臺灣博碩士論文加值系統

本篇論文主要提出整合多重資訊於霍式轉換之頭部偵測演算法，基於固定式單一CCD 環境下。整合形狀特徵與顏色特徵之優點，在於能夠提升霍式轉換(Hough Transform)偵測頭部的準確性。在本論文中，首先針對人頭部呈現橢圓形狀之特徵，利用霍式轉換在一張邊緣影像中偵測出橢圓也就是頭部位置。但是在複雜背景或有遮蔽環境下，霍式轉換無法準確偵測頭部。在整合部分，經由觀察人頭部與肩膀不論在正面、側面或背面皆會呈現Ω形狀，透過兩組Ω樣板與邊緣影像做比對，找出極有可能是頭部輪廓的像素點並提高其權重，接著透過膚色高斯模型，將原始影像切割出人臉位置。最後將兩種資訊整合在霍式轉換上，進而改善選擇邊緣點之效率與投票之規則。在最後的實驗結果中，我們將會呈現克服各種環境且有不錯偵測率的結果。

In this work, we propose an algorithm of integrating multiple cues on Hough transform for head detection from a fixed CCD camera. The idea of combining shape and color features is to increase the accuracy of Hough transform in head detection. It is obvious that the shape of human head is nearly an ellipse. Although Hough transform is effective to estimate the parameters of head ellipse, it is failed in cases of occlusion and complex background. To overcome this problem, the template matching algorithm using six Ω-shape templates with different scales and orientations is applied to compute the existence possibility of head contour at every pixel. In addition, the skin color is also detected. Finally, we propose a framework to integrate head contour and skin color into Hough transform for facilitating the head detection. In the experiments, we exhibit that our proposed head detection method has higher detection rate in various
environments.

中文摘要................................................I
英文摘要...............................................II
致謝..................................................III
目錄...................................................IV
表目錄.................................................VI
圖目錄................................................VII
第一章緒論..............................................1
1.1 研究動機............................................1
1.2 頭部偵測困難與挑戰..................................2
1.3 頭部偵測相關研究....................................4
1.4 系統概述............................................6
第二章頭部偵測相關技術..................................9
2.1 邊緣偵測............................................9
2.1.1 Absolute Difference Mask.........................10
2.2 霍氏轉換...........................................12
2.3 膚色偵測...........................................13
2.3.1 正規劃RGB 色彩空間...............................13
2.3.2 HSV 色彩空間.....................................14
2.3.3 YCbCr 色彩空間...................................15
第三章頭部輪廓與膚色偵測...............................16
3.1 Ω樣板比對..........................................16
3.2 YCbCr 高斯模型.....................................22
第四章橢圓偵測.........................................24
4.1 基於幾何性質之橢圓偵測.............................24
4.2 霍式轉換偵測結果觀察...............................31
第五章整合多重資訊之頭部偵測...........................34
5.1 頭部輪廓資訊整合...................................34
5.2 修改霍式轉換投票規則...............................37
第六章 實驗結果........................................40
6.1 環境描述...........................................40
6.2 影片資料庫.........................................40
6.3 實驗結果...........................................41
第七章 結論與未來展望..................................47
參考文獻...............................................48

[1] Si-Cheng Zhang, Zhiqiang Liu, “A Robust Real Time. Ellipse Detector”, Pattern Recognition, 38, 273-287, 2005 .
[2] J. Prakash, K. Rajesh, “Human Face Detection and Segmentation using Eigenvalues of Covariance Matrix, Hough Transform and Raster Scan Algorithms”, World Academy of Science, Engineering and Technology, 39, 2008 .
[3] 徐士中, “Real-time Face Detection Under Arbitrary Background”, 國立臺灣科技大學電機工程所, 臺北, 2006 .
[4] Jianping Wu, “Robust Real-Time Ellipse Detection by Direct Least-Square-Fitting”, International Conference on Computer Science and Software Engineering, 2008 .
[5] Samuel Inverso, “Ellipse Detection Using Randomized Hough Transform,”Technical Report, May 20, 2002 .
[6] M. Soriano, S. Huovinen, B. Martinkauppi, and M. Laaksonen, “Using the skin locus to cope with changing illumination conditions in color-based face tracking”,
IEEE Nordic Signal Processing Symposium, pp.383-386, 2000 .
[7] 周建豐, “Robust Face Detection and Facial Feature Extraction based on Spatial Relationship”, 國立中央大學資訊工程研究所, 桃園, 2002 .
[8] Matthias Wimmer, Bernd Radig, Michael Beetz, “A Person and Context Specific Approach for Skin Color Classification”, International Conference on Pattern
Recognition, pp.39-42, 2006 .
[9] Guoshing Huang, Jiahong Su, “A Real-time Face Detection and Tracking”, 49 International Conference on Audio, Language and Image Processing, 2008 .
[10] 王健權, “The Study on Real-Time Human Face Tracking by the Technology of Digital Image Processing with Active Camera”, 國立成功大學電機工程所, 臺南, 2002 .
[11] Sanjay Kr. Singh, D.S. Chauhan, Mayank Vatsa, Richa Singh, “A Robust Skin Color Based Face Detection Algorithm”, Tamkang Journal of Science and Engineering, Vol. 6, No. 4, pp. 227-234, 2003 .
[12] 蔡欣明, “Human Detection by Combining Histograms of Oriented Gradients with Global Feature”, 國立高雄第一科技大學電腦與通訊工程所, 高雄, 2008 .
[13] 張博智,”Hand Recognition Based on Geometric Shape and Contour Difference”, 國立高雄第一科技大學電腦與通訊工程所, 高雄, 2009 .
[14] Robert A. McLaughlin, “Randomized Hough Transform: better ellipse detection”, Digital Signal Processing Applications, vol. 1, pp. 409-414, Nov. 1996 .
[15] Viola, P, Jones, M., “Rapid object detection using a boosted cascade of simple features,” Computer Vision and Pattern Recognition, 2001.
[16] Open Source Computer Vision Library http://www.opencv.org.cn/index.php
[17] Haar Cascades
http://www.alereimondo.com.ar/OpenCV/34
[18] P.V.C. Hough, “Method and Means for Recognizing Complex Patterns”, US patent 3069654, 1962.