Academic project
Big Data Multibiometric
Research on preclassification of large-scale biometric databases using supplementary biometric attributes.
Big Data Multibiometric
This project (2015–2019) investigated approaches for narrowing the search space in identification systems that use large biometric databases. Its hypothesis was to employ supplementary biometric attributes (soft biometrics) to shortlist records before applying recognition techniques based on more discriminative characteristics, such as facial features or fingerprints.
The proposal considered visual attributes estimated from images, including appearance and movement. Such attributes cannot uniquely establish identity and may be subject to uncertainty, bias, and variable quality across populations and acquisition conditions.
The investigation concerned preclassification techniques, not guaranteed individual identification. Surveillance images and biometric data also raise questions about lawful collection, privacy, transparency, and discrimination. Available project documentation does not establish which safeguards were implemented in practice.