Detection of skin cancer using support vector machine classifier compare with convolutional neural network classifier based on accuracy

Machine learning algorithms are efficient to improve accuracy in prediction and detection of various critical diseases. So, these are widely used in medical imaging. The main aim of work is to predict Skin cancer using Novel Machine Learning classifier. In this research, convolutional neural network classifier is used to predict Skin cancer disease to improve the a ccuracy, sensitivity, and specificity it is compared with novel support vector machine classifier. A total of twenty samples are collected from the Skin cancer Predefined classification set from Kaggle. Initially, the pores and skin image are used by filtering and segmented Gaussian using lively contour. Segmented images are fed as input to extraction feature. Accuracy is calculated by using the convolutional neural network classifier. Whereas G power considered as 0.8. From the MATLAB simulation results, Novel support vector machine (SVM) classifier achieved accuracy of 94.30 %, and convolutional neural network classifier achieved accuracy of 95.91 %. The 2 tailed significance value is.000 and p less than 0.05. From results it is observed that the convolutional neural network classifier appears significantly better than novel support vector machine classifier.