13 papers
InfraPatch: Cross-Task Targeted Grayscale Patch Attacks on Infrared-Adapted Vision-Language Models
Chengyin Hu, Dingyi Lu, Jiaju Han +5
Infrared vision-language models (IR-VLMs) have emerged as a promising paradigm for multimodal perception under low-visibility conditions, yet their robustness to targeted adversari…
Radar4D-VLM: Proposal-Grounded Temporal 4D Radar Reasoning Across Frozen Language Models
Jiaju Han, Xuemeng Sun, Qike Zhang +6
Vision-language models for autonomous driving primarily rely on cameras and LiDAR, leaving 4D radar largely unexplored as a standalone perceptual modality despite its robustness to…
QR-Structured Thermal Triggers for Targeted Semantic Attacks on Infrared Vision-Language Models
Xiang Chen, Yingying Zhao, Chao Li +7
Infrared vision-language models (IR-VLMs) extend thermal perception to open-vocabulary classification, image captioning, and visual question answering. However, their robustness to…
AirflowAttack: Thermal-Airflow Adversarial Perturbations against Infrared Remote-Sensing Vision-Language Models
Cong Su, Jiaju Han, Xuemeng Sun +5
Vision-language models (VLMs) are increasingly deployed on infrared (IR) remote sensing imagery in security-critical settings, yet their adversarial robustness remains unexamined.…
FusionRS: A Large-Scale RGB-Infrared-Style Remote Sensing Dataset for Cross-Modal Vision-Language Learning
Jiaju Han, Ben Zhang, Xuemeng Sun +8
Remote sensing vision-language models have advanced Earth observation, but available large-scale vision-language resources remain RGB-centered, leaving complementary infrared infor…
Exposing Vulnerabilities in Visible-Infrared VLMs: A Unified Geometric Adversarial Framework with Cross-Task Transferability
Xiang Chen, Yuxian Dong, Chao Li +6
Vision-language models (VLMs) have achieved strong performance across diverse multimodal tasks, but their adversarial robustness in visible-infrared (VIS-IR) scenarios remains unde…
From Clouds to Hallucinations: Atmospheric Retrieval Hijacking in Remote Sensing Vision-Language RAG
Jiaju Han, Chao Li, Chengyin Hu +8
Multimodal RAG systems increasingly rely on vision-language retrievers to ground visual queries in external textual evidence. Existing adversarial studies on RAG mainly manipulate…
Challenging Vision-Language Models with Physically Deployable Multimodal Semantic Lighting Attacks
Yingying Zhao, Chengyin Hu, Qike Zhang +7
Vision-Language Models (VLMs) have shown remarkable performance, yet their security remains insufficiently understood. Existing adversarial studies focus almost exclusively on the…
CoDA: Exploring Chain-of-Distribution Attacks and Post-Hoc Token-Space Repair for Medical Vision-Language Models
Xiang Chen, Fangfang Yang, Chunlei Meng +6
Medical vision--language models (MVLMs) are increasingly used as perceptual backbones in radiology pipelines and as the visual front end of multimodal assistants, yet their reliabi…
Revealing Physical-World Semantic Vulnerabilities: Universal Adversarial Patch for Infrared Vision-Language Models
Chengyin Hu, Yuxian Dong, Yikun Guo +8
Infrared vision-language models (IR-VLMs) are becoming important for semantic perception in low-visibility environments, yet their robustness to physical semantic attacks remains u…
XSPA: Crafting Imperceptible X-Shaped Sparse Adversarial Perturbations for Transferable Attacks on VLMs
Chengyin Hu, Jiaju Han, Xuemeng Sun +6
Vision-language models (VLMs) share visual-textual representations across zero-shot classification, image captioning, and visual question answering (VQA), creating a pathway throug…
Thermal Topology Collapse: Universal Physical Patch Attacks on Infrared Vision Systems
Chengyin Hu, Yikun Guo, Yuxian Dong +7
Infrared pedestrian detectors are increasingly deployed in all-weather perception systems, but their robustness against physical adversarial attacks remains insufficiently understo…
A Semantic Decoupling-Based Two-Stage Rainy-Day Attack for Revealing Weather Robustness Deficiencies in Vision-Language Models
Chengyin Hu, Xiang Chen, Zhe Jia +4
Vision-Language Models (VLMs) are trained on image-text pairs collected under canonical visual conditions and achieve strong performance on multimodal tasks. However, their robustn…