Aligning Big Brains & Atlases


Documentation of the Aligning Big Brains & Atlases Fiji plugin.

View the Project on GitHub BIOP/ijp-imagetoatlas

[EXPERIMENTAL] ABBA - Aligning Big Brains & Atlases

A Fiji plugin for the registration of thin brain slices to various atlases (3D mouse Allen Brain atlas, Waxholm Space Atlas of the Sprague Dawley Rat Brain, and BrainGlobe atlases (WIP)) + QuPath associated tools.

Aligning Big Brains & Atlases or ABBA for short, is a Fiji plugin which allows to register serial sections to several atlases, in coronal, sagittal and horizontal orientations.

It uses BigDataViewer and BigWarp for the display and on-the-fly computation of spline-transformed multiresolution images (typical output of Whole Slide Imaging).

It has been developed by the BioImaging & Optics Platform at EPFL. This page contains the documentation of ABBA (installation and usage). If you require additional help, please post your question on the forum and tag your question / issue with abba and fiji or qupath. If you have already installed ABBA, you can also click on Help > Ask for help in the forum from the plugin (some helpful information from your local installation will be included in your post).


There’s currently three forms of documentation available:

Demo datasets

If you want to test this plugin, you can download one of the following test dataset:

Dataset 1, provided by Lucie Dixsaut, Johannes Gräff lab, EPFL

One animal, 85 serial sections, 3 fluorescent channels (DAPI - nuclei, FITC - autofluorescence, mCherry - labelled sparse cells)

Dataset 2, provided by Bianca A. Silva, Johannes Gräff lab, EPFL

One animal, 87 serial sections, 2 fluorescent channels (nuclei and autofluorescence)

Important note on File formats!

TL; DR: Use calibrated VSI, CZI, OME-TIFF, NDPI, a few others (please read the text below). Otherwise convert you files to pyramidal OME-Tiff files. You can use the Fiji plugin Kheops for instance.

All files need to be properly calibrated (microns, millimeters, etc, but not pixels!). ABBA takes advantage of the calibration to set appropriate registration parameters.

:bulb: It is strongly recommended to work with multiresolution file formats (VSI, OME-TIFF, SVS), since brain slices are usually very big 2d images. ABBA, like QuPath, uses pre-computed downsampled images of these files to speed-up (very significantly) the display and processing of these images. Downsampled images also help for registration, since the registration is made with large scale features (size above a few cells), which are incorrectly sampled if no downsampled image pre-exists.

:warning: Because Fiji is used is the workflow, only Bio-Formats supported formats are correctly handled. You can check on the Bio-Formats documentation if your file formats will be correctly handled. This will be the case if Pyramid is checked. File which can be opened only via OpenSlide are not supported.

Tested file formats for ABBA :

CZI, NDPI, OME-TIFF should work perfectly. Let us know if that’s the case in practice and we’ll update the list.

RGB images as well as 8-bits and 16-bits images have also been successfully tested.