Paper Image

Neural indoor surface reconstruction from images

Published on:

18 September 2023

Primary Category:

Computer Vision and Pattern Recognition

Paper Authors:

Xinyi Yu,

Liqin Lu,

Jintao Rong,

Guangkai Xu,

Linlin Ou

Bullets

Key Details

Represents indoor scene geometry using implicit neural networks

Proposes a two-stage training process with novel constraints

Introduces mask-guided consistency losses for robust optimization

Achieves high-quality reconstruction of complex geometric details

Outperforms previous state-of-the-art methods on indoor datasets

AI generated summary

Neural indoor surface reconstruction from images

This paper proposes a neural network method to reconstruct high-quality 3D models of indoor scenes using only posed 2D images as input. It represents the 3D surface implicitly using neural networks, and trains them using novel consistency constraints and masking techniques to achieve robust and detailed reconstructions.

Answers from this paper

Comments

No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up