<![CDATA[Reliable human identification is a challenging area in forensic science especially where conventional biometrics are not available. This is addressed in this study by coming up with a new multimodal system. The study contribution is the Adaptive Convergence-Triggered Mutation Particle Swarm Optimization (ACTM-PSO) algorithm of optimal fusion regarding tongue and dental biometrics. The combination leverages durability regarding dental structure as well as the specialized texture of tongue prints, which in most cases preserved in post-mortem cases. The technique extracts the texture and morphological features of tongue and dental images. The proposed ACTM-PSO, which has non-linear adaptive inertia weight and stagnation-triggered mutation, optimizes the weighting of features for fusion before classification using an SVM. Experimentally, the system attains accuracy of 97.3% and low Equal Error Rate (EER) of 2.1% in comparison to traditional PSO (88.3% accuracy) and unimodal systems. It has an accuracy of more than 95% in rotations of the image (0°, 90°, 180°, 270°), which is forensically practical. The False Match Rate (FMR) and the False Non-Match Rate (FNMR) are 2.0% and 3.3%. This paper provides a strong optimization based backbone that goes ahead to develop multimodal biometric integration to be used in a forensic application to increase the reliability of evidence in legal contexts.]]>
Copyrights © 2026
