The principle of image processing is based on a one-way passage through the modules, where each module processes the image and sends new information to the next module. In some cases, a single module may be sufficient. The order of modules is called Flow and can be configured in the ‘Modules’ tab by clicking the ‘Edit’ button.
In the edit mode, you can add modules by using the ‘+’ button on a desired position in the Flow, which shows the following pop up window containing brief information about each module. You can also remove, rename or add notes to existing modules in the Flow.
Passing through the modules has top to bottom direction. The passing can be branched for parallel processing. Models communicate with each other via context.
The following modules are available
- Anomaly of Surface – searches for anomalies in the image
- Surface Detection – surface recognition
- Detector – searches for objects (and optionally classifies them)
- Classifier – object classification
- Measure - measuring dimensions of objects
- Preprocess – tool for easy image editing before the next processing
- Mask – covering parts of the image to be ignored
- Unifier – unifies the position and rotation of the objects in the images for further processing
- Code – image processing by a custom code
Artificial Intelligence modules are trained based on user-annotated models. The user then selects which model should be used for the module.