OKIRAKU search: Leaf images based visual tree search system

This paper addresses leaf images based visual tree search system called OKIRAKU Search. A user photographs an isolated leaf on a white background, and inputs the photographs to the system. The system automatically extracts the leaf region and computes shape features and color features, and searches it in the known species. The system shows the user the top matched species. The point of our approach is to accept either front leaf image or rear leaf image, or both images. We investigate the effectiveness of the features and images. In our experiment, 297 classes (278 species) are uniquely collected in the fields around our campus. The experimental results reveal that our proposed method obtains high performance. Moreover, we have developed a prototype tree search systems for mobile phone version and web-based version implemented the proposed method.