This Indico installation shall not be used to organise MPI-CBG courses and events from beginning of 2024 on. Please use MPG Indico instead.
12-15 August 2018
MPI-CBG
Europe/Berlin timezone

BigStitcher: Reconstructing high-resolution image datasets of cleared and expanded samples

13 Aug 2018, 18:15
15m
MPI-CBG

MPI-CBG

Pfotenhauerstraße 108 01307 Dresden Germany
Short Talk Light sheet fluorescence microscopy Image analysis of light sheet data

Speaker

David Hörl (Department of Biology II, Ludwig-Maximilians-Universität München)

Description

Continuous advancements in microscopy and sample preparation methods such as clearing or expansion allow for the investigation of ever larger samples at high resolution. This entails increasingly large datasets that may consist of hundreds of images of one sample that are not aligned, suffer from optical disturbances and often cannot even be opened as a whole, which can pose a serious bottleneck to scientific inquiries. With terabyte-sized datasets becoming more and more common, development of software tools that make handling and analysis of large and complex image data available to the broader scientific community is an urgent issue.

To allow efficient handling and reconstruction of large multi-tile and multi-view image data, we developed the BigStitcher software. It enables import from most file formats, interactive handling, fast and precise alignment, as well as deconvolution and real-time fusion of large image datasets. We additionally support the alignment of multi-tile acquisitions taken from different orientations, effectively doubling the size of objects that can be imaged. We also enable the correction of a variety of optical distortions, e.g. via automatic illumination selection, flat-field correction and interest-point-based correction of chromatic aberrations.

We implemented BigStitcher using the state-of-the-art frameworks ImgLib2 and BigDataViewer. By combining multi-resolution data representation and sub-pixel accurate registration algorithms, even very large datasets can be reconstructed on conventional, off-the-shelf hardware. In an effort to make large sample reconstruction available as a routine task, we provide a user-friendly graphical user interface (GUI) to manually guide the alignment and interactively display the intermediate results using BigDataViewer. BigStitcher is open-source and provided as a Fiji-plugin, making it a powerful, scalable tool for automated processing of very large image datasets.

Terms and Conditions Yes
Affiliation Department of Biology II, Ludwig-Maximilians-Universität München

Primary authors

David Hörl (Department of Biology II, Ludwig-Maximilians-Universität München) Dr Fabio Rojas Rusak (Cardiovascular and Metabolic Sciences, Max Delbrück Center for Molecular Medicine in the Helmholtz Association (MDC)) Mr Friedrich Preusser (Berlin Institute for Medical Systems Biology (BIMSB), Max Delbr ü ck Center for Molecular Medicine in the Helmholtz Association (MDC),) Dr Paul Tillberg (Janelia Research Campus, Howard Hughes Medical Institute) Dr Nadine Randel (Janelia Research Campus, Howard Hughes Medical Institute) Dr Raghav K. Chhetri (Janelia Research Campus, Howard Hughes Medical Institute) Prof. Albert Cardona (Janelia Research Campus, Howard Hughes Medical Institute) Dr Philipp J. Keller (Janelia Research Campus, Howard Hughes Medical Institute) Dr Hartmann Harz (Department of Biology II, Ludwig-Maximilians-Universität München) Prof. Heinrich Leonhardt (Department of Biology II, Ludwig-Maximilians-Universität München) Prof. Mathias Treier (Cardiovascular and Metabolic Sciences, Max Delbrück Center for Molecular Medicine in the Helmholtz Association (MDC)) Dr Stephan Preibisch (Max-Delbrück-Centrum )

Presentation Materials

There are no materials yet.
Your browser is out of date!

Update your browser to view this website correctly. Update my browser now

×