臺灣博碩士論文加值系統

本論文提出新的演算法運用在單鏡頭二維攝影機上追蹤肩膀與手肘點的位置。機於人體的姿勢，本於在線建立人體的骨架。在所提出的方法中，通過每張影像的差建立背景，然後從背景分割出前景（人體），而人體的輪廓是通過鏈碼找到的，在先前研究中提出的輪廓和不同特徵用於定位頭部，手部和腳部，為了標示出手肘，我們使用四方向轉換找到骨架，找到的骨架會斷斷續續，所以我們使用Arcelli-Baja 演算法來改善此問題。根據骨架和身體輪廓，手肘位於連接手掌和肩膀的曲線上。 在手肘的定位方法中，根據手掌的不同姿勢提出了不同的規則。根據手掌與身體的佔用情況(手掌在身體裡面或外面)，手掌在左側或右側的位置以及肩膀和手掌的相對位置，將姿勢分為不同的情況。為了減少這些標示手肘的錯誤，我們使用粒子濾波器改善。在三個影片中進行不同的方法比較，並驗證分析此方法的準確性。

This thesis proposes a new algorithm to track the two-dimensional shoulders and elbows of a human body using a monocular camera. Based on the located significant points of a human body, this thesis also online builds the skeleton of a human body. In the proposed method, a background is registered by frame difference, and then the foreground (human body) is segmented from the background. The contour of the human body are found by the chain code. The contour and different features proposed in previous studies are used to locate the head, hands, and feet. To locate the elbows, body skeleton is found using the four-point distance transform. The Arcelli-Baja algorithm is then applied to fix the intermittent skeleton. The elbows is localized on the curve connecting the palm and the shoulder based on the skeleton and body contour. In the elbow localization method, different rules are proposed depending different postures of the hand palm. The posture is divided into different cases depending the occupation condition of the palm with body, the position of the palm on the left or right side, and the relative positions of the shoulders and palms. To minimize the tracking error caused by some incorrectly localized elbows and smoothen the trajectory, the particle filter is applied to the localized elbows. Experiments in three videos with comparisons with different elbow tracking methods are performed to verify the effectiveness and accuracy of the proposed elbow tracking method.

摘要 i
Abstract ii
Content iii
List of Figure vi
List of Table xi
Chapter 1 Introduction 1
Chapter 2 Human Body Segmentation 5
2.1 RGB-based Background Registration and Update 6
2.1.1 Frame Difference 6
2.1.2 Background Registration 7
2.1.3 Background Update 8
2.2 Object Detection 9
2.2.1 Background Difference 9
2.2.2 Shadow Removal 10
2.2.3 Morphological Operator 11
2.2.4 Post Processing 13
Chapter 3 Significant Point Estimation 14
3.1 4-distance Transform 15
3.2 Contour Convex Points 17
3.3 Best Fitting Ellipse(BFE) 18
3.4 Localization of Significant Points 20
3.5 Localization of the Head 20
Chapter 4 Localization of the Head and Palms..22
4.1 Localziation of the palms 22
4.1.1 Optical Flow 23
4.1.2 Background Likelihood 24
4.1.3 Skin Color 25
4.1.4 Search Area 27
4.2 Localization Using Mixture of Features 27
Chapter 5 Localization of the Shoulders and Elbows 29
5.1 Skelton Extraction Using the Arcelli-Baja Algorithm 30
5.2 Localization of the Shoulders 31
5.3 Localization of the Elbows 34
5.3.1 Finding the first point of palm 34
5.3.2 Search Along the Skeleton Points 36
5.3.3 Search Along the Contour Points 39
5.3.4 Search When the Palm is Within the Silhouette 42
5.3.5 Search Directly Toward the Shoulder Endpoint 43
Chapter 6 Particle Filter Tracking 45
6.1 Particle Filter 45
6.1.1 Dynamic Model and Observation Model 46
6.1.2 Resampling 48
6.1.3 Particle Estimation 49
6.2 Tracking of the Head 49
6.3 Tracking of the Elbows 50
Chapter 7 Experimental Results 52
7.1 Experimental Results 54
Chapter 8 Conclusion 71
References 72

[1]E. Chen, X. Tang, B. Fu, “A Modified Pedestrian Retrieval Method Based on Faster R-CNN with Integration of Pedestrian Detection and Re-Identification, ” Proc. 2018 Int. Conf. Audio, Language and Image Processing (ICALIP), pp. 63-66, 2018
[2]W. Chen, M. Zhang, Z. Pan, G. Liu, H. Shen, S. Chen and Y. Liu, “Animations, Games, and Virtual Reality for the Jing-Hang Grand Canal, ” IEEE Computer Graphics and Applications, vol. 30, no. 3, pp. 84–88, 2010.
[3]X. Hu, R. Su and L. He, “The Design and Implementation of the 3D Educational Game Based on VR Headsets,” Proc. 2016 International Symposium on Educational Technology (ISET), pp. 53–56, 2016.
[4]M. Yuan, I. R. Khan, F. Farbiz, S. Yao, A. Niswar, M.-H. Foo, “A Mixed Reality Virtual Clothes Try-On System, ” IEEE Transactions on Multimedia, vol. 15 , no. 8, pp. 1958–1968, 2013.
[5]K. Bai, Chengyin, Liuqin and Gaoming, “Utilization of DLT Theory in 2D Image Analysis System,” Proc.2010 Second World Congress on Software Engineering, vol. 1, pp. 279–282, 2010
[6]J. Yu, Z. Xie, D. Huang, Y. Ding, S. Tang and L. Ma, “Real-Time 2.5D Facial Cartoon Animation Based on Pose and Expression Estimation,” Proc. 2015 IEEE International Symposium on Multimedia (ISM), pp. 53–56, 2015.
[7]C. Lee, H. Song, B. Choi and Y.-S. Ho, “3D scene capturing using stereoscopic cameras and a time-of-flight camera, ” IEEE Transactions on Consumer Electronics, vol. 57,no. 3, pp. 1370–1376, 2011.
[8]C. F. Juang, C. M. Chang, J. R. Wu, and D. M. Lee, “Computer vision-based human body segmentation and posture estimation,” IEEE Trans. Syst., Man, and Cyber., Part A: Systems and Humans, vol.39, no.1, pp. 119-133, January 2009.
[9]J. H. Holland, “Genetic algorithms,” Scientific American, vol. 267, no. 1, pp. 44-50m 1992.
[10]P. C. Hernandez, J. Czyz, F. arques, T. Umeda, X. Marichal and B. Macq, “Bayesian Approach for Morphology-Based 2-D Human Motion Capture,” IEEE Transactions on multimedia, vol. 9, no. 4, pp. 754-765, June 2007.
[11]H. J. Jiang, T. C. Chen, and C. F. Juang, “Stereo camera-based real-time 3D character construction and human behavior following with interactive entertainment application,” Proc. Nat. Symp. System Science and Engineering, pp. 501-506, Taiwan, R.O.C., June 2012.
[12]OpenNiTE: http://openni.ru/files/nite/index.html
[13]C. W. Chang “Real-Time Hand Motion Tracking in 2D Image by Mixture of Fuzzy Features and Particle Filter,” Master Thesis, National Chung Hsing University, Taiwan, R.O.C., July 2007
[14]C. Arcelli and G. S. di Baja, “A one-pass two-operation process to detect the skeletal pixels on the 4-distance transform,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 11, no. 4, pp. 411-414, April 1989.
[15]C. F. Juang, J. R. Wu, and W. J. Lin, “Automatic Human Body Extraction and Posture Analysis in Consecutive Images,” Master Thesis, National Chung Hsing University, Taiwan, R.O.C., July 2007.
[16]M. S. Arulampalam, S. Maskell, N. Gordon, and T. Clapp, “A tutorial on particle filters for online nonlinear/non-gaussian Bayesian tracking,” IEEE Trans. Signal Processing, vol. 50, no. 2, pp. 174-188, Feb. 2002.
[17]F. Fang, K. Qian, B. Zhou and X. Ma “Real-time RGB-D based people detection and tracking system for mobile robots,” Proc. 2017 IEEE International Conference on Mechatronics and Automation (ICMA), pp. 1937–1941, 2017.
[18]H.-Y. Zhang, “Multiple moving objects detection and tracking based on optical flow in polar-log images,” Proc. 2010 International Conference on Machine Learning and Cybernetics, vol. 3, pp. 1577–1582, 2010.
[19]H. Wu, L. Meng and J. Pan “Research of people counting based on face detection and area-tracking,” Proc. IEEE Proceedings of 2013 3rd International Conference on Computer Science and Network Technology, pp. 525–528, 2013.
[20]P. J. Toivanen, H. Elmongui, “Sequential local transform algorithms for gray-level distance transforms,” Proc. 9th European Signal Processing Conference (EUSIPCO 1998), pp. 1–4, 1998.
[21]Y. Xu, K. Xu, J. Wan, Z. Xiong and Y. Li, “ Research on Particle Filter Tracking Method Based on Kalman Filter,” Proc. 2018 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation pp. 156 –1568, 2018.