I2V-DETACH limits unauthorized video creation from images
I2V-DETACH: Source Grounding Detachment for Unauthorized Image-to-Video Generation
Computer Vision and Pattern Recognition
Summary
People can turn images into videos using advanced AI, but this can be misused without permission. The authors studied ways to stop AI from copying important parts of the original image when making videos, instead of making the videos look worse. Their new method, I2V-DETACH, breaks the way AI keeps using clues from the image while generating videos. Tests show it works better than earlier tricks and can work even on AI systems it wasn’t directly designed for.
What this means in practice
- •For digital content protection teams: Prevent unauthorized video generation from protected images by disrupting source-content preservation in AI workflows.
- •For media editing software developers: Integrate image immunization methods to reduce illicit use of images in automated video creation tools.
Authors
Chanhui Lee, Yeonghwan Song, Yewon Kang, Jeany Son
Abstract
Recent diffusion models have advanced image-conditioned video generation, but their accessibility raises risks of unauthorized manipulation. Image immunization has emerged as a promising defense for Image-to-Video (I2V) protection, yet existing methods often rely on degrading visual quality or temporal coherence. In this work, we revisit I2V immunization as reducing source-content preservation, shifting the focus from output degradation to weakening source-conditioned content propagation. To this end, we propose I2V-DETACH, an image immunization method that prevents unauthorized I2V generation through source grounding detachment. I2V models preserve source content by propagating source-image cues into video latents during denoising. Accordingly, I2V-DETACH disrupts this process in two complementary ways: it suppresses source-video grounding to reduce source cue propagation, and strengthens intra-video interactions to shift denoising toward video-side dynamics with reduced reliance on the source image. To obtain informative optimization signals, we initialize auxiliary video latents with source-oriented states that remain within the source-conditioned generation regime while deviating from source-preserving trajectories. Extensive experiments across diverse diffusion architectures show that I2V-DETACH substantially reduces source-content preservation, improves over prior immunization methods, and exhibits strong black-box transferability.