Skin Cancer Detection and Classification Using DWT-GLCM with Probabilistic Neural Networks

Abstract

Skin cancer is a leading type of cancer which causes millions of deaths of human beings. Early identification and appropriate medications for new harmful skin malignancy cases are fundamental to guarantee a low death rate as the survival rate. Most of the related works focus on machine learning-based algorithms, but they failed to provide the maximum accuracy and specificity. Thus, to overcome this problem, integrated unsupervised learning with probabilistic neural network (PNN)-based skin cancer detection and classification mechanism is proposed. Initially, the unsupervised learning algorithm, i.e., k-means clustering is used for efficient detection of cancer-affected region from skin images. Then, various features are extracted from the segmented outcome by employing the gray-level co-occurrence matrix (GLCM), discrete wavelet transform (DWT), and computation of image statistics for texture, and low-level and color features, respectively. Finally, to archive the maximum efficiency of the system, the PNN was developed for classification of skin cancer with the gray level co-occurrence matrix (GLCM)-based texture features, discrete wavelet transform (DWT)-based low level features, and statistical color features, respectively. Thus, the research work can be effectively used for the classification of benign and malignant skin cancers. The simulation analysis shows that the proposed method shows better qualitative and quantitative analysis compared to the state of art approaches.