Physics-guided in-line holographic reconstruction via complex-valued multi-scale wavelet decomposition and cross-frequency gated fusion

In-line holography enables compact and stable quantitative phase imaging, yet its reconstruction is constrained by twin-image interference and noise sensitivity. In this work, we propose a physics-constrained reconstruction framework that achieves spectrum-adaptive modulation of the complex amplitude via multi-scale wavelet decomposition with time-frequency priors and a cross-frequency gated fusion mechanism, while a differentiable Huber-TV model provides stable, physically interpretable variational constraints on the reconstructed phase. Simulations and experiments demonstrate that the proposed method improves phase fidelity and structural consistency, providing a physically consistent and practical solution for in-line holographic reconstruction under complex conditions.