FASTMAP (Flexible Atlas Segmentation Tool for Multi-Area Processing) is a tool for the registration of biological images to custom atlas plastes and the segmentation of labels of interest within atlas regions. This tool operates as an ImageJ Plugin that draws upon versatile and powerful image analyses tools in ImageJ and presents them as a clean, concise, and easy to follow plugin.
FASTMAP was created by Dylan Terstege, a Neuroscience PhD candidate in the Epp Lab at the University of Calgary.
FASTMAP was demonstrated at the 2021 Tissue Clearing and Expansion Workshop hosted by the Dynamic Brain Circuits UBC and BCREGMED. This virtual workshop was recorded, with the FASTMAP presentation beginning at the 13:55 mark in the linked YouTube video. Tech demo begins at 30:14.
Current Version: V 1.1.1
Updates since V 1.0.1:
- No more need to manually combine ROIs in ROI manager
- Automatically saves registered ROI Sets
Section | Description |
---|---|
1. Installation Instructions | How to install on macOS and Windows. |
2. Image Processing Tutorial | How to process a sample image set |
3. Atlas Plate Customization | How to build your own atlas plates for custom projects |
4. Troubleshooting | Potential complications and how to overcome them |
5. Citation | How to cite FASTMAP |
6. Contact Us | Where to reach us with questions |
FASTMAP was designed with ease-of-use at the forefront of our minds. The installation process reflects this with a simple 3-step approach:
-
1.1 Install ImageJ. Ensure that ImageJ is installed. If it was not previously installed, it can be downloaded here.
-
1.2 Download the Appropriate Version of FASTMAP. Download the appropriate version of FASTMAP for your operating system: macOS or Windows. Linux has not yet been tested.
-
1.3 Install Plugin in ImageJ. Before installing, see settings section to determine whether the plugin should be modified in any way to better suit your project prior to installation. Once satisfied with the settings, open ImageJ and select "Plugins > Install...". Navigate to the newly downloaded version of FASTMAP and allow this to save to the ImageJ Plugins folder.
Note: If installing on FIJI, ensure you use the "Install..." command rather than the "Install Plugin Command".
The following guide will outline how to process a dataset using FASTMAP:
INITIALIZATION
- 2.1 Images. FASTMAP has been optimized for 8- and 16-bit .tif files, with the label of interest being in an image file which is separate folder from the channel to be used as a reference during atlas registration (DAPI, propidium iodide, autofluorescence, etc.).
If your images are not already in this format, ImageJ can rapidly convert image formats by selecting "Process > Batch > Convert...".
-
2.2 File Organization. Files should be organized under a common "parent folder". This parent folder should contain at least two subfolders: one containing the "label images" (the to-be-segmented labels) and a second containining the "registration images" (images to be used as references during atlas registration). Folders can contain all images for an entire project, as long as the images for each subject are in continuous strings and corresponding files appear in the same alphabetical order in both folders.
-
2.3 Atlas Organization. FASTMAP is highlighted by its applicability to a nearly limitless range of biological samples, irrespective of sample type or orientation. This is due to the flexibility of the atlas plate registration and generation. To facilitate immediate use, sample atlases have been provided.
To use these atlas plates, download the provided atlases and unzip only the main folders (ex. "Sagittal.zip") and not the nested zipped folders (ex. "RoiSet1.zip"). Keep these unzipped main folders in an easilly accessible "Plates" folder.
RUNNING THE PLUGIN
-
2.4 Run Plugin. With image files and atlas plates in proper formats and organizations, the plugin may now be ran (open ImageJ, "Plugins > FASTMAP"). The user will be prompted to navigate to the parent folder containing both image channels. A second prompt will ask the user to navigate to a folder containing the desired atlas plates for the analysis.
-
2.5 Image Assignment. A dialogue window titled FASTMAP will populate and prompt the user to identify which of the subfolders in the parent folder contains the registration images and which contains label images.
-
2.6 Range and Analysis Type. The plugin will then display the number of files in the registration and label subfolders. Output data will be appended based upon the range inputted during this step, so if images from multiple samples are in the same folder the the range of the current sample should be defined.This window also asks which type(s) of analysis should be applied to the image set. Analysis types are as follows:
- Volumetric Analysis: reports the total area of the region, the summed area of all labels within the region, and the percentage of the region that is comprised of labels.
- Object Counts: reports the total area of the region and number of objects counted within each region.
Both the volumetric analysis and object counts analysis types have been written to use the default auto-thresholding parameters of ImageJ. If working with already binarized labels as inputs or raw images in which the default thresholding is insufficient, see the settings section for information on how to modify the plugin to better suit the input images.
At the bottom of this window, the user will be asked whether ROIs have already been gathered for this range of images. If images have previously been processed through FASTMAP and the ROIs have been saved, this option can be chosen to conserve the previously registered ROIs. For more information on saving ROIs, click here.
- 2.7 Choose Registration Plate. A composite image displaying all plates from the selected atlas will populate on the screen. Drag this to the right side of the screen before selecting 'OK', otherwise it will be hidden by the next image to populate the screen. This image is the working registration image and a prompt will ask which atlas plate it most closely resembles.
- 2.8 Automated resizing. FASTMAP can be applied to images collected using a wide variety of microscopy techniques. Furthermore, it can be applied to images of different sizes and resolutions. If the user is looking at the same group of regions using multiple microscopy types or imaging magnifications, the plugin can scale atlas plates to limit the extent to which regions need to be manually resized and manipulated. This linear transform is initiated by drawing a rectangle around the sample in the image.
- 2.9 Adjusting Registration. Missized and often off-centered regions will then populate on the registration image and a window titled ROIManager will appear. Using this ROI Manager, select all regions (click on the region at the top of the list, hold the 'shift' key, then click on the region at the bottom of the list), next select "More... > OR(Combine)", then click "add(t)". Finally, click 'OK'.
One-by-One, regions will resize and come to the center of the image. Move, adjust, or delete the ROIs as needed - clicking 'OK' on the dialogue window once satisfied with each region. No need to adjust any dummy regions, but do not delete until after all regions have been adjusted.
With all regions adjusted, any ROIs which aren't essential to the analysis may now be deleted from the ROI Manager.
After adjusting all ROIs, select all ROIs in the ROI Manager window and save the ROI set using a name specific to that image. This will allow for the adjusted ROIs to be called upon at a later date should the user wish to apply different thresholding parameters or a different analysis type to the image.
OTHER NOTES
- 2.10 Settings.
There are a few settings which may be customized to better suit particular images. These can be adusted by dragging and dropping the .ijm file into ImageJ and editting lines as needed. The default settings upon download are set to:
- Thresholding: "Default"
- Threshold Background Colour: "Dark"
These settings are found in both the volumetric analysis and the object counts sections of the code and should be adjusted according to the analysis type.
Tip: If using an already binarized image, the following lines can simply be commented out using "//":
setAutoThreshold("Default dark"); //may comment out this line if using an already binary image
run("Convert to Mask"); //may comment out this line if using an already binary image
2.11 Outputs. After all images from the specified file range have been processed, output files will be appear in the parent folder. Depending on the analysis type, they will either be named VolumetricAnalysis.csv or RegionalCounts.csv. There is no sample identifier in this file name, so it is advised that the filename is edited before the user starts processing the next range of images to avoid overwritting the previous file.
A benefit to FASTMAP over other image registration tools is its atlas flexibility. With minimal up-front tracing, atlases can be generated for application to any image set.
Sample plates can be found here.
The process of generating a custom atlas is outlined in the following steps:
-
3.1 Collecting Reference Images. All atlases should be based off of something. This could be a subset of regions in an existing atlas or clearly delineated regions in an all-encompassing range of your own samples. These images do not need to be of any particular format or resolution, but should be clear enough for easy tracing of regions of interest.
-
3.2 Region Tracing. Open the first of the reference images using ImageJ. With the Polygon Tool, trace a region of interest. By hitting "t" on the keyboard, add this ROI to the ROI Manager applet. Rename this newly logged ROI with the name of the region.
Tips for region tracing:
- Clicking "Show All" in the ROI Manager applet is recommended.
- Ensure that regions are not overlapping at all during initial tracing
- If ROIs do not cover the entire X and Y range of the sample, add a dummy ROI as a space filler. This will ensure that regions resize correctly when running the plugin. For an example of this, see the example coronal plates.
Continue tracing and renaming until all regions of interest present in the reference image have been traced.
- 3.3 Saving Plates. With all regions traced, highlight all ROIs in the ROI Manager applet then click "More > Save". Plates should be saved to a designated atlas folder as "RoiSet_xx.zip", with xx being an identifier unique to that particular plate. Leading zeros are not required for single digit identifiers.
With regions saved, the current reference image can be closed and the next reference image can be loaded. ROIs aligned to the previous reference image can then be modified, deleted, or added to as needed.
- 3.4 Composite Atlas Image. When running the plugin, a composite image of all of the plates in the selected atlas populates on the screen and the user is asked to identify which plate most closely aligns with the image they are registering. This composite image can be generated using a wide variety of programs (Adobe Illustrator and Microsoft Office's PowerPoint both work really well for putting together quick composite images). The important notes with this composite are the following:
- Composite image files must be in .tif format. Images can initially be generated as a .PNG or .JPG before being converted to a .tif file in ImageJ
- The name of the composite image file must match the name of the atlas folder. For example, the sample mouse sagittal atlas is in a folder called "Sagittal", so the composite image is named "Sagittal.tif".
-4.1 Macro Error, Statement cannot begin with '<' in line X << DOCTYPE html >
(Updated June 02, 2022) Some users have been met with the above error message when attempting to run FASTMAP after installing the plug-in on a Windows device. In these cases, the error was resolved by following these steps:
- Users located the plugin folder in the file directory where their copy of FIJI was installed.
- Upon, inspecting the FASTMAP_.ijm file in this folder by dragging and dropping the file into the FIJI UI, it was noted that several lines of text had been appended to the file.
- The text was then deleted and replaced with copied code accessed from the appropriate GitHub page.
- The issue was resolved upon FIJI restart
-**4.2 ROIs reverting to original size after clicking outside of their bounds **
(Updated July 4, 2022) When using intricate atlas plate organizations, some users have experienced frequent issues wherein upon clicking outside of the bounds of the ROI it reverts to its original size. In the case where this occurs, one possible solution is to select the appropriate ROI in the RoiManager window and then use the commands in Edit>Selection>Scale to approximate the resizing of the ROI before continuing with the free-form ROI fitting. If FASTMAP users are finding that they are running into this issue frequently, it is recommeneded that they click ‘OK’ through all but the last region before making any adjustments. For example, when working with a plate which has 10 regions I would suggest clicking ‘Ok’ through the first 9 and then ‘Update’ in the RoiManager window when the 10th region is resized. This will allow you to morph each region without consequence. Afterwards, you can click 'Ok' on the final ROI and FASTMAP will continue processing.
Thank you to the FASTMAP users who came forward with these issues. Other troubleshooting procedures will be appended to this section should common issues arise.
-- Updated March 2, 2022 --
If you find FASTMAP to be useful, and apply it in your research, please cite the following article outlining this flexible open-source atlas registration tool:
FASTMAP: Open-Source Flexible Atlas Segmentation Tool for Multi-Area Processing of Biological Images Dylan J. Terstege, Daniela O. Oboh, Jonathan R. Epp eNeuro 28 February 2022, ENEURO.0325-21.2022; DOI: 10.1523/ENEURO.0325-21.2022
The manuscript outlining this tool can be found here.
Contributors:
- Dylan Terstege* (code/tool conceptualization/written documentation/atlas plates)
- received support from the Canadian Open Neuroscience Platform
@dterstege -
dylan.terstege@ucalgary.ca
- Daniela Oboh (atlas plates)
Principal Investigator:
- Jonathan Epp (tool conceptualization)
*corresponding author