The concept of BioID is the simultaneous integration of several different biometric channels that can be extracted from various devices. Two (or more) channels may share the same input device, but the preprocessing which performs the feature extraction is individual.

Currently, from the video device, face (static) and mimic features (dynamic) can be used. The audio device offers voice features.

After recording each biometric feature is preprocessed separately. In the preprocessing step the biometric features are extracted from the recordings.

The extracted features are used to calculate a user's biometric template (training), which afterwards is used in identification or verification tasks (classification). Both, training and classification, are performed by one object, the classifier. Equivalent to preprocessing, every channel works with an individual classifier.

In the classification case every channel's classifier produces exactly one classification result. The results of the classifiers are combined by a special strategy - the fusion - depending on the level of security.

This concept ensures that BioID is modular and extensible. The combination of used channels easily can be adopted, depending on the application. Additional biometric channels (e.g. iris recognition) can be integrated without any problems.