Weak Segmentation-Guided GAN for~Realistic Color Edition
To overcome these challenges, we propose a novel GAN architecture leveraging any differentiable segmentation model.
ProofPaper ↗
Key points
- Editing the color of images in a realistic way finds many applications such as changing the perception of an image, data augmentation or film post processing.
- Conventional methods using segmentation and histogram matching maximize the controllability but also introduce a lack of realism or are complex to automate.
- On the contrary, GANs that specialize in realism are difficult to control.
- It generates images that look realistic while offering a precise color control.
Sources (1)
- [1]Weak Segmentation-Guided GAN for~Realistic Color EditionarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 12, 11:36 AM
To overcome these challenges, we propose a novel GAN architecture leveraging any differentiable segmentation model.
Editing the color of images in a realistic way finds many applications such as changing the perception of an image, data augmentation or film post processing.
Extractive summary: sentences quoted from the sources.