Presentation
30 May 2022 AI-powered 3D shape measurement through a phase-to-phase network
Author Affiliations +
Abstract
Recently, researchers have found that the power of deep learning can be leveraged to perform high quality 3D shape measurement by directly learning from a single-shot fringe image. However, such end-to-end fringe-to-depth learning has limited flexibility given that its trained deep neural network can only be used for patterns with a certain frequency. This research proposes a phase-to-phase learning approach to address such limitation. By establishing a phase-to-phase training network from phase obtained from Fourier transform to phase obtained from phase shifting, this proposed network can be flexibly applied to measurements with fringe images of different pattern frequencies.
Conference Presentation
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Beiwen Li and Vignesh Suresh "AI-powered 3D shape measurement through a phase-to-phase network", Proc. SPIE PC12098, Dimensional Optical Metrology and Inspection for Practical Applications XI, PC1209804 (30 May 2022); https://doi.org/10.1117/12.2622761
Advertisement
Advertisement
KEYWORDS
3D metrology

Phase measurement

3D image processing

3D image reconstruction

Phase shifting

Fourier transforms

Machine learning

Back to Top