Research Article
Artificial Intelligence-Based Breast and Cervical Cancer Diagnosis and Management System
@INPROCEEDINGS{10.1007/978-3-031-28725-1_6, author={Elbetel Taye Zewde and Mizanu Zelalem Degu and Gizeaddis Lamesgin Simegn}, title={Artificial Intelligence-Based Breast and Cervical Cancer Diagnosis and Management System}, proceedings={Artificial Intelligence and Digitalization for Sustainable Development. 10th EAI International Conference, ICAST 2022, Bahir Dar, Ethiopia, November 4-6, 2022, Proceedings}, proceedings_a={ICAST}, year={2023}, month={3}, keywords={Breast cancer Cervical cancer Decision support system Screening Histopathological images Cancer management}, doi={10.1007/978-3-031-28725-1_6} }- Elbetel Taye Zewde
Mizanu Zelalem Degu
Gizeaddis Lamesgin Simegn
Year: 2023
Artificial Intelligence-Based Breast and Cervical Cancer Diagnosis and Management System
ICAST
Springer
DOI: 10.1007/978-3-031-28725-1_6
Abstract
Breast cancer and cervical cancer are two of the most common and deadly malignancies in women. Early diagnosis and treatment can save lives and improve quality of life. However, there is a shortage of pathologists and physicians in most developing countries, including Ethiopia, preventing many breast and cervical cancer patients from early cancer screening. Many women, particularly in low resource settings, have limited access to early diagnosis of breast and cervical cancer and receive poor treatment which in turn increases the morbidity and mortality due to these cancers. In this paper, an integrated intelligent decision support system is proposed for the diagnosis and management of breast and cervical cancer using multimodal im-age data. The system includes breast cancer type, sub-type and grade classification, cervix type (transformation zone) detection and classification, pap smear image classification, and histopathology-based cervical cancer type classification. In addition, patient registration, data retrieval, and storage as well as cancer statistical analysis mechanisms are integrated into the proposed system. A ResNet152 deep learning model was used for classification tasks and satisfactory results were achieved when testing the model. The developed system was deployed to an offline web page which has added the advantage of storing the digital medical images and the labeled results for future use by the physicians or other researchers.
