Leaf recognition based on leaf tip and leaf base using centroid contour gradient

In this paper, we suggest to normalize the leaf tip and leaf base as both of them may incline to one direction which able to influence the data extraction process. The features extraction method we used is Centroid Contour Gradient (CCG) which calculate the gradient between pairs of boundary point corresponding to interval angle, θ. CCG had outperformed its competitors which is Centroid Contours Distance (CCD) as it is successfully captures the curvature of leaf tip and leaf base. The accuracy to classify the leaf tip using CCG is 99.47%, and CCD is only 80.30%. For the accuracy of leaf base classification, CCG (98%) also outperforms CCD (88%). The average accuracy to recognize the 5 classes of plant is 96.6% for CCG and 74.4% for CCD. In this research, we utilized the Feed-forwad Back-propagation as our classifier.