Efficient infrared image processing and machine learning algorithm for breast cancer screening

Breast cancer is a significant global health concern, and early detection is crucial for improving survival rates. The article reviews various studies that investigate different methods for detecting breast cancer using non-invasive imaging techniques with a focus on thermal imaging, including feature extraction, image segmentation, and machine learning. It concludes that developing efficient and accurate breast cancer screening software requires an integrated approach to data collection, image processing, and machine learning algorithms. The article presents a novel technique for developing breast cancer screening software using Rotational Thermographic Imaging, dynamic temperature-based data collection, Colour-based Infrared Image Processing, and a Machine Learning algorithm to provide a complete breast imaging in sitting position to reduce the chances of missing abnormalities. Image processing and machine learning techniques are utilised to extract a comprehensive relevant feature set from the captured images and used to train a machine learning model. The system was tested on an increasing patient population in a clinical setting deployed at a hospital. The algorithm's performance was evaluated using several metrics, including sensitivity (82.14%), specificity (98.33%), and accuracy (93.27%). The results demonstrate that the proposed algorithm achieved high accuracy and sensitivity, making it a promising tool for breast cancer screening.