Publication: 3D computerised forensic facial reconstruction using blender
No Thumbnail Available
Date
2026-02
Authors
Shyan, Lee Sy
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This research focuses on the three-dimensional (3D) artistic reconstruction of facial appearance from unknown human skulls, examining the relationship between cranial morphology and overlying soft tissues that define facial form. Facial reconstruction remains a critical approach in forensic identification when conventional methods fail, and no antemortem information is available. The primary aim of this study was to develop and evaluate a 3D digitised facial reconstruction methodology using Blender software and to assess its performance through skull–face comparison. An additional objective was to assess the facial similarity between the 3D reconstructed face and the 2D reconstructed face from previous studies (Kanesa, 2020; Nurul Hikmah, 2022). Cicero Moraes’s reconstruction framework was adopted as the primary methodology and was further modified using updated and established techniques for individual facial features, including the eyes, nose, ears, and lips, to improve reconstruction accuracy. A control study was first conducted using two computed tomography (CT) skull scans, and the resulting faces were compared with corresponding facial photographs to validate and refine the reconstruction workflow. The revised methodology was then applied to five unknown physical skulls as experimental samples. The reconstruction workflow comprised skull acquisition and analysis, skull digitisation, importation into Blender, placement of Facial Soft Tissue Thickness (FSTT) markers, tissue modelling, individual facial feature modelling, and final rendering. The reconstructed faces were evaluated using facial superimposition, fade and wipe mode examination, and facial mapping to assess the reconstruction performance. Additional comparisons were made with previously published two-dimensional (2D) facial reconstructions to examine the facial similarity between 2D and 3D methods. The results demonstrated that the proposed 3D workflow enabled consistent placement of facial features and improved control over facial proportions and surface morphology. The use of a complete population-specific FSTT dataset enhanced anatomical coherence. Nevertheless, discrepancies were observed in facial similarity examination due to methodological variation, biological variability, and practitioner-related factors, including artistic subjectivity and the cross-race effect. In conclusion, this study demonstrates the viability of Blender as a platform for 3D forensic facial reconstruction and highlights the importance of methodological standardisation, population-specific data, and practitioner training in improving reconstruction accuracy and reliability.