About this document
Enhancing Low-Light Images with Normalizing Flow by fanwu0409 is a document available to read on EtoBox.
The paper presents LLFlow, a low-light image enhancement method that utilizes a normalizing flow model to address the challenges of mapping low-light images to normally exposed ones. By capturing complex conditional distributions, LLFlow improves image quality by reducing noise and artifacts while enhancing color richness. Experimental results demonstrate its effectiveness over traditional pixel-wise loss methods, achieving better visual fidelity and consistency with human perception.
- Author
- fanwu0409
- Language
- EN