Image Task Detection of Microcalcification on Mammogram

Texture analysis has been very much used in medical image problems as well as related areas such as computer vision and pattern recognition. Among all medical image task detection of microcalcification on mammograms is the most difficult one because breast cancer is the most prevalent cancer that leads to death in women today. More over microcalcification are deposits of calcium that can be seen in mammograms, which is the best way to detect breast cancer in the earliest stage and also to reduce death from breast cancer. Owing to the small size of micro calcification with a diameter of less than 0.5 mm level and are in the form of groups as clusters and in homogeneous back ground it is very difficult to detect. In this paper, we propose to develop a supervised texture mammography technique for image classification that includes supervised and unsupervised methods. In the unsupervised method for the detection of microcalcification, the prior information is required and in the case of supervised method information on microcalcification is very much needed for the processing. Previously many methods have been developed for the detection of microcalcification on mammogram. This paper analyses the texture analysis of mammography images using supervised and unsupervised classification methods and the results identifies that the unsupervised method has high accuracy than the supervised methods.