Hossam, A., Harb, H., Abd El Kader, H. (2018). AUTOMATIC IMAGE SEGMENTATION METHOD FOR BREAST CANCER ANALYSIS USING THERMOGRAPHY. JES. Journal of Engineering Sciences, 46(No 1), 12-32. doi: 10.21608/jesaun.2017.114377
Aya Hossam; Hany M. Harb; Hala M. Abd El Kader. "AUTOMATIC IMAGE SEGMENTATION METHOD FOR BREAST CANCER ANALYSIS USING THERMOGRAPHY". JES. Journal of Engineering Sciences, 46, No 1, 2018, 12-32. doi: 10.21608/jesaun.2017.114377
Hossam, A., Harb, H., Abd El Kader, H. (2018). 'AUTOMATIC IMAGE SEGMENTATION METHOD FOR BREAST CANCER ANALYSIS USING THERMOGRAPHY', JES. Journal of Engineering Sciences, 46(No 1), pp. 12-32. doi: 10.21608/jesaun.2017.114377
Hossam, A., Harb, H., Abd El Kader, H. AUTOMATIC IMAGE SEGMENTATION METHOD FOR BREAST CANCER ANALYSIS USING THERMOGRAPHY. JES. Journal of Engineering Sciences, 2018; 46(No 1): 12-32. doi: 10.21608/jesaun.2017.114377
AUTOMATIC IMAGE SEGMENTATION METHOD FOR BREAST CANCER ANALYSIS USING THERMOGRAPHY
2Computers and Systems Eng. Depart, Faculty of Engineering, Azhar university, Cairo, Egypt
Abstract
Breast Cancer (BC) is considered one from various diseases that has got great attention in the last decades. This is because the high mortality rates among young women in the whole world according to this disease. The early detection of this disease is necessary to enhance the opportunity of survival. Thermography is an efficient screening tool that can help in detecting BC by indicating parts of the body where an abnormal temperature variation is found. To realize an effective BC detection system using thermography, it is necessary to locate the region of interest (ROI) in the thermograms prior to analysis. This paper introduces a new automatic segmentation method (SM) for identifying the ROI image from breast thermograms. It depends on the statistics of image in DMR-IR database. After detecting the boundaries of ROI using this new SM, two different approaches for accurate identifying and enhancing the breast boundaries are investigated: the Hough transform (HT) algorithm to locate and define the parabola curves in the breast image, and a second approach introducing the enhancement of the contrast of detected boundaries in the breast ROI image. To evaluate the results of this method, statistical features have been extracted from the segmented ROI image. Then, we used the Support Vector Machine (SVM) and the Artificial Neural Network (ANN) classifiers to detect the normal and abnormal breasts based on these features. The experimental results prove that the accuracy of SVM, and ANN classifiers reached to 96.67% and 96.07%, respectively. This denotes that the proposed automatic SM is a favorable technique for extracting the breast ROI image from breast thermograms.
View on SCiNiTO
Top