Leaf image retrieval using combined feature of vein and contour

Abstract

This paper proposes a novel method of recognizing plant leaf image using a combination of venation and contour features. In this method, the shape is cut into pieces under different scales to describe the leaf image in a multiscale way. Both leaf venation and contour features extracted from the cut pieces are utilized to provide comprehensive description under each scale. The performance of the proposed method has been evaluated on two leaf datasets. And the experimental results indicates that the proposed method can achieve higher (or similar) recognition accuracies than the state-of-the-art approaches with acceptable speed.