Detection of Skin Cancer Image by Feature Selection Methods Using New Buzzard Optimization (BUZO) Algorithm

Abstract

Feature selection is used in machine learning as well as in statistical pattern recognition. This is important in many applications, such as classification. There are so many extracted features in these applications which are either useless or do not have much information. If not removing these features, make raises the computational burden for the main application. In different methods of feature selection, a subset is selected as the answer, which can optimize the value of an evaluation function. In this study, a new algorithm for classification of Dermoscopy images into two types of malignant and benign are presented. To develop the general skin cancer detection system, at first a pre-processing step is applied to enhance image quality. Then the lesion area is removed from the healthy areas using the Otsu threshold method. Nine shape feature and nine color features are extracted from the segmented image using different optimization schema. At the end of the operation, classification was done by SVM, KNN and Decision Tree methods. The results show that combination of buzzard optimization algorithm for feature extraction and SVM classifier accuracy is 94.3%. This result shows the high potential of buzzard optimization algorithm for feature extraction.

Optical Technologies for the Improvement of Skin Cancer Diagnosis: A Review

“…Most of ML approaches try to reproduce physicians’ assessments, the well-known ABCDE rule, and because of that, the next step consists in feature extraction of morphological (e.g. area, perimeter, symmetry, and eccentricity) and colorimetric (e.g., hue, saturation, lightness, and homogeneity) data from lesions [ 110 , 111 , 112 ]. Textural parameters have also shown to be very useful, since they quantify image features from a statistical approach considering image histogram, binary patterns, Gabor filters, or gray-level co-occurrence [ 113 , 114 ].…”

“…There is not an ideal classification algorithm that outperforms the others as it depends on the dataset, image segmentation, and feature extraction (see Figure 9 ); however, in the literature, some of them have been found to provide better discrimination among skin lesions. Support Vector Machine (SVM) is the most applied classifier, as it shows the best performance [ 107 , 109 , 110 , 111 , 112 , 114 , 115 , 116 , 117 , 118 ] followed by k-Nearest Neighbors (KNN) [ 65 , 106 , 107 , 108 , 109 , 110 , 111 , 116 , 117 , 118 ], Neural Networks (NN) [ 110 , 114 , 115 , 117 , 119 ], Random Forest (RF) [ 107 , 109 , 113 , 114 ] and Decision Trees (DT) [ 65 , 109 , 110 , 111 , 117 , 118 ]. SVMs are based on statistics to build hyperplanes that maximize the distance between sets of data points [ 120 , 121 ].…”

Skin Lesions Classification Using Deep Learning Techniques: Review

“…By comparison, this feature selection approach selects numerous features but their impact on the process is important. Global optimization and rapid convergence are excellent features of the BUZ algorithm [51].…”

“…The pros of CNN approaches are that they usually have higher PSNR values and can outperform other filters (Adaptive Wiener filter, Median filter, Adaptive Median filter, and Wiener filter). Arshaghi et al [51] a new classification algorithm for dermabbed images designed for classification skin lesions malignant from benign conditions. The image quality is improved by a pre-processing stage.…”

A Markov-Based Intrusion Tolerance Finite Automaton

“…In view of the lack of an intelligent detection algorithm for intrusion tolerance systems, Hajisalem and Babaie [11] combined artificial bee colony (ABC) [12,13] and artificial fish swarm (AFS) [14,15] into a hybrid classification method. First, the training dataset was segmented by fuzzy mean clustering and correlation-based feature selection, respectively; the Classification and Regression Tree (CART) was implemented to generate If-Then rules based on the selected features, so as to distinguish between normal records and abnormal intrusion records.…”