Flow Map Denoisers: Traversing the Distortion-Perception Plane for Inverse Problems
Researchers have introduced a novel method called Flow Map Denoisers, which addresses the fundamental tradeoff in image restoration between minimizing error and maximizing perceptual quality. This new approach utilizes flow map models, an extension of flow matching, to implicitly define a one-parameter family of denoisers. By adjusting a lookahead parameter, users can continuously span the distortion-perception frontier, offering a flexible way to balance reconstruction fidelity with sharpness. The method has been validated through extensive experiments on datasets like CelebA and AFHQ for various inverse tasks. AI
IMPACT Offers a flexible approach to image restoration by allowing continuous control over the distortion-perception tradeoff, potentially improving results for various inverse problems.