A deep-shallow and global–local multi-feature fusion network for photometric stereo

作者：

Highlights：

• We propose a photometric stereo network that integrates global–local features and deep-shallow features, namely MF-PSN.

• Extensive ablation studies demonstrate the effectiveness of the architecture of the proposed MF-PSN.

• Compared to other photometric stereo networks, the robustness of the MF-PSN is enhanced.

• Extensive experiments show the state-of-the-art performance of the proposed MF-PSN.

摘要

•We propose a photometric stereo network that integrates global–local features and deep-shallow features, namely MF-PSN.•Extensive ablation studies demonstrate the effectiveness of the architecture of the proposed MF-PSN.•Compared to other photometric stereo networks, the robustness of the MF-PSN is enhanced.•Extensive experiments show the state-of-the-art performance of the proposed MF-PSN.

论文关键词：Photometric stereo,3D reconstruction,Deep neural networks,Convolutional neural network,Multi-feature fusion

论文评审过程：Received 18 October 2021, Revised 13 December 2021, Accepted 19 December 2021, Available online 24 December 2021, Version of Record 4 January 2022.

论文官网地址：https://doi.org/10.1016/j.imavis.2021.104368