Xu Zheng

FERD: Fairness-Enhanced Data-Free Adversarial Robustness Distillation

Greedy Distill: Efficient Video Generative Modeling with Linear Time Complexity

Rethinking Expressivity and Degradation-Awareness in Attention for All-in-One Blind Image Restoration

EgoNight: Towards Egocentric Vision Understanding at Night with a Challenging Benchmark

DiMeR: Disentangled Mesh Reconstruction Model with Normal-only Geometry Training

Efficient Degradation-agnostic Image Restoration via Channel-Wise Functional Decomposition and Manifold Regularization

PhysToolBench: Benchmarking Physical Tool Understanding for MLLMs

Learning Robust Anymodal Segmentor with Unimodal and Cross-modal Distillation

Understanding-in-Generation: Reinforcing Generative Capability of Unified Model via Infusing Understanding into Generation

Consistent3DGen: Bridging Stochastic Generation and Deterministic Reconstruction for Image-to-3D Diffusion Models

E$^2$AT: Multimodal Jailbreak Defense via Dynamic Joint Optimization

Towards Omnidirectional Reasoning: A Dataset, Benchmark, and GRPO-based Method

Don't Just Chase “Highlighted Tokens” in MLLMs: Revisiting Visual Holistic Context Retention

Domain-RAG: Retrieval-Guided Compositional Image Generation for Cross-Domain Few-Shot Object Detection

RealRAG: Retrieval-augmented Realistic Image Generation via Self-reflective Contrastive Learning

CLIP-to-Seg Distillation for Inductive Zero-shot Semantic Segmentation

Semantic-Centric Alignment for Zero-shot Panoptic and Semantic Segmentation