Thermal Time Constant CNN-Based Spectrometry for Biomedical Applications.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Artificial Intelligence

Background:

• Accurate thermal property estimation is crucial for diagnosing skin conditions.
• Traditional methods for analyzing thermal data can be complex and time-consuming.

Purpose of the Study:

• To develop a novel deep learning approach for recovering thermal time constants.
• To utilize these constants for the pathological detection of skin states.

Main Methods:

• A convolutional neural network (CNN) was employed to model the thermal system as a Foster Network.
• The CNN learns to retrieve time constants and amplitudes from temperature-time curves.
• The method was validated using simulated data and real thermographic signals.

Main Results:

• The deep learning system accurately estimates thermal time constants and temperature profiles.
• Recovered time constant errors were below 1% for ideal data and under 5% for noisy signals.
• The method demonstrated effectiveness in analyzing thermographic data from a psoriasis patient.

Conclusions:

• The proposed CNN-based method offers a robust and accurate way to extract thermal parameters.
• This technique shows significant potential for non-invasive skin disease diagnosis using thermal imaging.
• The findings support the use of AI in quantitative thermal analysis for medical applications.