Skip to main content
SearchLogin or Signup

Fourier Ptychography Microscopy for Biosignature Detection using Deep Learning

Presentation #001.11 in the session “Future Missions, Instruments, and Facilities”.

Published onOct 26, 2020
Fourier Ptychography Microscopy for Biosignature Detection using Deep Learning

We present a proof-of-concept Fourier Ptychography Microscopy (FPM) system set-up on the benchtop, Fig. 1, to perform wide Field-of-View (FoV), high spatial resolution imaging (<1 µm), for biosignature detection and motility in liquid samples. A proof-of concept FPM used a Raspberry Pi V2.1 (without IR filter) camera module (8 megapixels, 1.12µm pixel size) with a 3-mm focal-length lens (0.15 NA) was positioned to achieve ~1.5× magnification. Spatial frequency overlap of ~70% was obtained by placing the Unicorn HAT HD (16×16) LED array (3.3 mm pitch) positioned 65 mm (to provide 0.4 NA) below the sample stage. This resulted in a larger synthetic NA of 0.55. The control and computation were performed through a NVIDIA Jetson Nano board (embedded AI computing platform). This proof-of-concept was improved by adding a tube lens used in an infinite-conjugate configuration with an objective lens to enhance the magnification while avoiding aliasing. Furthermore, we have developed a customized algorithm for the Fourier ptychographic reconstruction and have successfully tested it on a well-known data set using Python, a typical result of the developed FPM reconstruction is shown in Fig. 2. Along with FPM reconstruction algorithm, a deep learning model has also been generated, trained, and tested on a workstation platform. Its inferencing model has been developed and was successfully deployed on a NVIDIA Jetson Nano board to perform on-edge reconstruction on open source datasets. Our near term objectives are to improve the noise performance of our FPM system and eliminate lens curvature effects. The outcome will be a compact, wide FoV, high spatial resolution FPM system, ready for further maturation, with future applications in planetary lander explorations of the Ocean Worlds.

  1. T. Nguyen, Y. Xue, Y. Li, L. Tian, and G. Nehmetallah, “Deep learning approach for Fourier ptychography microscopy,” Opt. Express 26, 26470-26484 (2018)

  2. T. Nguyen, S. Aslam, D. Bower, J. L. Eigenbrode, N. Gorius, T. Hewagama, L. Miko, G. Nehmetallah, “Portable flow device using Fourier ptychography microscopy and deep learning for detection in biosignatures,” Proc. SPIE 11401, Real-Time Image Processing and Deep Learning 2020, 114010H (22 April 2020); doi: 10.1117/12.2557316


No comments here