Gamma Distribution PCA-Enhanced Feature Learning for Angle-Robust SAR Target Recognition

Motivation: • Fully consider the unique statistical characteristics of SAR data and capture the angle-invariant feature to alleviate deep model's sensitivity to angle variations. • Simple, effective and readily compatible, so as to ensure its greater applicability. Problem: The imaging angle of synthetic aperture radar (SAR) significantly impact the scattering characteristics of targets, thereby angle-inadequate training samples leads to poor robustness of deep networks.