臺灣博碩士論文加值系統

摘要 i
ABSTRACT ii
目錄 iii
圖目錄 v
表目錄 vii
第一章 序論 1
1.1 研究動機 1
1.2 論文章節架構 1
第二章 相關文獻 2
2.1 影像實例分割 ( Instance Segmentation ) 2
2.2 光流 ( Optical Flow ) 3
2.3 時序記憶網路與注意力機制 4
2.3.1 長短期記憶網路 ( Long Short-Term Memory ) 4
2.3.2 注意力機制 ( Attention Mechanism ) 6
2.4 行為辨識 ( Action Recognition ) 7
2.4.1 2D-CNN + LSTM 7
2.4.2 3D-CNN 8
2.4.3 Two-Stream Convolutional Networks 10
第三章 研究方法 11
3.1 研究困難點 11
3.2 系統架構 12
3.3 影像空間處理 13
3.3.1 骨架 ( Skeleton ) 14
3.3.2 分割影像 14
3.4 時序特徵擷取—光流圖 ( FlowNet ) 15
3.5 深度特徵擷取 18
3.6 Attention-Based LSTM 19
3.7 評分機制 22
第四章 實驗結果 24
4.1 實驗環境 24
4.2 實驗資料集 24
4.3 評估方法 25
4.4 實驗結果 26
4.4.1 模型準確度 26
4.4.2 模型速度 29
第五章 結論 30
第六章 參考文獻 31

[1]S. Liu, J. Jia, S. Fidler, and R. Urtasun, “SGN: Sequential Grouping Networks for Instance Segmentation,” ICCV, pp. 3496-3504, 2017.
[2]B. D. Brabandere, D. Neven, and L. V. Gool, “Semantic Instance Segmentation with a Discriminative Loss Function,” arXiv:1708.02551, 2017.
[3]J. Dai, K. He, Y. Li, S. Ren, and J. Sun, “Instance-sensitive Fully Convolutional Networks,” arXiv:1603.08678, 2016.
[4]K. He, G. Gkioxari, P. Dollár, and R. Girshick, “Mask R-CNN,” ICCV, pp. 2961-2969, 2017.
[5]D. Bolya, C. Zhou, F. Xiao, and Y. J. Lee, “YOLACT: Real-time Instance Segmentation,” ICCV, pp. 9157-9166, 2019.
[6]X. Wang, T. Kong, C. Shen, Y. Jiang, and L. Li, “SOLO: Segmenting Objects by Locations,” arXiv:1912.04488, 2020.
[7]OpenCV Official documentation, https://docs.opencv.org/
[8]B. D. Lucas and T. Kanade, “An Iterative Image Registration Technique with an Application to Stereo Vision,” IJCAI, 1981.
[9]Gunnar Farneback, “Two-Frame Motion Estimation Based on Polynomial Expansion,” LNCS, 2003.
[10]National Taiwan University Ph.D Hung-yi Lee Official website, https://speech.ee.ntu.edu.tw/~hylee/
[11]S. Hochreiter and J. Schmidhuber, “Long Short-Term Memory,” Neural Computation, 1997.
[12]J. Carreira and A. Zisserman, “Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset,” arXiv:1705.077503v, 2018.

[13]J. Y.H. Ng, M. Hausknecht, S. Vijayanarasimhan, O. Vinyals, R. Monga, and G. Toderici , “Beyond Short Snippets: Deep Networks for Video Classification,” arXiv:1503.08909, 2015.
[14]J. Donahue, L. A. Hendricks, M. Rohrbach, S. Venugopalan, S. Guadarrama, K. Saenko, and T. Darrell, “Long-term Recurrent Convolutional Networks for Visual Recognition and Description,”arXiv:1411.4389, 2016.
[15]G. Thung and H. Jiang, “A torch library for action recognition and detection using CNNs and LSTMs,” ,2016.
[16]D. Tran, L. Bourdev, R. Fergus, L. Torresani, and M. Paluri, “Learning Spatiotemporal Features with 3D Convolutional Networks,” arXiv:1412.0767v4, 2015.
[17]D. Tran, H. Wang, L. Torresani, J. Ray, Y. LeCun, and M. Paluri, “A Closer Look at Spatiotemporal Convolutions for Action Recognition,” CVPR, pp. 6450-6459, 2018.
[18]K. Simonyan and A. Zisserman, “Two-Stream Convolutional Networks for Action Recognition in Videos,” NIPS, 2014.
[19]E. Ilg, N. Mayer, T. Saikia, M. Keuper, A. Dosovitskiy, and T. Brox, “FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks,” CVPR, pp. 2462-2470, 2017.
[20]Pytorch Offical website, https://pytorch.org/
[21]K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” CVPR, pp. 770-778, 2016.
[22]Shi-Jie Zhu, “A Distilled 2D CNN-LSTM Framework with Temporal Attention Mechanism for Action Recognition,” 2021.
[23]MediaPipe Official website, https://google.github.io/mediapipe/
[24]L. Fan, Z. Xia, X. Zhang, and X. Feng, “Lung nodule detection based on 3D convolutional neural networks,” 2017.
[25]K. Simonyan and A. Zisserman “Very Deep Convolutional Networks for Large-Scale Image Recognition,” arXiv:1409.1556v6, 2014.
[26]S. Crabtree and P. Beudert. “Scenic Art for the Theatre: History, Tools, and Techniques,” 1998.