Unsupervised multi-view stereo network based on multi-stage depth estimation

作者：

Highlights：

• We propose a novel unsupervised MVS network for dense 3D reconstruction.

• A multi-view group-wise correlation module is proposed to improve the performance.

• The multi-stage architecture is introduced to increase the depth map resolution.

• The experiment results demonstrate the potential of the unsupervised MVS method.

摘要

•We propose a novel unsupervised MVS network for dense 3D reconstruction.•A multi-view group-wise correlation module is proposed to improve the performance.•The multi-stage architecture is introduced to increase the depth map resolution.•The experiment results demonstrate the potential of the unsupervised MVS method.

论文关键词：3D reconstruction,Depth estimation,Multi-view stereo

论文评审过程：Received 13 February 2022, Revised 30 March 2022, Accepted 4 April 2022, Available online 7 April 2022, Version of Record 26 April 2022.

论文官网地址：https://doi.org/10.1016/j.imavis.2022.104449