You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: index.html
+56-25Lines changed: 56 additions & 25 deletions
Original file line number
Diff line number
Diff line change
@@ -5,7 +5,7 @@
5
5
<!-- Meta tags for social media banners, these should be filled in appropriatly as they are your "business card" -->
6
6
<!-- Replace the content tag with appropriate information -->
7
7
<metaname="description" content="DiffVax is a scalable, lightweight, and fast framework for image and video immunization, preventing diffusion-based editing with imperceptible noise and achieving unprecedented speed and robustness.">
8
-
<metaproperty="og:title" content="Optimization-Free Image Immunization Against Diffusion-Based Editing"/>
8
+
<metaproperty="og:title" content="DiffVax: Optimization-Free Image Immunization Against Diffusion-Based Editing"/>
9
9
<metaproperty="og:description" content="DiffVax is a scalable, lightweight, and fast framework for image and video immunization, preventing diffusion-based editing with imperceptible noise and achieving unprecedented speed and robustness."/>
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X630-->
@@ -14,7 +14,7 @@
14
14
<metaproperty="og:image:height" content="630"/>
15
15
16
16
17
-
<metaname="twitter:title" content="Optimization-Free Image Immunization Against Diffusion-Based Editing">
17
+
<metaname="twitter:title" content="DiffVax: Optimization-Free Image Immunization Against Diffusion-Based Editing">
18
18
<metaname="twitter:description" content="DiffVax is a scalable, lightweight, and fast framework for image and video immunization, preventing diffusion-based editing with imperceptible noise and achieving unprecedented speed and robustness.">
19
19
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X600-->
<spanclass="author-block"><ahref="https://scholar.google.com/citations?user=262ASa4AAAAJ&hl=en" target="_blank">Sandeep P. Chinchali</a><sup>2</sup></span>
77
78
<spanclass="author-block"><ahref="https://scholar.google.com/citations?hl=en&user=8kA3eDwAAAAJ" target="_blank">James M. Rehg</a><sup>1</sup></span>
Current image immunization defense techniques against diffusion-based editing embed imperceptible noise in target images to disrupt editing models. However, these methods face scalability challenges, as they require time-consuming re-optimization for each image—taking hours for small batches. To address these challenges, we introduce DiffVax, a scalable, lightweight, and optimization-free framework for image immunization, specifically designed to prevent diffusion-based editing. Our approach enables effective generalization to unseen content, reducing computational costs and cutting immunization time from days to milliseconds—achieving a 250,000$\times$ speedup. This is achieved through a loss term that ensures the failure of editing attempts and the imperceptibility of the perturbations. Extensive qualitative and quantitative results demonstrate that our model is scalable, optimization-free, adaptable to various diffusion-based editing tools, robust against counter-attacks, and, for the first time, effectively protects video content from editing. Our code and qualitative results are provided in the supplementary.
152
+
Current image immunization defense techniques against diffusion-based editing embed imperceptible noise into target images to disrupt editing models. However, these methods face scalability challenges, as they require time-consuming optimization for each image separately, taking hours for small batches. To address these challenges, we introduce DiffVax, a scalable, lightweight, and optimization-free framework for image immunization, specifically designed to prevent diffusion-based editing. Our approach enables effective generalization to unseen content, reducing computational costs and cutting immunization time from days to milliseconds, achieving a speedup of 250,000×. This is achieved through a loss term that ensures the failure of editing attempts and the imperceptibility of the perturbations. Extensive qualitative and quantitative results demonstrate that our model is scalable, optimization-free, adaptable to various diffusion-based editing tools, robust against counter-attacks, and, for the first time, effectively protects video content from editing.
<strong>(a)</strong> An attacker performs malicious editing on an original image.
177
+
178
+
<strong>(b)</strong> Existing defenses immunize images by solving a costly optimization problem for each image individually, taking over 10 minutes per image.
179
+
<strong>(c)</strong> DiffVax enables scalable protection by first training an immunizer model (green box) on a diverse dataset. Once trained, the model can immunize unseen images with a single forward pass, producing effective perturbations in approximately 70 milliseconds per image. </p>
Our process begins with the immunizer model $f(\cdot;\theta)$ which generates imperceptible noise $\epsilon_{im}$ to be applied to original image $I$. This noise is applied to the masked region $M$ of the image, resulting in immunized image $I_{im}$. The immunized image is then processed by a diffusion-based editing model $SD(\cdot)$ using a text prompt $P$ and the complementary mask $\sim M$ to edit the background of the original image. The training aims to minimize two loss terms $\mathcal{L}_{noise}$ and $\mathcal{L}_{edit}$, which penalizes the applied noise magnitude, and if the edit is successful, respectively. During training, the immunizer learns to generalize across diverse images, ensuring editing attempts fail while preserving visual fidelity. This end-to-end framework enables robust, scalable immunization against diffusion-based editing for both images and videos.</p>
0 commit comments