Cancer detection using image processing and machine/deep learning methods
Breast cancer is one of the highest mortality cancers among women. The breast tumors can be classified into two categories, benign and malignant. Benign is the non-cancerous tumor; While the other variant, malignant is the cancerous tumor. These tumors are dangerous and mostly life-threatening due t...
| Main Author: | |
|---|---|
| Format: | Final Year Project / Dissertation / Thesis |
| Published: |
2022
|
| Subjects: | |
| Online Access: | http://eprints.utar.edu.my/4813/ http://eprints.utar.edu.my/4813/1/fyp_EE_LZZ_2022.pdf |
| _version_ | 1848886249903882240 |
|---|---|
| author | Leong, Zeh Zen |
| author_facet | Leong, Zeh Zen |
| author_sort | Leong, Zeh Zen |
| building | UTAR Institutional Repository |
| collection | Online Access |
| description | Breast cancer is one of the highest mortality cancers among women. The breast tumors can be classified into two categories, benign and malignant. Benign is the non-cancerous tumor; While the other variant, malignant is the cancerous tumor. These tumors are dangerous and mostly life-threatening due to the characteristics of the recurrence of the tumor. This is because the traditional classification methods are time-consuming, costly, labor-intensive and has reached their bottleneck. Integrating deep learning technology with medicinal solutions could improve the efficiency in early detection and treatment to improve the survival rates of breast cancer. Therefore, this paper researched the application of CNNs on the open-source Mendeley Breast Ultrasound dataset (MBU) by Rodrigues (2018) and the Breast Ultrasound Image dataset (BUSI) by Al-Dhabyani (2020). Moreover, the image preprocessing methods are implemented to refine the ultrasound image quality. Furthermore, the DCGAN model is used for data augmentation and to increase the data quantity. Subsequently, transfer learning-based approach is proposed for differentiating breast tumors. The proposed models, CNN-AlexNet, TL-Inception-V3 and TL-DenseNet are fine-tuned and trained on the MBU dataset. Moreover, the proposed classifier models are tested and evaluated on the BUSI dataset. The finetuned TL-DenseNet exhibited the finest performance among all proposed models by achieving an accuracy of 91.46% and F1-score of 0.9144, followed by the fine-tuned TL-Inception-V3 with accuracy of 91.04% and F1-score of 0.9100. The CNNAlexNet also performs decently on the testing set with accuracy of 90.42% and F1- score of 0.9038. |
| first_indexed | 2025-11-15T19:35:30Z |
| format | Final Year Project / Dissertation / Thesis |
| id | utar-4813 |
| institution | Universiti Tunku Abdul Rahman |
| institution_category | Local University |
| last_indexed | 2025-11-15T19:35:30Z |
| publishDate | 2022 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | utar-48132022-12-29T10:23:18Z Cancer detection using image processing and machine/deep learning methods Leong, Zeh Zen RC0254 Neoplasms. Tumors. Oncology (including Cancer) T Technology (General) TK Electrical engineering. Electronics Nuclear engineering TR Photography Breast cancer is one of the highest mortality cancers among women. The breast tumors can be classified into two categories, benign and malignant. Benign is the non-cancerous tumor; While the other variant, malignant is the cancerous tumor. These tumors are dangerous and mostly life-threatening due to the characteristics of the recurrence of the tumor. This is because the traditional classification methods are time-consuming, costly, labor-intensive and has reached their bottleneck. Integrating deep learning technology with medicinal solutions could improve the efficiency in early detection and treatment to improve the survival rates of breast cancer. Therefore, this paper researched the application of CNNs on the open-source Mendeley Breast Ultrasound dataset (MBU) by Rodrigues (2018) and the Breast Ultrasound Image dataset (BUSI) by Al-Dhabyani (2020). Moreover, the image preprocessing methods are implemented to refine the ultrasound image quality. Furthermore, the DCGAN model is used for data augmentation and to increase the data quantity. Subsequently, transfer learning-based approach is proposed for differentiating breast tumors. The proposed models, CNN-AlexNet, TL-Inception-V3 and TL-DenseNet are fine-tuned and trained on the MBU dataset. Moreover, the proposed classifier models are tested and evaluated on the BUSI dataset. The finetuned TL-DenseNet exhibited the finest performance among all proposed models by achieving an accuracy of 91.46% and F1-score of 0.9144, followed by the fine-tuned TL-Inception-V3 with accuracy of 91.04% and F1-score of 0.9100. The CNNAlexNet also performs decently on the testing set with accuracy of 90.42% and F1- score of 0.9038. 2022-06 Final Year Project / Dissertation / Thesis NonPeerReviewed application/pdf http://eprints.utar.edu.my/4813/1/fyp_EE_LZZ_2022.pdf Leong, Zeh Zen (2022) Cancer detection using image processing and machine/deep learning methods. Final Year Project, UTAR. http://eprints.utar.edu.my/4813/ |
| spellingShingle | RC0254 Neoplasms. Tumors. Oncology (including Cancer) T Technology (General) TK Electrical engineering. Electronics Nuclear engineering TR Photography Leong, Zeh Zen Cancer detection using image processing and machine/deep learning methods |
| title | Cancer detection using image processing and machine/deep learning methods
|
| title_full | Cancer detection using image processing and machine/deep learning methods
|
| title_fullStr | Cancer detection using image processing and machine/deep learning methods
|
| title_full_unstemmed | Cancer detection using image processing and machine/deep learning methods
|
| title_short | Cancer detection using image processing and machine/deep learning methods
|
| title_sort | cancer detection using image processing and machine/deep learning methods |
| topic | RC0254 Neoplasms. Tumors. Oncology (including Cancer) T Technology (General) TK Electrical engineering. Electronics Nuclear engineering TR Photography |
| url | http://eprints.utar.edu.my/4813/ http://eprints.utar.edu.my/4813/1/fyp_EE_LZZ_2022.pdf |