CoChat
Sign Up
Cited by 405

The DOI matches its Crossref record; 7 of the 7 references it cites were found in a registry or an index.

LABKIT: Labeling and Segmentation Toolkit for Big Image Data

Arzt, Matthias, Deschamps, Joran, Schmied, Christopher, Pietzsch, Tobias, Schmidt, Deborah, Tomancak, Pavel, Haase, Robert, Jug, Florian·2022·Frontiers in Computer Science
DOI 10.3389/fcomp.2022.777728

Published version of a preprint

Open this paper in CoChat — free account

Abstract

We present LABKIT, a user-friendly Fiji plugin for the segmentation of microscopy image data. It offers easy to use manual and automated image segmentation routines that can be rapidly applied to single- and multi-channel images as well as to timelapse movies in 2D or 3D. LABKIT is specifically designed to work efficiently on big image data and enables users of consumer laptops to conveniently work with multiple-terabyte images. This efficiency is achieved by using ImgLib2 and BigDataViewer as well as a memory efficient and fast implementation of the random forest based pixel classification algorithm as the foundation of our software. Optionally we harness the power of graphics processing units (GPU) to gain additional runtime performance. LABKIT is easy to install on virtually all laptops and workstations. Additionally, LABKIT is compatible with high performance computing (HPC) clusters for distributed processing of big image data. The ability to use pixel classifiers trained in LABKIT via the ImageJ macro language enables our users to integrate this functionality as a processing step in automated image processing workflows. Finally, LABKIT comes with rich online resources such as tutorials and examples that will help users to familiarize themselves with available features and how to best use LABKIT in a number of practical real-world use-cases.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.3389/fcomp.2022.777728

Cites

7 of 7 references verified0 citation errors · 0 not checkable

Statistics

Figures

Image 1
CC-BY · From the PDF
FIGURE 2 Semantic segmentation of microscopy images with LABKIT's pixel classification: (A) Maximum intensity projection of a confocal stack showing HeLa cells expressing C1-GFP (left), next to the sparse labeling (scribbles, center) and resulting cell segmentation (right). (B) Bright field microscopy image of E.coli , sparse labeling discriminating cells and background and the resulting segmentation. Arrowheads show that segmentation of out-of-focus objects can be reduced by including pixels of such objects in the background class. (C) Fixed mouse liver tissue section stained with immunofluorescence and imaged in multiple channels with a spinning disk confocal microscope, showing Hepatocyte nuclei stained with antibody against HNF-4 α a transcription factor expressed in hepatocytes, hepatocyte cytoplasm (autofluorescence) and all nuclei stained with DAPI. (D) Labeling and resulting segmentation of the liver tissue section shown in (A) , segmenting Hepatocyte cytoplasm (green), Hepatocyte nuclei (blue), nuclei of non-parenchymal cells (yellow) and sinusoids (magenta). Scale bars 20 µ m (A) , (C) , and 5 µ m (B) .
FIGURE 2 Semantic segmentation of microscopy images with LABKIT's pixel classification: (A) Maximum intensity projection of a confocal stack showing HeLa cells expressing C1-GFP (left), next to the sparse labeling (scribbles, center) and resulting cell segmentation (right). (B) Bright field microscopy image of E.coli , sparse labeling discriminating cells and background and the resulting segmentation. Arrowheads show that segmentation of out-of-focus objects can be reduced by including pixels of such objects in the background class. (C) Fixed mouse liver tissue section stained with immunofluorescence and imaged in multiple channels with a spinning disk confocal microscope, showing Hepatocyte nuclei stained with antibody against HNF-4 α a transcription factor expressed in hepatocytes, hepatocyte cytoplasm (autofluorescence) and all nuclei stained with DAPI. (D) Labeling and resulting segmentation of the liver tissue section shown in (A) , segmenting Hepatocyte cytoplasm (green), Hepatocyte nuclei (blue), nuclei of non-parenchymal cells (yellow) and sinusoids (magenta). Scale bars 20 µ m (A) , (C) , and 5 µ m (B) .CC-BY · From the PDF
FIGURE 3 LABKIT labeling tools used for curation: labels generated by manual dense labeling or automatic segmentation can be efficiently curated with drawing, filling, erasing, or deletion of entire objects. Scale bar 10 µ m.
FIGURE 3 LABKIT labeling tools used for curation: labels generated by manual dense labeling or automatic segmentation can be efficiently curated with drawing, filling, erasing, or deletion of entire objects. Scale bar 10 µ m.CC-BY · From the PDF
FIGURE 4 Comparing automatic and manual ground-truth generation with LABKIT: (A) Fluorescence image of nuclei (out of 26 images) extracted from the 2018 Data Science Bowl (Caicedo et al., 2019). (B) Results from LABKIT automated segmentation of (A) after extracting connected components and giving each instance a unique pixel value. The arrows point to various segmentation errors. On the top right corner, the total time necessary to obtain the corresponding segmentation of all 26 images (including labeling) is indicated. Below the timing is the average precision (see Methods) as compared to a dense manual labeling performed by another observer. (C) Curation of (B) with same post-processing. The arrows point to the corrected errors mentioned in (B) . The timing information includes (B) . (D) Dense manual labeling of (A) and the same post-processing as in (B) . No scale bar was available for the images.
FIGURE 4 Comparing automatic and manual ground-truth generation with LABKIT: (A) Fluorescence image of nuclei (out of 26 images) extracted from the 2018 Data Science Bowl (Caicedo et al., 2019). (B) Results from LABKIT automated segmentation of (A) after extracting connected components and giving each instance a unique pixel value. The arrows point to various segmentation errors. On the top right corner, the total time necessary to obtain the corresponding segmentation of all 26 images (including labeling) is indicated. Below the timing is the average precision (see Methods) as compared to a dense manual labeling performed by another observer. (C) Curation of (B) with same post-processing. The arrows point to the corrected errors mentioned in (B) . The timing information includes (B) . (D) Dense manual labeling of (A) and the same post-processing as in (B) . No scale bar was available for the images.CC-BY · From the PDF
FIGURE 5 Pixel classification failure modes: (A) Image of a contractile vacuole pore in paramecium caudatum acquired with transmission electron microscopy (left), in which inner pore (blue), cytoplasm (green) and background (magenta) were labeled with scribbles (center). LABKIT's classifier fails to distinguish background and inner pore classes (right). Image provided by Richard Allen (Cell Image Library, 38894). (B) Mitotic spindle of a Ptk2 cell expressing GFP-tubulin imaged in wide-field fluorescence (left). Left and right halves of the mitotic spindle are labeled in different classes (blue and green, respectively), while background is labeled in magenta (center). LABKIT's classifier is incapable of discriminating similar structures (right). Image provided by Sophie Dumont and Timothy J. Mitchison (Cell Image Library, 6568). (C) Bright field image of C. elegans (left), labeled with two classes (center): foreground (green) and background (magenta). After applying connected components analysis to the classification result, contiguous worms belong to the same object (right). Image provided by Fred Ausubel (Broad Bioimage Benchmark Collection, BBBC010). Scale bars (A) 500 nm, (B) 10 µ m, and (C) 500 µ m.
FIGURE 5 Pixel classification failure modes: (A) Image of a contractile vacuole pore in paramecium caudatum acquired with transmission electron microscopy (left), in which inner pore (blue), cytoplasm (green) and background (magenta) were labeled with scribbles (center). LABKIT's classifier fails to distinguish background and inner pore classes (right). Image provided by Richard Allen (Cell Image Library, 38894). (B) Mitotic spindle of a Ptk2 cell expressing GFP-tubulin imaged in wide-field fluorescence (left). Left and right halves of the mitotic spindle are labeled in different classes (blue and green, respectively), while background is labeled in magenta (center). LABKIT's classifier is incapable of discriminating similar structures (right). Image provided by Sophie Dumont and Timothy J. Mitchison (Cell Image Library, 6568). (C) Bright field image of C. elegans (left), labeled with two classes (center): foreground (green) and background (magenta). After applying connected components analysis to the classification result, contiguous worms belong to the same object (right). Image provided by Fred Ausubel (Broad Bioimage Benchmark Collection, BBBC010). Scale bars (A) 500 nm, (B) 10 µ m, and (C) 500 µ m.CC-BY · From the PDF
Image 6
CC-BY · From the PDF

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 7 references

Research works printed but not linked: 44

  1. Alemi Koohbanani, 2020, Nuclick: a deep learning framework for interactive segmentation of microscopic images, Med. Image Anal, 65, 101771, 10.1016/j.media.2020.101771 · DOI · Paper page
  2. Arganda-Carreras, 2017, Trainable weka segmentation: a machine learning tool for microscopy pixel classification, Bioinformatics, 33, 2424, 10.1093/bioinformatics/btx180 · DOI · Paper page
  3. Arshadi, 2021, SNT: a unifying toolbox for quantification of neuronal anatomy, Nat. Methods, 18, 374, 10.1038/s41592-021-01105-7 · DOI
  4. Baltissen, 2018, Comparison of segmentation methods for tissue microscopy images of glioblastoma cells, 396, 10.1109/ISBI.2018.8363601 · DOI
  5. Berg, 2019, Ilastik: interactive machine learning for (bio) image analysis, Nat. Methods, 16, 1226, 10.1038/s41592-019-0582-9 · DOI
  6. Betzig, 2006, Imaging intracellular fluorescent proteins at nanometer resolution, Science, 313, 1642, 10.1126/science.1127344 · DOI
  7. Bornes, 2021, Scratch-induced partial skin wounds re-epithelialize by sheets of independently migrating keratinocytes, Life Sci. Alliance, 4, e202000765, 10.26508/lsa.202000765 · DOI
  8. Breiman, 2001, Random forests, Mach. Learn, 45, 5, 10.1023/A:1010933404324 · DOI · Paper page
  9. Buchholz, 2020, DenoiSeg: joint denoising and segmentation. arXiv:2005.02987 [cs] · arXiv
  10. Dodt, 2007, Ultramicroscopy: three-dimensional visualization of neuronal networks in the whole mouse brain, Nat. Methods, 4, 331, 10.1038/nmeth1036 · DOI
  11. Folk, 2011, An overview of the hdf5 technology suite and its applications, 36, 10.1145/1966895.1966900 · DOI
  12. Gustafsson, 2000, Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy, J. Microsc, 198, 82, 10.1046/j.1365-2818.2000.00710.x · DOI · Paper page
  13. Haase, 2020, Clij: Gpu-accelerated image processing for everyone, Nat. Methods, 17, 5, 10.1038/s41592-019-0650-1 · DOI
  14. Hama, 2011, Scale: a chemical approach for fluorescence imaging and reconstruction of transparent mouse brain, Nat. Neurosci, 14, 1481, 10.1038/nn.2928 · DOI
  15. Hell, 1994, Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy, Opt. Lett, 19, 780, 10.1364/OL.19.000780 · DOI
  16. Hess, 2006, Ultra-high resolution imaging by fluorescence photoactivation localization microscopy, Biophys. J, 91, 4258, 10.1529/biophysj.106.091116 · DOI
  17. Hörl, 2019, Bigstitcher: reconstructing high-resolution image datasets of cleared and expanded samples, Nat. Methods, 16, 870, 10.1038/s41592-019-0501-0 · DOI
  18. Horlava, 2020, A comparative study of semi- and self-supervised semantic segmentation of biomedical microscopy data. arXiv:2011.08076 [cs, stat] · arXiv
  19. Huisken, 2004, Optical sectioning deep inside live embryos by selective plane illumination microscopy, Science, 305, 1007, 10.1126/science.1100035 · DOI
  20. Icha, 2016, Using light sheet fluorescence microscopy to image zebrafish eye development, J. Vis. Exp, 110, e53966, 10.3791/53966 · DOI
  21. Jain, 2020, Regionalized tissue fluidization is required for epithelial gap closure during insect gastrulation, Nat. Commun, 11, 5604, 10.1038/s41467-020-19356-x · DOI
  22. Jordan, 2019, Chapter 9-In situ cryo-electron tomography and subtomogram averaging of intraflagellar transport trains, 179, 10.1016/bs.mcb.2019.04.005 · DOI
  23. Knott, 2008, Serial section scanning electron microscopy of adult brain tissue using focused ion beam milling, J. Neurosci, 28, 2959, 10.1523/JNEUROSCI.3189-07.2008 · DOI · Paper page
  24. Lalit, 2021, Embedding-based Instance Segmentation in Microscopy. arXiv:2101.10033 [cs, eess] · arXiv
  25. Leite, 2021, Paintera
  26. Linkert, 2010, Metadata matters: access to image data in the real world, J. Cell Biol, 189, 777, 10.1083/jcb.201004104 · DOI
  27. Mais, 2020, Patchperpix for instance segmentation, 288, 10.1007/978-3-030-58595-2_18 · DOI
  28. Maninis, 2018, Deep extreme cut: From extreme points to object segmentation, 616, 10.1109/CVPR.2018.00071 · DOI
  29. Maška, 2014, A benchmark for comparison of cell tracking algorithms, Bioinformatics, 30, 1609, 10.1093/bioinformatics/btu080 · DOI · Paper page
  30. McQuin, 2018, Cellprofiler 3.0: Next-generation image processing for biology, PLoS Biol, 16, e2005970, 10.1371/journal.pbio.2005970 · DOI
  31. Morone, 2020, Deep learning approach for quantification of organelles and misfolded polypeptide delivery within degradative compartments, Mol. Biol.Cell, 31, 1512, 10.1091/mbc.E20-04-0269 · DOI
  32. 2019, napari: A Multi-Dimensional Image Viewer For Python
  33. Pietzsch, 2012, Imglib2–generic image processing in java, Bioinformatics, 28, 3009, 10.1093/bioinformatics/bts543 · DOI
  34. Pietzsch, 2015, Bigdataviewer: visualization and processing for large image data sets, Nat. Methods, 12, 481, 10.1038/nmeth.3392 · DOI
  35. Preibisch, 2014, Efficient Bayesian-based multiview deconvolution, Nat. Methods, 11, 645, 10.1038/nmeth.2929 · DOI
  36. Rueden, 2017, Imagej2: imagej for the next generation of scientific image data, BMC Bioinformatics, 18, 529, 10.1186/s12859-017-1934-z · DOI
  37. Rust, 2006, Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (storm), Nat. Methods, 3, 793, 10.1038/nmeth929 · DOI
  38. Schmidt, 2018, Cell detection with star-convex polygons, 265
  39. Stringer, 2021, Cellpose: a generalist algorithm for cellular segmentation, Nat. Methods, 18, 100, 10.1038/s41592-020-01018-x · DOI · Paper page
  40. Tischer, 2020, Bigdataprocessor2: a free and open-source fiji plugin for inspection and processing of tb sized image data, bioRxiv, 10.1101/2020.09.23.244095 · DOI
  41. Ulman, 2017, An objective comparison of cell-tracking algorithms, Nat. Methods, 14, 1141, 10.1038/nmeth.4473 · DOI
  42. Wolff, 2018, Multi-view light-sheet imaging and tracking with the mamut software reveals the cell lineage of a direct developing arthropod limb, Elif, 7, e34410, 10.7554/eLife.34410 · DOI
  43. Wolny, 2020, Accurate and versatile 3d segmentation of plant tissues at cellular resolution, Elife, 9, e57613, 10.7554/eLife.57613 · DOI
  44. Zhang, 2020, Interactive object segmentation with inside-outside guidance, 12231, 10.1109/CVPR42600.2020.01225 · DOI · Paper page

Cited by

Sorted by citations

Metadata

Authors
Arzt, M., Deschamps, J., Schmied, C., Pietzsch, T., Schmidt, D., Tomancak, P., Haase, R., & Jug, F.
Venue
Frontiers in Computer Science
Publisher
Frontiers Media SA
Published
2022-02-10
Volume
4
ISSN
2624-9898
DOI
10.3389/fcomp.2022.777728
Cited by
405
References
7
Licence
creativecommons.org/licenses/by/4.0/ (vor)
Funders
  • Max-Planck-Gesellschaft (M.IF.A.MOZG8106)
  • Bundesministerium für Bildung und Forschung (01IS18026C, 031L0102)
  • Deutsche Forschungsgemeinschaft (EXC2068, JU3110/1-1, TO563/8-1)
  • European Regional Development Fund (CZ.02.1.01/0.0/0.0/16_013/0001791)

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review