SPIRiT

Single-Pixel Image Reconstruction Toolbox (SPIRiT) Version 2.1, 1st April 2020.

This package contains Matlab scripts and functions that simulate the acquisition and reconstruction of an image with a single-pixel camera.

New Features!

• Recontruction by completion of missing data

help datcomp

• Fast Hadamard 2D transform

help fwht2

• Pre-processing and playing with the STL-10 database

help preprocess_stl10
help loadprep_stl10

Content of version 2.1

SPIRiT 2.1 implements:

• (from v2.1) The Bayesian completion method described in [3]. See main_completion_stl10.m.
• (from v2.0) The SNMF pattern generalization method described in [2]. See main_abswp_simulation.m and main_abswp_experimental.m.
• (from v2.0) The ABS-WP adaptive acquisition method described in [1]. See main_abswp_simulation.m and main_abswp_experimental.m.

The .\function\ folder contains the functions that are called in the above scripts. The .\data\ folder contains:

• three PNG images that are processed in main_abswp_simulation.m
• two experimental datasets (MAT-files) that can be processed by main_abswp_experimental.m. For details, see .\data\Readme.txt.

The .\reference\ folder contains the PDF of [1], [2], and [3].

Datasets

We provide:

• A function to preprocess the STL-10 database that can be downloaded at https://ai.stanford.edu/~acoates/stl10/
• Two experimental datasets (Department of Physics, Politecnico di Milano, Italy) of the Jaszczak target acquired using wavelet patterns, initially published in [1]. For details, see .\data\Readme.txt.

Illustration

• Hadamard acquisition for a 10% sampling ratio (see main_completion_stl10.m)

Installation

Just make sure to add .\function\ to your Matlab search path.

path(fullfile(pwd,'function'),path);

SPIRiT may require one of the following toolboxes to run:

• Image Processing Toolbox (MathWorks)
• Statistics and Machine Learning Toolbox (MathWorks), to corrupt data with Poisson noise
• Wavelab850.

Note: If required, make sure Wavelab850 appears at the top of your search path to avoid conflits

Contact

nicolas.ducros@insa-lyon.fr, University of Lyon, France.

License

SPIRiT is distributed freely under Creative Commons Attribution-ShareAlike 4.0 International license (CC-BY-SA 4.0)

References

[3] N. Ducros et al., "A completion network for reconstruction from compressed acquisition', IEEE ISBI, 2020. https://hal.archives-ouvertes.fr/hal-02342766/document

[2] F. Rousset et al., "A semi nonnegative matrix factorization technique for pattern generalization in single-pixel imaging", IEEE Transactions on Computational Imaging, 4(2), 284-294, 2018. https://doi.org/10.1109/TCI.2018.2811910 https://hal.archives-ouvertes.fr/hal-01635461/document

[1] F. Rousset et al., "Adaptive basis scan by wavelet prediction for single-pixel imaging", IEEE Transactions on Computational Imaging, 3(1), 36-46, 2017. http://dx.doi.org/10.1109/TCI.2016.2637079 https://hal.archives-ouvertes.fr/hal-01314314/document