Part orientation for handwritten character recognition

• Main Authors: LI, HUEY JIUAN, 李惠娟
• Other Authors: LI, ZI CAI
• Format: Others
• Language: en_US
• Published: 1996
• Online Access: http://ndltd.ncl.edu.tw/handle/89931911197348736715

碩士 === 國立中山大學 === 應用數學研究所 === 84 === A regional decomposition method and a hierarchical model without consideration of orientation have been proposed to investigate the recognition rates of patterns in [1-4]. In this thesis, new models with orientations at various positions have been proposed to produce a more complete character recognizer, based on the probability of occurrence of the patterns.New formulas are developed to evaluate the recognition rates of non- fix orientation from those of fix orientation. Also, mathematical analysis is made to discover some important properties of character recognition versus the sample shapes, rectangle and square, which may facilitate their recognition and analysis. Moverover, the recognition rates have been analyzed and compared to those obtained from the previous studies in [1-4]. Numerical experiments have also been conducted for 89 patterns of the most frequently used alphanumeric handprints. This study displays a deeper, inherent similarity and distinctness among different patterns and characters, which include part symmetry and part resemblance in possibly different positions. Hence, the results of this thesis should be useful to pattern analysis and recognition, character understanding while missing in writing, gambling, poor scanning, noise or various kinds of distortion.