🏢 Fudan University
EMR-Merging: Tuning-Free High-Performance Model Merging
·3173 words·15 mins·
loading
·
loading
Image Classification
🏢 Fudan University
EMR-MERGING: A tuning-free model merging technique achieves high performance by electing a unified model and generating lightweight task-specific modulators, eliminating the need for additional data …
Efficient Combinatorial Optimization via Heat Diffusion
·2280 words·11 mins·
loading
·
loading
AI Theory
Optimization
🏢 Fudan University
Heat Diffusion Optimization (HeO) framework efficiently solves combinatorial optimization problems by enabling information propagation through heat diffusion, outperforming existing methods.
Disentangled Style Domain for Implicit $z$-Watermark Towards Copyright Protection
·1999 words·10 mins·
loading
·
loading
Computer Vision
Image Generation
🏢 Fudan University
This paper introduces a novel implicit Zero-Watermarking scheme using disentangled style domains to detect unauthorized dataset usage in text-to-image models, offering robust copyright protection via …
DG-SLAM: Robust Dynamic Gaussian Splatting SLAM with Hybrid Pose Optimization
·1723 words·9 mins·
loading
·
loading
AI Applications
Robotics
🏢 Fudan University
DG-SLAM achieves robust real-time visual SLAM in dynamic scenes using 3D Gaussian splatting and a novel hybrid pose optimization, significantly outperforming existing methods.
CSPG: Crossing Sparse Proximity Graphs for Approximate Nearest Neighbor Search
·2426 words·12 mins·
loading
·
loading
AI Theory
Optimization
🏢 Fudan University
CSPG: a novel framework boosting Approximate Nearest Neighbor Search speed by 1.5-2x, using sparse proximity graphs and efficient two-staged search.
Automated Label Unification for Multi-Dataset Semantic Segmentation with GNNs
·3190 words·15 mins·
loading
·
loading
AI Generated
Computer Vision
Image Segmentation
🏢 Fudan University
GNNs automate multi-dataset semantic segmentation label unification, improving model training efficiency and performance by resolving conflicts across label spaces.
3DET-Mamba: Causal Sequence Modelling for End-to-End 3D Object Detection
·1690 words·8 mins·
loading
·
loading
Computer Vision
3D Vision
🏢 Fudan University
3DET-Mamba: A novel end-to-end 3D object detector leveraging the Mamba state space model for efficient and accurate object detection in complex indoor scenes, outperforming previous 3DETR models.