Learning-based 3D imaging from single structured-light image

• Published in: Graphical models Vol. 126; p. 101171
• Main Authors: Nguyen, Andrew-Hieu, Rees, Olivia, Wang, Zhaoyang
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.04.2023; Elsevier
• Subjects: Deep learning; Fringe-to-phase transformation; Single-shot imaging; Structured light; Three-dimensional image acquisition; Three-dimensional sensing; Three-dimensional image acquisition; Deep learning; Structured light; Single-shot imaging; Fringe-to-phase transformation; Three-dimensional sensing
• ISSN: 1524-0703; 1524-0711
• DOI: 10.1016/j.gmod.2023.101171

Integrating structured-light technique with deep learning for single-shot 3D imaging has recently gained enormous attention due to its unprecedented robustness. This paper presents an innovative technique of supervised learning-based 3D imaging from a single grayscale structured-light image. The proposed approach uses a single-input, double-output convolutional neural network to transform a regular fringe-pattern image into two intermediate quantities which facilitate the subsequent 3D image reconstruction with high accuracy. A few experiments have been conducted to demonstrate the validity and robustness of the proposed technique. [Display omitted] •Integration of structured-light technique with convolutional neural network allows accurate single-shot 3D imaging.•Convolutional neural network transforms a single image into four phase-shifted images and a map of integer fringe orders.•Intermediate quantities produced by the network help the 3D imaging scheme yield higher accuracy than existing techniques.•The proposed technique provides a faster speed than classic accurate 3D imaging techniques.