FeatureJ is a powerful suite of plugins for ImageJ and Fiji designed to extract differential geometric features from multi-dimensional grayscale images. Developed by Erik Meijering, it allows bioimage analysts and computer vision researchers to easily compute image derivatives, find edges, and extract structural elements.
Here is everything you need to master the fundamentals of FeatureJ in less than 10 minutes. 🔌 Quick Installation (1 Minute)
Getting FeatureJ ready takes just a few clicks depending on your primary software base:
If using Fiji: Go to Help > Update…, click Manage Update Sites, check the box for ImageScience, and click Apply changes.
If using vanilla ImageJ: Manually download FeatureJ_.jar and imagescience.jar from the official ImageScience FeatureJ Page and drop them into your plugins folder.
Accessing the tool: Restart your software and navigate to Plugins > FeatureJ. You can open the floating FeatureJ Panel to keep all options one click away. 🛠️ Core Tools and How to Use Them (6 Minutes)
The suite is broken into several primary image processing functions. For all of these tools, the key setting is the Smoothing Scale (
), which represents the standard deviation of the Gaussian kernel. A smaller scale captures fine, noisy details, while a larger scale smooths out noise but might obscure subtle transitions.
┌────────────────────────────────────────────────────────┐ │ FEATUREJ SUITE │ ├──────────────┬──────────────┬────────────┬─────────────┤ │ Derivatives │ Laplacian │ Hessian │ Edges │ │ (Gradients) │ (Transitions)│ (Shapes) │ (Canny) │ └──────────────┴──────────────┴────────────┴─────────────┘ 1. FeatureJ Derivatives
What it does: Computes multi-dimensional, Gaussian-scaled image gradients.
When to use: When you need the underlying math for custom filter designs or mapping directionality.
How to use: Specify the differentiation order (up to 10) for X, Y, or Z. Setting an order to 0 applies only Gaussian smoothing to that dimension. 2. FeatureJ Edges
What it does: Executes a highly precise Canny Edge Detection algorithm.
When to use: Highlighting boundaries or cell walls in microscopy images. How to use: Input your smoothing scale (
) to control noise suppression, then apply the lower and upper hysteresis thresholds to lock onto true boundaries. 3. FeatureJ Laplacian
What it does: Highlights rapid areas of intensity change and maps zero-crossings.
When to use: Advanced computer vision tasks mimicking human visual pathways.
How to use: Run the plugin to output a secondary derivative map. Check the option to detect zero-crossings to create a binary mask where sign changes occur. 4. FeatureJ Hessian
What it does: Calculates the eigenvalues of the Hessian matrix for every individual pixel or voxel.
When to use: Discriminating between structural shapes like plates, lines/vessels, or blobs.
How to use: Select the plugin, choose your scale, and FeatureJ will generate eigenvalue images. Small values across all eigenvalues mean a flat background, while high values on specific axes pinpoint ridges or points. 5. FeatureJ Statistics
FeatureJ: An ImageJ Plugin Suite for Image Feature Extraction
Leave a Reply