- Title
- A model for the detection of breast cancer using machine learning and thermal images in a mobile environment
- Creator
- Lennox, Nicholas
- Subject
- Breast -- Cancer -- Imaging
- Subject
- Breast -- Imaging Mobile communication systems
- Date Issued
- 2020
- Date
- 2020
- Type
- Thesis
- Type
- Masters
- Type
- MIT
- Identifier
- http://hdl.handle.net/10948/46097
- Identifier
- vital:39492
- Description
- Breast cancer is the most common cancer amongst women and one of the deadliest. Various modalities exist which image the breasts, all with a focus on early detection; thermography is one such method. It is a non-invasive test, which is safe and can be used for a wide variety of breast densities. It functions by analysing thermal patterns captured via an infrared camera of the surface of the breast. Advances in infrared and mobile technology enable this modality to be mobile based; allowing a high degree of portability at a lower cost. Furthermore, as technology has improved, machine learning has played a larger role in medical practices by offering unbiased, consistent, and timely second opinions. Machine learning algorithms are able to classify medical images automatically if offered in the correct format. This study aims to provide a model, which integrates breast cancer detection, thermal imaging, machine learning, and mobile technology. The conceptual model is theorised from three literature studies regarding: identifiable aspects of breast cancer through thermal imaging, the mobile ecosystem, and classification using machine learning algorithms. The model is implemented and evaluated using an experiment designed to classify automatically thermal breast images of the same quality that mobile attachable thermal cameras are able to capture. The experiment contrasts various combinations of segmentation methods, extracted features, and classification algorithms. Promising results were shown in the experiment with a high degree of accuracy obtained. The successful results obtained from the experimentation process validates the feasibility of the model.
- Format
- xiii,168 leaves
- Format
- Publisher
- Nelson Mandela University
- Publisher
- Faculty of Engineering, the Built Environment and Technology
- Language
- English
- Rights
- Nelson Mandela University
- Hits: 1273
- Visitors: 1343
- Downloads: 301
Thumbnail | File | Description | Size | Format | |||
---|---|---|---|---|---|---|---|
View Details Download | SOURCE1 | Lennox, N 212231081 Dissertation April 2020.pdf | 2 MB | Adobe Acrobat PDF | View Details Download |