Japanese / English

Detail of Publication

Text Language Japanese
Authors Masakazu Iwamura, Yoshio Furuya, Koichi Kise, Shinichiro Omachi, and Seiichi Uchida
Title A General Assignment of Supplementary Information
Journal Trans. IEICE
Vol. J93-D
No. 5
Pages pp.579-587
Reviewed or not Reviewed
Month & Year May 2010
Abstract {\it Pattern recognition with supplementary information} is a new pattern recognition framework to avoid unavoidable errors. It determines an output class by combining a feature vector extracted from the pattern and a small {\it hint} on the true class. The hint called {\it supplementary information} is freely designed so as to minimize the error rate. Under the assumption that supplementary information is infallible, an approximately best assignment of the information is derived. In this paper, we derive a general assignments methodology of both fallible and infallible supplementary information.
URL http://search.ieice.org/bin/summary.php?id=j93-d_5_579&category=D&year=2010&lang=J&abst=
Back to list