Breast Cancer Diagnosis by Negative Association Rule Classifier from Mammogram

Abstract

Breast cancer is the major cause of cancer death among female. Screening mammography is the primary method is used for the reliable detection of early and potentially curable breast cancer. Research indicates that the mortality rate could decrease by 30% if women age 50 and older have regular mammograms. The detection rate can be increased 5-15% by providing the radiologist with results from a computer-aided diagnosis (CAD) system acting as a second opinion.

It would be substantial advantage if an accurate computer aided diagnostic system is existed to diagnose normal cases of mammograms and thus allowing the oncologist to focus on suspicious cases. This strategy could reduce the radiologist's workload and to confirm the accurate screening performance. The texture is one of the major classical features of image data which is used for recognizing regions of interest in an image. In image analysis, textural features are those features in which a specific pattern of data distribution is repeated sequentially throughout the image. Feature selection is a key function in various image processing techniques. A feature is an image contents that can capture certain visual characteristics of the image. Texture is a concept of important feature of many image types, which is the pattern of data or arrangement of the structure found in a picture. Texture features are used in different applications such as image mining, remote sensing and content-based image retrieval. These features can be extracted in many different ways. The most general and usual way is using a Gray Level Co-occurrence Matrix (GLCM). GLCM contains the second order statistical attribute of an image. Textural features can be calculated from GLCM to calculate the details about the image pattern. The texture statistical second order method considered is spatial gray level dependence method, gray level run length method and gray level difference method. Features are extracted from the first-order statistical method and second-order statistical method and are combined. It is observed that the result of these combined features provides higher accuracy when compared with the features from the first-order statistical method and second-order statistical method alone.

 The purpose of our experiments is to explore the feasibility approach to extract patterns and whether that pattern will be helpful to diagnose breast cancer and tissue as well as increase the diagnostic accuracy for optimum classification between normal and abnormalities in digital mammograms. Result shows very reliable and the accuracy level which is very encouraging in compared to other techniques. It is well understood that data mining techniques are more reliable for larger databases than the one used for these preliminary tests. Computer-aided diagnostic method using association rule mining may assist medical professionals and improve the accuracy of mammogram detection. In particular, a Computer aided method based on association rules may be more précised for a larger dataset .Experimental results show that this proposed method can quickly and effectively mine potential association rules.