Introduction

The Problem

Quantitative phase imaging (QPI) is a fundamental imaging technique that visualizes the retardation of electromagnetic radiation as it passes through an object. The parameter that governs this retardation is called refractive index. In biological imaging, QPI is an important tool to measure the dry mass or the refractive index (related to mass density [Bar52] [DW52]) of single cells and tissues, which enables a profound characterization of the investigated samples.

Why qpimage?

In the Guck group, we make heavy use of QPI and thus require a reliable and well-documented software library that, independent of the particular QPI setup used, allows us to address QPI-related research questions. Qpimage attempts to unify QPI analysis by providing a unique and user-friendly API for working with QPI data, including the choice of input data (complex field, phase with amplitude or intensity, hologram), memory-efficient and fast storage of large data sets (using HDF5, phase and amplitude data are stored separately), or robust and extendable background correction techniques (tilt and second order polynomial fits, binary mask). The main reason for the development of qpimage is our QPI analysis software DryMass.

What are the alternatives?

There are other open-source Python libraries that address quantitative phase imaging analysis with varying scopes and motivations.

  • HoloPy is an established Python library for digital holographic microscopy (DHM) that comes with several additional features such as scattering calculations and model fitting. The overlap between HoloPy and qpimage is the computation of phase and amplitude from raw hologram data. The main difference is that HoloPy is focused on DHM analysis with a rich set of tools while qpimage is only focused on managing quantitative phase data (data conversion and storage as well as an extended set of background correction algorithms). However, there is a broad set of additional tools in the “qpimage universe”, including qpformat for loading experimental data, qpsphere for scattering calculations and model fitting (focus is on cell-sized objects), and DryMass as a user interface to these libraries.
  • The Python package shampoo focuses on DHM reconstruction and detection and tracking of biological cells. The overlap between shampoo and the “qpimage universe” (see above) is quite large. The difference is mostly the scope of the projects; While shampoo is an optimized library for DHM analysis, the “qpimage universe” encompasses other quantitative phase imaging (QPI) techniques with the aim to becoming a generic tool in QPI analysis. Experimental .tif files from the shampoo project can be opened with qpformat (see Hologram from tif file).
  • If you are using electron holography, HyperSpy might be worth looking at. If you are storing your hologram data in the HyperSpy file format, you can still load it with qpformat (see HyperSpy hologram file format) and analyze it with qpimage.

Citing qpimage

If you are using qpimage in a scientific publication, please cite it with:

(...) using qpimage version X.X.X (available at
https://pypi.python.org/pypi/qpimage).

or in a bibliography

Paul Müller (2017), qpimage version X.X.X: Phase image analysis
[Software]. Available at https://pypi.python.org/pypi/qpimage.

and replace X.X.X with the version of qpimage that you used.

Furthermore, several ideas implemented in qpimage have been described and published in scientific journals:

  • Phase retrieval from holographic images with a gaussian filter is implemented according to [SSM+15].
  • Phase background image correction with a tilt fitted to a border of the image data was used in [SSM+15] and [SSM+16].
  • Phase background image correction with a polynomial fitted to known background regions was introduced for DHM in [CCC+06] (in this reference the phase correction is applied to the hologram data before field reconstruction).
  • Intensity background correction by dividing by a reference intensity image for tomographic imaging was used in [SCG+17].