early breast cancer detection in thermogram images using adaboost classifier and fuzzy c-means clustering algorithm

Background: in this paper we compare a highly accurate supervised to anunsupervised technique that uses breast thermal images with the aim ofassisting physicians in early detection of breast cancer.methods: first, we segmented the images and determined the region ofinterest. then, 23 features that included statistical, morphological, frequencydomain, histogram and gray-level co-occurrence matrix based features wereextracted from the segmented right and left breasts. to achieve the bestfeatures, feature selection methods such as minimum redundancy and maximumrelevance, sequential forward selection, sequential backward selection,sequential floating forward selection, sequential floating backward selection,and genetic algorithm were used. contrast, energy, euler number, and kurtosiswere marked as effective features.results: the selected features were evaluated by fuzzy c-means clusteringas the unsupervised method and compared with the adaboost supervisedclassifier which has been previously studied. as reported, fuzzy c-meansclustering with a mean accuracy of 75% can be suitable for unsupervisedtechniques.conclusion: fuzzy c-means clustering can be a suitable unsupervisedtechnique to determine suspicious areas in thermal images compared toadaboost as the supervised technique with a mean accuracy of 88%.