Biometric System Enhancements

Securics believes in the failure prediction paradigm. This provides a uniquely effective core for fusion of multiple algorithms or modalities, and significantly enhances recognition system performance. This fusion via failure prediction is comparable to, or better than, previously published fusion techniques while utilizing less data to achieve its results.

The utility of multi-modal failure prediction is clear - if one modality has failed, we can fuse information from another modality that has succeeded, and achieve high recognition performance.  From a pure quality assessment for fusion, the failure condition might not be evident.  We encourage an the alternative to quality assessment especially for systems and problems that can benefit from a per instance prediction of recognition failure.