Printed Chinese character recognition using Fuzzy theory and hidden Markov model approaches
碩士 === 國立中央大學 === 資訊及電子工程研究所 === 81 === In this thesis, we propose a two - stage method for Chinese character recognition. The Fuzzy c-means ( FCM ) clustering algorithm is used to train the cluster centers , and the hidden Markov model (HMM) techni...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
1993
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/13506475379906497172 |
| Summary: | 碩士 === 國立中央大學 === 資訊及電子工程研究所 === 81 === In this thesis, we propose a two - stage method for Chinese
character recognition. The Fuzzy c-means ( FCM ) clustering
algorithm is used to train the cluster centers , and the
hidden Markov model (HMM) technique is used to train the
reference character template in the learning procedure. The
mean square error(MSE) criterion is used to determine which
cluster the input character belongs to . Again . we use HMM
technique to make direct recognition. Feature extraction is
another important part for character recognition . We extract
the global and ocal features by the reflection method .
They can represent the character contour and
structure information. By the FCM clustering algorithm,elements
of set can belong to several clusters and to different
degrees. It provide a much more adequate tool for
representing real - data structures. The states of the HMM
correspond to subdivisions of character - images. This
permits recognition of Chinese character even thought the
images are damaged by noises. For future research , some
suggestions are offered.
|
|---|