author: Gherardo Varando (gherardo.varando@gmail.com)

The use of obj detection toolset require a basic knowledge of ImageJ/Fiji. Please refer to its online documentation for further explanations (http://imagej.net/Introduction).

Dependencies and Requirements

obj detection toolset is an extension for ImageJ, so obviously a version of ImageJ is required. It is recommend to install FIJI, which, citing its website, is a “batteries-included” distribution of ImageJ, bundling a lot of plugins which facilitate scientific image analysis. Moreover, obj detection toolset require two plugins that are not available through the ImageJ updater and have to be manually installed. That is to simply drag-and-drop the .jar files in the plugin directory of the ImageJ installation (http://imagej.net/Installing_3rd_party_plugins). Both plugins, ObjCounter and MaxLoGs filter, are available through github:

The toolset and all the dependencies are tested with ImageJ 1.51g and Java 1.8 but they should be compatibles also with older versions of ImageJ running on Java 1.6.


obj detection toolset is distributed as an imageJ custom toolset, that is a set of macros accessible via buttons in the tool bar (https://imagej.nih.gov/ij/docs/guide/146-20.html). To download the toolset files clone or download the entire repository:

To install the toolset copy the obj_detection_toolest.ijm file in the ImageJ/macros/toolsets/ directory and the folder obj_detection in the plugin directory. After restarting ImageJ the toolset can be made visible with the More Tools Menu (usually on the right of the ImageJ interface). The toolset actions are also available through the Plugins –> obj detection menu.

First run

After the installation open ImageJ/FIJI and click the More Tools Menu image and select obj detection toolset. Now you can access the toolset buttons in the ImageJ tool bar image.

You can now open an image or a stack and count objects. If you have FIJI installed you can open the blob sample image in File –> Open Samples. Then click in the first button of the toolset image for detection and counting of the objects in the image. You can try different choices for the parameters of the algorithm (e.g. Rmin=1, Rmax=4, By=1, ThrMethod=Moments, Min=0 ,Max=0, Toll=0 and if you are using a stack Fraction=0.45).

Parameter choice menu for **obj detection image**<span data-label="fig:parameter"></span>

While the algorithm is running the various operations will be shown in the ImageJ status bar, and when the algorithm is tagging and counting the objects a text window with the results will appear with the output of the ongoing operations. When the algorithm finish all the operations the outputs will appear.


As explained in the ObjCounter plugin help page it is better to change some Java configuration parameter to avoid thread stack overflow due to recursive calls. Java options are defined in the ImageJ.cfg file in the main directory of your ImageJ installation (see ImageJ help on example of Java configurations for ImageJ ). The thread stack size can be changed with the option (-XssSIZE) where (SIZE) is the desired size of the thread stack. Refer to Oracle documentation for a complete explanation of the Java options(unix, windows.

Basic use

In this section we explain the basic usage of the toolset.

Object detection

The object detection work-flow can be accessed with the relative button image once the obj detection toolset is active or with the Plugins –> obj detection –> obj detection image menu entry. The user is asked for the following set of parameters:

Parameter selection in **obj detection image**<span data-label="fig:guiDetection_1"></span>

An image is needed to be able to run the algorithm, otherwise a warning message is shown. The image is considered to represent bright object on a dark background, that is objects have in general higher pixels values than background.

Schematic of the object detection workflow<span data-label="fig:schema"></span>

  1. MaxLoGs filter of the original image with parameters Rmin,Rmax,By.

  2. Segmentation of the MaxLoGs filter with an automatic threshold method (ThrMethod) in foreground and background pixels.

  3. Denoising of the original image with a two-scale median filter.

  4. Retaining of the foreground pixels.

  5. Object tagging with ObjCounter plugin using parameter Min, Max, Fraction, Toll.

Output of **obj detection image**<span data-label="fig:guiDetection_2"></span>

When the algorithm finishes the operations the results are shown. In particular four more windows will open:

Validation with overlay

If in the image an overlay is present the ObjCounter algorithm will try to validate the object detections comparing it with the overlay. The overlay should be composed by points indicating approximately the center of the objects, thus one point for every object should be present (Be careful in multi slice stack of images, in this case even if an object span through different slices there must be just one point in one of the slices). If a polygon or square selection is present in the overlay the validation will be restricted to the selected region. The result of the validation is shown in the log text window and it reports the number of points of the overlay, the number of objects detected by the algorithm, the precision, recall and F1 measure (https://en.wikipedia.org/wiki/Precision_and_recall).

Example of overlay for validation in `blobs.gif` sample image<span data-label="fig:overlay"></span>

Batch processing

The toolset offers the possibilities of batch processing large number of images with the Plugins –> obj detection –> obj detection folder command or with the respective button image. All the images (and nothing more) have to be placed in a single folder, the user is then asked to select the parameters as in the obj detection image and additionally:

In the output folders the algorithm will write the output files as soon as they have been computed, working one image at a time. ImageJ will be unavailable for additional operations until the end of the process.



This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.