Zilei Wang

33

Papers

61

Total Citations

Papers (33)

Task Preference Optimization: Improving Multimodal Large Language Models with Vision Task Alignment

R-TPT: Improving Adversarial Robustness of Vision-Language Models through Test-Time Prompt Tuning

ClearSight: Visual Signal Enhancement for Object Hallucination Mitigation in Multimodal Large Language Models

A Dynamic Learning Method towards Realistic Compositional Zero-Shot Learning

GCD: Advancing Vision-Language Models for Incremental Object Detection via Global Alignment and Correspondence Distillation

Target Semantics Clustering via Text Representations for Robust Universal Domain Adaptation

Semantic-guided Robustness Tuning for Few-Shot Transfer Across Extreme Domain Shift

Infer from What You Have Seen Before: Temporally-dependent Classifier for Semi-supervised Video Segmentation

Connecting the Dots: Collaborative Fine-tuning for Black-Box Vision-Language Models

Highway Vehicle Counting in Compressed Domain

Learning a Unified Classifier Incrementally via Rebalancing

Meta-SR: A Magnification-Arbitrary Network for Super-Resolution

RPN Prototype Alignment for Domain Adaptive Object Detector

Semi-Supervised Video Semantic Segmentation With Inter-Frame Feature Reconstruction

SimpleNet: A Simple Network for Image Anomaly Detection and Localization

Boundary-Enhanced Co-Training for Weakly Supervised Semantic Segmentation

Class Relationship Embedded Learning for Source-Free Unsupervised Domain Adaptation

Look and Think Twice: Capturing Top-Down Visual Attention With Feedback Convolutional Neural Networks

DualNet: Learn Complementary Features for Image Recognition

VegFru: A Domain-Specific Dataset for Fine-Grained Visual Categorization

Revisiting Foreground and Background Separation in Weakly-supervised Temporal Action Localization: A Clustering-based Approach

Preparing the Future for Continual Semantic Segmentation

Towards Effective Instance Discrimination Contrastive Loss for Unsupervised Domain Adaptation

Polynomial Regression Network for Variable-Number Lane Detection

Collaborating Domain-Shared and Target-Specific Feature Clustering for Cross-Domain 3D Action Recognition

Cross-Domain Cross-Set Few-Shot Learning via Learning Compact and Aligned Representations

Continual Semantic Segmentation via Structure Preserving and Projected Feature Alignment

Lifting the Veil on Visual Information Flow in MLLMs: Unlocking Pathways to Faster Inference

Towards Continual Universal Segmentation

Progressive Distribution Bridging: Unsupervised Adaptation for Large-scale Pre-trained Models via Adaptive Auxiliary Data

Cooperative Pseudo Labeling for Unsupervised Federated Classification

Exploring Vacant Classes in Label-Skewed Federated Learning

Protecting Model Adaptation from Trojans in the Unlabeled Data