Development of real environment character string recognition method that can handle curved character strings

2017 Fiscal Year Final Research Report

• Project/Area Number: 15K00258
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Perceptual information processing
• Research Institution: Sasebo National College of Technology
• Principal Investigator: SHIKU OSAMU 佐世保工業高等専門学校, 電子制御工学科, 教授 (00235516)
• Project Period (FY): 2015-04-01 – 2018-03-31
• Keywords: 文字認識 / 文字正規化 / カメラ入力

Outline of Final Research Achievements

We developed a character normalization method for camera based character recognition. In this method, the character normalization process is executed for each character candidates extracted from images. The details are as follows. (1) a part of the character string is taken out as a substring. (2) rotation and shear normalization are applied for the substring. (3) characters are extracted from the normalized substring. (4) shear and size normalization are applied for the extracted characters. We tested the proposed method for many images. The results show that the proposed method is more accuracy than a conventional method.

Free Research Field

画像認識