Publications (14)
Training-free Motion Factorization for Compositional Video Generation
Zixuan Wang, Ziqin Zhou, Feng Chen +4
Compositional video generation aims to synthesize multiple instances with diverse appearance and motion. However, current approaches mainly focus on binding semantics, neglecting t…
VisualRouter: Query-Grounded Visual Sampling for Long Video Understanding
Haiyue Zhang, Yi Bin, Xun Jiang +5
VisualRouter is a training-free, plug‑and‑play framework that classifies queries as global or local and applies tailored visual sampling strategies to select informative frames, im…
Unified Prompt Attack Against Text-to-Image Generation Models
Duo Peng, Qiuhong Ke, Mark He Huang +2
Text-to-Image (T2I) models have advanced significantly, but their growing popularity raises security concerns due to their potential to generate harmful images. To address these is…
UPAM: Unified Prompt Attack in Text-to-Image Generation Models Against Both Textual Filters and Visual Checkers
Duo Peng, Qiuhong Ke, Jun Liu
Text-to-Image (T2I) models have raised security concerns due to their potential to generate inappropriate or harmful images. In this paper, we propose UPAM, a novel framework that…
Unsupervised Domain Adaptation via Domain-Adaptive Diffusion
Duo Peng, Qiuhong Ke, Yinjie Lei +1
Unsupervised Domain Adaptation (UDA) is quite challenging due to the large distribution discrepancy between the source domain and the target domain. Inspired by diffusion models wh…
PaperSearchQA: Learning to Search and Reason over Scientific Papers with RLVR
James Burgess, Jan N. Hansen, Duo Peng +5
Search agents are language models (LMs) that reason and search knowledge bases (or the web) to answer questions; recent methods supervise only the final answer accuracy using reinf…
Diffusion-based Image Translation with Label Guidance for Domain Adaptive Semantic Segmentation
Duo Peng, Ping Hu, Qiuhong Ke +1
Translating images from a source domain to a target domain for learning target models is one of the most common strategies in domain adaptive semantic segmentation (DASS). However,…
Global and Local Texture Randomization for Synthetic-to-Real Semantic Segmentation
Duo Peng, Yinjie Lei, Lingqiao Liu +2
Semantic segmentation is a crucial image understanding task, where each pixel of image is categorized into a corresponding label. Since the pixel-wise labeling for ground-truth is…
Hierarchical Paired Channel Fusion Network for Street Scene Change Detection
Yinjie Lei, Duo Peng, Pingping Zhang +2
Street Scene Change Detection (SSCD) aims to locate the changed regions between a given street-view image pair captured at different times, which is an important yet challenging ta…
Semantic-Aware Domain Generalized Segmentation
Duo Peng, Yinjie Lei, Munawar Hayat +2
Deep models trained on source domain lack generalization when evaluated on unseen target domains with different data distributions. The problem becomes even more pronounced when we…
Visual Prompting for One-shot Controllable Video Editing without Inversion
Zhengbo Zhang, Yuxi Zhou, Duo Peng +4
One-shot controllable video editing (OCVE) is an important yet challenging task, aiming to propagate user edits that are made -- using any image editing tool -- on the first frame…
Training-free Dense-Aligned Diffusion Guidance for Modular Conditional Image Synthesis
Zixuan Wang, Duo Peng, Feng Chen +2
Conditional image synthesis is a crucial task with broad applications, such as artistic creation and virtual reality. However, current generative methods are often task-oriented wi…
Diff-Tracker: Text-to-Image Diffusion Models are Unsupervised Trackers
Zhengbo Zhang, Li Xu, Duo Peng +2
We introduce Diff-Tracker, a novel approach for the challenging unsupervised visual tracking task leveraging the pre-trained text-to-image diffusion model. Our main idea is to leve…
Sparse-to-dense Feature Matching: Intra and Inter domain Cross-modal Learning in Domain Adaptation for 3D Semantic Segmentation
Duo Peng, Yinjie Lei, Wen Li +2
Domain adaptation is critical for success when confronting with the lack of annotations in a new domain. As the huge time consumption of labeling process on 3D point cloud, domain…