGardnera, S.RDorlinga "Artificial neural networks (the multilayer perceptron)—a review of applications in the atmospheric sciences "
[6] Samarth Tripathi, Shrinivas Acharya, Ranti Dev Sharma, Sudhanshu Mittal, and Samit Bhattacharya, “Using deep and convolutional neural networks for accurate emotion classification on deap dataset.,” in AAAI, 2017, pp. 4746–4752.
[7] Andrew Senior, Françoise Beaufays "Long Short-Term Memory Recurrent Neural Network Architectures for Large Scale Acoustic Modeling. Hasim Sak "
[8] Kohavi, R. A "Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection" (1995) .
[9] S. Alhagry, A. A. Fahmy, and R. A. El-Khoribi, “Emotion recognition based on eeg using lstm recurrent neural network,” Emotion, vol. 8, no. 10, 2017.
[10] Van Der Walt, S. Chris Colbert, Gaël Varoquaux "The NumPy array: a structure for efficient numerical computation,Stefan"
[11] https://github.com/jbmouret/matplotlib_for_papers
[12] A. Grossman and J. Morlet (1984) Decomposition of hardy functions into square integrable wavelets of constant shape, SIAM J. Appl. Math., vol. 15, pp. 723-736
[13] SpaceYasar Dasdemir1, Esen Yildirim ,Serdar Yildirim" Emotion Analysis using Different Stimuli with EEG Signals in Emotional "
[14] 情緒標準刺激與反應常模之基礎研究-子計畫「情緒短片」 共同主持人：謝淑 蘭 教授、孫蒨如教授、翁嘉英教授
[15] Jeong-Hoon, ShinDae-Hyeon Park, Analysis for Characteristics of Electroencephalogram (EEG) and Influence of Environmental Factors According to Emotional Changes
[16] Rich Caruana, Steve Lawrence, Lee Giles,Overfitting in Neural Nets: Backpropagation, Conjugate Gradient, and Early Stopping
[17] Chih-Wei Hsu, Chih-Chung Chang, and Chih-Jen Lin, A Practical Guide to Support Vector Classification
[18] K. Rustan M. Leino 0 and Peter Muller, A basis for verifying multi-threaded programs
[19] https://www.jetsonhacks.com/nvidia-jetson-tx2-j21-header-pinout/
[20] Dennis J McFarland1 , Muhammad A Parvaz2 , William A Sarnacki1 , Rita Z Goldstein2 and Jonathan R Wolpaw1"Prediction of subjective ratings of emotional pictures by EEG features
[21] SVM and kernel machines: linear and non-linear classification
[22] A Practical Guide to Support Vector Classification
[23] Emotion recognition based on EEG features in movie clips with channel selection
|
[Endnote RIS 格式]
[相關文章]
[文章引用]
[完整記錄]
[館藏目錄]
至系統瀏覽論文 ( 永不開放)
plurk
funp
live
udn
HD
myshare
netvibes
friend
youpush
delicious
baidu
Google bookmarks
del.icio.us
hemidemi
facebook
twitter
google
reddit