Scabies Skin Classification Using SVM Method With GLCM Feature Extraction

The skin is the first organ in the human body, pain and other factors. Lack of knowledge about the types of scabies and not knowing how to prevent it can cause a person to get this skin disease. To perform image analysis based on the statistical distribution of the pixel intensity, it can be done by extracting features. In the feature extraction process using Contrast, Correlation, Energy, Homogeneity is good in determining the value of skin texture or defects while in the results of segmentation and feature extraction based on shape and texture characteristics, feature extraction based on skin texture can be a variable in classifying skin images. The increase in accuracy value in the processing of skin image classification of artificial neural networks has a less high accuracy value compared to the Support Vector Machine method based on class or skin labels. At an angle of 0° precision gets 91%, recall 90%, and accuracy 95%. For an angle of 45° precision 96%, recall 96% accuracy 95%. At the GLCM angle to 90° precision 92%, recall 92% and accuracy 93%. Angle 135° precision 87% Recall 86% and Accuracy 91%, it can be concluded that in ANN the best accuracy is obtained at a GLCM angle of 45° with an accuracy of 98%. As for the results of the support vector machine at an angle of 0° the precision gets 98%, 98% recall, and 99% accuracy. For an angle of 45° precision 98%, recall 98% accuracy 99%. At the GLCM angle to 90° precision 96%, recall 96% and accuracy 98%. Angle 135° precision 98%, Recall 98% and Accuracy 99%. The highest accuracy is obtained at GLCM angles of 0°, 45°, 135° Support Vector Machine obtains the highest accuracy with an accuracy of 99%.