Leaf classification based on shape and edge feature with k-NN classifier

摘要

In this paper a new approach is proposed to classify leafs in efficient and effective manner. In this approach, the features are extracted from edges of the leaf images and it is used two kinds of features which are edge based and shape based features for leaf classification. When the present method is tested on Flavia dataset, it gives the average classification accuracy rate of 94.37%. The Flavia dataset contains 32 kinds of plant leaves. The experimental result shows that the method gives better performance results compared with existing methods.