Implementation and assessment of new hybrid model using CNN for flower image classification

Convolutional Neural Networks (CNN) is an advanced technique for image classification. Lots of CNN models have been used for the classification of objects in the images. CNN is trained using profound learning algorithms that have made some enormous achievements in the recognition of large-scale identification methods in the field of machine learning. This paper proposes hybrid models for flower classification in order to achieve better classification accuracy. The study implemented four different hybrid models; the first is VGG16+SVM, the second is ResNet50+SVM, then AlexNet + SVM, and the last hybrid model is GoogleNet + SVM. The ordered dataset conveys 6027 images of various species of flowers. The first execution model result the accuracy of 80.67%, the second model accuracy is of 90.01%, the third model result the accuracy of 80.27%, and the last model carried out an 82.54% total accuracy.