AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Beyond Bit Matching: Orthogonal Watermarks for Collusion-Resistant Image Fingerprinting

Zhakshylyk Nurlanov, Tobias Weißberg, Florian Bernard

NeurIPS 2026 · Image generation › Autoregressive and latent generation · approach: Robust diffusion watermarking · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes OrthoMark, an image fingerprinting method that maps user keys to near-orthogonal watermark patterns to resist averaging collusion attacks.

Problem

Image fingerprinting for source tracing is vulnerable to averaging collusion, where several recipients average differently marked copies to suppress individual watermarks; discrete bit-string watermarks are especially fragile because averaging pushes bit evidence toward ambiguous decisions.

Approach

Maps each user key to a near-orthogonal watermark pattern instead of a discrete bit string, so that averaged copies retain identifiable evidence for colluders rather than becoming ambiguous.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use continuous near-orthogonal vector keys with cosine retrieval instead of binary payload bit matchingVectorMark vs TrustMark · MS-COCO, K=5 averagingstrict all-colluder TPR: 0.000 -> 1.00Improved
train a neural encoder/extractor instead of spread-spectrum embeddingVectorMark vs SS-DFT · MS-COCO, K=10 averagingstrict all-colluder TPR: .171 -> .996Improved
train with collusion objective (K_train=20) (ablation)VectorMark · short runs, K=20 averagingstrict all-colluder TPR: .234 -> .650Improved
add collusion objective at K=2..10 (ablation)VectorMark · short runs, K<=10strict all-colluder TPR: 1.00 -> 1.00No clear effect
add collusion training (ablation)VectorMark · short runs, geometric single-key TPRTPR: .926 -> .901Hurt
increase key dimension from 100 to 1024 (ablation)VectorMark · short runs, K=10 averagingstrict all-colluder TPR: .000 -> 1.00Improved
increase key dimension from 100 to 576 (ablation)VectorMark · short runs, single-key geometricTPR: .969 -> .995Improved
use soft cosine scores instead of hard codeword decisions (ablation)VectorMark · MS-COCO registry N=1e6, K=10exact-set recovery: .624 -> .974Improved
retain logits for cosine matching instead of hard agreement (ablation)ChunkySeal · registry N=1e6, K=10exact-set recovery: .002 -> .441Improved
two-pass decoding with horizontally flipped image (ablation)VectorMark payload decoding · MS-COCO geometric payloadbit accuracy: .827 -> .975Improved
scale registry from 1e3 to 1e6 keysVectorMark · K=10 exact-set recoveryexact-set recovery: .989 -> .980No clear effect
increase coalition size from 10 to 20 at N=1e6VectorMark · exact-set recoveryexact-set recovery: .980 -> .007Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy