AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

SERUM: Simple, Efficient, Robust, and Unifying Marking for Diffusion-based Image Generation

Jan Kociszewski, Hubert Jastrzębski, Tymoteusz Stępkowski, Filip Manijak, Krzysztof Rojek, Franziska Boenisch, Adam Dziedzic

ICLR 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. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add dynamic augmentation sampler (m=4) on top of precomputed augmentations (ablation)SERUM detector · Perturbation robustness avg, SD 2.1TPR @ 1% FPR: 97.16 -> 99.75Improved
add augmentations during detector training (precomputed) (ablation)SERUM detector · Perturbation robustness avg, SD 2.1TPR @ 1% FPR: 78.68 -> 97.16Improved
increase watermark strength alpha from 0.1 to 0.5 (ablation)SERUM · Perturbation robustness avg, SD 2.1TPR @ 1% FPR: 78.49 -> 99.75Improved
increase watermark strength alpha from 0.5 to 0.7 (ablation)SERUM · COCO FID, SD 2.1FID: 19.24 -> 20.82Hurt
set sampler temperature to 1 versus 0 or infinity (ablation)SERUM augmentation sampler · Perturbation robustness avg, SD 2.1TPR @ 1% FPR: 99.48 (tau=0), 99.34 (tau=inf) -> 99.75 (tau=1)Improved
increase sampler batch size m from 4 to 32 (ablation)SERUM · Perturbation robustness avg, SD 2.1TPR @ 1% FPR: 99.75 -> 99.82; training 9.5h -> 60hNo clear effect
train detector with perturbation terms removed (ablation)SERUM · Regen-30 attackTPR @ 1% FPR: 88.50 -> 52.28Hurt
omit rotation from training augmentations (ablation)SERUM detector · RotateTPR @ 1% FPR: 99.34 -> 18.46Hurt
use SERUM watermark instead of GaussMarker/RingIDStable Diffusion · COCO 2014 FID avgFID: 20.16 (GaussMarker), 19.70 (RingID) -> 18.96Improved
use SERUM instead of RingID/GaussMarker on strict-FPR evaluationStable Diffusion · Perturbation robustness avg, SD 2.1TPR @ 0% FPR: 23.59 (GaussMarker) -> 96.34Improved
apply SERUM to DiT-based SD 3.0 instead of UNet SD 2.1SERUM · Perturbation robustness avgTPR @ 1% FPR: 99.75 -> 95.24Hurt
fine-tune on watermarked images with LoRA instead of full fine-tuningSD 1.4 · RadioactivityTPR @ 1% FPR: 77.12 -> 52.30Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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