InterLight: Leveraging Intrinsic Illumination Priors for Low-Light Image Enhancement

Researchers introduced InterLight, a new low-light image enhancement framework that uses intrinsic illumination priors to restore dark images more faithfully.

The method combines physics-guided augmentation, illumination-conditioned prompts, a luminance-gated memory mechanism, and self-supervised consistency.

It aims to reduce common failures in low-light enhancement such as over-enhancement, color shifts, and noise.

According to the paper, InterLight improves texture recovery and distortion control across multiple benchmarks.