Efficient feature fusion, selection and classification technique for Plant Leaf Image Retrieval system

Abstract

In recent days, Digitized image retrieval system is becoming more popular for retrieval of plant leaf species by botanist. There are many problems in using other retrieval system and is also a time consuming process. The digitized image retrieval system is still a challenging process due to lack of efficient pre-processing, feature extraction and classification technique. In order to overcome the difficulties this paper developed an approach for Plant Leaf Image Retrieval(PLIR) system using boundary extraction, feature fusion (shape, color and texture feature) extraction, feature selection using Genetic Algorithm(GA) and an efficient classification using Support Vector Machine (SVM) is proposed. The ultimate goal of our approach is to develop a system where user can easily process a unknown plant leaf with efficient image processing and computing technique within fraction of seconds. The effectiveness of the system is proved by quantitative approach. The performance of the proposed work is compared with the existing traditional classification algorithm and real time unknown plant leaf image.