An efficient feature extraction technique for breast cancer diagnosis using curvelet transform and swarm intelligence

Breast cancer is one of the major cause of death among women in the whole world. Accurate detection and early diagnosis can increase a survival of the patients. In this paper, an efficient feature extraction method using curvelet transform and feature selection using swarm intelligence has been proposed. In this work, coefficients are extracted from the Region of Interest (ROI) mammogram images by fast discrete curvelet transform via wrapping method. The extracted curvelet coefficients are too high to be classified and in turn time complexity is also very high. In order to reduce the time complexity and to select the prominent features, optimization algorithm using swarm intelligence has been proposed. Then the selected features are given to the SVM classifier. A set of images (182) from Mammographic Image Analysis Society (MIAS) database is taken for evaluating the system. The performance is analysed using classification accuracy rate. The experimental results are compared with the wavelet transform. The results shows that curvelet transform with particle swarm optimization (PSO) has high classification accuracy rates than the other existing methods for diagnosing the breast cancer.