Deformable convolutions in multi-view stereo

作者：

Highlights：

• The use of the AGC in the cost volume, yield a reduction in GPU memory usage without noticeable penalties in the result.

• The use of deformable convolutions in the feature extraction network improved the network ability to model transformations.

摘要

•The use of the AGC in the cost volume, yield a reduction in GPU memory usage without noticeable penalties in the result.•The use of deformable convolutions in the feature extraction network improved the network ability to model transformations.

论文关键词：Multi-view stereo,Depth map,Deep learning

论文评审过程：Received 11 November 2021, Accepted 22 December 2021, Available online 27 December 2021, Version of Record 8 January 2022.

论文官网地址：https://doi.org/10.1016/j.imavis.2021.104369