Flower Species Classification using Statistical and Gist Descriptors

This Paper aims at classification of flower images by means of Gist descriptor and statistical features using SVM classifier. With advances in digital image processing, automated classification of flower images over large categories of dataset is possible by knowing the name of the flower. In case of unknown species name, classification of flower image can be performed by its visual content. The purpose of the paper is to identify different species of flower with the help of visual content. Here hundred and two categories of flower dataset are used and it is obtained from Visual Geometry Group, university of Oxford. In this paper ten classes are taken from dataset and classified. The Gist descriptor is combined with statistical features such as mean, standard deviation, skewness, Kurtosis are given to SVM classifier. It uses multiple kernel frame work to classify the images. The recognition rate is 79.36%.