NCJ Number
186074
Journal
International Journal of Forensic Document Examiners Volume: 5 Dated: December/January 1999 Pages: 102-104
Date Published
1999
Length
3 pages
Annotation
Computerized seal imprint processing is becoming essential and critical in today's financial-based society of oriental countries; the two most important issues for effective seal processing and management are seal imprint verification and compression, which are discussed in this paper to assist the manual imprint check process.
Abstract
Sealing is widely used as an official authentication method in Chinese society. The traditional manual and the computer-assisted seal verification approaches can be integrated effectively if they are conducted carefully; however, to improve efficiency these integrated semiautomatic processes must be included in the whole system. A generalized Euclidean distance concept is used to measure the distance of two seal imprints that will be insensitive to stroke width, deformation, broken lines, etc. Experiments show that the verification rate has improved from 65 percent to 75 percent compared with popular Euclidean distance measurement. Furthermore, a compression ratio of 0.4188 is derived from the authors' modified JBIG algorithm, in which the initial transition table of the coding algorithm is replaced by a table that is statistically computed from the imprint image set. This modification shortens the adaptation phase and results in a more effective compression. For comparison, a compression ratio of 0.607 is obtained from the popular Huffman coding method. These two algorithms will be extended as the automatic part of a computerized seal processing system to enhance the system quality and performance. 6 references