Dual-Polarization SAR Ship Target Recognition Based on Mini Hourglass Region Extraction and Dual-Channel Efficient Fusion Network

A novel Dual-polarization SAR ship target recognition method based on feature and loss fusion deep network is proposed in this paper, to improve the generalization ability, accuracy of imbalance samples, real-time performance general learning network. The combined composed two parts. first part region extraction lightweight mini Hourglass network, eliminate impact data background noise identification accuracy. second a two-channel feature/loss function Efficient B2 backbone aiming solve problem dual-polarization image iterative convergence acceleration. tested slices from OpenSAR sets. experimental results indicates that, achieves rate 110FPS with 87.72%, exceeds SOTA by 3.72% speed improved 75.68%. can be applied dual polarization, samples low-resolution condition for reference.