HAL will be down for maintenance from Friday, June 10 at 4pm through Monday, June 13 at 9am. More information
Skip to Main content Skip to Navigation
Conference papers

Handwritten Chinese character recognition: Effects of shape normalization and feature extraction

Abstract : The ¯eld of handwritten Chinese character recog- nition (HCCR) has seen signi¯cant advances in the last two decades, owing to the e®ectiveness of many techniques, especially those for character shape normalization and feature extraction. This paper reviews the major methods of normalization and feature extraction, and evaluates their perfor- mance experimentally. The normalization meth- ods include linear normalization, nonlinear normal- ization (NLN) based on line density equalization, moment normalization (MN), bi-moment normaliza- tion (BMN), modi¯ed centroid-boundary alignment (MCBA), and their pseudo-two-dimensional (pseudo 2D) extensions. As to feature extraction, we fo- cus on some e®ective variations of direction features: chaincode feature, normalization-cooperated chain- code feature (NCCF), and gradient feature. We have compared the normalization methods previ- ously, but in this study, will compare them with better implementation of features. As results, the current methods perform superiorly on handprinted characters, but are insu±cient for unconstrained handwriting.
Document type :
Conference papers
Complete list of metadata

Cited literature [43 references]  Display  Hide  Download

Contributor : Chine Publications Liama Connect in order to contact the contributor
Submitted on : Thursday, December 14, 2006 - 6:07:22 PM
Last modification on : Friday, October 8, 2021 - 4:26:19 PM
Long-term archiving on: : Thursday, September 20, 2012 - 4:06:40 PM


Files produced by the author(s)


  • HAL Id : inria-00120408, version 1
  • PRODINRA : 252101



Cheng-Lin Liu. Handwritten Chinese character recognition: Effects of shape normalization and feature extraction. Arabic and Chinese Handwriting Recognition (SACH06), Sep 2006, Maryland / USA, United States. ⟨inria-00120408⟩



Record views


Files downloads