Perfomance Analysis of Support Vector Machine (SVM) and Probabilistic Neural Network (PNN) on Identification System of Medicinal Plant and House Plant Based on Image

Abstract

This research analyzed perfomance of clasification for plant identification using Support Vector Machine (SVM) and Probabilistic Neural Network (PNN). In this research we used kernel linear, polynomial and RBF for clasifier SVM. In thus research, we use 1.440 medicinal plant images and 300 house plant images belong to 30 are extracted using Fuzzy Local Binary Patern based on texture feature.The expermental result shows that SVM kernel polynomial is superior compare to PNN with accuracy 73.57% to medicinal plant identification and SVM kernel RBF is superior compare to PNN with accuracy 81.11% to house plant identification. The proposed system is promising to improve identification medicinal plant and house plant.