Full-field burn depth detection based on near-infrared hyperspectral imaging and ensemble regression

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Computational Biology

Background:

• Accurate burn severity diagnosis is vital for effective wound management and clinical treatment.
• Visual inspection for burn depth assessment lacks quantitative measurement and accuracy.
• Current methods for burn depth evaluation present significant limitations in clinical practice.

Purpose of the Study:

• To develop and validate a full-field burn depth detection system.
• To utilize near-infrared hyperspectral imaging (HSI) combined with ensemble regression for quantitative burn depth analysis.
• To provide a more accurate and practical tool for clinicians in assessing burn severity.

Main Methods:

• Proposed a full-field burn depth detection system utilizing near-infrared hyperspectral imaging (HSI).
• Introduced rotational feature subspace ensemble regression to model the complex relationship between HSI data and burn depth.
• Validated the system through in vivo measurements on a porcine model.

Main Results:

• The developed system achieved an average relative error of approximately 7% in burn depth measurement.
• Demonstrated the capability of the method for accurate, full-field assessment of burn depth.
• The system provides quantitative data, overcoming limitations of visual inspection.

Conclusions:

• The proposed near-infrared hyperspectral imaging system with ensemble regression offers an accurate method for burn depth assessment.
• This technology provides a quantitative and practical reference for clinicians, improving burn wound management.
• The system has the potential to significantly enhance the diagnosis and treatment of burn injuries.