Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{202021,
        author = {Siddhesh Chavan and Aditya Chavan and Dhruv Daberao and Kshitij Pokharapurkar and Naman Buradkar},
        title = {Diabetes Detection using Thermal Footprints: AnAI-Driven Framework for Early Risk Stratification of Diabetic Foot Complications},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {5833-5840},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=202021},
        abstract = {Diabetic foot ulceration (DFU) is one of the most severe and costly complications of diabetes mellitus, often progressing to infection, hospitalization, and lower-limb amputation when risk is not detected early. This project proposes an AI-assisted diabetic detection framework using plantar thermal footprints, motivated by evidence that abnormal plantar temperature distributions and asymmetry appear before visible tissue breakdown. The proposed pipeline integrates standardized thermal image acquisition, preprocessing (denoising, contrast normalization, foot alignment, and region-of-interest segmentation), handcrafted thermal descriptors (including Thermal Change Index and inter-foot asymmetry), and machine/deep learning classifiers for risk stratification. The design is grounded in published thermographic studies and benchmark datasets, especially the Plantar Thermogram Database. Based on findings reported in the analyzed literature, feature-based models (e.g., AdaBoost over optimized thermal features) and lightweight convolutional models (e.g., MobileNetV2 on enhanced thermograms) achieve strong discriminative performance, with F1-scores around 97% in diabetic-vs-control classification settings. The proposed system prioritizes non-invasive screening, computational efficiency, and practical deployment at clinics/community settings, while acknowledging current limitations such as small public datasets, domain shift across sensors, and limited longitudinal labels for ulcer progression. This work contributes a structured BE-level implementation blueprint and a clinically motivated roadmap toward proactive, scalable DFU risk monitoring.},
        keywords = {Diabetic Foot Ulcer, Infrared Thermography, Thermal Footprints, Machine Learning, Deep Learning, Early Detection, Biomedical Signal Processing.},
        month = {May},
        }