Automatic Detection of Infection in Diabetic Foot Ulcer Images Using Improved-CNN-SVM Approach

• Published in: 2023 7th International Conference on Electronics, Communication and Aerospace Technology (ICECA) pp. 505 - 510
• Main Authors: G, Sivashankar, Avinash, Bhagyalakshmi, Shabbir Alam, Mohammad, Vyankatesh Ghamande, Manasi, Gupta, Keerat Kumar, Rastogi, Ravi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 22.11.2023
• Subjects: Diabetes; Diabetic Foot Ulcer (DFU); Feature extraction; Gaussian Filter; Image color analysis; Image segmentation; Medical services; Noise reduction; Red Green Blue (RGB); Training
• DOI: 10.1109/ICECA58529.2023.10395225

Unregulated glucose levels in the blood lead to the chronic disease diabetes. Diabetic foot ulcers (DFUs) and other devastating outcomes may be avoided with early detection. The lower limb of a diabetic patient may need to be amputated if they experience a DFU. DFU is difficult to diagnose and usually requires a number of expensive and time-consuming clinical investigations for the treating physician. Applying deep learning, machine learning, and computer vision techniques in today's age of data deluge has resulted in a number of solutions that can help doctors make more accurate diagnoses in less time. As a result, researchers have recently focused more on developing methods for automatically identifying DFU. Preprocessing, segmentation, feature extraction, and model training are all used in the suggested method. It does noise reduction and RGB to HSI color space conversion during preprocessing. OSTU thresholding segmentation is used for the separation. It uses histogram for feature extraction and Improved CNN-SVM for model training. The new method is compared to two common approaches, including CNN and CNN-SVM, and fares better than both.