Chinese Journal of Computers   Full Text
Title	Automatic Fragment Re-Assembly Method Based on DDTW Match
Authors	GAO Jian1) ZHANG Cai-Ming1) MENG Xiang-Xu1) FENG Zhi-Quan2)
Address	1)(College of Computer Science and Technology, Shandong University, Jinan 250101) 2)(School of Information and Science Engineering, Jinan Universtiy, Jinan 250022)
Year	2009
Issue	No.2(342—349)
Abstract &   Background	Abstract This paper presents an automatic re-assembly method for matching 3D fragments, which can compute the best match of a pair of broken pieces. To improve speed and accuracy, this paper gives a torsion estimation method based on anti-noise section and a fixed time cost 3D overlap test method. First, the contour curves of fragments are found, then the corner points. Contour curves can be divided into several sub-contours. Second, the torsion sequences are obtained for the DDTW based matching work. Thus, the scale and transformation matrixes can be computed to change size and position of one fragment. Best matching result will be finally chosen according to matching test and extra evaluations. Experiment results show that this method is simple and robust, and can get proper matching results quickly. Keywords 3D fragment; assembly; torsion; DDTW; overlapping detection Background This work is supported by the National Key Basic Research 973 Program of China (2006CB303102), the National Nature Science Foundation of China (Nos.60673003, 60633030). In this project, 3D objects are scanned into 3D point cloud. Among these objects, lots of them are fragmented. If we assembly them by hand, it would take a long period of time. Thus, we scan the fragments separately and plan to assembly them with computer. There are two key points in this problem. One is to find a feature from the 3D point data set, and it must be robust, invariant and computable. Here the authors choosed the torsion to be the feature parameter. The other is to find the best match within each fragment pair. With respect to the abnormity of the data, the matching method must be robust and can give error matches as few as possible. The DDTW based process appears to be appropriate, with the help of overlapping check method. Finally, the matching work is left to computer. Now the authors are continuing their work on 3D point data processing.