Abstract
Photorealistic style transfer aims to match the color and tone of an input image to that of a style target while preserving content and fine details. Existing large image-editing models can perform stylization, but their computational cost, hallucinations, and limited precision control make them impractical for high-resolution, interactive workflows. Hist2Style addresses this by distilling a large editor into a lightweight model constrained to locally affine transformations in bilateral space. The model conditions on a histogram-based style embedding, enabling an interpretable and editable interface for color and tone control. Hist2Style delivers real-time, high-resolution photorealistic stylization, avoids structural drift by construction, and supports interactive user-guided adjustments.
Key Contributions
- A photorealistic, histogram-guided stylization network that is robust to hallucinations and scales to high resolutions.
- An interactive real-time editing framework with direct histogram manipulation for intuitive control of color and tone.
- A Stylization Quality Assessment (SQA) metric that correlates strongly with human preference.
BibTeX
@InProceedings{Galor_2026_CVPR,
author = {Galor, Dekel and Pikielny, Adam and Zhang, Zhoutong and Wang, Ke and Waller, Laura and Chen, Jiawen and Chugunov, Ilya},
title = {Hist2Style: Histogram-Guided Stylization with Bilateral Grids},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2026},
pages = {29717-29726}
}