|
Title:
|
AN APPEARANCE BASED METHOD FOR 3D MODEL COMPARISON |
|
Author(s):
|
Hironobu Gotoda |
|
ISBN:
|
972-99353-6-X |
|
Editors:
|
Nuno Guimarães and Pedro Isaías |
|
Year:
|
2005 |
|
Edition:
|
1 |
|
Keywords:
|
Shape similarity, shape matching, shape retrieval, 3D modeling. |
|
Type:
|
Full Paper |
|
First Page:
|
178 |
|
Last Page:
|
184 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
Comparing and matching 3D models has become a major subject of research in computer vision, graphics, and database systems. It is also a key technology that enables us to perform 3D model search on the WWW to find ones that meet particular requirements. Previous approaches to this problem have been based on the idea that some set of features extracted from the 3D models will be useful to distinguish between dissimilar models while identifying similar ones. Although various kinds of features have been proposed and examined, none have been proven to be robust enough or to fit well with the common sense of human beings. This paper presents an alternative approach to 3D model comparison. Our method is based on the simple observation that humans will perceive and identify 3D objects from their appearances. This leads to an appearance-based 3D model comparison method that evaluates the similarities of 3D models using multiple images representing their views in various angles. More specifically, an image of a 3D model is morphed to an image of the other model, where the degree of distortion involved in the morphing is computed. Such degree of distortion can be considered to be a measure of similarity between the two images, and thus to be a partial measure between the two models. Experimental results show that our method is suitable for comparing shapes of living creatures such as animals and plants. |
|
|
|
|
|
|
