5 papers across 3 sessions
In this work, we propose CLIPGaussians, a novel approach to style transfer as a plug-in model to the Gaussian splatting
We present a dataset of 210,000 triplets (content, style, stylized images) and an end - to - end stylization framework tailored for it, ensuring efficient style transfer.
Given unposed sparse-view images and an arbitrary style image, our method predicts stylized 3D Gaussians in less than a second using a feed-forward network.
