An Efficient Texture Feature Selection and Classification of Mammographic Image using AQPSO

In computer-aided diagnosis systems, Image processing algorithms can be used to extract features directly from digitized mammograms. In general, two classes of features are extracted from mammograms with these algorithms, such as morphological and non-morphological features. Image texture analysis is one of an important technique that represents gray level properties of images used to illustrate non-morphological features. This technique has made known to be a promising technique in analyzing mammographic lesions caused by masses. The texture descriptor namely entropy, energy, sum average, sum variance, and cluster tendency has been analyzed for texture pattern ROI. These textures features are derived from co-occurrence matrices, wavelet and ridgelet transforms of mammographic images. Earlier work used Genetic algorithm and Random Forest algorithm for selection and classification of these features. In order to improve the performance, proposed system uses Adaptive Quantum-behaved Particle swarm optimization for feature selection process. Comparison of AQPSO with Genetic Algorithm can be done experimentally and proves that the proposed system provides better result when compare with existing work