Presentation 1993/11/12
Reliability Evaluation on character recognition result using a neural network
Akimichi Tanaka, Osamu Nakamura, Tadashi Kitamura,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a neural network which outputs reliability that each category is correct from the distances between a feature vector extracted from an input character image and the mean feature vectors of all categories.When it learns by the back propagation method,we classify learning samples by whether the nearest- distance category equals to an input category and give a different kind of teaching signals to each class.The experiments using handwritten numbers show that the reliability the neural network outputs is effective in recognition and rejection.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Character Recognition / Pattern matching method / Neural network / Reliability evaluation
Paper # AI93-66,PRU93-83
Date of Issue

Conference Information
Committee AI
Conference Date 1993/11/12(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Artificial Intelligence and Knowledge-Based Processing (AI)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Reliability Evaluation on character recognition result using a neural network
Sub Title (in English)
Keyword(1) Character Recognition
Keyword(2) Pattern matching method
Keyword(3) Neural network
Keyword(4) Reliability evaluation
1st Author's Name Akimichi Tanaka
1st Author's Affiliation NTT Human Interface Laboratories()
2nd Author's Name Osamu Nakamura
2nd Author's Affiliation NTT Human Interface Laboratories
3rd Author's Name Tadashi Kitamura
3rd Author's Affiliation NTT Human Interface Laboratories
Date 1993/11/12
Paper # AI93-66,PRU93-83
Volume (vol) vol.93
Number (no) 317
Page pp.pp.-
#Pages 5
Date of Issue