AI Paper Atlas

Research map / Video / Video segmentation and tracking

Preserving Forgery Artifacts: AI-Generated Video Detection at Native Scale

Zhengcen Li, Chenyang Jiang, Hang Zhao, Shiyang Zhou, Yunyang Mo, Feng Gao, Fan Yang, Qiben Shan, Shaocong Wu, Jingyong Su

ICLR 2026 · Video › Video segmentation and tracking · approach: Video restoration with temporal priors

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
process frames at dynamic native resolution up to 720p instead of random crop to 224p (ablation)Qwen2.5-ViT · Magic Videos + GenVideo avg ACCavg ACC: 78.06 -> 89.92Improved
use dynamic resolution instead of random resize to 224p (ablation)Qwen2.5-ViT · Magic Videos + GenVideo avg ACCavg ACC: 84.61 -> 88.60Improved
raise the maximum resolution from 448p to 720p (ablation)Qwen2.5-ViT · Magic Videos (avg ACC)ACC: 81.19 -> 83.20Improved
increase sampled frames T from 2 to 8 (ablation)Qwen2.5-ViT · Magic Videos + GenVideo avg ACCavg ACC: 82.93 -> 88.60Improved
full fine-tuning instead of LoRA (r=16) (ablation)Qwen2.5-ViT · Magic Videos + GenVideo avg ACCavg ACC: 86.84 -> 88.60Improved
LoRA instead of linear probing (ablation)Qwen2.5-ViT · Magic Videos + GenVideo avg ACCavg ACC: 81.26 -> 86.84Improved
higher spatial/temporal fidelity on GenVideo (low-res short videos) versus Magic (ablation)Qwen2.5-ViT · GenVideoavg ACC: 93.50 -> 96.64Mixed
use image-based AI-generated image detectors trained on image data for video detectionRINE, FatFormer, B-Free, Effort, WaveRep · Magic VideosmACC: 49.47 for RINE vs 83.20 for Qwen2.5-ViTHurt
train on higher-quality generators (higher VBench score)cross-generator detectors · Cross-validation (14 generators)cross-validation accuracy: Pearson rho = 0.86Improved
restrict generator quality correlation analysis to six DiT-based modelscross-generator detectors · Cross-validation (DiT models)Pearson correlation: 0.86 -> 0.92Improved
evaluate at a resolution different from that used in trainingvideo detectors · cross-resolution evaluationdetection performanceHurt

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