AN APPLICATION TO WRITER RECOGNITION USING LETTER-SHAPES INFORMATION OF NON-LINEAR NORMALIZATION

スポンサーリンク

概要

• 論文の詳細を見る

• Generally, in the field of ofT-line hand written character recognition, we attempt to improve the recognition rate by compensating for varied letter-shapes of the same character using non-linear normalization. Therefore, information about personal writing style may be included in stored information about varied letter-shapes in non-linear normalization. To investigate this supposition, we examined on-line writer recognition using compensated varied letter-shapes information in non-linear normalization. Consequently, it was proved that a writer could be identified when stroke variation in a two-dimensional pattern was regarded as the writer's personal feature.

• 社団法人電子情報通信学会の論文
• 2003-01-14

著者

• Takigawa Yu

  Graduate School Of Science And Technology Nagasaki University

• Matsuda Kazuo

  Graduate School of Science and Technology Nagasaki University

• Miyahara Sueharu

  Graduate School of Science and Technology Nagasaki University

• Miyahara S

  Graduate School Of Science And Technology Nagasaki University

• Matsuda K

  Graduate School Of Science And Technology Nagasaki University

• Nakamura Chiaki

  Graduate School of Science and Technology Nagasaki University

関連論文

• RESTORATION PROCESSING OF CHARACTER PATTERNS WHICH AFFECTS RECOGNITION ACCURACY OF PRINTED KANJI CHARACTERS
• AN APPLICATION TO WRITER RECOGNITION USING LETTER-SHAPES INFORMATION OF NON-LINEAR NORMALIZATION

スポンサーリンク