activity
20242026
collaborators

6 papers

cs.CV2026

MedLoc-R1: Performance-Aware Curriculum Reward Scheduling for GRPO-Based Medical Visual Grounding

Guangjing Yang, Ziyuan Qin, Chaoran Zhang +6

Medical visual grounding serves as a crucial foundation for fine-grained multimodal reasoning and interpretable clinical decision support. Despite recent advances in reinforcement…

cs.CV2025

iDPA: Instance Decoupled Prompt Attention for Incremental Medical Object Detection

Huahui Yi, Wei Xu, Ziyuan Qin +4

Existing prompt-based approaches have demonstrated impressive performance in continual learning, leveraging pre-trained large-scale models for classification tasks; however, the ti…

cs.CV2025

Guiding Medical Vision-Language Models with Explicit Visual Prompts: Framework Design and Comprehensive Exploration of Prompt Variations

Kangyu Zhu, Ziyuan Qin, Huahui Yi +4

While mainstream vision-language models (VLMs) have advanced rapidly in understanding image level information, they still lack the ability to focus on specific areas designated by…

cs.CV2024

Evaluating Hallucination in Text-to-Image Diffusion Models with Scene-Graph based Question-Answering Agent

Ziyuan Qin, Dongjie Cheng, Haoyu Wang +5

Contemporary Text-to-Image (T2I) models frequently depend on qualitative human evaluations to assess the consistency between synthesized images and the text prompts. There is a dem…

eess.SP2024

EEG-DIF: Early Warning of Epileptic Seizures through Generative Diffusion Model-based Multi-channel EEG Signals Forecasting

Zekun Jiang, Wei Dai, Qu Wei +3

Multi-channel EEG signals are commonly used for the diagnosis and assessment of diseases such as epilepsy. Currently, various EEG diagnostic algorithms based on deep learning have…

cs.CV2024

TV-SAM: Increasing Zero-Shot Segmentation Performance on Multimodal Medical Images Using GPT-4 Generated Descriptive Prompts Without Human Annotation

Zekun Jiang, Dongjie Cheng, Ziyuan Qin +10

This study presents a novel multimodal medical image zero-shot segmentation algorithm named the text-visual-prompt segment anything model (TV-SAM) without any manual annotations. T…