An automatic flower classification approach using machine learning algorithms

This work aims to develop an effective flower classification approach using machine learning algorithms. Eight flower categories were analyzed in order to extract their features. Scale Invariant Feature Transform (SIFT) and Segmentation-based Fractal Texture Analysis (SFTA) algorithms are used to extract flower features. The proposed approach consists of three phases namely: segmentation, feature extraction, and classification phases. In segmentation phase, the flower region is segmented to remove the complex background from the images dataset. Then flower image features are extracted. Finally for classification phase, the proposed approach applied Support Vector Machine (SVM) and Random Forests (RF) algorithms to classify different kinds of flowers. An experiment was carried out using the proposed approach on a dataset of 215 flower images. It shows that Support Vector Machine (SVM) based algorithm provides better accuracy compared to the Random Forests (RF) algorithm when using the SIFT as a feature extraction algorithm. While, Random Forests (RF) algorithm provides its better accuracy with SFTA. Moreover, the system is capable of automatically recognize the flower name with a high degree of accuracy.