6 papers
Pay More Attention To Text In High-Resolution MLLMs
Zhongkuan Mao, Wenzhuo Zhao, Xianjie Liu +7
Failures of high-resolution MLLMs are commonly attributed to a visual problem, motivating zooming, cropping, and related visual interventions to recover fine-grained evidence or su…
TrustRoboReward: Preference-Ordered Isotonic Score Editing for Multi-Paradigm Robot Reward Models
Yidong Wang, Yan Zhan, Ziteng Feng +16
Reward models are a bottleneck for reinforcement learning in embodied AI. Long-horizon robotic manipulation requires scalable vision feedback beyond handcrafted rewards or task-spe…
Is It Time for the Renaissance of Salient Object Detection in the Era of MLLMs?
Wenzhuo Zhao, Xiuzhi Li, Zhongkuan Mao +6
The zero-shot capabilities of multimodal large language models (MLLMs) are pushing salient object detection (SOD) beyond task-specific supervision. To disentangle MLLMs beyond conv…
Thinking Once Is Enough: Intermediate-Layer Evidence Routing for High-Resolution VQA
Zhongkuan Mao, Xianjie Liu, Tianyu Meng +9
High-resolution visual question answering (HR-VQA) is often treated as a problem of insufficient evidence acquisition, where failing multimodal large language models must inspect i…
From Stealthy Data Fabrication to Unsafe Driving: Realistic Scenario Attacks on Collaborative Perception
Qingzhao Zhang, Runting Zhang, Z. Morley Mao
Collaborative perception allows connected and autonomous vehicles (CAVs) to improve perception by sharing sensory data, but it also introduces security risks from manipulated input…
Camouflage-aware Image-Text Retrieval via Expert Collaboration
Yao Jiang, Zhongkuan Mao, Xuan Wu +2
Camouflaged scene understanding (CSU) has attracted significant attention due to its broad practical implications. However, in this field, robust image-text cross-modal alignment r…