Custom Classification
- Updated2026-02-18
- 1 minute(s) read
You can define a custom feature extraction process for specific machine vision applications using Vision.
When to Use
Typical applications include sorting and inspection applications for which you can define a feature descriptor to represent the different classes in a specific application. Examples of such feature descriptors include statistics about the grayscale pixel distribution in an image, measurements from a Vision gauging tool, or color spectra from Vision color learning algorithms.
Concepts
With custom classification, you create a classifier by training it with prelabeled training feature vectors. NI Vision custom classification uses the same classification algorithms as the Particle Classifier, including the Minimum Mean Distance, Nearest Neighbor, and K-Nearest Neighbor classifications.