Martin Böckeler has successfully on April 20, 2015 defended his Master thesis titled `Automated fingerprint quality assessment as a replacement for dactyloscopic expert assessment`, after working with da/sec in period October 2014 to April 2015.
The popularity of biometrics heavily increased in the last decades. A wide range of commercial applications like border control, access control, verification and identification of human beings are among convenience applications such as online identification only a fraction of the wide use case scenario of biometrics. One of the most explored, permanent, unique and widely accepted biometric modalities is the fingerprint. In the forensic sector fingerprints are of special interest as they are often left unintentionally as latents on objects at crime scenes. Gathering and reconstructing fingerprints with magnetic powder or superglue fuming is useful to reconstruct who was at the scene. To use the reconstructed latents in court cases for identification, they have to go through an analysis, comparison, evaluation, and verification (ACE-V) process that heavily relies on subjective dacyloscopic expert opinion which is itself influenced by human factors like health problems, stress or inadequate training. Various studies had shown that the fingerprint quality computed by an algorithm highly correlates with its comparison score. Nonetheless, no known study has compared the correlation between automatic and dactyloscopic expert fingerprint quality assessment. This thesis addresses the lack, explores correlations between automatic and expert fingerprint quality assessment and further investigates if dactyloscopic expert assessments can be predicted.