Advanced Fringe Patterns Denoising and Processing via Ensemble Deep Learning Model

Abstract

The strength of deep learning methods has shown substantial achievements in optical metrology, especially in fringe denoising, fringe analysis, and phase unwrapping. Despite extensive research efforts for decades, the challenge of accurately extracting desired phase distribution information from recorded fringes remains one of the most challenging open problems. This study introduces an ensemble methodology that leverages two deep neural models, such as Single Shot Detector (SSD), and You Only Look Once (YOLO), for automated fringe pattern analysis in decision-making support for the determination of parameters from the calculated phase distribution. The assessment of the proposed methodology, as detailed in this study, was conducted using heterogeneous datasets encompassing data recorded in Kösters interferometer placed in the laboratory of the Polish Central Office of Measures and computer-generated fringe sample images. We benchmark our methodology for the fringe analysis task and find generalization behavior and robustness to noisy recorded data.

Bibliography

[1] Eckle K. and Schmidt-Hieber J. A comparison of deep networks with relu activation function and linear spline-type methods. Neural Networks, 110:232–242, 2019.
[2] Feng S., Chen Q., Gu G., Tao T., Zhang L., Hu Y., Yin W., and Zuo C. Fringe pattern analysis using deep learning. Advanced photonics, 1(2):0250 01–025001, 2019.
[3] Feng S., Zhang L., Zuo C., Tao T., Chen Q., and Gu G. High dynamic range 3d measurements with fringe projection profilometry: a review. Measurement Science and Technology, 29(12):1220 01, 2018.
[4] Girshick R. Fast r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 1440–1448, 2015.
[5] Harding K. Handbook of optical dimensional metrology. CRC Press, 2013.
[6] Jackson N. Principles of interferometry. Jets from Young Stars II: Clues from High Angular Resolution Observations, pages 193–218, 2008.
[7] Kim E.K., Lee H., Kim J.Y., and Kim S. Data augmentation method by applying color perturbation of inverse psnr and geometric transformations for object recognition based on deep learning. Applied Sciences, 10(11):3755, 2 020.
[8] Liu H., Yan N., Shao B., Yuan S., and Zhang X. Deep learning in fringe projection: a review. Neurocomputing, page 127493, 2024.
[9] Liu W., Anguelov D., Erhan D., Szegedy C., Reed S., Fu C.-Y., and Berg A.C. Ssd: Single shot multibox detector. In Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14, pages 21–37. Springer, 2016.
[10] Michalecki G. Automatic calibration of gauge blocks measured by optical interferometry. Science Review, 1:15–17, 2001.
[11] Pandey D., Ramaiah J., Ajithaprasad S., and Gannavarpu R. Subspace analysis based machine learning method for automated defect detection from fringe patterns. Optik, 270:170 026, 2022.
[12] Postalcıoğlu S. Performance analysis of different optimizers for deep learning-based image recognition. International Journal of Pattern Recognition and Artificial Intelligence, 34(02):20510 03, 2020.
[13] Ren S., He K., Girshick R., and Sun J. Faster r-cnn: Towards real-time object detection with region proposal networks. Advances in neural information processing systems, 28, 2015.
[14] Roy A.M., Bose R., and Bhaduri J. A fast accurate fine-grain object detection model based on yolov4 deep neural network. Neural Computing and Applications, 3 4(5):3895–3921, 2022.
[15] Schnars U., Falldorf C., Watson J., Jüptner W. Digital holography. Springer, 2015.
[16] Servin M., Quiroga J.A., Padilla J.M., et al. Fringe pattern analysis for optical metrology. Wiley Online Library, 2023.
[17] Vishnoi A., Madipadaga A., Ajithaprasad S., and Gannavarpu R. Automated defect identification from carrier fringe patterns using wignerville distribution and a machine learning-based met od. Applied Optics, 6 0(15):4391–4397, 2021.
[18] Wang C.-Y., Bochkovskiy A., and Liao H.-Y. M. Scaled-yolov4: Scaling cross stage partial network. In Proceedings of the IEEE/cvf conference on computer vision and pattern recognition, pages 13029–13038, 2021.
[19] Wang X. and Song J. Iciou: Improved loss based on complete intersection over union for bounding box regression. IEEE Access, 9:105686–105695, 2021.
[20] Wengierow M., Salbut L., Ramotowski Z., and Szumski R. Measurement system based on multi-wavelength interferometry for long gauge block calibration. Metrology Journal, 15:234–245, 2020.
[21] Xiao Y., Decencière E., Velasco-Forero S., Burdin H., Bornschlögl T., Bernerd F., Warrick E., and Baldeweck T. A new color augmentation method for deep learning segmentation of histological images. In 2019 IEEE 16th international symposium on biomedical imaging (ISBI 2019), pages 886–890. IEEE, 2019.
[22] Zhang J. Xu and S. Status, challenges, and future perspectives of fringe projection profilometry. Optics and Lasers in Engineering, 135:106193, 2020.
[23] Zhang Z. Improved adam optimizer for deep neural networks. In 2018 IEEE/ACM 26th international symposium on quality of service (IWQoS), pages 1–2. Ieee, 2018.
[24] Zhu Y., Liu L., Luan Z., and Sun J. Discussions about fft-based two-step phase-shifting algorithm. Optik, 119(9):424–428, 2008.
[25] Zuo C., Feng S., Huang L., Tao T., Yin W., and Chen Q. Phase shifting algorithms for fringe projection profilometry: A review. Optics and lasers in engineering, 109:23–59, 2018.

Metrology & Hallmark Central Office of Measures

Access panel

Search

Current volume