VGGT-SLAM: Dense RGB SLAM Optimized on the SL(4) Manifold

Unified Codebase for Advanced World Models.

VGGT-X: When VGGT Meets Dense Novel View Synthesis

Track4World: Feedforward World-centric Dense 3D Tracking of All Pixels

This is a list of relevant papers for 3D Geometric Foundation Models and Applications.

VGGT 3D Vision Agent optimized for Apple Silicon with Metal Performance Shaders

😎 A curated list of CVPR 2025 Oral paper. Total 96

Optimizing Monocular Depth Estimation with TensorRT: Model Conversion, Inference Acceleration, and 3D Reconstruction

🔥RayletDF in PyTorch (ICCV 2025 Highlight)

Code for "Improving Robotic Manipulation with Efficient Geometry-Aware Vision Encoder"

Large-scale accurate indoor gaussian splatting 3D reconstruction based on video stream

🛠️ 3D-Assembly-Vision: Real-time 3D monitoring for industrial assembly. Powered by YOLOv11, DOPE, VGGT and LSTM for synchronized 6D pose estimation and scene reconstruction.

🍎 Accelerate 3D reconstructions from images using VGGT, optimized for Apple Silicon with Metal Performance Shaders for fast and efficient processing.

VGGT Global Attention Map Extractor for the paper "OracleGS: Grounding Generative Priors for Sparse-View Gaussian Splatting"

[WACV 2026 Oral] Reference Implementation of the paper "OracleGS: Grounding Generative Priors for Sparse-View Gaussian Splatting"

Implement cross-modal coarse-to-fine fusion for accurate, consistent, and metric-scale dense 3D mapping using LiDAR data.

Add skymask function for VGGT

Tooth-NVS uses an end-to-end feedforward model to process tooth/oral videos: extracting frames into temporal RGB sequences, leveraging the vggt-x 3D geometric base model for parameter extraction and point cloud generation, jointly optimizing via 3DGS for NVS, with ongoing camera-parameter-free Nope-Nerf optimization (results pending).

This repository acquires continuous frame sequences through cameras mounted on drones and uses an end-to-end feedforward deep learning architecture to directly regress three-dimensional geometric representations from two-dimensional images.

Photos, robotics logs, and external SLAM outputs to browser-viewable 3D Gaussian Splats