Credible Recognition of Radar Images: Interpretability Metric and Classification Score

Amir Hosein Oveis; Elisa Giusti; Selenia Ghio; Giulio Meucci; Marco Martorella

doi:10.1109/IGARSS52108.2023.10281648

Credible Recognition of Radar Images: Interpretability Metric and Classification Score

Amir Hosein Oveis^*, Elisa Giusti, Selenia Ghio, Giulio Meucci, Marco Martorella

^*Corresponding author for this work

Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Automatic target recognition (ATR) is one of the most demanding applications of synthetic aperture radar (SAR) in the field of radar reconnaissance and surveillance. Convolutional neural networks (CNNs) have been extensively employed for SAR-ATR and obtained remarkable accuracy. However, regarding the black-box nature and non-transparency in decision-making, CNN's reliability is unsatisfactory. Recently, some progress has been made toward providing a visual explanation of CNN's classification procedure. In this paper, we employ the Local Interpretable Model-agnostic Explanation (LIME) algorithm to propose an interpretability metric that can be helpful to evaluate the overall robustness of CNNs. Using the proposed framework, the user can infer what proportion of the results are based on target features, while the remainder is based on irrelevant correlations from the background clutter. The theoretical findings are validated by the public MSTAR database.

Original language	English
Title of host publication	IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	1084-1087
Number of pages	4
ISBN (Electronic)	9798350320107, 9798350320091 (USB)
ISBN (Print)	9798350331745 (PoD)
DOIs	https://doi.org/10.1109/IGARSS52108.2023.10281648
Publication status	Published - 20 Oct 2023
Event	2023 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023 - Pasadena, United States Duration: 16 Jul 2023 → 21 Jul 2023

Publication series

Name	IEEE International Geoscience and Remote Sensing Symposium proceedings
Publisher	IEEE
ISSN (Print)	2153-6996
ISSN (Electronic)	2153-7003

Conference

Conference	2023 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023
Country/Territory	United States
City	Pasadena
Period	16/07/23 → 21/07/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Keywords

Automatic Target Recognition (ATR)
Convolutional Neural Network (CNN)
eXplainable Artificial Intelligence (XAI)
Local Interpretable Model-agnostic Explanation (LIME)
Synthetic Aperture Radar (SAR)

ASJC Scopus subject areas

Computer Science Applications
General Earth and Planetary Sciences

Access to Document

10.1109/IGARSS52108.2023.10281648

Cite this

Oveis, A. H., Giusti, E., Ghio, S., Meucci, G., & Martorella, M. (2023). Credible Recognition of Radar Images: Interpretability Metric and Classification Score. In IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium (pp. 1084-1087). (IEEE International Geoscience and Remote Sensing Symposium proceedings). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/IGARSS52108.2023.10281648

Oveis, Amir Hosein ; Giusti, Elisa ; Ghio, Selenia et al. / Credible Recognition of Radar Images : Interpretability Metric and Classification Score. IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium. Institute of Electrical and Electronics Engineers (IEEE), 2023. pp. 1084-1087 (IEEE International Geoscience and Remote Sensing Symposium proceedings).

@inproceedings{1fd12938d1d14a52b757e43846c9b5d0,

title = "Credible Recognition of Radar Images: Interpretability Metric and Classification Score",

abstract = "Automatic target recognition (ATR) is one of the most demanding applications of synthetic aperture radar (SAR) in the field of radar reconnaissance and surveillance. Convolutional neural networks (CNNs) have been extensively employed for SAR-ATR and obtained remarkable accuracy. However, regarding the black-box nature and non-transparency in decision-making, CNN's reliability is unsatisfactory. Recently, some progress has been made toward providing a visual explanation of CNN's classification procedure. In this paper, we employ the Local Interpretable Model-agnostic Explanation (LIME) algorithm to propose an interpretability metric that can be helpful to evaluate the overall robustness of CNNs. Using the proposed framework, the user can infer what proportion of the results are based on target features, while the remainder is based on irrelevant correlations from the background clutter. The theoretical findings are validated by the public MSTAR database.",

keywords = "Automatic Target Recognition (ATR), Convolutional Neural Network (CNN), eXplainable Artificial Intelligence (XAI), Local Interpretable Model-agnostic Explanation (LIME), Synthetic Aperture Radar (SAR)",

author = "Oveis, {Amir Hosein} and Elisa Giusti and Selenia Ghio and Giulio Meucci and Marco Martorella",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023 ; Conference date: 16-07-2023 Through 21-07-2023",

year = "2023",

month = oct,

day = "20",

doi = "10.1109/IGARSS52108.2023.10281648",

language = "English",

isbn = "9798350331745 (PoD)",

series = "IEEE International Geoscience and Remote Sensing Symposium proceedings",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

pages = "1084--1087",

booktitle = "IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium",

}

Oveis, AH, Giusti, E, Ghio, S, Meucci, G & Martorella, M 2023, Credible Recognition of Radar Images: Interpretability Metric and Classification Score. in IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium. IEEE International Geoscience and Remote Sensing Symposium proceedings, Institute of Electrical and Electronics Engineers (IEEE), pp. 1084-1087, 2023 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023, Pasadena, United States, 16/07/23. https://doi.org/10.1109/IGARSS52108.2023.10281648

Credible Recognition of Radar Images: Interpretability Metric and Classification Score. / Oveis, Amir Hosein; Giusti, Elisa; Ghio, Selenia et al.
IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium. Institute of Electrical and Electronics Engineers (IEEE), 2023. p. 1084-1087 (IEEE International Geoscience and Remote Sensing Symposium proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Credible Recognition of Radar Images

T2 - 2023 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2023

AU - Oveis, Amir Hosein

AU - Giusti, Elisa

AU - Ghio, Selenia

AU - Meucci, Giulio

AU - Martorella, Marco

PY - 2023/10/20

Y1 - 2023/10/20

N2 - Automatic target recognition (ATR) is one of the most demanding applications of synthetic aperture radar (SAR) in the field of radar reconnaissance and surveillance. Convolutional neural networks (CNNs) have been extensively employed for SAR-ATR and obtained remarkable accuracy. However, regarding the black-box nature and non-transparency in decision-making, CNN's reliability is unsatisfactory. Recently, some progress has been made toward providing a visual explanation of CNN's classification procedure. In this paper, we employ the Local Interpretable Model-agnostic Explanation (LIME) algorithm to propose an interpretability metric that can be helpful to evaluate the overall robustness of CNNs. Using the proposed framework, the user can infer what proportion of the results are based on target features, while the remainder is based on irrelevant correlations from the background clutter. The theoretical findings are validated by the public MSTAR database.

AB - Automatic target recognition (ATR) is one of the most demanding applications of synthetic aperture radar (SAR) in the field of radar reconnaissance and surveillance. Convolutional neural networks (CNNs) have been extensively employed for SAR-ATR and obtained remarkable accuracy. However, regarding the black-box nature and non-transparency in decision-making, CNN's reliability is unsatisfactory. Recently, some progress has been made toward providing a visual explanation of CNN's classification procedure. In this paper, we employ the Local Interpretable Model-agnostic Explanation (LIME) algorithm to propose an interpretability metric that can be helpful to evaluate the overall robustness of CNNs. Using the proposed framework, the user can infer what proportion of the results are based on target features, while the remainder is based on irrelevant correlations from the background clutter. The theoretical findings are validated by the public MSTAR database.

KW - Automatic Target Recognition (ATR)

KW - Convolutional Neural Network (CNN)

KW - eXplainable Artificial Intelligence (XAI)

KW - Local Interpretable Model-agnostic Explanation (LIME)

KW - Synthetic Aperture Radar (SAR)

UR - http://www.scopus.com/inward/record.url?scp=85178334378&partnerID=8YFLogxK

U2 - 10.1109/IGARSS52108.2023.10281648

DO - 10.1109/IGARSS52108.2023.10281648

M3 - Conference contribution

AN - SCOPUS:85178334378

SN - 9798350331745 (PoD)

T3 - IEEE International Geoscience and Remote Sensing Symposium proceedings

SP - 1084

EP - 1087

BT - IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium

PB - Institute of Electrical and Electronics Engineers (IEEE)

Y2 - 16 July 2023 through 21 July 2023

ER -

Oveis AH, Giusti E, Ghio S, Meucci G, Martorella M. Credible Recognition of Radar Images: Interpretability Metric and Classification Score. In IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium. Institute of Electrical and Electronics Engineers (IEEE). 2023. p. 1084-1087. (IEEE International Geoscience and Remote Sensing Symposium proceedings). doi: 10.1109/IGARSS52108.2023.10281648

Credible Recognition of Radar Images: Interpretability Metric and Classification Score

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this