Digital Image Analysis

Abstract

Light microscopy offers life sciences the possibility to study a variety of organisms on widely differing spatial and temporal scales. Recent years have seen the appearance of a multitude of important light microscopy techniques. To characterise and describe biological processes, however, sheer visualisation of images is inadequate. The assembly of image‐processing algorithms, so‐called workflows, is needed to extract meaningful quantitative data. These workflows are as important as image acquisition and, therefore, are critical components for data acquisition and sample preparation. Here, we describe a variety of image‐processing workflows covering a large part of the biological landscape pointing out common themes and outlining a generic template for workflow assembly.

Key Concepts

  • Digital image analysis is an important tool to extract meaningful data out of an image.
  • Digital image analysis is the last component of a workflow comprising sample preparation and image acquisition.
  • Segmentation allows to identify objects in an image.
  • Filtering is needed to reduce noise in images from biological samples.
  • Digital image analysis is the key for reproducible workflows in life sciences.

Keywords: image processing; image analysis; segmentation; filtering; data representation; analysis workflow

Figure 1. Generic workflow for image analysis. The idea of image analysis is to extract meaningful data out of an image. The image can be seen as a filtered representation of the object. In order to identify the object in the image, a segmentation step is needed which can be typically facilitated by filtering steps. The identified object can be used to describe the object by a variety of features. All or a subset of features can be later used for a graphical representation.
Figure 2. Three different analysis workflows for biological samples. Most digital image processing with biological images follow the steps outlined here. The imaging modality is coupled to the biological question, to allow for a simple pipeline to be derived, either in 2D or 3D as needed. Object detection or ‘segmentation’ is usually done on filtered images of channels whose characteristics do not vary with the experimental conditions being tested (A: transmission image, B: average channel intensity, C: DAPI). The detection can be 1D (B,C), 2D (B) or 3D (B,C) as required. Finally, the metric being extracted can be straightforward such as an area (A) or more complex such as clustering (B) or Distance maps (C). Visualisation allows this data to be observed and understood, and the visualisation method should help to emphasize the expected behaviour of the data, while retaining something about the underlying distribution of the individual data points (dotplots vs barplots).
close

References

Carpenter AE, Jones TR, Lamprecht MR, et al. (2006) CellProfiler: image analysis software for identifying and quantifying cell phenotypes. Genome Biology 7 (10): R100.

Eliceiri KW, Berthold MR, Goldberg IG, et al. (2012) Biological imaging software tools. Nature Methods 9 (7): 697–710.

Esser‐von Bieren J, Mosconi I, Guiet R, et al. (2013) Antibodies trap tissue migrating helminth larvae and prevent tissue damage by driving IL‐4Ralpha‐independent alternative differentiation of macrophages. PLoS Pathogens 9 (11): e1003771.

Esser‐von Bieren J, Volpe B, Kulagin M, et al. (2015) Antibody‐mediated trapping of helminth larvae requires CD11b and Fcgamma receptor I. Journal of Immunology 194 (3): 1154–1163.

Kanca O, Caussinus E, Denes AS, et al. (2014) Raeppli: a whole‐tissue labeling tool for live imaging of Drosophila development. Development 141 (2): 472–480.

Kreshuk A, Koethe U, et al. (2014) Automated detection of synapses in serial section transmission electron microscopy image stacks. PloS One 9 (2): e87351.

Kreshuk A, Straehle CN, Pax E, et al. (2011) Automated detection and segmentation of synaptic contacts in nearly isotropic serial electron microscopy images. PloS One 6 (10): e24899.

Luisier F, Blu T and Unser M (2011) Image denoising in mixed Poisson‐Gaussian noise. IEEE Transactions on Image Processing: A Publication of the IEEE Signal Processing Society 20 (3): 696–708.

Schneider CA, Rasband WS and Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9 (7): 671–675.

van Steensel B and Dekker J (2010) Genomics tools for unraveling chromosome architecture. Nature Biotechnology 28 (10): 1089–1095.

Unser M (2014) Wavelets: on the virtues and applications of the mathematical microscope. Journal of Microscopy 255 (3): 123–127.

Further Reading

Andrews PD, Harper IS and Swedlow JR (2002) To 5D and beyond: quantitative fluorescence microscopy in the postgenomic era. Traffic 3 (1): 29–36.

Carpenter AE (2009) Extracting rich information from images. Methods in Molecular Biology 486: 193–211.

Conrad C and Gerlich DW (2010) Automated microscopy for high‐content RNAi screening. The Journal of Cell Biology 188 (4): 453–461.

Linkert M, Rueden CT, Allan C, et al. (2010) Metadata matters: access to image data in the real world. The Journal of Cell Biology 189 (5): 777–782.

Zimmer C (2012) From microbes to numbers: extracting meaningful quantities from images. Cellular Microbiology 14 (12): 1828–1835.

Contact Editor close
Submit a note to the editor about this article by filling in the form below.

* Required Field

How to Cite close
Burri, Olivier, Guiet, Romain, and Seitz, Arne(May 2016) Digital Image Analysis. In: eLS. John Wiley & Sons Ltd, Chichester. http://www.els.net [doi: 10.1002/9780470015902.a0005781.pub2]