Fast 3D Foundation Model Initialized Gaussian Splatting
This work enables rapid 3D reconstruction for near real-time applications in robotics, VR, and autonomous navigation by eliminating the time-consuming SfM step.
The paper presents a fast 3D Gaussian Splatting method that bypasses traditional SfM by using 3D Foundation Models for initialization, achieving competitive quality (23.61 dB PSNR, 0.19 LPIPS) in about three minutes per scene.
This paper introduces a fast method for high-quality 3D Gaussian Splatting (3DGS) reconstruction without traditional Structure-from-Motion (SfM). The proposed approach leverages 3D Foundation Models (3DFMs) for camera pose and point-cloud initialization, then jointly optimizes both camera poses and Gaussian primitives using a depth-guided loss function. This enables fast convergence even from rough initialization with as few as 50-60 input views. To further improve reconstruction quality in sparse-view scenarios, an MLP-based pose refinement module is introduced alongside depth-guided supervision from the foundation model. Extensive experiments on Mip-NeRF 360, Tanks and Temples, and RobustNeRF demonstrate that the proposed method achieves competitive reconstruction quality (23.61 dB PSNR, 0.19 LPIPS) while reducing training time to approximately three minutes per scene. The proposed method produces ready-to-use 3DGS models at a fraction of the time required by existing pipelines, making it suitable for near real-time applications in robotics, VR, and autonomous navigation.