StyleDistillation: A New Insight of Image Style Enables Personalized Aesthetic Manipulation

Yuxin Wang, Xiaoyu Geng, Yuke Li, Zheng Wang
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:126332-126357, 2026.

Abstract

Text-guided stylized image generation has yielded promising advances by leveraging the powerful capabilities of text-to-image diffusion models. However, the inherent coupling of style and content information within the reference image presents a significant challenge. To address this, we propose StyleDistillation, a novel approach grounded in two key observations about the CLIP embedding space from a style perspective. By leveraging a lightweight StyleDistiller module, combined with carefully designed optimization objectives based on geometric and semantic priors, we can extract fine-grained style representation from the reference image. Additionally, we introduce a Prompt Alignment Enhancement mechanism during inference, which significantly improves the control that text prompts exert over the generated images. Extensive experiments demonstrate that our method achieves outstanding performance in both style reproduction and prompt alignment. Furthermore, StyleDistillation supports various personalized operations, including style editing and style fusion, highlighting its substantial potential for diverse applications.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-wang26bi, title = {{S}tyle{D}istillation: A New Insight of Image Style Enables Personalized Aesthetic Manipulation}, author = {Wang, Yuxin and Geng, Xiaoyu and Li, Yuke and Wang, Zheng}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {126332--126357}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/wang26bi/wang26bi.pdf}, url = {https://proceedings.mlr.press/v306/wang26bi.html}, abstract = {Text-guided stylized image generation has yielded promising advances by leveraging the powerful capabilities of text-to-image diffusion models. However, the inherent coupling of style and content information within the reference image presents a significant challenge. To address this, we propose StyleDistillation, a novel approach grounded in two key observations about the CLIP embedding space from a style perspective. By leveraging a lightweight StyleDistiller module, combined with carefully designed optimization objectives based on geometric and semantic priors, we can extract fine-grained style representation from the reference image. Additionally, we introduce a Prompt Alignment Enhancement mechanism during inference, which significantly improves the control that text prompts exert over the generated images. Extensive experiments demonstrate that our method achieves outstanding performance in both style reproduction and prompt alignment. Furthermore, StyleDistillation supports various personalized operations, including style editing and style fusion, highlighting its substantial potential for diverse applications.} }
Endnote
%0 Conference Paper %T StyleDistillation: A New Insight of Image Style Enables Personalized Aesthetic Manipulation %A Yuxin Wang %A Xiaoyu Geng %A Yuke Li %A Zheng Wang %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-wang26bi %I PMLR %P 126332--126357 %U https://proceedings.mlr.press/v306/wang26bi.html %V 306 %X Text-guided stylized image generation has yielded promising advances by leveraging the powerful capabilities of text-to-image diffusion models. However, the inherent coupling of style and content information within the reference image presents a significant challenge. To address this, we propose StyleDistillation, a novel approach grounded in two key observations about the CLIP embedding space from a style perspective. By leveraging a lightweight StyleDistiller module, combined with carefully designed optimization objectives based on geometric and semantic priors, we can extract fine-grained style representation from the reference image. Additionally, we introduce a Prompt Alignment Enhancement mechanism during inference, which significantly improves the control that text prompts exert over the generated images. Extensive experiments demonstrate that our method achieves outstanding performance in both style reproduction and prompt alignment. Furthermore, StyleDistillation supports various personalized operations, including style editing and style fusion, highlighting its substantial potential for diverse applications.
APA
Wang, Y., Geng, X., Li, Y. & Wang, Z.. (2026). StyleDistillation: A New Insight of Image Style Enables Personalized Aesthetic Manipulation. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:126332-126357 Available from https://proceedings.mlr.press/v306/wang26bi.html.

Related Material