• Telagarapu Prabhakar

Distinguishing between malignant and benign tumours is difficult and time-consuming task for radiologists. This leads to a high rate of unnecessary biopsies that can be avoided. So, we are proposing a method to effectively differentiate benign and malignant tumours with optimal feature set. Statistical texture features are extracted from tumour region by recent shearlet transform and then to reduce the dimensionality of the feature in order to reduce the computational complexity feature selection methods are applied. Feature selection methods are such as Principal component analysis (PCA), Genetic algorithm and ReliefF based feature selection methods are applied. Using the new reduced feature set abnormalities are classified using KNN and SVM classifier. By comparing the Performance of the feature selection methods ReliefF based feature selection with SVM classification method gives with the Accuracy of 78.33%, Sensitivity is 84.16% and 80.16% as specificity.

PCA, Genetic algorithm, Feature selection, Shearlet Transform,k-NN,SVM and ReliefF.

"Combining Support Vector Machine with Relief F algorithm to classify ultrasound breast tumour images", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.2, Issue 6, page no.2095-2100, July-2015, Available :http://www.jetir.org/papers/JETIR1506069.pdf

"Combining Support Vector Machine with Relief F algorithm to classify ultrasound breast tumour images", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.2, Issue 6, page no. pp2095-2100, July-2015, Available at : http://www.jetir.org/papers/JETIR1506069.pdf