Analysis of Performance of Two Wavelet Families Using GLCM Feature Extraction for Mammogram Classification of Breast Cancer

Abstract

: Mammography is the technique to detect breast cancer abnormal tissues using digital screening. It is the most efficient method to detect the cancerous tissues in the breast. But as the data for detecting, the abnormal tissue is very large, so it is a very inappropriate method for some radiologists to detect the abnormal tissues correctly. Therefore, computer-aided diagnosis is useful for detecting the cancerous tissues. For this, feature extraction and selection is considered an important and efficient method for mammogram classification of breast cancer. In this proposed work, the focus is made on wavelet family performance named db8 and bior3.7 used for extracting the features using GLCM feature extraction technique and 27 texture features are extracted at each level of decomposition and then the classification is done using different classifiers named Fuzzy-NN, Naive Bayes, MLP and Genetic programming. After this, the feature selection method is also applied to the extracted features named as PCA and Wavelet and then the comparison is made with different classification algorithms for both the wavelet family.

Breast cancer diagnosis through an optimization‐driven multispectral gamma correction (ODMGC)

“…Two factors are used to calculate this matrix: θ, which means the relative orientation or rotational angle (0°, 45°, 90°, … 315°), and d, which expresses the distance between pixel pairs in terms of pixel numbers (1, 2, 3, etc.). In this article, the GLCM with d = 1 and θ = 0° is applied and then subjected to PCA analysis for feature dimension reduction 37 …”