Color Image Edge Detection and Segmentation: A Comparison of the Vector Angle and the Euclidean Distance Color Similarity Measures

ii I hereby declare that I am the sole author of this thesis. This is a true copy of the thesis, including any required final revisions, as accepted by my examiners. I understand that my thesis may be made electronically available to the public. iii Acknowledgements I would like to thank Dr. Ed Jernigan for his continued support and help in the long years that lead to the preparation of this thesis. I am also grateful to him for allowing me to choose my own topic of study and for letting me carry out the research at my own pace. I would also like to thank Dr. Bob Dony for the rich discussions on color image processing and in particular for the insight of using the vector angle as a color similarity measure. I would like to thank him, as well as Dr. Catherine Burns for taking the time to review this thesis and for providing very helpful comments. To NCR belongs the credit of allowing me to attend the University of Waterloo on a part-time basis while I worked there full time and for the tuition reimbursement program which provided much needed financial support. I would also like to acknowledge the warm support of my dear companion, my family and my friends in encouraging me to finish this work. Abstract This work is based on Shafer's Dichromatic Reflection Model as applied to color image formation. The color spaces RGB, XYZ, CIELAB, CIELUV, rgb, l 1 l 2 l 3 , and the new h 1 h 2 h 3 color space are discussed from this perspective. Two color similarity measures are studied: the Euclidean distance and the vector angle. The work in this thesis is motivated from a practical point of view by several shortcomings of current methods. The first problem is the inability of all known methods to properly segment objects from the background without interference from object shadows and highlights. The second shortcoming is the non-examination of the vector angle as a distance measure that is capable of directly evaluating hue similarity without considering intensity especially in RGB. Finally, there is inadequate research on the combination of hue-and intensity-based similarity measures to improve color similarity calculations given the advantages of each color distance measure. These distance measures were used for two image understanding tasks: edge detection, and one strategy for color image segmentation, namely color clustering. Edge …