JavaScript is disabled for your browser. Some features of this site may not work without it.
| dc.contributor.author | Fikadu, Mossie | |
| dc.date.accessioned | 2021-08-16T12:13:39Z | |
| dc.date.available | 2021-08-16T12:13:39Z | |
| dc.date.issued | 2021-03-03 | |
| dc.identifier.uri | http://ir.bdu.edu.et/handle/123456789/12443 | |
| dc.description.abstract | In our world, breast cancer(BC) is the greatest cause of death next to lung cancer, but if this cancer type is diagnosed early, it is curable. Different imaging tools are well established to screen breast cancer including the current golden standard (mammography). However, due to expos to radiation, the non-availability of the rural areas, low sensitivity in young women’s and it’s being a painful procedure, the complement imaging tools have been explored and researched. In the work of literature, several efforts have been reported on detecting breast cancer using thermography. However, accuracy is still challenging tasks because of the difficulty of quality data collection, breast region of interest(ROI) segmentation, and optimal feature selection techniques. In this thesis, the possibility of cancer detection using thermography imaging tool has been investigated by giving special attention to the above three issues. To realize this, by selecting the breast thermograms that obtained from dynamic acquisition protocol new automatic segmentation technique using horizontal projection profile (HPP) and height standardization approach has been applied.And also binary particle swarm optimization( BPSO) based wrapper approach feature section methods have been proposed and 11 best features were selected from 26 attributes. The effect of using the proposed (BPSO) technique as a feature selection has been evaluated. Finally, the Support Vector Machine (SVM) has been used as a classifier to distinguish breast thermograms as normal and abnormal classes. MATLAB environment has been used to train and test the proposed approach using percentage-split (80% for training and 20% for testing) evaluation method. In this study, the overall classification accuracy shows that 96.2264%. | en_US |
| dc.language.iso | en_US | en_US |
| dc.subject | ELECTRICAL AND COMPUTER ENGINEERING | en_US |
| dc.title | Thermography Breast Cancer Detection Using Appropriate Segmentation and Feature Selection Approach | en_US |
| dc.type | Thesis | en_US |
