Detection of Microcalcification Using Mammograms

Mammography is one of the most common and useful techniques used for early detection of the breast cancer. It is the low-dose x-ray examination performed to patient to detect the primary mass when it is still small and confined to the breast. The present of microcalcification is a highly indication of the cancerous tissues. Microcalcification is a tiny specks of calcium deposited in the breast. The problem encountered in detecting the microcalcification by using this method is the limitation of the mammogram image (x-ray) to detect the microcalcification due to mainly to their small size, low contrast, and the similarity of their radiographic appearance to dense tissue. Statistic had shown that approximately 10%-30% of breast cancers retrospectively visible on the mammograms were missed ormisinterpreted due tohuman or technical factors [1]. This project focuses on the enhancement of the mammograms image by applying the image processing techniques to assist doctors in detecting the breast cancer disease. The aim is to provide a low-cost technology in detecting the breast cancer at the early stage. This project develops the program using MATLAB and Borland C++ to enhance the digitized mammograms image by using the image processing technique. The mammogram is first digitized and processed by the program developed to detect the microcalcification deposited in the breast. The morphological operation was a simple and suitablemethod in identifying the microcalcification. The top-hat algorithm method that is a morphological operation was developed using MATLAB and successfully obtained the output image that shows the candidate microcalcification. The top hat method consists of four stages which are digitization of mammograms, image enhancement, image segmentation and feature extraction. Various image processing techniques were applied including filters, histogram generation, thresholding and edge detection. The top hat method was applied to mammograms samples of eight patients and able to detect the microcalcification. The results obtained were defined into three categories, below expectation, meet expectation and above expectation. In conclusion, the project had met an acceptable degree ofaccuracy level