Preliminary evaluation of gunshot residue pattern analysis using flash-pulse infrared thermography and multi-task...

Area of Science:

• Forensic Science
• Computer Science
• Materials Science

Background:

• Traditional gunshot residue (GSR) analysis for forensic reconstruction, especially shooting distance estimation, faces challenges including time consumption, cost, and subjective interpretation on diverse substrates.
• Existing methods lack a comprehensive approach to simultaneously analyze multiple ballistic attributes from GSR patterns.
• No prior research has integrated flash-pulse infrared thermography with multi-task deep learning for ballistic analysis.

Purpose of the Study:

• To investigate the feasibility of using flash-pulse infrared thermography and multi-task deep learning for ballistic screening.
• To simultaneously predict multiple ballistic attributes, including weapon category, firearm model, and ammunition type, from thermographic deposition patterns.
• To estimate shooting distance using the developed deep learning model.

Main Methods:

• A proof-of-concept study utilizing a dataset of 312 ballistic samples on textile targets under controlled laboratory conditions.
• Development of a hybrid multi-task neural network integrating flash-pulse infrared thermography data.
• Training the model to classify weapon category, firearm model, and ammunition type, and estimate shooting distance.

Main Results:

• The multi-task deep learning model achieved high accuracies: 94.3% for weapon category, 85.1% for firearm model, and 92.6% for ammunition type.
• Shooting distance was estimated with a mean absolute error of 7.92 cm.
• The proposed approach demonstrated superior performance compared to traditional machine learning baselines and single-task neural networks.

Conclusions:

• Combining flash-pulse infrared thermography with multi-task deep learning presents a viable, rapid, non-destructive, and objective complementary tool for ballistic screening and shooting distance assessment.
• The method shows potential for enhancing forensic analysis by providing objective data from thermographic residue patterns.
• The study is a preliminary screening approach, requiring further validation across a broader range of firearms, ammunition, and textile conditions before consideration as a confirmatory method.