Accurate hand tracking is crucial in various applications, from rehabilitation to human–computer interaction. This study presents a comparative evaluation of two distinct hand-tracking systems, Leap Motion Controller version 2 and StretchSense ProFidelity gloves, aimed at assessing their performance in capturing hand kinematics. The systems were integrated within a Unity-based framework for synchronized data acquisition, followed by quantitative analysis in MATLAB. Twenty participants, clustered according to hand size (small, medium or large), performed standardized hand gestures (pinch, fist, and finger abduction/adduction) and object-interaction tasks. Spatial accuracy and temporal fidelity were evaluated using metrics such as joint angles, fingertip positions, and finger curvature during grip movements. Results indicate that neither system is universally superior; rather, their suitability depends on task requirements, such as the need for spatial precision, tolerance to occlusions, or ease of deployment. The Leap Motion Controller demonstrated high accuracy in non-occluded conditions and minimal setup time but was sensitive to occlusions, mainly during object-interaction tasks. StretchSense gloves offered discrete performance under occlusion but require time-intensive calibration and were susceptible to sensor drift and glove displacement.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Comparative Evaluation of Hand Tracking Systems: Gloves vs. Optical Sensors

  • Luca Rota,
  • Samuele Sonzogni,
  • Ivana Mostachetti,
  • Daniel Lanzoni,
  • Daniele Regazzoni

摘要

Accurate hand tracking is crucial in various applications, from rehabilitation to human–computer interaction. This study presents a comparative evaluation of two distinct hand-tracking systems, Leap Motion Controller version 2 and StretchSense ProFidelity gloves, aimed at assessing their performance in capturing hand kinematics. The systems were integrated within a Unity-based framework for synchronized data acquisition, followed by quantitative analysis in MATLAB. Twenty participants, clustered according to hand size (small, medium or large), performed standardized hand gestures (pinch, fist, and finger abduction/adduction) and object-interaction tasks. Spatial accuracy and temporal fidelity were evaluated using metrics such as joint angles, fingertip positions, and finger curvature during grip movements. Results indicate that neither system is universally superior; rather, their suitability depends on task requirements, such as the need for spatial precision, tolerance to occlusions, or ease of deployment. The Leap Motion Controller demonstrated high accuracy in non-occluded conditions and minimal setup time but was sensitive to occlusions, mainly during object-interaction tasks. StretchSense gloves offered discrete performance under occlusion but require time-intensive calibration and were susceptible to sensor drift and glove displacement.