Self-DFUNet Thermogram Analysis for Diabetic Foot Ulcer Stratification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting and quantifying the risk of diabetic foot ulcers from thermal images are limited by their inability to provide accurate, real-time, and non-invasive solutions, especially for patients with obesity or visual impairment, and lack effective severity grading for early intervention.
Innovation Solution
A machine learning model, Self-DFUNet, utilizing thermogram images from IR cameras, which includes a Self Organized Operational Neural Network (SelfONN) and inverted bottleneck residual blocks, classifies diabetic foot ulcers into Mild, Moderate, and Severe categories, enabling early detection and intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used for detecting diabetic foot ulcers, then the detection process is simple, but the accuracy and real-time capability are insufficient
Solution Approach 1:
The patent replaces traditional mechanical/visual inspection methods with a thermal imaging system that detects infrared radiation from the foot. This substitution enables non-contact, real-time temperature mapping that significantly improves detection accuracy for early diabetic foot complications while maintaining operational simplicity through automated image analysis
Solution Approach 2:
The system transitions from visual inspection to thermal parameter detection by measuring temperature distributions across the foot surface. This parameter change enables quantitative assessment of tissue health through temperature patterns, providing early detection capability without requiring complex invasive procedures
2Measurement precision
If invasive detection methods are used, then accurate diagnosis can be obtained, but patient comfort and accessibility are reduced
Solution Approach 1:
The patent replaces invasive physical examinations with non-contact thermal imaging technology. The system captures infrared emissions from the foot surface to create temperature maps, providing accurate diagnosis of underlying tissue conditions without requiring physical contact, needle insertion, or complex patient preparation procedures
Solution Approach 2:
The system creates a thermal copy or map of the foot's temperature distribution pattern. This thermal fingerprint serves as a diagnostic indicator that accurately reflects underlying tissue health without requiring direct manipulation or invasion of the patient's body, enabling repeated measurements for monitoring purposes
3Loss of time
If early detection systems are implemented, then intervention timing is improved, but the complexity of analysis increases
Solution Approach 1:
The system performs preliminary thermal mapping and pattern recognition to identify early signs of diabetic foot complications before clinical symptoms manifest. By detecting abnormal temperature patterns in the pre-ulcerative stage, the system enables timely intervention while using automated analysis algorithms to manage the complexity of early detection criteria
Solution Approach 2:
The thermal imaging system incorporates automated image processing and analysis capabilities that independently evaluate temperature patterns and generate diagnostic assessments. This self-service functionality reduces the burden on healthcare providers by performing preliminary screening and triage, allowing early detection without requiring complex manual analysis protocols
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The model provides a highly accurate and real-time severity classification of diabetic foot ulcers, outperforming existing non-invasive solutions, allowing for timely intervention and reducing the burden on healthcare systems by enabling remote monitoring and risk stratification.
Implementation Method 1
thermal images of the patient's feet
Implementation Method 2
thermogram images
Data Source
AI summary
A machine learning model for stratification of early diabetic foot complication using thermogram images is provided that is operable to predictively diagnose a risk for diabetic foot ulceration formation via receiving a thermogram of a foot; identifying, via a machine learning model, a risk factor of diabetic foot ulceration on the foot; and outputting, from the machine learning model, the risk factor as a diagnosis.