{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,13]],"date-time":"2026-09-13T10:14:58Z","timestamp":1789294498676,"version":"build-2803163510"},"reference-count":33,"publisher":"Association for Computing Machinery (ACM)","issue":"6","content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Graph."],"published-print":{"date-parts":[[2025,12]]},"abstract":"<jats:p>In text-to-image models, consistent character generation is the task of achieving text alignment while maintaining the subject's appearance across different prompts. However, since style and appearance are often entangled, the existing methods struggle to preserve consistent subject characteristics while adhering to varying style prompts. Current approaches for consistent text-to-image generation typically rely on large-scale fine-tuning on curated image sets or per-subject optimization, which either fail to generalize across prompts or do not align well with textual descriptions. Meanwhile, training-free methods often fail to maintain subject consistency across different styles. In this work, we introduce a training-free method that, for the first time, jointly achieves style preservation and subject consistency across varied styles. The attention matrices are manipulated such that Queries and Keys are obtained from the anchor image(s) that are used to define the subject, while the Values are imported from a parallel copy that is not subject-anchored. Additionally, cross-image components are added to the self-attention mechanism by expanding the Key and Value matrices. To do without shifting from the target style, we align the statistics of the Value matrices. As is demonstrated in a comprehensive battery of qualitative and quantitative experiments, our method effectively decouples style from subject appearance and enables faithful generation of text-aligned images with consistent characters across diverse styles.<\/jats:p>\n                  <jats:p>Code will be available at our project page: jbruner23.github.io\/consistyle.<\/jats:p>","DOI":"10.1145\/3763303","type":"journal-article","created":{"date-parts":[[2025,12,4]],"date-time":"2025-12-04T17:15:39Z","timestamp":1764868539000},"page":"1-16","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["ConsiStyle: Style Diversity in Training-Free Consistent T2I Generation"],"prefix":"10.1145","volume":"44","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-0574-6546","authenticated-orcid":false,"given":"Yohai","family":"Mazuz","sequence":"first","affiliation":[{"name":"Tel Aviv University, Ofaqim, Israel"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-8582-6182","authenticated-orcid":false,"given":"Janna","family":"Bruner","sequence":"additional","affiliation":[{"name":"Tel Aviv University, Herzliya, Israel"},{"name":"Amazon, Tel Aviv, Israel"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5578-8892","authenticated-orcid":false,"given":"Lior","family":"Wolf","sequence":"additional","affiliation":[{"name":"Tel Aviv University, Herzliya, Israel"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2025,12,4]]},"reference":[{"key":"e_1_2_2_1_1","unstructured":"Karlheinz Agsteiner. 2023. The 77 Styles of Stable Diffusion's SDXL. https:\/\/medium.com\/@karlheinz.agsteiner\/the-77-styles-of-stable-diffusions-sdxl-28303582fcf7. Accessed: 2025-08-24."},{"key":"e_1_2_2_2_1","doi-asserted-by":"crossref","unstructured":"Yuval Alaluf Daniel Garibi Or Patashnik Hadar Averbuch-Elor and Daniel Cohen-Or. 2023. Cross-Image Attention for Zero-Shot Appearance Transfer. arXiv:2311.03335 [cs.CV]","DOI":"10.1145\/3641519.3657423"},{"key":"e_1_2_2_3_1","doi-asserted-by":"publisher","DOI":"10.1145\/3641519.3657430"},{"key":"e_1_2_2_4_1","doi-asserted-by":"publisher","DOI":"10.1109\/FG61629.2025.11099464"},{"key":"e_1_2_2_5_1","volume-title":"Muse: Text-To-Image Generation via Masked Generative Transformers. ArXiv abs\/2301.00704","author":"Chang Huiwen","year":"2023","unstructured":"Huiwen Chang, Han Zhang, Jarred Barber, AJ Maschinot, Jos\u00e9 Lezama, Lu Jiang, Ming Yang, Kevin P. Murphy, William T. Freeman, Michael Rubinstein, Yuanzhen Li, and Dilip Krishnan. 2023. Muse: Text-To-Image Generation via Masked Generative Transformers. ArXiv abs\/2301.00704 (2023). https:\/\/api.semanticscholar.org\/CorpusID:255372955"},{"key":"e_1_2_2_6_1","volume-title":"British Machine Vision Conference. https:\/\/api.semanticscholar.org\/CorpusID:232428225","author":"Cong Wenyan","year":"2021","unstructured":"Wenyan Cong, Junyan Cao, Li Niu, Jianfu Zhang, Xuesong Gao, Zhiwei Tang, and Liqing Zhang. 2021. Deep Image Harmonization by Bridging the Reality Gap. In British Machine Vision Conference. https:\/\/api.semanticscholar.org\/CorpusID:232428225"},{"key":"e_1_2_2_7_1","first-page":"50742","article-title":"DreamSim: Learning New Dimensions of Human Visual Similarity using Synthetic Data","volume":"36","author":"Fu Stephanie","year":"2023","unstructured":"Stephanie Fu, Netanel Tamir, Shobhita Sundaram, Lucy Chai, Richard Zhang, Tali Dekel, and Phillip Isola. 2023. DreamSim: Learning New Dimensions of Human Visual Similarity using Synthetic Data. In Advances in Neural Information Processing Systems, Vol. 36. 50742\u201350768.","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_2_8_1","volume-title":"An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion. ArXiv abs\/2208.01618","author":"Gal Rinon","year":"2022","unstructured":"Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patashnik, Amit H. Bermano, Gal Chechik, and Daniel Cohen-Or. 2022. An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion. ArXiv abs\/2208.01618 (2022). https:\/\/api.semanticscholar.org\/CorpusID:251253049"},{"key":"e_1_2_2_9_1","volume-title":"StyleShot: A Snapshot on Any Style. ArXiv abs\/2407.01414","author":"Gao Junyao","year":"2024","unstructured":"Junyao Gao, Yanchen Liu, Yanan Sun, Yinhao Tang, Yanhong Zeng, Kai Chen, and Cairong Zhao. 2024. StyleShot: A Snapshot on Any Style. ArXiv abs\/2407.01414 (2024). https:\/\/api.semanticscholar.org\/CorpusID:270870813"},{"key":"e_1_2_2_10_1","volume-title":"A neural algorithm of artistic style. arXiv preprint arXiv:1508.06576","author":"Gatys Leon A.","year":"2015","unstructured":"Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge. 2015. A neural algorithm of artistic style. arXiv preprint arXiv:1508.06576 (2015)."},{"key":"e_1_2_2_11_1","volume-title":"Image Style Transfer Using Convolutional Neural Networks. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2016","author":"Gatys Leon A.","year":"2016","unstructured":"Leon A. Gatys, Alexander S. Ecker, and Matthias Bethge. 2016. Image Style Transfer Using Convolutional Neural Networks. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2016), 2414\u20132423. https:\/\/api.semanticscholar.org\/CorpusID:206593710"},{"key":"e_1_2_2_12_1","volume-title":"Interactive Story Visualization with Multiple Characters. SIGGRAPH Asia 2023 Conference Papers","author":"Gong Yuan","year":"2023","unstructured":"Yuan Gong, Youxin Pang, Xiaodong Cun, Menghan Xia, Haoxin Chen, Longyue Wang, Yong Zhang, Xintao Wang, Ying Shan, and Yujiu Yang. 2023. Interactive Story Visualization with Multiple Characters. SIGGRAPH Asia 2023 Conference Papers (2023). https:\/\/api.semanticscholar.org\/CorpusID:258960665"},{"key":"e_1_2_2_13_1","doi-asserted-by":"publisher","DOI":"10.1109\/PROC.1979.11328"},{"key":"e_1_2_2_14_1","unstructured":"Amir Hertz Ron Mokady Jay Tenenbaum Kfir Aberman Yael Pritch and Daniel Cohen-Or. 2022. Prompt-to-prompt image editing with cross attention control. (2022)."},{"key":"e_1_2_2_15_1","volume-title":"2024 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Hertz Amir","year":"2023","unstructured":"Amir Hertz, Andrey Voynov, Shlomi Fruchter, and Daniel Cohen-Or. 2023. Style Aligned Image Generation via Shared Attention. 2024 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023), 4775\u20134785. https:\/\/api.semanticscholar.org\/CorpusID:265608730"},{"key":"e_1_2_2_16_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.595"},{"key":"e_1_2_2_17_1","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.167"},{"key":"e_1_2_2_18_1","volume-title":"StyleShot: A Snapshot on Any Style. arXiv preprint arxiv:2407.01414","author":"Junyao Gao","year":"2024","unstructured":"Gao Junyao, Liu Yanchen, Sun Yanan, Tang Yinhao, Zeng Yanhong, Chen Kai, and Zhao Cairong. 2024. StyleShot: A Snapshot on Any Style. arXiv preprint arxiv:2407.01414 (2024)."},{"key":"e_1_2_2_19_1","unstructured":"Maxime Oquab Timoth\u00e9e Darcet Theo Moutakanni Huy V. Vo Marc Szafraniec Vasil Khalidov Pierre Fernandez Daniel Haziza Francisco Massa Alaaeldin El-Nouby Russell Howes Po-Yao Huang Hu Xu Vasu Sharma Shang-Wen Li Wojciech Galuba Mike Rabbat Mido Assran Nicolas Ballas Gabriel Synnaeve Ishan Misra Herve Jegou Julien Mairal Patrick Labatut Armand Joulin and Piotr Bojanowski. 2023. DINOv2: Learning Robust Visual Features without Supervision."},{"key":"e_1_2_2_20_1","volume-title":"SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis. In The Twelfth International Conference on Learning Representations, ICLR 2024","author":"Podell Dustin","year":"2024","unstructured":"Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M\u00fcller, Joe Penna, and Robin Rombach. 2024. SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis. In The Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7\u201311, 2024. OpenReview.net. https:\/\/openreview.net\/forum?id=di52zR8xgf"},{"key":"e_1_2_2_21_1","volume-title":"High-Resolution Image Synthesis with Latent Diffusion Models. 2022 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Rombach Robin","year":"2021","unstructured":"Robin Rombach, A. Blattmann, Dominik Lorenz, Patrick Esser, and Bj\u00f6rn Ommer. 2021. High-Resolution Image Synthesis with Latent Diffusion Models. 2022 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2021), 10674\u201310685. https:\/\/api.semanticscholar.org\/CorpusID:245335280"},{"key":"e_1_2_2_22_1","doi-asserted-by":"crossref","unstructured":"Nataniel Ruiz Yuanzhen Li Varun Jampani Yael Pritch Michael Rubinstein and Kfir Aberman. 2022. DreamBooth: Fine Tuning Text-to-image Diffusion Models for Subject-Driven Generation. (2022).","DOI":"10.1109\/CVPR52729.2023.02155"},{"key":"e_1_2_2_23_1","unstructured":"Simo Ryu. 2023. Low-rank Adaptation for Fast Text-to-Image Diffusion Fine-tuning. https:\/\/github.com\/cloneofsimo\/lora. Accessed: 2025-05-19."},{"key":"e_1_2_2_24_1","volume-title":"Burcu Karagol Ayan, Seyedeh Sara Mahdavi, Raphael Gontijo Lopes, Tim Salimans, Jonathan Ho, David J. Fleet, and Mohammad Norouzi.","author":"Saharia Chitwan","year":"2022","unstructured":"Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L. Denton, Seyed Kamyar Seyed Ghasemipour, Burcu Karagol Ayan, Seyedeh Sara Mahdavi, Raphael Gontijo Lopes, Tim Salimans, Jonathan Ho, David J. Fleet, and Mohammad Norouzi. 2022. Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding. ArXiv abs\/2205.11487 (2022). https:\/\/api.semanticscholar.org\/CorpusID:248986576"},{"key":"e_1_2_2_25_1","volume-title":"Cheng Perng Phoo, and Bharath Hariharan","author":"Tang Luming","year":"2023","unstructured":"Luming Tang, Menglin Jia, Qianqian Wang, Cheng Perng Phoo, and Bharath Hariharan. 2023. Emergent Correspondence from Image Diffusion. ArXiv abs\/2306.03881 (2023). https:\/\/api.semanticscholar.org\/CorpusID:259089017"},{"key":"e_1_2_2_26_1","doi-asserted-by":"crossref","unstructured":"Yoad Tewel Omri Kaduri Rinon Gal Yoni Kasten Lior Wolf Gal Chechik and Yuval Atzmon. 2024. Training-Free Consistent Text-to-Image Generation. arXiv:2402.03286 [cs.CV] https:\/\/arxiv.org\/abs\/2402.03286","DOI":"10.1145\/3658157"},{"key":"e_1_2_2_27_1","volume-title":"Deep Image Harmonization. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Tsai Yi-Hsuan","year":"2017","unstructured":"Yi-Hsuan Tsai, Xiaohui Shen, Zhe L. Lin, Kalyan Sunkavalli, Xin Lu, and Ming-Hsuan Yang. 2017. Deep Image Harmonization. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2017), 2799\u20132807. https:\/\/api.semanticscholar.org\/CorpusID:1033001"},{"key":"e_1_2_2_28_1","doi-asserted-by":"crossref","unstructured":"Ye Wang Tongyuan Bai Xuping Xie Zili Yi Yilin Wang and Rui Ma. 2025. SigStyle: Signature Style Transfer via Personalized Text-to-Image Models. https:\/\/wangyephd.github.io\/projects\/sigstyle.html. arXiv preprint.","DOI":"10.1609\/aaai.v39i8.32868"},{"key":"e_1_2_2_29_1","doi-asserted-by":"publisher","DOI":"10.1145\/3728461"},{"key":"e_1_2_2_30_1","unstructured":"Hu Ye Jun Zhang Sibo Liu Xiao Han and Wei Yang. 2023. IP-Adapter: Text Compatible Image Prompt Adapter for Text-to-Image Diffusion Models. (2023)."},{"key":"e_1_2_2_31_1","doi-asserted-by":"crossref","unstructured":"Richard Zhang Phillip Isola Alexei A Efros Eli Shechtman and Oliver Wang. 2018. The Unreasonable Effectiveness of Deep Features as a Perceptual Metric. In CVPR.","DOI":"10.1109\/CVPR.2018.00068"},{"key":"e_1_2_2_32_1","volume-title":"StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation. NeurIPS 2024","author":"Zhou Yupeng","year":"2024","unstructured":"Yupeng Zhou, Daquan Zhou, Ming-Ming Cheng, Jiashi Feng, and Qibin Hou. 2024. StoryDiffusion: Consistent Self-Attention for Long-Range Image and Video Generation. NeurIPS 2024 (2024)."},{"key":"e_1_2_2_33_1","volume-title":"Image Harmonization by Matching Regional References. ArXiv abs\/2204.04715","author":"Zhu Ziyue","year":"2022","unstructured":"Ziyue Zhu, Zhao Zhang, Zheng Lin, Ruiqi Wu, Zhi Chai, and Chunle Guo. 2022. Image Harmonization by Matching Regional References. ArXiv abs\/2204.04715 (2022). https:\/\/api.semanticscholar.org\/CorpusID:248085801"}],"container-title":["ACM Transactions on Graphics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3763303","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T21:16:31Z","timestamp":1764969391000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3763303"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,12]]},"references-count":33,"journal-issue":{"issue":"6","published-print":{"date-parts":[[2025,12]]}},"alternative-id":["10.1145\/3763303"],"URL":"https:\/\/doi.org\/10.1145\/3763303","relation":{},"ISSN":["0730-0301","1557-7368"],"issn-type":[{"value":"0730-0301","type":"print"},{"value":"1557-7368","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12]]},"assertion":[{"value":"2025-05-23","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-08-09","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2025-12-04","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}