Do Not Break the Vessels: Structure-Preserving Mean Flow for Vascular Image Translation
For medical imaging applications requiring anatomically faithful vascular reconstruction, this work addresses the critical problem of structural discontinuities in cross-modal translation.
The paper tackles structure preservation in cross-modal vascular image translation, proposing a Structure-Preserving Mean Flow (SPMF) framework that enforces topology invariance via an orthogonality constraint on the flow velocity field. The method achieves state-of-the-art results with peak PSNR of 24.96 dB on NIRII-to-2PF and 24.83 dB on fundus datasets.
Reconstructing anatomically faithful vascular structures from clinically accessible imaging modalities is of substantial clinical significance. However, existing cross-modal translation methods mainly emphasize pixel-level fidelity or visual realism and treat structure preservation as a property of the final output rather than an invariant of the generative process. This limitation often leads to structural discontinuities and artifacts, compromising anatomical coherence and clinical reliability. In this work, we propose a Structure-Preserving Mean Flow (SPMF) framework that formulates vascular image translation as a topology-invariant transport process. Based on a structural invariance principle, we derive an orthogonality constraint on the flow velocity field that formally separates appearance transport from topological distortion. We implement this constraint as a time-weighted surrogate objective within a Brownian bridge diffusion model to preserve topology at every diffusion step. Moreover, we propose a Prototype-Guided Structural Refinement (PGSR) module to align degraded inference-time structures with reliable training-time structures. Experiments on paired NIRII-to-2PF and fundus datasets demonstrate consistent improvements over state-of-the-art methods, achieving peak PSNR values of 24.96 dB and 24.83 dB, respectively.