Rapid Reconstruction of Real-World 3D Models for 3D Printing Using NeRF

This study introduces the use of Neural Radiance Fields (NeRF) for efficient and accurate 3D model reconstruction, reducing costs. Images are captured with a smartphone and processed using incremental Structure-From-Motion (SfM) to determine camera poses. The Instant NGP algorithm reconstructs implicit models of 3D objects, achieving a 186-239x speed increase and over 1.5x fidelity improvement over traditional methods. NeRF 2 Mesh is then used to generate high-quality meshes, outperforming Colmap. Results show NeRF's practicality in 3D printing.