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

Applied Sciences

Roberto Corizzo¹  Paolo Mignone²  Petre Lameski³  Eftim Zdravevski³  Tatjana Atanasova-Pacemska⁴  Biserka Petrovska⁵ 

[1] Department of Computer Science, American University, 4400 Massachusetts Ave NW, Washington, DC 20016, USA;Department of Computer Science, University of Bari Aldo Moro, Via E. Orabona, 4, 70125 Bari, Italy, [e-mail];Faculty of Computer Science and Engineering, Ss. Cyril and Methodius University in Skopje, Rugjer Boshkovik 16, 1000 Skopje, North Macedonia;Faculty of Computer Science, University Goce Delcev, 2000 Stip, North Macedonia;Ministry of Defence, 1000 Skopje, North Macedonia;

关键词: remote sensing;    convolutional neural network;    fine-tuning;    learning rate scheduler;    cyclical learning rates;    label smoothing;   

DOI  :  10.3390/app10175792

来源: DOAJ

【 摘 要 】

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.

【 授权许可】

Unknown