Early Detection of Breast Cancer using Mammogram and Wavelet Transform for Woman Care

Mammography is the vital screening testfor early diagnosis of Breast cancer and thereby the death rate caused by this disease is greatly reduced. It is visually tough to recognize malignant micro-calcifications as it is a highly complicated issue. Current research in this area is facing challenges to afford accurate and best solution for the prediction of calcifications. The proposed system is developed for efficient classification of the images as normal and abnormal using wavelet transform and Neural Network. The noise present in the test images are removed by Oriented Rician Noise-Reducing Anisotropic Diffusion (ORNRAD) filter, which makes use of anisotropic diffusion for filtering images. Then the images are decomposed into sub-bands using Symlet filter and the crucial sub band features are extracted by employing statistical metrics such as Mean, Standard Deviation and Variance to make the classification results precise. Levenberg Marquardt is exploited as a training algorithm to train the neural networks and to classify the images.