Three-dimensional Model Retrieval Technology And Systems To Achieve

3D model databases will become ubiquitous due to following facts. Firstly, improved digital multimedia technology, computer vision theory and graphics processing unit have expanded the applications of 3D graphics. Secondly, popularized modeling tools and 3D scanning devices are making acquisition of 3D models easier and less expensive, creating a large supply of public available 3D data sets. So retrieval is becoming more and more necessary.The two new 3D model search methods are put forward in this thesis. These two methods all belong to shape-based retrieval techniques. Before extracting feature from 3D model, we need to do some pre-processing. Then the research of algorithms of 3D model retrieval is devised from the following three aspects: (1) The classification of point-pairs is put forward based on shape distribution. This approach separates the pairs of points based on geometric properties of line connecting them. The different kinds of lines construct different histograms.(2)Thickness histogram adds the normal of triangle into shape feature. Because of this new feature, this approach approves to perform better than shape distribution. (3)The combined distribution calculate the angles between the normal. The feature represents the distribution at certain normal and distance.The experimental results show that the proposed retrieval methods are effective and efficient. As for similarity measurement of 3D model retrieval, this paper demonstrates two data models related to the similarity measurement and proposes two kinds of multi-feature combination structures. Some of the distance functions for measuring model similarity used in this paper are compared in the paper too.This paper designs a 3D retrieval system—Measure3D as the test bed for retrieval algorithms, which is an experimental frame system. Therefore, some key issues of 3D model retrieval for future research are also discussed at the end of this thesis.