SCM: A Searched Convolutional Metaformer for SAR Ship Classification

Ship classification technology using synthetic aperture radar (SAR) has become a research hotspot. Many deep-learning-based methods have been proposed with handcrafted models or transplanted computer vision networks. However, most of these are designed for graphics processing unit (GPU) platforms, leading to limited scope application. This paper proposes novel mini-size searched convolutional Metaformer (SCM) classifying SAR ships. Firstly, network architecture searching (NAS) algorithm progressive data augmentation is find an efficient baseline network. Then, transformer classifier employed improve the spatial awareness capability. Moreover, ConvFormer cell by filling normal into block. further improves feature-extracting Experimental results obtained show that SCM provides best accuracy only 0.46×106 weights, achieving good trade-off between performance and model size.