Documentation and analysis of plastic fingerprint impressions involving contactless three-dimensional surface scanning
MetadataShow full item record
Fingerprint impressions are frequently encountered during the investigation of crime scenes, and may establish a crucial linkage between the suspect and the crime scene. Plastic fingerprint impressions found at crime scenes are often transient and delicate, leaving photography the sole means of documentation. A traditional photography approach can be inadequate in documenting impressions that contain three-dimensional (3D) details due to the limitations of camera and lighting conditions on scene. In this study, 3D scanning was proposed as a novel method for the documentation of plastic fingerprints. Structured-light 3D scanning (SLS) captures the distortion of projected light patterns on the subject to obtain its 3D profile, which allows fast acquisition of the complete 3D geometric information of the surface. The contactless operation of SLS also eliminates the risk of destroying fragile evidence, making it a sound choice for forensic applications. In this study, the feasibility of 3D scanning of plastic fingerprint impressions was evaluated and compared with traditional photography regarding the quantity and quality of perceptible friction ridge features. Attempts were made to develop a procedure to extract curvature features from 3D scanned fingerprints and flatten the friction ridge features into two-dimensional (2D) images to allow direct comparison with the traditional photography method in the CSIpix® Matcher and NFIQ 2.0 software. One of the developed methods (3DR) utilizing a discrete geometry operator and convexity features outperformed traditional photography, both in minutiae count and match quality, while traditional photography could not always capture enough high-quality minutiae for comparisons, even after digital enhancement. The reproducibility of the 3D scanning process was evaluated using 3D point cloud statistics. The pair-wise mean distance and standard deviation were calculated for four levels of comparisons with theoretically increasing disparity, including pairs of scans of the same impressions. The results showed minimal shape deviation from scan to scan for the same impression, but high variations for different impressions.
RightsAttribution-NonCommercial-ShareAlike 4.0 International