TY - GEN
T1 - Computational interpretability of multilayer preceptron used for SAR image target recognition
AU - Zheng, Tong
AU - Lei, Peng
AU - Wang, Jun
AU - Liu, Chunsheng
AU - Wang, Jianing
N1 - Publisher Copyright:
© 2021 IEEE.
PY - 2021
Y1 - 2021
N2 - The achievement of deep neural networks (DNNs) in the computer vision has aroused great concerns in the synthetic aperture radar (SAR) automatic target recognition (ATR) field. As a simple but effective model, the multilayer perceptron (MLP) is widely used in SAR image target recognition. However, the black-box problem could limit the development of DNNs in SAR ATR. In this paper, we explore the interpretability of MLP from the perspective of computation process of its forward propagation. By using the matrix representation, the function is studied that the angles between parameters and features as well as features magnitudes. Besides, the feature statistics is adopted to discuss the effect of nonlinear activation functions. Finally, some experiments on the MSTAR datasets are carried out and analyzed to demonstrate the effectiveness of the proposed method.
AB - The achievement of deep neural networks (DNNs) in the computer vision has aroused great concerns in the synthetic aperture radar (SAR) automatic target recognition (ATR) field. As a simple but effective model, the multilayer perceptron (MLP) is widely used in SAR image target recognition. However, the black-box problem could limit the development of DNNs in SAR ATR. In this paper, we explore the interpretability of MLP from the perspective of computation process of its forward propagation. By using the matrix representation, the function is studied that the angles between parameters and features as well as features magnitudes. Besides, the feature statistics is adopted to discuss the effect of nonlinear activation functions. Finally, some experiments on the MSTAR datasets are carried out and analyzed to demonstrate the effectiveness of the proposed method.
KW - SAR image target recognition
KW - convolutional neural networks
KW - interpretability
KW - multilayer perceptron
UR - https://www.scopus.com/pages/publications/85181125877
U2 - 10.1109/Radar53847.2021.10028293
DO - 10.1109/Radar53847.2021.10028293
M3 - 会议稿件
AN - SCOPUS:85181125877
T3 - Proceedings of the IEEE Radar Conference
SP - 1371
EP - 1374
BT - 2021 CIE International Conference on Radar, Radar 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2021 CIE International Conference on Radar, Radar 2021
Y2 - 15 December 2021 through 19 December 2021
ER -