Computer-Aided Detection of Diabetic Foot via Infrared Imaging using Machine and Deep Learning Approaches: A Survey

Abstract

Diabetes mellitus (DM) is a major health problem and the most prevalent worldwide. Diabetic foot (DF) is the most common complication of DM, which can lead to death, amputation, and plantar ulcers. Early detection of these complications protects the diabetic patient from dangerous stages that lead to amputation. This study explores the effectiveness of computer-aided diagnostic systems, particularly those that utilize the power of artificial intelligence (AI) and deep learning (DL). AI and DL may provide promising means of early detection and diagnosis of DF complications. In addition, these have the potential to revolutionize patient care by providing tools that can analyze complex medical data with remarkable accuracy and speed. Using thermal imaging is an innovative approach that has recently gained more attention. Infrared thermal imaging captures heat emanating from the body, providing a non-invasive way to detect abnormal plantar temperatures that indicate underlying inflammation or infection. Machine learning (ML) and DL classification techniques improve the effectiveness of computer-aided detection (CAD) of DF. By training algorithms on huge data sets, these systems can learn to identify patterns and anomalies that otherwise would elude human detection. This study explores several ML techniques, with a particular emphasis on DL classification to accurately identify the feet of diabetic patients. The findings from this research will contribute to future studies aimed at improving detection processes and helping medical professionals deliver timely and effective care to their patients. By automating the initial screening process, healthcare providers can prioritize patients who need immediate attention. This will improve resource allocation and potentially reduce the incidence of serious complications such as ulcers and amputations.