Book Zhangbing.indb

1 Anthropometric landmarks lend themselves to an effective posture prediction and, at the same time, facilitate the reconstruction of a geometric model of the human body. 2 Back-propagation multi-layer perceptron artifi cial neural networks (BP-MLP-ANN) can process multi-dimensional variables (such as 3D coordinates of landmarks, demographic characteristics, and posture data) in an integral way. 3 There is an optimum number of layers, and an optimum number of neurons on the layers, for an optimum BP-MLP-ANN architecture, but it always depends on the application. 4 Even though there are general rules for fi nding an optimum architecture of a BP-MLP-ANN, it has to be found by a trial and error experimentation in each case. 5 Due to the phenomenon of over-training, a larger number of training epochs and a larger number of neurons will result in an over-fi tted generalization in terms of the learned prediction rule. 6 Because certain groups of landmarks show similar behavior when the posture of the human body changes, a landmark cluster-oriented posture prediction method is more practical than a method handling all the body landmarks together. 7 The sensitivity of artifi cial neural networks is at least as diffi cult to measure as the sensitivity of human beings. 8 Our love of life should always be standing even when we have to take a seat. 9 Ph.D. students should be encouraged to do research in artifi cial neural networks that can predict the future of our life. 10 Looking backward is the best way of going forward. These propositions are considered defendable and as such have been approved by the supervisors, prof. dr. He for their taking care of and helping me. Yu for his friendship, and for sharing the knowledge of artifi cial neural networks with me. Finally, I appreciate indeed the opportunities provided for me by my family. My parents, Huimu Li and Jiaxing Zhang, provided me a warm and loving home and made me feel safe and happy. Their love and encouragement help me fi nish this thesis and go further. My husband, Jose, who is with me when I need somebody, and who helped me with his support and good suggestions to achieve the best in my research. My young brother, Jie, and my sisters, Xin, Ming, Fan, Lu, who always are of help for me to overcome the obstacles in my life.