A Comparative Study of Leaf Disease Classification using SVM Kernel Functions

Abstract

This leaf disease classification is an important task to be performed. So, leaf extraction from the complex background is a difficult task and to achieve this different algorithm are used for segmentation. Active contour is used for the segmentation and in this paper three different algorithms are used Gradient Vector Flow (GVF), Vector Flow Convolution (VFC) and Adaptive Diffusion Flow (ADF). Using these methods segmentation is performed and later 7 features are extracted from the image. Finally, it will be fed to the Support Vector Machine (SVM) classifier and in this comparison of different kernel functions is used.  The linear kernel function gives 90 % accuracy result when compare to other kernel functions.