Comparative Analysis of CNN Models and Bayesian Optimization-Based Machine Learning Algorithms in Leaf Type Classification

Authors : Muhammet Fatih ASLAN

Pages : 13-24

Doi:10.17694/bajece.1174242

View : 10 | Download : 4

Publication Date : 2023-01-30

Article Type : Research Paper

Abstract :In this study, the leaves are classified by various Machine Learning insert ignore into journalissuearticles values(ML); and Deep Learning insert ignore into journalissuearticles values(DL); based Convolutional Neural Networks insert ignore into journalissuearticles values(CNN); methods. In the proposed method, first, image pre-processing is performed to increase the accuracy of the posterior process. The obtained image is a grayscale image without noise as a result of the pre-processing. These preprocessed images are used in classification with ML and DL. The Speeded Up Robust Features insert ignore into journalissuearticles values(SURF); are extracted from the grayscale image for ML-based learning. The features are restructured as visual words using the Bag of Visual Words insert ignore into journalissuearticles values(BoVW); method. Then, histograms are generated for each image according to the frequency of the visual word. Those histograms represent the new feature data. The histogram features are classified by four different ML methods, Decision Tree insert ignore into journalissuearticles values(DT);, k-Nearest Neighbor insert ignore into journalissuearticles values(KNN);, Naive Bayes insert ignore into journalissuearticles values(NB); and Support Vector Machine insert ignore into journalissuearticles values(SVM);. Before using the ML methods, Bayesian Optimization insert ignore into journalissuearticles values(BO); method, which is one of the Hyperparameter Optimization insert ignore into journalissuearticles values(HO); algorithms, is applied to determine hyperparameters. In the classification process performed with four different ML algorithms, the best accuracy is achieved with the KNN algorithm as 98.09%. Resnet18, ResNet50, MobileNet, GoogLeNet, DenseNet, which are state-of-the-art CNN architectures, are used for DL-based learning. CNN models have higher accuracy than ML algorithms.
Keywords : Bag of Visual Words, Bayesian Optimization, Convolutional Neural Networks, Deep Learning, Speeded Up Robust Features