This repo contains implementations of methods for style transfer using natural language descriptions, which allow to specify the transformation up to the object in the image. The methods are based on GroupViT [1] and CLIPstyler [2].
- Xu, Jiarui, et al. "GroupViT: Semantic Segmentation Emerges from Text Supervision." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022.
- Kwon, Gihyun, and Jong Chul Ye. "Clipstyler: Image style transfer with a single text condition." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022.