Performance Comparison of CNN Based Hybrid Systems Using UC Merced Land-Use Dataset

Fatma Yaşar Çıklaçandır; Semih Utku

Article Details

Performance Comparison of CNN Based Hybrid Systems Using UC Merced Land-Use Dataset

Authors:

Fatma Yaşar Çıklaçandır | Semih Utku

Publisher:

Dokuz Eylul University

Main Topics:

Bilgisayar Görüşü ve Çoklu Ortam Hesaplama (Diğer)

Journal:

Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 75, pp.725-737

Document Type:

Article

Publication Date:

9/2023

Document Language:

English

Document ID (Org No.):

2001298690

pISSN:

1302-9304

Keywords:

Abstract:

 Remote sensing is the technology of collecting and examining data about the earth with special sensors. The data obtained are used in many application areas. The classification success of remote sensing images is closely related to the accuracy and reliability of the information to be used. For this reason, especially in recent studies, it is seen that Convolutional Neural Network CNN, which has become popular in many fields, is used and high successes have been achieved. However, it is also an important need to obtain this information quickly. Therefore, in this study, it is aimed to get results as successful as CNN and in a shorter time than CNN. Hybrid systems in which features are extracted with CNN and then classification is performed with machine learning algorithms have been tested. The successes of binary combinations of two different CNN architectures ResNet18, GoogLeNet and four different classifiers Support Vector Machine, K Nearest Neighbor, Decision Tree, Discriminant Analysis have been compared with various metrics. GoogLeNet  Support Vector Machine 93.33 is the method with the highest accuracy rate, while ResNet18  Decision Tree 50.95 is the method with the lowest accuracy rate.

 Remote sensing is the technology of collecting and examining data about the earth with special sensors. The data obtained are used in many application areas. The classification success of remote sensing images is closely related to the accuracy and reliability of the information to be used. For this reason, especially in recent studies, it is seen that Convolutional Neural Network CNN, which has become popular in many fields, is used and high successes have been achieved. However, it is also an important need to obtain this information quickly. Therefore, in this study, it is aimed to get results as successful as CNN and in a shorter time than CNN. Hybrid systems in which features are extracted with CNN and then classification is performed with machine learning algorithms have been tested. The successes of binary combinations of two different CNN architectures ResNet18, GoogLeNet and four different classifiers Support Vector Machine, K Nearest Neighbor, Decision Tree, Discriminant Analysis have been compared with various metrics. GoogLeNet  Support Vector Machine 93.33 is the method with the highest accuracy rate, while ResNet18  Decision Tree 50.95 is the method with the lowest accuracy rate.

 Remote sensing is the technology of collecting and examining data about the earth with special sensors. The data obtained are used in many application areas. The classification success of remote sensing images is closely related to the accuracy and reliability of the information to be used. For this reason, especially in recent studies, it is seen that Convolutional Neural Network CNN, which has become popular in many fields, is used and high successes have been achieved. However, it is also an important need to obtain this information quickly. Therefore, in this study, it is aimed to get results as successful as CNN and in a shorter time than CNN. Hybrid systems in which features are extracted with CNN and then classification is performed with machine learning algorithms have been tested. The successes of binary combinations of two different CNN architectures ResNet18, GoogLeNet and four different classifiers Support Vector Machine, K Nearest Neighbor, Decision Tree, Discriminant Analysis have been compared with various metrics. GoogLeNet  Support Vector Machine 93.33 is the method with the highest accuracy rate, while ResNet18  Decision Tree 50.95 is the method with the lowest accuracy rate.

 Uzaktan algılama, yeryüzü ile ilgili verilerin özel sensörler aracılığıyla toplanması ve incelenmesi teknolojisidir. Elde edilen veriler birçok uygulama alanında kullanılmaktadır. Uzaktan algılama görüntülerinin sınıflandırma başarısı, kullanılacak bilgilerin doğruluğu ve güvenilirliği ile yakından ilgilidir. Bu nedenle özellikle son yıllarda yapılan çalışmalarda birçok alanda popüler hale gelen Konvolüsyonel Sinir Ağlarının CNN kullanıldığı ve yüksek başarılar elde edildiği görülmektedir. Ancak bu bilgilerin hızlı bir şekilde elde edilmesi de önemli bir ihtiyaçtır. Dolayısıyla bu çalışmada CNN kadar başarılı ve CNNden daha kısa sürede sonuç alınması amaçlanmaktadır. CNN ile özniteliklerin çıkarıldığı ve daha sonra makine öğrenmesi algoritmaları ile sınıflandırmanın yapıldığı hibrit sistemler test edilmiştir. İki farklı CNN mimarisi ResNet18, GoogLeNet ve dört farklı sınıflandırıcının Support Vector Machine, K Nearest Neighbor, Decision Tree, Discriminant Analysis ikili kombinasyonlarının başarıları çeşitli metriklerle karşılaştırılmıştır. GoogLeNet  SVM 93,33 en yüksek doğruluk oranına sahip yöntem olurken, ResNet18  DT 50,95 en düşük doğruluk oranına sahip yöntemdir.

Article Details

Performance Comparison of CNN Based Hybrid Systems Using UC Merced Land-Use Dataset

Authors:

Fatma Yaşar Çıklaçandır | Semih Utku

Publisher:

Dokuz Eylul University

Main Topics:

Bilgisayar Görüşü ve Çoklu Ortam Hesaplama (Diğer)

Journal:

Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 75, pp.725-737

Document Type:

Article

Publication Date:

9/2023

Document Language:

English

Document ID (Org No.):

2001298690

pISSN:

1302-9304

Remote Sensing

Convolutional Neural Network

Machine Learning

Hybrid Systems

Uzaktan Algılama

Konvolüsyonel Sinir Ağı

Makine Öğrenmesi

Hibrit SistemlerUzaktan Algılama

Konvolüsyonel Sinir Ağı

Makine Öğrenmesi

Hibrit Sistemler

Issues

Article Details

Authors: Fatma Yaşar Çıklaçandır | Semih Utku

Publisher: Dokuz Eylul University

Main Topics: Bilgisayar Görüşü ve Çoklu Ortam Hesaplama (Diğer)

Journal: Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 75, pp.725-737

Document Type: Article

Publication Date: 9/2023

Document Language: English

Document ID (Org No.): 2001298690

pISSN: 1302-9304

Issues

Authors:

Fatma Yaşar Çıklaçandır | Semih Utku

Publisher:

Dokuz Eylul University

Main Topics:

Bilgisayar Görüşü ve Çoklu Ortam Hesaplama (Diğer)

Journal:

Dokuz Eylül Üniversitesi Mühendislik Fakültesi Fen ve Mühendislik Dergisi Vol.25 Issue 75, pp.725-737

Document Type:

Article

Publication Date:

9/2023

Document Language:

English

Document ID (Org No.):

2001298690

pISSN:

1302-9304