FreePath: Neural Mean Free Path for Image Dehazing

Researchers posted arXiv:2611.07398v1, introducing FreePath, a neural mean free path framework for single image dehazing that models dehazing as feature space path state restoration rather than estimating atmospheric scattering parameters. FreePath uses a Mean Free Path Residual Block inside a residual encoder and decoder to learn path and collision fields, plus a FreePath State Conservation Integration module that preserves restored path states during deep feature aggregation. Experiments on paired CDD-11 haze images and real RTTS scenes show FreePath consistently improves full-reference fidelity, no-reference perceptual quality and visual faithfulness over representative dehazing baselines.

arXiv:2611.07398v1 Announce Type: new Abstract: Single image dehazing requires recovering scene radiance from visual responses that are weakened and veiled by participating media. Existing methods either invert explicit scattering variables, which can be unstable under complex scenes, or learn direct restoration mappings, which often discard the physical organization of haze formation. We present \emph{FreePath}, a neural mean free path framework for image dehazing. Instead of estimating atmospheric parameters, FreePath represents dehazing as feature space path state restoration, inspired by the expected distance a visual response can propagate before interacting with haze. Within a residual encoder and decoder, the Mean Free Path Residual Block learns path and collision fields to guide haze constrained correction, while FreePath State Conservation Integration preserves restored path states during deep feature aggregation. Experiments on paired CDD-11 haze images and real RTTS scenes show that FreePath consistently improves full reference fidelity, no reference perceptual quality, and visual faithfulness over representative dehazing baselines.