AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

TTP: Test-Time Padding for Adversarial Detection and Robust Adaptation on Vision-Language Models

Zhiwei Li, Yitian Pang, Weining Wang, Zhenan Sun, Qi Li

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Test-time adaptation of VLMs · 1 citations (OpenAlex)

Read the paper

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 test-time padding defense (TTP) over R-TPTCLIP ViT-B/32 · 8 fine-grained datasets, PGD eps=4.0robust accuracy (%): 35.3 -> 39.7Improved
TTP vs R-TPT on larger backboneCLIP ViT-L/14 · 8 fine-grained datasets, PGDrobust accuracy (%): 49.6 -> 51.6Improved
TTP applied under CW, DeepFool and FGSM attacksCLIP ViT-B/32 · Flowers and DTDrobust accuracy (%): 51.8 -> 54.1 (Flowers avg); 34.2 -> 38.7 (DTD avg)Improved
detect-then-adapt keeps clean inputs unchangedTTP · avg clean accuracy ViT-B/32clean accuracy (%): 57.4 -> 57.1No clear effect
add TPT on inputs detected as cleanTTP · avg clean accuracy ViT-B/32clean accuracy (%): 57.1 -> 57.9Improved
use all-black or all-white padding instead of random padding for detection (ablation)TTP detector · detection accuracy ViT-B/32detection accuracy (%): 95.8 -> 98.5 (0), 98.7 (255)Improved
add fixed spatial padding to CLIP (ablation)TTP · 8 fine-grained datasets, PGDrobust accuracy (%): 0.0 -> 37.5 (B/32); 0.0 -> 38.0 (B/16)Improved
add entropy minimization training of padding (ablation)TTP · 8 fine-grained datasets, PGDrobust accuracy (%): 38.3 -> 39.7 (B/32); 39.5 -> 42.9 (B/16)Improved
add similarity-aware ensemble (ablation)TTP · 8 fine-grained datasets, PGDrobust accuracy (%): 39.0 -> 39.7 (B/32); 40.8 -> 42.9 (B/16)Improved
increase padding size from 8 to 64 then to 256 (ablation)TTP · DTD robust accuracy ViT-B/32robust accuracy (%): 19.6 -> 33.9 (size 64) -> 21.9 (size 256)Mixed
use TTC detect-then-defend instead of TTP detection under strong attackTTC · 8 fine-grained datasets, PGD eps=4.0 ViT-B/32robust accuracy (%): 39.7 -> 6.8Hurt

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