Exploring hand wrist feature as biometric identifier: A deep learning perspective using Yolov11
Gokulakrishnan Elumalai, G Malathi · Results in Engineering · 2025
The integration of advanced biometric systems, such as fingerprint, palm, and wrist recognition, is becoming increasingly significant in modern security applications. However, it suffers from environmental limitations, such as spoofing risks, user discomfort, and there is a growing need for secure and reliable identity verification which has led to the exploration of alternative biometric modalities beyond traditional methods. To address these challenges, this study proposes the identification of surface characteristics in the hand wrist, such as bifurcations, crossings, ridge ends, and bridges, as distinctive biometric identifiers. These anatomical characteristics, which resemble fingerprint ridge patterns, present a promising direction for biometric recognition systems, especially in environments where conventional identifiers may be ineffective.The study introduces a novel deep learning framework tailored to detect hand wrist surface characteristics using a customized YOLOv11 architecture. In particular, we assess the detection capabilities of the proposed YOLOv11x model and benchmark its performance against the widely adopted YOLOv8 models. Experimental evaluations demonstrate that the custom YOLOv11x achieves a superior detection accuracy of 94.3% , indicating its robustness and accuracy in recognizing fine-grained biometric traits. The findings highlight the uniqueness and reliability of the features of the hand wrist for secure and contactless biometric authentication, and establish HWF-YOLOv11x as a promising framework for next-generation biometric systems.