Discovering and Explaining the Non-Causality of Deep Learning in SAR ATR
This work addresses the issue of non-causal learning in SAR ATR for researchers and practitioners, providing a quantitative analysis that is incremental over existing qualitative methods.
The paper tackles the problem of deep learning models in SAR ATR overfitting to spurious background correlations in the MSTAR dataset, quantifying this non-causality using Shapley values to measure clutter contributions and explaining how data and model biases lead to overfitting.
In recent years, deep learning has been widely used in SAR ATR and achieved excellent performance on the MSTAR dataset. However, due to constrained imaging conditions, MSTAR has data biases such as background correlation, i.e., background clutter properties have a spurious correlation with target classes. Deep learning can overfit clutter to reduce training errors. Therefore, the degree of overfitting for clutter reflects the non-causality of deep learning in SAR ATR. Existing methods only qualitatively analyze this phenomenon. In this paper, we quantify the contributions of different regions to target recognition based on the Shapley value. The Shapley value of clutter measures the degree of overfitting. Moreover, we explain how data bias and model bias contribute to non-causality. Concisely, data bias leads to comparable signal-to-clutter ratios and clutter textures in training and test sets. And various model structures have different degrees of overfitting for these biases. The experimental results of various models under standard operating conditions on the MSTAR dataset support our conclusions. Our code is available at https://github.com/waterdisappear/Data-Bias-in-MSTAR.