A Frequency-Coupled Latent Space Facilitates Synthetic-to-Measured SAR Image Translation

Deep learning (DL)-based synthetic aperture radar (SAR) target recognition depends on measured data that are costly to collect, and synthetic images produced by electromag netic simulation offer an economical substitute whose direct use is limited by the synthetic-to-measured (S2M) domain gap. Owing to the coherent imaging mechanism of SAR, this gap manifests not only in spatial appearance but also in the dis tribution of image energy across spatial-frequency bands and orientations, a structure that latent representations learned solely for reconstruction encode only implicitly and therefore cannot align selectively. To reduce this deep-feature gap, we propose LFCS2M, a latent diffusion framework for S2M SAR image translation in a frequency-coupled latent space. A frequency domain feature refinement module (FDFRM) imposes a learnable two-dimensional spectral mask on the latent representation, allowing bandwise and orientationwise feature adjustment with out assuming a predefined spectral discrepancy. A bidirectional diffusion bridge guided by Measured Information Guided Cross Attention (MIGCA) further maps synthetic latent features to ward the measured distribution. Qualitative and quantitative experiments show that LFCS2M generates high-fidelity SAR images that closely resemble measured data, supporting DL based SAR target recognition by narrowing the domain gap and reducing data acquisition cost. The code of LFCS2M is accessible at https://github.com/Jordan-Liao/LFCS2M.