FLEG: Feed-Forward Language Embedded Gaussian Splatting from Any Views via Compact Semantic Representation
cs.CV
Submitted: 2025-12-19
Updated: 2026-11-01
Project page: https://fangzhou2000.github.io/projects/fleg
Terminology
Sources
- MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images
- AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views
- Language-driven Semantic Segmentation
- IGGT: Instance-Grounded Geometry Transformer for Semantic 3D Reconstruction
- SemanticSplat: Feed-Forward 3D Scene Understanding with Language-Aware Gaussian Fields
- LangSplatV2: High-dimensional 3D Language Gaussian Splatting with 450+ FPS
- Depth Anything 3: Recovering the Visual Space from Any Views
- LangScene-X: Reconstruct Generalizable 3D Language-Embedded Scenes with TriMap Video Diffusion
- DINOv2: Learning Robust Visual Features without Supervision
- SpatialSplat: Efficient Semantic 3D from Sparse Unposed Images
- Splatt3R: Zero-shot Gaussian Splatting from Uncalibrated Image Pairs
- The Replica Dataset: A Digital Replica of Indoor Spaces
- Uni3R: Unified 3D Reconstruction and Semantic Understanding via Generalizable Gaussian Splatting from Unposed Multi-View Images
- UniForward: Unified 3D Scene and Semantic Field Reconstruction via Feed-Forward Gaussian Splatting from Only Sparse-View Images
- GSemSplat: Generalizable Semantic 3D Gaussian Splatting from Uncalibrated Image Pairs
- No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Images
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs