Guangming Lu

PointRePar : SpatioTemporal Point Relation Parsing for Robust Category-Unified 3D Tracking

Prune Redundancy, Preserve Essence: Vision Token Compression in VLMs via Synergistic Importance-Diversity

DiffTrans: Differentiable Geometry-Materials Decomposition for Reconstructing Transparent Objects

DiffNCL: Diffusion-Driven Weakly-Noisy Correspondence Learning

Video-ToC: Video Tree-of-Cue Reasoning

A Unified Framework for Multimodal Secret Data Steganography

Amodal SAM: Open-World Amodal Segmentation

General Cross-Attack Backdoor Detector Based on Disturbance Immunity of Triggers

EditInfinity: Image Editing with Binary-Quantized Generative Models

A Set of Generalized Components to Achieve Effective Poison-only Clean-label Backdoor Attacks with Collaborative Sample Selection and Triggers

Efficient and Separate Authentication Image Steganography Network

SSHR: More Secure Generative Steganography with High-Quality Revealed Secret Images

SR$^2$: BOOSTING 3D LARGE LANGUAGE MODEL WITH SPATIAL RELATION REASONING