Shape Rule Setting for Recognizing Handprinted Character
碩士 === 國立臺灣大學 === 資訊工程學研究所 === 105 === This work show that how to set shape rules to skip incorrect templates and constraints to limit the training in spatial topology distortion measurement. We use logical rules to implement shape rule and reduce the number of templates into the following recogniti...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2017
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/61552078189432370121 |
| Summary: | 碩士 === 國立臺灣大學 === 資訊工程學研究所 === 105 === This work show that how to set shape rules to skip incorrect templates and constraints to limit the training in spatial topology distortion measurement. We use logical rules to implement shape rule and reduce the number of templates into the following recognition steps. The constraints can enlarge the difference of correct and incorrect templates and make classifying easier. The shape rules and constraints will improve the execution time and recognition rate.
|
|---|