A Hybrid Approach on Mammograms High Density Ratio Using Data Mining Techniques for Breast Abnormalities

Dr.C. Victoria Priscilla and N.M. Sangeetha

Diagnosing various breast diseases in the medical field is a difficult task even for medical experts and radiologist. The objective of the proposed hybrid approach is to find the breast abnormalities and its stages. In this research work, we concentrate on five important stages. The initial stage involves preprocessing using standard median filter ,the second stage involves density determination using formulation of density pixel and density ratio, the third stage involves the clustering procedure using K-means clustering , the fourth stage contains the computation of clustering and density ratio , the fifth stage rule based classification for finding the stages of abnormalities and detect the disease as well the patient’s curability level using the medical treatments. Performance attributes sensitivity, specificity, accuracy, precision and recall are calculated. The result shows to 94.1% accuracy and proposed method is used the decision making of the drug, and identification of the stages of breast abnormalities which helps the radiologist for getting better accuracy rate and results.

Volume 11 | Issue 1

Pages: 397-403