Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2509.24798 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918501901402112 |
|---|---|
| author | Tong, Lei Liu, Zhihua Lu, Chaochao Oglic, Dino Diethe, Tom Teare, Philip Tsaftaris, Sotirios A. Jin, Chen |
| author_facet | Tong, Lei Liu, Zhihua Lu, Chaochao Oglic, Dino Diethe, Tom Teare, Philip Tsaftaris, Sotirios A. Jin, Chen |
| contents | We present Causal-Adapter, a modular framework that adapts frozen text-to-image diffusion backbones for counterfactual image generation. Our method supports causal interventions on target attributes and consistently propagates their effects to causal dependents while preserving the core identity of the image. Unlike prior approaches that rely on prompt engineering without explicit causal structure, Causal-Adapter leverages structural causal modeling with two attribute-regularization strategies: (i) prompt-aligned injection, which aligns causal attributes with textual embeddings for precise semantic control, and (ii) a conditioned token contrastive loss that disentangles attribute factors and reduces spurious correlations. Causal-Adapter achieves state-of-the-art performance on both synthetic and real-world datasets, including up to a 91% reduction in MAE on Pendulum for accurate attribute control and up to an 87% reduction in FID on ADNI for high-fidelity MRI generation. These results demonstrate robust, generalizable counterfactual editing with faithful attribute modification and strong identity preservation. Code and models will be released at: https://leitong02.github.io/causaladapter/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_24798 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Causal-Adapter: Taming Text-to-Image Diffusion for Faithful Counterfactual Generation Tong, Lei Liu, Zhihua Lu, Chaochao Oglic, Dino Diethe, Tom Teare, Philip Tsaftaris, Sotirios A. Jin, Chen Computer Vision and Pattern Recognition Artificial Intelligence We present Causal-Adapter, a modular framework that adapts frozen text-to-image diffusion backbones for counterfactual image generation. Our method supports causal interventions on target attributes and consistently propagates their effects to causal dependents while preserving the core identity of the image. Unlike prior approaches that rely on prompt engineering without explicit causal structure, Causal-Adapter leverages structural causal modeling with two attribute-regularization strategies: (i) prompt-aligned injection, which aligns causal attributes with textual embeddings for precise semantic control, and (ii) a conditioned token contrastive loss that disentangles attribute factors and reduces spurious correlations. Causal-Adapter achieves state-of-the-art performance on both synthetic and real-world datasets, including up to a 91% reduction in MAE on Pendulum for accurate attribute control and up to an 87% reduction in FID on ADNI for high-fidelity MRI generation. These results demonstrate robust, generalizable counterfactual editing with faithful attribute modification and strong identity preservation. Code and models will be released at: https://leitong02.github.io/causaladapter/. |
| title | Causal-Adapter: Taming Text-to-Image Diffusion for Faithful Counterfactual Generation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2509.24798 |