Automating Information Extraction from 3-D Scan Data.
Abstract
The object of the research reported here was to evaluate the success of ARNScan versions 7.1 and 7.2, newly developed software for extracting body measurements from 3-D scans. Investigators used traditional methods to measure 123 male and female subjects for 21 dimensions associated with the sizing and design of military clothing. The same subjects were then scanned, and the same or comparable measurements extracted from the 3-D images. A comparison of the results obtained by each method yielded a number of differences. A variety of statistical procedures were then undertaken to establish whether these differences were important or significant; that is, were they large enough to place subjects in different sizes? Differences were tested against three standards: (1) acceptable measurer error established in the 1988 U.S. Army survey (ANSUR), (2) acceptable error estimated by three experienced tailors, and (3) garment grade in traditionally sized dress clothing. ARNScan fared well against the garment grade criterion; it did not do well against the ANSUR standard. The comparison produced mixed results with respect to the tailors' standards. ARNScan is still in the development stage and is currently being refined.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 14, 1998
- Accession Number
- ADA361314
Entities
People
- Bruce Bradtmiller