Paper Abstract and Keywords |
Presentation |
2005-03-17 11:30
A shape-based part retrieval method for mechanical assembly drawings Takayuki Baba (Fujitsu Laboratories), Rujie Liu (Fujitsu Research and Development Center), Susumu Endo, Shuichi Shiitani, Yusuke Uehara, Daiki Masumoto, Shigemi Nagata (Fujitsu Laboratories) |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Designers in machinery manufactures frequently retrieve a desired drawing among stored drawings to design a new product efficiently by changing parts of previous products or derive hints from previous work. In this paper, we propose a shape-based part retrieval method to find drawings containing parts similar to a query part from among mechanical assembly drawings depicting several parts. In this method, each drawing is automatically divided into parts in advance. Firstly some parts similar to a query part are retrieved, and then drawings containing their parts are displayed. We verified the performance for this method through experiments using 190 mechanical assembly drawings with three frequently-used query parts. For two query parts, both precision rate and recall rate in the break-even point are 90% or more. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
similar shape-based retrieval / part / assembly drawing / parts catalog / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 104, no. 741, PRMU2004-225, pp. 79-84, March 2005. |
Paper # |
PRMU2004-225 |
Date of Issue |
2005-03-10 (TL, PRMU) |
ISSN |
Print edition: ISSN 0913-5685 |
Download PDF |
|
|