A morphological comparison framework for writer's palm impressions based on white line intersections and bifurcations
Year: 2026
Type: article
Abstract: Writer's palm impressions, deposited from the ulnar-lateral aspect of the palm, are encountered in forensic casework but have received limited systematic attention as a distinct trace type. This region represents a transitional area between friction ridge skin and non-ridge skin, where conventional friction ridge details such as cores and deltas are infrequently observed in practice, limiting the direct applicability of fingerprint-based comparison systems. This study proposes a preliminary interpretive framework for morphological comparison of writer's palm impressions by defining intersections and bifurcations of white lines as morphological feature points and establishing six comparison criteria. To examine applicability under deformation, a baseline image similarity algorithm (ORB) and an evaluation by forensic science specialists were applied. ORB maintained stable correspondence under mild transformations but showed limited ability to retrieve corresponding impressions for flexed writer's palm impressions involving strong non-linear deformation, with only 2 of 40 flexed impressions returning the correct match within the top five results. In contrast, all seven forensic science specialists successfully identified the corresponding impression by interpreting relative spatial relationships among a small number of white line feature points with reference to the distal transverse crease. These observations suggest that white line-based morphological features may be feasible for structured interpretive comparison of writer's palm impressions under deformation conditions that challenge automated matching approaches, particularly where conventional friction ridge details are limited or absent. The proposed framework is intended not as a definitive comparison system, but as an initial framework to support further investigation.
Source: Journal of Forensic Sciences
Institution: Sungkyunkwan University
Language: English
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
Funder: National Forensic Service
Award: G9001396265
OpenAlex ID: W7234729076
