LW-CMDANet: A Novel Attention Network for SAR Automatic Target Recognition
نویسندگان
چکیده
Deep learning based synthetic aperture radar automatic target recognition (SAR-ATR) plays an significant role in the military and civilian fields. However, data limitation large computational cost are still severe challenges actual application of SAR-ATR. To improve performance CNN model with limited samples SAR-ATR, this paper proposes a novel multi-domain feature subspaces fusion representation method, i.e., lightweight cascaded attention network, namely LW-CMDANet. First, we design four-layer to perform hierarchical features via hinge loss function, which can efficiently alleviate overfitting problem by non-greedy training style small dataset. Then, module, on discrete cosine transform wavelet transform, is embedded into previous further complete class-specific extraction from both frequency domains input maps. Thus, enhance ability manner, effectively accuracy model. Experimental results SAR datasets show that our proposed method achieve better or competitive than many current existing state-of-the-art methods terms cost.
منابع مشابه
Pose Estimation for SAR Automatic Target Recognition
This paper explores statistically pose estimation in SAR ATR. Based on our proposed method of maximizing mutual information, further experiments are conducted by using the new MSTAR/ IU Database. Different pose estimator topologies and training criteria are also employed. Experimental results show that our proposed method reduces the average pose estimation error to within 10 degrees of the tru...
متن کاملKernel generalized neighbor discriminant embedding for SAR automatic target recognition
In this paper, we propose a new supervised feature extraction algorithm in synthetic aperture radar automatic target recognition (SAR ATR), called generalized neighbor discriminant embedding (GNDE). Based on manifold learning, GNDE integrates class and neighborhood information to enhance discriminative power of extracted feature. Besides, the kernelized counterpart of this algorithm is also pro...
متن کاملA Multi-scale Local Phase Quantization plus Biomimetic Pattern Recognition Method for Sar Automatic Target Recognition
Synthetic aperture radar (SAR) automatic target recognition (ATR) has been receiving more and more attention in the past two decades. But the problem of how to overcome SAR target ambiguities and azimuth angle variations has still left unsolved. In this paper, a multi-scale local phase quantization plus biomimetic pattern recognition (BPR) method is presented to solve these two difficulties. By...
متن کاملAutomatic SAR Target Recognition Using Random Subspace Ensemble Classifiers
Automatic Target Recognition (ATR) has become a widely researched problem due to its applicability in domains such as surveillance systems. It can be implemented on various types of images captured by different sensors. In this paper a novel framework for Automatic Target Recognition (ATR) in Synthetic Aperture Radar (SAR) imagery using Ensemble Classifier is presented. A combination of Princip...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
سال: 2022
ISSN: ['2151-1535', '1939-1404']
DOI: https://doi.org/10.1109/jstars.2022.3195074