Our Software Step-by-Step

 

SETUP OF A CLASSIFICATION TASK

 

Before the software is able to classify automatically, a few settings have to be adjusted once, using some sample images. Compared to conventional software, this is very easy and intuitive to achieve when working with Falco and Iris.

 

1. Setup of the Object Generation

  • Load a sample image (1a).
  • Define the background with a few clicks.
  • Choose the image pre-processing settings for your application.
  • Save the settings.

The software is now able to generate the objects automatically for the sample image (1b) as well as for other images (when recorded under the same circumstances).
Each object can now be selected by the user, simply by clicking on it – visualized by a frame around its contour (1c).

step01

1a sample image; 1b object generation; 1c objects are marked when clicked

 


 

2 . Setup of the Classifier

  • Load one or several sample images for your different classes (2a).
  • Choose the relevant features from a list.
  • Choose the classifier type from a list.
  • Define the number of classes (in this example: seven), their color and names.
  • For each class, choose several sample objects from the images, simply by clicking on them. They are marked in their respective class color (2b).

The software will now define the suitable classifier automatically by means of machine learning.

  • Save the classifier.
2

2a sample image; 2b sample objects and their class colors

 


 

3. Classification

Once the setups of the Object Generation and the Classifier have been completed, you can start a classification.

  • Load the respective classifier.
  • Define the image source (e.g. single image from harddrive, RGB-camera in real-time etc.; 3a)
  • Start the classification.

The software will output the object stream (feature data of each object) and a visulization of the results (3b).

3

3a image source; 3b classification result

 

 

To see some examples of what our software can achieve visit the section Examples.