Shengfeng He

24

Papers

174

Total Citations

Papers (24)

OneRestore: A Universal Restoration Framework for Composite Degradation

Delving into Multimodal Prompting for Fine-Grained Visual Classification

Rethinking Multi-view Representation Learning via Distilled Disentangling

NexusGS: Sparse View Synthesis with Epipolar Depth Priors in 3D Gaussian Splatting

RePOSE: 3D Human Pose Estimation via Spatio-Temporal Depth Relational Consistency

Learning with Unreliability: Fast Few-shot Voxel Radiance Fields with Relative Geometric Consistency

Cross-Subject Mind Decoding from Inaccurate Representations

Stable Score Distillation

Keep the Balance: A Parameter-Efficient Symmetrical Framework for RGB+X Semantic Segmentation

OmniVTON: Training-Free Universal Virtual Try-On

Instance-Level Video Depth in Groups Beyond Occlusions

ViewSRD: 3D Visual Grounding via Structured Multi-View Decomposition

Seeing 3D Through 2D Lenses: 3D Few-Shot Class-Incremental Learning via Cross-Modal Geometric Rectification

Dr. Tongue: Sign-Oriented Multi-label Detection for Remote Tongue Diagnosis

Occlusion-Insensitive Talking Head Video Generation via Facelet Compensation

PersonaMagic: Stage-Regulated High-Fidelity Face Customization with Tandem Equilibrium

Adversarial Attacks on Event-Based Pedestrian Detectors: A Physical Approach

Stroke2Sketch: Harnessing Stroke Attributes for Training-Free Sketch Generation

D3still: Decoupled Differential Distillation for Asymmetric Image Retrieval

MODfinity: Unsupervised Domain Adaptation with Multimodal Information Flow Intertwining

Drag Your Noise: Interactive Point-based Editing via Diffusion Semantic Propagation

Knowledge Bridger: Towards Training-Free Missing Modality Completion

Beyond Textual Constraints: Learning Novel Diffusion Conditions with Fewer Examples

Instruct2See: Learning to Remove Any Obstructions Across Distributions