Get Started
Home
Topics
Search
Library
Research questionHow can text-guided image editing explain edits while keeping provenance resistant to white-box removal?Image-editing provenance should identify not only an image’s origin but also what instruction produced the edit. Separate watermarks may fail to carry this semantic information and can be easier to locate and remove when model weights are open.
AI
Computer Vision
Diffusion Models
Evaluation & Benchmarks
Image & Video Processing
Image Generation
Machine Learning
Research Paper
Technology
Latest papersRecent research connected to this question, newest first.AngelFingerprint: A Traceable, Explainable, and White-Box Stealthy Watermark for Text-Guided Image EditingThe source embeds the editing prompt’s CLIP text representation into diffusion-model weights and extracts it from generated image pixels. It evaluates this approach for text-guided editing on MagicBrush, including prompt-retrieval evidence; broader robustness and security guarantees are not established.research paper · Sep 4, 2026
Related questions
How can image editors infer edit regions and preserve non-target content without introducing artifacts?How can text-to-image diffusion models erase unwanted concepts while preserving benign concepts and resisting re-emergence?How can image forensics detect manipulated text, locate altered regions, and ground explanations in evidence?How can multi-view image editing preserve geometric and photometric consistency across views of a 3D scene?