AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Invisible Shortcuts: Why Vision Encoders Know Your Camera

Vladan Stojnic, Ryan Ramos, Giorgos Kordopatis-Zilos, Noa Garcia, Giorgos Tolias

ECCV 2026 · Detection and segmentation › Open-set and few-shot detection · approach: Object-level representation learning · 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
use ImageNet21k instead of ImageNet1k pretraining (stronger metadata-semantics correlation)timm models · IN1k/IN21k diagnosticsmetadata sensitivity (MP, SPD): Cramer's V 0.047 -> 0.067Hurt
increase JPEG-semantics correlation p_i during training (ablation)ResNet50 · IN1k diagnosticsSPD_p and MPHurt
increase JPEG-semantics correlation p_i during training (ablation)ResNet50 · IN1kSP accuracyHurt
use stronger acquisition-topic correlation in Re-LAION-2B subset (ablation)ResNet50 CLIP · Re-LAION-2B diagnosticsSPD: Cramer's V 0.255 -> 0.396Hurt
add color jitter, grayscale and blur augmentations (ablation)ResNet50 · IN1kSPD_p JPEG: 2.7 -> 1.4Improved
add color jitter, grayscale and blur augmentations (ablation)ResNet50 · IN1kSP accuracy: 78.2 -> 77.3No clear effect
add post-hoc adversarial linear layer (ablation)ResNet50 · IN1kSPD_p JPEG: 2.7 -> 0.8Improved
combine augmentations and post-hoc mitigation (ablation)ResNet50 · IN1kSPD_p JPEG: 2.7 -> 0.3Improved
combine augmentations and post-hoc mitigation (ablation)ResNet50 · IN1kSP accuracy: 78.2 -> 76.5Hurt
add Gaussian blur augmentation (ablation)ResNet50 · IN1kprocessing vs acquisition sensitivityMixed
train with strong JPEG-semantics correlation (ablation)ResNet50 · Generated image detectionavg accuracy: 57.3 -> 60.4Improved
apply post-hoc metadata mitigationfoundation models · ImageNet-C/R/SketchaccuracyMixed

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