國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
圖表(5-3) 不同模型以及輸入幀數調整的實驗結果
圖表(5-4) 睡意偵測相關文獻的實驗結果比較
5.4 睡意辨識的執行效能分析
圖表(5-5)為在多種模型架構下對辨識睡意的處理效能的分析結果，其中測試 環境的作業系統為 Ubuntu 16.04、CPU 使用 Intel Xeon E5、GPU 使用 GeForce GTX 1080，圖表中的處理效能為從影片中擷取 frame 到輸出睡意辨識結果的所 需平均處理時間，在此種測試環境下，我們提出的睡意偵測方法 E 所得到的處理 速度大約為 13.33fps，對睡意的辨識速度來說不算是快，不過仍然可以滿足辨識
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
所需時間為 real time 的需求，我們認為辨識速度不快的主要原因是我們合併多種 子模型來處理睡意辨識，圖表(5-5)中的方法 C 和方法 D 相對於方法 E 來說，因 為合併模型較少因此偵測睡意速度較快，但是相對的睡意偵測的準確率就較低。
圖表(5-5) 不同辨識模型的效能處理速度
‧
攝角度下來錄製的影片，而並非是在開放空間下的環境下錄製(in the wild)，因此 在未來將會把研究的重點放在把模型輕量化來減少運算量，藉此來增加偵測的處‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
理速度並藉由使用多種拍攝角度下的影像或者是採用 data augmentation 等方式 來分析對於睡意偵測的效能影響。
‧
[1] Reza Ghoddoosian, Marnim Galib, and Vassilis Athitsos. A Realistic Dataset and Baseline Temporal Model for Early Drowsiness Detection. arXiv:1904.07312, 2019.
[2] Tun-Huai Shih and Chiou-Ting Hsu. MSTN: Multistage spatial-temporal network for driver drowsiness detection. Springer, 146–153, 2016.
[3] Park S, Pan F, Kang S, and Yoo CD. Driver drowsiness detection system based on feature representation learning using various deep networks. Springer, 154–164, 2016.
[4] Krizhevsky A, Sutskever I, and Hinton GE. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems : 1097–1105, 2012.
[5] Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556, CVPR 2014
[6] Heming Yao, Wei Zhang, Rajesh Malhan, Jonathan Gryak, and Kayvan Najarian.
Filter-Pruned 3D Convolutional Neural Network for Drowsiness Detection. IEEE, 1258–1262, 2018.
[7] Xuan-Phung Huynh, Sang-Min Park, and Yong-Guk Kim. Detection of driver drowsiness using 3d deep neural network and semi-supervised gradient boosting machine. Springer, pp. 134–145, ACCV 2016.
[8] Xuanhan Wang, Lianli Gao, Jingkuan Song, and Hengtao Shen. Beyond Frame-level CNN: Saliency-Aware 3-D CNN With LSTM for Video Action Recognition.
IEEE Signal Processing Letters, pp. 510–514, 2017.
[9] Liang Zhang, Peiyi Shen, and Juan Song. Multimodal Gesture Recognition Using 3-D Convolution and Convolutional LSTM. IEEE, 4517–4524, 2017.
[10] Koustav Mullick and Anoop M. Namboodiri. Learning Deep And Compact Models For Gesture Recognition. arXiv:1712.10136, 2017.
‧
Implementation of Training Convolutional Neural Networks. arXiv:1506.01195, 2015.
[12] Srivastava N, Hinton GE, Krizhevsky A, Sutskever I, and Salakhutdinov R.
Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 15(1):1929–1958, 2014
[13] https://en.wikipedia.org/wiki/Adaptive_histogram_equalization
[14] Zhang K, Zhang Z, Li Z, and Qiao Y. Joint face detection and alignment using multitask cascaded convolutional networks. IEEE Signal Process Lett 23(10):1499–
1503, 2016
[15] Weng C-H, Lai Y-H, and Lai S-H. Driver drowsiness detection via a hierarchical temporal deep belief network. Springer: 117–133, 2016
[16] Krizhevsky, Alex, Sutskever, Ilya, and Hinton. Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems, pp. 1097–1105, 2012.
[17] Chollet François. Keras. 2015.
[18] Martín A, et al. Tensorflow: a system for large-scale machine learning. OSDI. Vol.
16. 2016.
[19] Gao Huang, Zhuang Liu, Kilian Q. Weinberger, and Laurens van der Maaten.
Densely Connected Convolutional Networks. arXiv:1608.06993, 2016.
[20] K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition.
In CVPR, 2016.
[21] Jongmin Yu, Sangwoo Park, Sangwook Lee, and Moongu Jeon. Driver Drowsiness Detection Using Condition-Adaptive Representation Learning Framework. IEEE, 2018.
[22] Zaremba W and Sutskever I. Learning to execute. arXiv:1410.4615, 2014
[23] Jing-Ming Guo, Herleeyandi Markoni. Driver drowsiness detection using hybrid convolutional neural network and long short-term memory. Springer, 2018
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
[24] Jasper S.W, Jason T, Kerry A.N, Gideon D.P.A.A, Mark S. Real-time monitoring of driver drowsiness on mobile platforms using 3D neural networks. Springer, 2019 [25] Rateb Jabbar, Khalifa Al-Khalifa, Mohamed Kharbeche, Wael Alhajyaseen, Mohsen Jafari, Shan Jiang. Real-time Driver Drowsiness Detection for Android Application Using Deep Neural Networks Techniques. IEEE ANT, 2018