Sample Compression and Encryption in Biometric Systems

Contrasting to most current biometric systems which store template data (i.e. extracted features) it will become desireable in future systems to store sample data. One reason is to enable to change to a different set of features without requiring all users to re-enroll.When storing sample data the larger data amount requires the samples to be stored in compressed form. A second motivation for sample data compression is the application in distributed biometric systems where the sensor data acquisition is dislocated from data processing and feature extraction. The transmission over possibly weak network links is obviously more efficient in compressed form. In case the sensor data is stored as sample data instead of template data, privacy issue become even more important and the protection of the data may be accomplished employing sample data encryption. The current project deals with compression and encryption of fingerprints, face, iris, and retina image data. The aim is to develop compression schemes significantly outperforming the current baseline ISO/IEC 19794 standard by using JPEG2000 Part 2 and JPEG extended System technology. With respect to encryption, we will apply techniques from the multimedia security area to sample data encryption. The aim is to result in significantly lower computational complexity as compared to full AES encryption.