Comparative Analysis Classifiers for Abdominal Mass Detection

Abstract

One of the practical ways of testing the internal organs is through Ultrasound image, however, the unprocessed image usually embedded with noises, which leads to be more tedious in granting clear view of the particular region that has affected. This paper aims in developing an advanced model for abdominal mass diagnosing with US images. The proposed detection model includes two phases: (i) Feature extraction and (ii)Classification. In the first phase, the texture features are extracted using Adaptive Gradient Location and Orientation histogram (AGLOH). Then, the second phase uses Linear Collaborative Discriminant Regression Classification (LCDRC) model to classify whether the given image is normal or abnormal. Further, this paper adopts an improved diagnosis precision while detecting mass that presents in abdominal regions. Furthermore, the proposed LCDRC classifier compares its performance over other conventional techniques include Support Vector Machine(SVM), Neural Network (NN), K-Nearest Neighbor (K-NN), Naive Bayes (NB) in terms of measures like Accuracy, Sensitivity, Specificity, Precision, False Positive rate (FPR), False Negative rate (FNR), Negative Prediction Value (NPV), False Discovery rate (FDR), Matthews correlation coefficient (MCC)and F₁Score, and the betterment of proposed method is proven.