NSTL国家科技图书文献中心

1. ControlNet++: Improving Conditional Controls with Efficient Consistency Feedback NSTL国家科技图书文献中心

Ming Li | Taojiannan Yang ... - 《Computer Vision - ECCV 2024,Part VII》 - European Conference on Computer Vision - 2025, - 129～147 - 共19页

摘要： ControlNet incorporated image-based conditional controls | this end, we propose ControlNet++, a novel approach | experiments show that ControlNet++ significantly improves | , it achieves improvements over ControlNet* by 11.1* | To enhance the controllability of text-to

关键词： Controllable generation | Diffusion model | ControlNet

2. SmartControl: Enhancing ControlNet for Handling Rough Visual Conditions NSTL国家科技图书文献中心

Xiaoyu Liu | Yuxiang Wei ... - 《Computer Vision - ECCV 2024,Part XLVIII》 - European Conference on Computer Vision - 2025, - 1～17 - 共17页

摘要： of the backbone model (e.g., ControlNet), which | Recent text-to-image generation methods such | as Control-Net have achieved remarkable success in | controlling image layouts, where the generated images by the | default model are constrained to strictly follow the

关键词： Text-to-Image generation | ControlNet | Rough conditions

3. Line art Extraction of Chinese Building Structure based on ControlNet Model NSTL国家科技图书文献中心

Fan Wu | Junjie Cheng ... - 《Sixteenth International Conference on Graphics and Image Processing (ICGIP 2024)》 - International Conference on Graphics and Image Processing - 2025, - 135391H.1～135391H.12 - 共12页

摘要： ControlNet and LoRA. The experiments demonstrate that our | Drawing architectural plan is a crucial step | in the preservation of ancient Chinese building | . This process is often time-consuming and typically | carried out by skilled human. This paper presents a

关键词： Historic building | Image segmentation | Lineart

4. ControlNet-XS: Rethinking the Control of Text-to-Image Diffusion Models as Feedback-Control Systems NSTL国家科技图书文献中心

Denis Zavadski | Johann-Friedrich Fei...... - 《Computer Vision - ECCV 2024,Part LXXXVIII》 - European Conference on Computer Vision - 2025, - 343～362 - 共20页

摘要： take an existing controlling network (ControlNet) and | proposed network ControlNet-XS. When comparing with the | The field of image synthesis has made | tremendous strides forward in the last years. Besides | defining the desired output image with text-prompts, an

关键词： Text-to-image generation | Controlling image generation models | Feedback-Control systems

5. Mask-ControlNet: Higher-Quality Image Generation with an Additional Mask Prompt NSTL国家科技图书文献中心

Zhiqi Huang | Huixin Xiong ... - 《Pattern Recognition,Part VI》 - International Conference on Pattern Recognition - 2025, - 83～97 - 共15页

摘要： issue, we developed a framework termed Mask-ControlNet | Text-to-image generation has witnessed great | progress, especially with the recent advancements in | diffusion models. Since texts cannot provide detailed | conditions like object appearance, reference images are

关键词： Image generation | Diffusion model | Object reconstruction

6. Learning Diffusion Models for Multi-view Anomaly Detection NSTL国家科技图书文献中心

Chieh Liu | Yu-Min Chu ... - 《Computer Vision - ECCV 2024,Part XXXIII》 - European Conference on Computer Vision - 2025, - 328～345 - 共18页

摘要： involves training a view-invariant ControlNet* that can* | We are exploring an emerging formulation in | anomaly detection (AD) where multiple instances of the | same object are produced simultaneously and | distinctly to address the limitation that using only a

关键词： Anomaly detection | Diffusion model | ControlNet

7. Multi-exposure Driven Stable Diffusion for Shadow Removal NSTL国家科技图书文献中心

Zheng Yan | Wenhao Tan ... - 《Advances in Brain Inspired Cognitive Systems,Part II》 - International Conference on Advances in Brain Inspired Cognitive Systems - 2025, - 148～158 - 共11页

摘要： on the ControlNet-driven stable diffusion model | Shadow removal has long been a challenging | task in the field of computer vision. Despite the | significant progress since the advent of deep learning | , existing approaches still fall short of robustness in

关键词： Shadow removal | Stable diffusion | Multi-exposure | ControlNet

8. TDM: Temporally-Consistent Diffusion Model for All-in-One Real-World Video Restoration NSTL国家科技图书文献中心

Yizhou Li | Zihua Liu ... - 《MultiMedia Modeling,Part IV》 - International Conference on MultiMedia Modeling - 2025, - 155～169 - 共15页

摘要： a fine-tuned ControlNet. Our method can restore | In this paper, we propose the first diffusion | -based all-in-one video restoration method that | utilizes the power of a pre-trained Stable Diffusion and | various types of video degradation with a single unified

关键词： Multi-task video restoration | Diffusion models | ControlNet

9. Enriching Information and Preserving Semantic Consistency in Expanding Curvilinear Object Segmentation Datasets NSTL国家科技图书文献中心

Qin Lei | Jiang Zhong ... - 《Computer Vision - ECCV 2024,Part LXXXIV》 - European Conference on Computer Vision - 2025, - 233～250 - 共18页

摘要： Semantic Consistency Preserving ControlNet* (SCP* | ControlNet). This involves an adaptation of ControlNet with | Curvilinear object segmentation plays a | crucial role across various applications, yet datasets | in this domain often suffer from small scale due to

关键词： Curvilinear structure segmentation | Expanding dataset | Image synthesis

10. TCAN: Animating Human Images with Temporally Consistent Pose Guidance Using Diffusion Models NSTL国家科技图书文献中心

Jeongho Kim | Min-Jung Kim ... - 《Computer Vision - ECCV 2024,Part LIX》 - European Conference on Computer Vision - 2025, - 326～342 - 共17页

摘要： utilize the pre-trained ControlNet* without fine-tuning* | ControlNet frozen, we adapt LoRA to the UNet layers | introducing an additional temporal layer to the ControlNet | Pose-driven human-image animation diffusion | models have shown remarkable capabilities in realistic

摘要： ControlNet incorporated image-based conditional controls | this end, we propose ControlNet++, a novel approach | experiments show that ControlNet++ significantly improves | , it achieves improvements over ControlNet* by 11.1* | To enhance the controllability of text-to

关键词： Controllable generation | Diffusion model | ControlNet

摘要： of the backbone model (e.g., ControlNet), which | Recent text-to-image generation methods such | as Control-Net have achieved remarkable success in | controlling image layouts, where the generated images by the | default model are constrained to strictly follow the

关键词： Text-to-Image generation | ControlNet | Rough conditions

摘要： ControlNet and LoRA. The experiments demonstrate that our | Drawing architectural plan is a crucial step | in the preservation of ancient Chinese building | . This process is often time-consuming and typically | carried out by skilled human. This paper presents a

关键词： Historic building | Image segmentation | Lineart

摘要： take an existing controlling network (ControlNet) and | proposed network ControlNet-XS. When comparing with the | The field of image synthesis has made | tremendous strides forward in the last years. Besides | defining the desired output image with text-prompts, an

关键词： Text-to-image generation | Controlling image generation models | Feedback-Control systems

摘要： issue, we developed a framework termed Mask-ControlNet | Text-to-image generation has witnessed great | progress, especially with the recent advancements in | diffusion models. Since texts cannot provide detailed | conditions like object appearance, reference images are

关键词： Image generation | Diffusion model | Object reconstruction

摘要： involves training a view-invariant ControlNet* that can* | We are exploring an emerging formulation in | anomaly detection (AD) where multiple instances of the | same object are produced simultaneously and | distinctly to address the limitation that using only a

关键词： Anomaly detection | Diffusion model | ControlNet

摘要： on the ControlNet-driven stable diffusion model | Shadow removal has long been a challenging | task in the field of computer vision. Despite the | significant progress since the advent of deep learning | , existing approaches still fall short of robustness in

关键词： Shadow removal | Stable diffusion | Multi-exposure | ControlNet

摘要： a fine-tuned ControlNet. Our method can restore | In this paper, we propose the first diffusion | -based all-in-one video restoration method that | utilizes the power of a pre-trained Stable Diffusion and | various types of video degradation with a single unified

关键词： Multi-task video restoration | Diffusion models | ControlNet

摘要： Semantic Consistency Preserving ControlNet* (SCP* | ControlNet). This involves an adaptation of ControlNet with | Curvilinear object segmentation plays a | crucial role across various applications, yet datasets | in this domain often suffer from small scale due to

关键词： Curvilinear structure segmentation | Expanding dataset | Image synthesis

摘要： utilize the pre-trained ControlNet* without fine-tuning* | ControlNet frozen, we adapt LoRA to the UNet layers | introducing an additional temporal layer to the ControlNet | Pose-driven human-image animation diffusion | models have shown remarkable capabilities in realistic

关键词： Image-to-Video generation | Human animation | Diffusion model

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： ControlNet incorporated image-based conditional controls | this end, we propose ControlNet++, a novel approach | experiments show that ControlNet++ significantly improves | , it achieves improvements over ControlNet by 11.1 | To enhance the controllability of text-to

关键词： Controllable generation | Diffusion model | ControlNet

摘要： of the backbone model (e.g., ControlNet), which | Recent text-to-image generation methods such | as Control-Net have achieved remarkable success in | controlling image layouts, where the generated images by the | default model are constrained to strictly follow the

关键词： Text-to-Image generation | ControlNet | Rough conditions

摘要： ControlNet and LoRA. The experiments demonstrate that our | Drawing architectural plan is a crucial step | in the preservation of ancient Chinese building | . This process is often time-consuming and typically | carried out by skilled human. This paper presents a

关键词： Historic building | Image segmentation | Lineart

摘要： take an existing controlling network (ControlNet) and | proposed network ControlNet-XS. When comparing with the | The field of image synthesis has made | tremendous strides forward in the last years. Besides | defining the desired output image with text-prompts, an

关键词： Text-to-image generation | Controlling image generation models | Feedback-Control systems

摘要： issue, we developed a framework termed Mask-ControlNet | Text-to-image generation has witnessed great | progress, especially with the recent advancements in | diffusion models. Since texts cannot provide detailed | conditions like object appearance, reference images are

关键词： Image generation | Diffusion model | Object reconstruction

摘要： involves training a view-invariant ControlNet that can | We are exploring an emerging formulation in | anomaly detection (AD) where multiple instances of the | same object are produced simultaneously and | distinctly to address the limitation that using only a

关键词： Anomaly detection | Diffusion model | ControlNet

摘要： on the ControlNet-driven stable diffusion model | Shadow removal has long been a challenging | task in the field of computer vision. Despite the | significant progress since the advent of deep learning | , existing approaches still fall short of robustness in

关键词： Shadow removal | Stable diffusion | Multi-exposure | ControlNet

摘要： a fine-tuned ControlNet. Our method can restore | In this paper, we propose the first diffusion | -based all-in-one video restoration method that | utilizes the power of a pre-trained Stable Diffusion and | various types of video degradation with a single unified

关键词： Multi-task video restoration | Diffusion models | ControlNet

摘要： Semantic Consistency Preserving ControlNet (SCP | ControlNet). This involves an adaptation of ControlNet with | Curvilinear object segmentation plays a | crucial role across various applications, yet datasets | in this domain often suffer from small scale due to

关键词： Curvilinear structure segmentation | Expanding dataset | Image synthesis

摘要： utilize the pre-trained ControlNet without fine-tuning | ControlNet frozen, we adapt LoRA to the UNet layers | introducing an additional temporal layer to the ControlNet | Pose-driven human-image animation diffusion | models have shown remarkable capabilities in realistic

关键词： Image-to-Video generation | Human animation | Diffusion model

4008-161-200 800-990-8900

国家科技图书文献中心

摘要： ControlNet incorporated image-based conditional controls | this end, we propose ControlNet++, a novel approach | experiments show that ControlNet++ significantly improves | , it achieves improvements over ControlNet* by 11.1* | To enhance the controllability of text-to

摘要： involves training a view-invariant ControlNet* that can* | We are exploring an emerging formulation in | anomaly detection (AD) where multiple instances of the | same object are produced simultaneously and | distinctly to address the limitation that using only a

摘要： Semantic Consistency Preserving ControlNet* (SCP* | ControlNet). This involves an adaptation of ControlNet with | Curvilinear object segmentation plays a | crucial role across various applications, yet datasets | in this domain often suffer from small scale due to

摘要： utilize the pre-trained ControlNet* without fine-tuning* | ControlNet frozen, we adapt LoRA to the UNet layers | introducing an additional temporal layer to the ControlNet | Pose-driven human-image animation diffusion | models have shown remarkable capabilities in realistic