4. Conclusion
In this paper, we have proposed a probabilistic neural network-based flower classification method with the use of
texture features. Suitable texture features such as CTMs, GLCM, and Gabor responses are explored for the purpose of
flower classification. It is observed that using the proposed textural features one can achieve relatively a good classification
accuracy when compared to any other available features. We have created our own database of flowers of 35 classes, each
containing 50 flower images and conducted an experiment under varying database size and we studied the size effect on
the classification accuracy. The experimental results have shown that using combined features outperforms any individual
feature.