|
|
| CARVIEW |
Model Architecture
We propose a Semantic-Guided Zero-Shot (SGZ) network for low-light image enhancement. It consists of:
- An Enhancement Factor Extraction (EFE) network for estimating pixel-wise light deficiency.
- A Recurrent Image Enhancement (RIE) network to progressively enhance the low-light image.
- An Unsupervised Semantic Segmentation (USS) network for preserving the semantic information.
During training, both RIE and USS have frozen parameters and they output the loss to update EFE. During testing, EFE and RIE are used sequentially to enhance a low-light image.
Enhancement Factor Extraction (EFE)
EFE provides an efficient estimate of light deficiency. Darker regions below indicate lower enhancement factor values, which will be enhanced more.
Visual Results: Aerial Video
Visual Results: Aerial Image Enhancement
Visual Results: Natural Image Enhancement
Visual Results: Facial Image Enhancement
Visual Results: Low-Light Object Detection
Visual Results: Low-Light Semantic Segmentation
