Limited-Data SAR ATR Methods Based on Model Design

This chapter introduces limited-data SAR ATR methods based on model design. This kind of methods aims to leverage model architectures to extract discriminative features from limited SAR images, enabling accurate recognition under limited-data conditions. The limited-data SAR ATR method based on multi-scale feature weighting, employs specific model designs to extract features at different levels of abstraction. It identifies distinct discriminative features at each scale and dynamically integrates them using adaptive weighting to form effective final features for precise recognition. The limited-data SAR ATR methods based on model design explore discriminative target information from both intra-class and inter-class perspectives, optimizing feature extraction and representation to overcome the challenges posed by limited data scenarios in SAR ATR applications.