Breast Mass Abnormalities Classification using Architectural Distortions

In breast masses we can use architectural distortion for recounting the arrangement of breast huddle abnormalities in mammograms either as malignant otherwise benign categories. The proposed, designed system assists the radiologists to accurately tag the breast cancer as benign and malignant and also minimizes the false-positive classification rate of breast masses efficiently. Radiologist marks the detected mass abnormality on a mammogram and extracts two textural features architectural distortion and denseness at the marked space. The denseness features furnishes and compute the radiographic heaviness of the distinguished area and the architectural distortion quality gives an estimation of its abnormality and structure. Skew and kurtosis define the distortions in architectural dimension and structure, and with these parametric variations, a benign tumor is differentiated from a malignant tumors MIAS database is taken for this work. The proposed method verified with plenty of experiments and quantitative comparison with other methods achieved better performance, the precision accuracy values of the system 97.43% of the malignant, and 97.36% for the benign images.