Medical External Wound Image Classification Using Support Vector Machine Technique.

Diagnosis is an activity that refers to the examination of something. Diagnosis is often associated with medical activities as a determinant of a person's condition, in the health sector diagnosis means a procedure performed by a doctor to determine a patient's condition. Unfortunately, it is rare to diagnose disease using an object wound, whereas if the wound is not treated immediately it can lead to more serious illnesses such as ulcers and tetanus or in some cases it can cause infection which then becomes a complication, in the worst case amputation occurs. The skin protects the body from various threats, the skin is also the first fortress for the body. Before implementing a prototype external wound diagnosis, it is necessary to test the accuracy of the algorithm to be used so we know that the algorithm used is suitable for the system to be developed. The algorithm that can be used for diagnosis or classification is the Support Vector Machine or SVM which in the process goes through 3 stages, namely data collection, preprocessing, and classification. This research obtained the results of feature extraction on the wound image test data using GLCM with a contrast value of 0.0082, a correlation value of 0.9769, an energy value of 0.6391, and a homogeneity value of 0.9959 as well as the accuracy of using the SVM algorithm which was measured using a confusion matrix to get an accuracy value of 96.39%, 93.06% precision, recall 92.85%, and F1-score 92.58%. The results of the accuracy of the classification of external wound images using the SVM algorithm are 92.85%. [ABSTRACT FROM AUTHOR]