DSpace Repository

Adaptive resonance theory and the self-organizing map in Thai character recognition

Show simple item record

dc.contributor.advisor Prof. Huynh Ngoc Phien (Chairperson) en_US
dc.contributor.advisor Dr. Hoang Le Tien (Member) en_US
dc.contributor.advisor Dr. Aekavute Sujarae (Member) en_US
dc.contributor.author Suriya Natsupakpong en_US
dc.date December 2001 en_US
dc.date.accessioned 2015-01-12T10:39:28Z
dc.date.available 2015-01-12T10:39:28Z
dc.identifier.other AIT Thesis no.CS-01-9 en_US
dc.identifier.uri http://www.cs.ait.ac.th/xmlui/handle/123456789/208
dc.description 56 p. en_US
dc.description.abstract This study is concerned with neural networks. In particular, two unsupervised neural networks; Adaptive Resonance Theory (ART) and the Self-Organizing Map (SOM), used for the recognition of 44 Thai consonant characters. These characters are represented with four font faces, namely, AngsanaUPC, BrowalliaUPC, CordiaUPC, and EucrosiaUPC. The recognition process starts with converting an image of each character to a binary representation, then using a skeleton and normalization algorithm for preprocessing, and sending them to the concerned neural network. Moreover, two supervised neural networks; Learning Vector Quantization (LVQ) and Back Propagation (BP), are considered for the same purpose. From the application results, it was found that neural networks can be ranked in descending order from the highest accuracy rate to the lowest as follows: BP, LVQ, SOM, and ART. In addition, this study proposes an improved algorithm by incorporating the supervised learning in SOM. The results of this study show that this algorithm has the fastest training time and the accuracy rates in training set and testing set are 98.23% and 82.58%, respectively. en_US
dc.description.sponsorship The Development and Promotion of Science and Technology Talents Project (DPST) en_US
dc.relation.ispartof Asian Institute of Technology. Thesis no. CS-01-9 en_US
dc.relation.ispartof Thesis no. CS-01-9 en_US
dc.source SAT/CS en_US
dc.subject Pattern recognition systems en_US
dc.subject Self-organizing maps en_US
dc.title Adaptive resonance theory and the self-organizing map in Thai character recognition en_US
dc.type Thesis en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account