Segmentation of Plantar Foot Thermal Images Using Prior Information.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Diabetic Foot Research

Background:

• Diabetic foot (DF) complications are linked to foot temperature variations.
• Early detection of thermal anomalies can help prevent DF ulceration.
• Accurate segmentation of plantar foot thermal images is crucial for reliable analysis.

Purpose of the Study:

• To develop a novel segmentation method for plantar foot thermal images.
• To create a decision support tool for preventing DF ulcers.
• To enable contactless, non-constraining thermal analysis for DF patients.

Main Methods:

• A new snake-based segmentation method incorporating prior shape information was developed.
• The method utilizes a novel energy term to guide curvature and a downsized prior-shape contour for initialization.
• Performance was evaluated against existing methods using a database of 50 thermal foot images.

Main Results:

• The proposed method significantly outperformed two classical prior-shape snake methods.
• Achieved a root-mean-square error of 5.12 pixels and a Dice similarity coefficient of 94%.
• Successfully segmented plantar foot regions for temperature difference assessment and hyperthermia detection.

Conclusions:

• The novel segmentation method provides accurate and robust plantar foot thermal image analysis.
• This technique facilitates the detection of hyperthermia, a precursor to diabetic foot ulcers.
• It represents a significant step towards mobile smartphone-based thermal analysis for DF management.