DMSA-Net: Dual-Domain Multiscale Attentive Network for Robust SAR Ship Detection

Deep learning has achieved remarkable progress in synthetic aperture radar (SAR) ship detection. However, critical challenges remain in mitigating noise interference and addressing multiscale object variations. To solve this, this paper proposes a dual-domain multiscale attentive network for SAR ship detection (DMSA-Net). Our contributions are threefold. Firstly, we design a spatial-frequency dual-stage feature extraction network (SF-DualNet), which innovatively constructs a dual-stage feature optimization architecture, with noise suppression achieved by a frequency-domain decoupling-reconstruction module in the first stage, and multiscale perception enhancement using a global-local enhancement module in the second stage. Secondly, we develop a wavelet decomposition-based multiscale dilated fusion pyramid (WMSD-FPN), which embeds discrete wavelet transform sampling into bidirectional cross-scale interaction paths and introduces multiscale dilated fusion modules at critical hierarchical levels to enhance multiscale feature discrimination. Thirdly, a shared-kernel-based dynamically weighted group normalization detection head is employed to significantly reduce computational costs and parameters while maintaining competitive accuracy. Extensive experiments on the HRSID and SSDD datasets demonstrate the effectiveness of DMSA-Net, achieving AP50 values of 92.4% and 97.2% respectively, outperforming existing methods in complex scenarios.