Tampere University of Technology

TUTCRIS Research Portal

Performance Comparison of Learned vs. Engineered Features for Polarimetric SAR Terrain Classification

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Standard

Performance Comparison of Learned vs. Engineered Features for Polarimetric SAR Terrain Classification. / Ahishali, Mete; Ince, Turker; Kiranyaz, Serkan; Gabbouj, Moncef.

2019 41st Photonics & Electromagnetics Research Symposium (PIERS). 2019.

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Harvard

Ahishali, M, Ince, T, Kiranyaz, S & Gabbouj, M 2019, Performance Comparison of Learned vs. Engineered Features for Polarimetric SAR Terrain Classification. in 2019 41st Photonics & Electromagnetics Research Symposium (PIERS). Photonics & Electromagnetics Research Symposium, Rome, Italy, 17/06/19. https://doi.org/10.1109/PIERS-Spring46901.2019.9017716

APA

Vancouver

Author

Bibtex - Download

@inproceedings{4819a4671ea844ed90eac6ba169f652f,
title = "Performance Comparison of Learned vs. Engineered Features for Polarimetric SAR Terrain Classification",
abstract = "In this work, we propose to use learned features for terrain classification of Polarimetric Synthetic Aperture Radar (PolSAR) images. In the proposed classification framework, the learned features are extracted from sliding window regions using Convolutional Neural Networks (CNNs), and then they are used for the classification with the linear Support Vector Machine (SVM) classifier. The classification performance of the proposed approach is compared with numerous target decomposition theorems (TDs) as the engineered features tested with two classifiers: Collective Network of Binary Classifiers (CNBCs) and SVMs. The experimental evaluations over two commonly used benchmark AIRSAR PolSAR images, San Francisco Bay and Flevoland at L-Band, reveal that the classification performance of the learned features with CNNs outperforms the performance of the engineered features as TDs even the dimension of learned features is the quarter of the engineered features.",
author = "Mete Ahishali and Turker Ince and Serkan Kiranyaz and Moncef Gabbouj",
year = "2019",
doi = "10.1109/PIERS-Spring46901.2019.9017716",
language = "English",
booktitle = "2019 41st Photonics & Electromagnetics Research Symposium (PIERS)",

}

RIS (suitable for import to EndNote) - Download

TY - GEN

T1 - Performance Comparison of Learned vs. Engineered Features for Polarimetric SAR Terrain Classification

AU - Ahishali, Mete

AU - Ince, Turker

AU - Kiranyaz, Serkan

AU - Gabbouj, Moncef

PY - 2019

Y1 - 2019

N2 - In this work, we propose to use learned features for terrain classification of Polarimetric Synthetic Aperture Radar (PolSAR) images. In the proposed classification framework, the learned features are extracted from sliding window regions using Convolutional Neural Networks (CNNs), and then they are used for the classification with the linear Support Vector Machine (SVM) classifier. The classification performance of the proposed approach is compared with numerous target decomposition theorems (TDs) as the engineered features tested with two classifiers: Collective Network of Binary Classifiers (CNBCs) and SVMs. The experimental evaluations over two commonly used benchmark AIRSAR PolSAR images, San Francisco Bay and Flevoland at L-Band, reveal that the classification performance of the learned features with CNNs outperforms the performance of the engineered features as TDs even the dimension of learned features is the quarter of the engineered features.

AB - In this work, we propose to use learned features for terrain classification of Polarimetric Synthetic Aperture Radar (PolSAR) images. In the proposed classification framework, the learned features are extracted from sliding window regions using Convolutional Neural Networks (CNNs), and then they are used for the classification with the linear Support Vector Machine (SVM) classifier. The classification performance of the proposed approach is compared with numerous target decomposition theorems (TDs) as the engineered features tested with two classifiers: Collective Network of Binary Classifiers (CNBCs) and SVMs. The experimental evaluations over two commonly used benchmark AIRSAR PolSAR images, San Francisco Bay and Flevoland at L-Band, reveal that the classification performance of the learned features with CNNs outperforms the performance of the engineered features as TDs even the dimension of learned features is the quarter of the engineered features.

U2 - 10.1109/PIERS-Spring46901.2019.9017716

DO - 10.1109/PIERS-Spring46901.2019.9017716

M3 - Conference contribution

BT - 2019 41st Photonics & Electromagnetics Research Symposium (PIERS)

ER -