Breast cancer is very common among the World's women population. Early detection of breast cancer can save life. Mammography based diagnosis is considered as the most effective to detect breast cancer. But, mammogram images often lead to misdiagnosis due to their low contrast nature. Recent studies show that mammography based diagnosis fails to detect cancerous lumps in the breast in ten out of one hundred patients. In this paper we address this issue. We propose a computer aided breast cancer detection technique in this paper. In the proposed method mammogram images are enhanced and segmented to locate the malignant lumps in the women's breast. The segmented image is then compared with a template image in order to determine the stage of breast cancer. Some statistical characteristics have also been presented in this paper to identify the malignant lumps from the benign ones.