Residual Transferability in Neural Image Watermarking
- 类型：arxiv
- 标识：2609.32241
- 链接：https://arxiv.org/abs/2609.32241
- 主分类：engineering
- 形态：method
- TLDR：Neural image watermarks can be forged by extracting watermark-bearing residuals from released images and transferring them to unrelated content. While prior work has demonstrated this vulnerability, what makes these residuals transferable remains poorly understood. We formalize this vulnerability with residual transferability (RT), a metric that quantifies how well watermark evidence remains decodable after transfer across unrelated images. Through comparative analyses and controlled interventions, we find that common training-side variations do not account for the large RT differences across
- 待LLM分类：否
- 来源文件：
- /inbox/tom/_candidates/2026-09-29-agent-rag-longcontext-candidates.json