A Novel Approach to Optimize Crime Investigation Process Using Palmprint Recognition

Abstract

Latent partial palmprint recognition is of vital importance in forensics and law enforcement because more than 30% of evidences from crime scenes are palms. Manual matching of palmprints is an exhaustive and time consuming task compared to fingerprint matching because a palm is a large area compared to a fingerprint and most often only partial prints are recovered. Experienced latent examiners can determine the part of palm a partial latent print is from. This information reduces the manual palmprint matching time and increases the matching accuracy. This research introduces an approach to optimize the current process used to identify a latent palmprint. The optimization was achieved by the segmentation methodology which was proposed after a statistical data analysis and the NIST biometric image software. The proposed approach was evaluated using 60 partial palmprint against a gallery set of 60 full palmprint based on efficiency and effectiveness and compared with the current process and with another approach. Evaluation results show that the proposed approach is enhanced the current process both in accuracy and speed.