Recognition of leaf based on its tip and base using centroid contour gradient

Abstract

This paper suggests normalization of the tip and base of leaf as both of them incline to one direction which is able to influence data extraction process. The extraction method we used is Centroid Contour Gradient (CCG) which calculates the gradient between pairs of boundary points corresponding to interval angle. CCG had outperformed its competitor which is Centroid Contours Distance (CCD), as it successfully captures the curvature of the tip and base of leaf. The accuracy in classifying the tip of leaf using CCG is 99.47%, but CCD is only 80.30%. For accuracy of leaf base classification, CCG (98%) also outperforms CCD (88%). The average accuracy for recognizing the 5 classes of plant is 96.6% for CCG and 74.4% for CCD. In this research, we utilized the Feed-forward Back-propagation as our classifier.