Identification of Malignant Regions using Infrared Image Processing

The purpose of this research is to see how thermal imaging detects aberrant cells in the human body. The goal of this study is to look at the unsolved challenges of identifying malignant regions. The suggested system uses thermal pictures as an input, for image processing one must follow pre-processing process.. It removes data that isn't relevant to the image processing and boosts certain image qualities that are important for further processing. The region of interest (ROI) is then detected using binarization, edge detection, and picture segmentation. We are satisfied with the results obtained utilizing thermal infrared cameras to detect the temperature of bodies and physical entities in nature, especially in terms of relevance and higher meaningfulness when compared to other competing approaches. The study employed advanced pattern classification techniques such as SVM and Random Forest, Nave Bayes and Neural Network K-Nearest Neighbour, Adaboost, and Logistic Regression, as well as the Visual Lab Breast Cancer Database using 5 folds, 10 folds, 20 folds with training set size 50, 60, 70, 80, and 90 respectively. Pre-processing, feature extraction, segmentation, and classification of digital mammography and digital infrared thermal pictures are also proposed. These techniques achieved an accuracy of 97.4%, 98.40%, and 99.7% through the random sampling cross-validation of Support Vector Machine, Random Forest method, and Logistic Regression respectively. Such kind of performance analysis on the thermal image of breast cancer is not found in reputed journals