DAWN-FM: Data-Aware and Noise-Informed Flow Matching for Solving Inverse Problems

Researchers introduce DAWN-FM, a new AI method using Flow Matching to solve inverse problems in fields like medical imaging and signal processing. The approach incorporates data and noise embedding to provide robust solutions even with incomplete or noisy observations, outperforming pretrained diffusion models in highly ill-posed scenarios.

• →DAWN-FM uses Flow Matching framework to solve inverse problems by mapping Gaussian distributions to target distributions through deterministic processes.
• →The method incorporates explicit data and noise embedding, making it particularly robust for scenarios with noisy or incomplete observations.
• →Unlike pretrained diffusion models, DAWN-FM is trained specifically for each inverse problem and adapts to varying noise levels.
• →The approach enables uncertainty quantification by generating multiple plausible outcomes for given problems.
• →Validation through numerical experiments shows effectiveness in image deblurring and tomography applications.