Aerial Scene Classification through Fine-Tuning with Adaptive Learning Rates and Label Smoothing

Petrovska, Biserka and Atanasova-Pacemska, Tatjana and Corizzo, Roberto and Mignone, Paolo and Lameski, Petre and Zdravevski, Eftim (2020) Aerial Scene Classification through Fine-Tuning with Adaptive Learning Rates and Label Smoothing. Applied Sciences, 10 (5792). pp. 1-25. ISSN 2076-3417

[thumbnail of Trud so IF, 2020, mdpi.pdf]
Trud so IF, 2020, mdpi.pdf

Download (2MB) | Preview


Remote Sensing (RS) image classification has recently attracted great attention for its application
in different tasks, including environmental monitoring, battlefield surveillance, and geospatial object
detection. The best practices for these tasks often involve transfer learning from pre-trained Convolutional
Neural Networks (CNNs). A common approach in the literature is employing CNNs for feature extraction,
and subsequently train classifiers exploiting such features. In this paper, we propose the adoption
of transfer learning by fine-tuning pre-trained CNNs for end-to-end aerial image classification. Our
approach performs feature extraction from the fine-tuned neural networks and remote sensing image
classification with a Support Vector Machine (SVM) model with linear and Radial Basis Function (RBF)
kernels. To tune the learning rate hyperparameter, we employ a linear decay learning rate scheduler
as well as cyclical learning rates. Moreover, in order to mitigate the overfitting problem of pre-trained
models, we apply label smoothing regularization. For the fine-tuning and feature extraction process,
we adopt the Inception-v3 and Xception inception-based CNNs, as well the residual-based networks
ResNet50 and DenseNet121. We present extensive experiments on two real-world remote sensing image
datasets: AID and NWPU-RESISC45. The results show that the proposed method exhibits classification
accuracy of up to 98%, outperforming other state-of-the-art methods.

Item Type: Article
Impact Factor Value: 2.474
Subjects: Natural sciences > Computer and information sciences
Engineering and Technology > Electrical engineering, electronic engineering, information engineering
Natural sciences > Matematics
Divisions: Faculty of Computer Science
Depositing User: Tatjana A. Pacemska
Date Deposited: 06 Oct 2020 08:45
Last Modified: 06 Oct 2020 08:45

Actions (login required)

View Item View Item