AI Paper Atlas

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

From Masks to Pixels and Meaning: A New Taxonomy, Benchmark and Metrics for VLM Image Tampering

Xinyi Shang, Yi Tang, Jiacheng Cui, Ahmed Elhagry, Salwa Al Khatib, Sondos Bsharat, Jiacheng Liu, Xiaohan Zhao, Jing-Hao Xue, Hao Li, Salman Khan, Zhiqiang Shen

ECCV 2026 · Detection and segmentation › Open-set and few-shot detection · approach: Open-vocabulary 3D and world perception · 1 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
replace mask supervision with pixel-difference supervision (Lite)SIDA-7B · PIXAR test setIoU: 11.8 -> 15.0Improved
replace mask supervision with pixel-difference supervision (Lite)SIDA-13B · PIXAR test setIoU: 10.8 -> 15.4Improved
train on the full PIXAR training set instead of the mask-guided Lite subsetPIXAR-7B · PIXAR test setIoU: 15.0 -> 18.1Improved
scale model from 7B to 13BPIXAR · PIXAR test setIoU: 18.1 -> 19.3Improved
use PIXAR instead of FakeShieldFakeShield · PIXAR test setIoU: 9.3 -> 18.1Improved
use PIXAR-7B instead of SIDA-7B for real vs tampered detectionSIDA-7B · PIXAR test setOverall F1-score: 25.1 -> 69.7Improved
raise training tau from 0.05 to 0.1 (ablation)PIXAR-7B · PIXAR test set (eval tau=0.05)IoU: 18.1 -> 12.6Hurt
raise training tau from 0.05 to 0.2 (ablation)PIXAR-7B · PIXAR test set (eval tau=0.05)IoU: 18.1 -> 8.7Hurt
use tau=0.1 for both training and evaluation (ablation)PIXAR-7B · PIXAR test set (symmetric tau)IoU: 18.1 -> 12.0Hurt
add Dice loss (lambda_dice 1.0 versus 0.0) (ablation)PIXAR-7B · PIXAR test setIoU: 10.8 -> 18.1Improved
add Dice loss (lambda_dice 1.0 versus 0.0) (ablation)PIXAR-7B · PIXAR test setTop-1 accuracy: 35.3 -> 36.2Improved
set lambda_sem 0.5 instead of 0.1 or 1.0 (ablation)PIXAR-7B · PIXAR test setTop-1 accuracy: 35.1 -> 36.2Mixed

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