Using 2D video analysis and model based image matching to measure joint angles for forensic biomechanical analysis.

Area of Science:

• Forensic science
• Biomechanics
• Injury analysis

Background:

• Forensic injury biomechanics relies on comprehensive data analysis for incident investigations.
• Video recordings offer objective and rich data for forensic analysis.

Purpose of the Study:

• To compare the accuracy of different analysis techniques and camera views for measuring human joint angles.
• To evaluate reverse projection and model-based image matching (MBIM) for joint angle measurement.

Main Methods:

• Eight static human postures were recorded using four cameras.
• One camera was in-plane with movement; three were out-of-plane.
• Joint angles (knee, elbow) were measured using a goniometer and compared to reverse projection and MBIM results.

Main Results:

• Multi-camera solutions and single, in-plane camera views yielded the most accurate joint angle reconstructions.
• Single out-of-plane camera views were less accurate.
• No significant accuracy difference was found between MBIM and reverse projection techniques.

Conclusions:

• Multi-camera and in-plane views are crucial for accurate joint angle measurement in forensic biomechanics.
• Both MBIM and reverse projection are viable techniques for joint angle analysis.
• Accurate joint angle data supports robust forensic injury investigations.