FPD: Fast pixelated detector data storage, analysis and visualisation.


Keywords
microscopy, STEM, TEM, fast, pixelated, detector, fpd, data, storage, EMD, hdf5, analysis, differential, phase, contrast, DPC, segmented, virtual, lattice, analyis, strain, non-rigid, image, alignment
License
GPL-3.0
Install
pip install fpd==0.1.6

Documentation

News

7 April 2020 - Release 0.1.10 (see changelog for details), updated notebook, and a new preprint: Fast Pixelated Detectors in Scanning Transmission Electron Microscopy. Part II: Post Acquisition Data Processing, Visualisation, and Structural Characterisation, arXiv (2020).

1 March 2020 - Release 0.1.9 with many new features and improvements - see changelog for details.

18 November 2019 - New preprint: Fast Pixelated Detectors in Scanning Transmission Electron Microscopy. Part I: Data Acquisition, Live Processing and Storage, arXiv (2019).

4 November 2019 - Release 0.1.8: updates for compatibility with most recent skimage and numpy.

See https://fpdpy.gitlab.io/fpd/news.html for earlier news.

FPD package

The fpd package provides code for the storage, analysis and visualisation of data from fast pixelated detectors. The data storage uses the hdf5 based EMD file format, and the conversion currently supports the Merlin readout from Medipix3 detectors. Differential phase contrast imaging and several other common data analyses, like radial distributions, virtual apertures, and lattice analysis, are also implemented, along with many utilities.

The package is relatively lightweight, with most of its few dependencies being standard scientific libraries. All calculations run on CPUs and many use out-of-core processing, allowing data to be visualised and processed on anything from very modest to powerful hardware.

A degree of optimisation through parallelisation has been implemented. The development environment is Linux; the efficiency may vary on Windows because of how forking works, but the Windows 10 Linux subsystem has been reported to work well.

Citing

If you find this software useful and use it to produce results in a puplication, please consider citing the website or related paper(s).

An example bibtex entry with the date in the note field yet to be specified:

@Misc{fpd,
    Title                    = {{FPD: Fast pixelated detector data storage, analysis and visualisation}},
    howpublished             = {\url{https://gitlab.com/fpdpy/fpd}},
    note                     = {{Accessed} todays date}
}

Aspects of the library are covered in papers:

  • Fast Pixelated Detectors in Scanning Transmission Electron Microscopy. Part I: Data Acquisition, Live Processing and Storage, arXiv (2019).

  • Fast Pixelated Detectors in Scanning Transmission Electron Microscopy. Part II: Post Acquisition Data Processing, Visualisation, and Structural Characterisation, arXiv (2020).

Publications

Some of the known scientific papers that used the fpd library are listed below. If you use the library for results contributing to a publication, please pass the paper details to developers for inclusion in this list.

  • Order and disorder in the magnetization of the chiral crystal CrNb3S6, Phys. Rev. B 99, 224429 (2019).

  • Heisenberg pseudo-exchange and emergent anisotropies in field-driven pinwheel artificial spin ice, Phys. Rev. B 100, 174410 (2019).

  • Strain Anisotropy and Magnetic Domains in Embedded Nanomagnets, Small, 1904738 (2019).

  • Sub-100 nanosecond temporally resolved imaging with the Medipix3 direct electron detector, Ultramicroscopy, 210, 112917 (2020).

  • Tensile deformations of the magnetic chiral soliton lattice probed by Lorentz transmission electron microscopy, arXiv (2019).

  • Spontaneous creation and annihilation dynamics and strain-limited stability of magnetic skyrmions, arXiv (2019).

  • Fast Pixelated Detectors in Scanning Transmission Electron Microscopy. Part I: Data Acquisition, Live Processing and Storage, arXiv (2019).

  • Fast Pixelated Detectors in Scanning Transmission Electron Microscopy. Part II: Post Acquisition Data Processing, Visualisation, and Structural Characterisation, arXiv (2020).

Installation

The package currently supports both python versions 2.7 and 3.x. Hyperspy is used in a few places but most of the fpd module can be used without it being installed (simply install the package dependencies manually and ignore them when using pip by adding --no-deps to the install command).

Installation from source:

pip3 install --user .

Instalation from PyPI (https://pypi.org/project/fpd/):

pip3 install --user fpd

-U can be added to force an upgrade / reinstall; in combination with --no-deps, only the fpd package will be reinstalled.

The package can be removed with:

pip3 uninstall fpd

Usage

In python or ipython:

import fpd
d = fpd.DPC_Explorer(-64)
import fpd.fpd_processing as fpdp
rtn = fpdp.phase_correlation(data, 32, 32)

where data is any array-like object. For example, this can be an in-memory numpy array, an hdf5 object on disk, or a dask array, such as that used in 'lazy' hyperspy signals.

All functions and classes are documented and can be read, for example, in ipython by appending a ? to the object. E.g.:

import fpd
fpd.DPC_Explorer?

Documentation

Release: https://fpdpy.gitlab.io/fpd/

Development version: https://gitlab.com/fpdpy/fpd/builds/artifacts/master/file/pages_development/index.html?job=pages_development

Notebook demos: https://gitlab.com/fpdpy/fpd-demos.

Further documentation and examples will be made available over time.

Related projects

https://www.gla.ac.uk/schools/physics/research/groups/mcmp/researchareas/pixstem/

http://quantumdetectors.com/stem/

https://gitlab.com/fast_pixelated_detectors/merlin_interface

https://gitlab.com/fast_pixelated_detectors/fpd_live_imaging

https://gitlab.com/pixstem/pixstem

https://emdatasets.com/format

http://hyperspy.org/

http://gwyddion.net/

More packages will be added to the https://gitlab.com/fast_pixelated_detectors group as they develop.