27th International Conference on Distributed Computing Systems Workshops (ICDCSW'07)
Effective Retrieval of High Dimensional Vector by Class Coding in Vegetable Geographical Origin Identification System
Toronto, Canada
June 22-June 29
ISBN: 0-7695-2838-4
We proposed a vegetable geographical origin identification system using difference of trace element composition by cultivated places to prevent camouflaging geographical origin. The system compared trace element compositions by calculating correlation coefficient. This is one-to-one comparison and makes system response time to identify slow. To solve this problem, we employed Similarity Preserve Hash (SPH) to reduce the number of target of one-to-one comparison and realized efficient identification to some degree. Although SPH makes identification itself faster, however, SPH wastes long time to retrieve trace element data for one-to-one comparison from database. In this paper, we discuss on coding of trace element composition to efficient retrieval from databases and realize efficient identification.
Citation:
Nobuyoshi Sato, Minoru Uehara, Koichiro Shimomura, Hirobumi Yamamoto, Kenichi Kamijo, "Effective Retrieval of High Dimensional Vector by Class Coding in Vegetable Geographical Origin Identification System," icdcsw, pp.35, 27th International Conference on Distributed Computing Systems Workshops (ICDCSW'07), 2007