Tri-texture feature extraction and region growing-level set segmentation in breast cancer diagnosis

Computer Aided Diagnosis (CAD) systems utilises the computer technology to detect and classify the normal and abnormal levels in breast cancer. This paper employs the series of feature extraction and the novel segmentation methods to improve the performance of cancer detection in breast region. Tri-texture feature extraction method such as grey level co-occurrence matrix (GLCM), Gabor and wavelet texture features are extracted from the segmented output. This paper employs the hybrid Genetic Algorithm (GA)Particle Swarm Optimisation (PSO) for relevant features for classification. Besides, the proposed work employs the two classifiers such as Support Vector Machine (SVM) (to classify normal and abnormal level) and Neural Network (NN) (to label the architectural distortion, asymmetry, masses and micro calcification). The hybrid Region Growing and Level (RGL) set methods provides the segmented output to analyse the abnormal categories. The utilisation of multiple methods improves the abnormality analysis of breast cancer diagnosis applications.