Mango Leaf Classification with Boundary Moments of Centroid Contour Distances as Shape Features

The previous research in mango leaf classification which used 270 features consisted of 256 texture features, 2 color features, and 2 shape features, could not achieve high classification performance. In this study, we conduct improvement by combining the previous features with the Boundary Moments of Centroid Contour Distance (CCD) and classify the combination features using Support Vector Machine with Linear and RBF kernels. The experiment results show that the combination features achieve higher classification performance compared to the previous features.