An Effective Classification System for Dental Bite-wing images Using Entire Tooth Shape
碩士 === 中興大學 === 資訊科學系所 === 95 === We present an automatic segmentation and classification system to segment and classify teeth in bite-wing images. Unlike other methods, we extract the complete tooth contour. Our final goal is to classify teeth in bite-wing images more correctly and efficiently. We...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2007
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/72673565691709907553 |
| Summary: | 碩士 === 中興大學 === 資訊科學系所 === 95 === We present an automatic segmentation and classification system to segment and classify teeth in bite-wing images. Unlike other methods, we extract the complete tooth contour. Our final goal is to classify teeth in bite-wing images more correctly and efficiently. We use Support Vector Machine to classify teeth in bite-wing images into molars and premolars. After classifying teeth, we utilize their spatial relationships to automatically number the teeth based on the universal numbering system. Furthermore, we use Sequential Forward Selection (SFS) method to rank the features and analyze the degree of importance for each feature in classification. In a set of 62 bite-wing images containing 375 teeth, experimental results show that our method is capable of classifying and numbering the teeth with high accuracy.
|
|---|