International Journal of Scientific Research and Engineering Development

Diabetic Foot Ulcer (DFU) is a critical complication of diabetes that requires early detection to prevent severe outcomes such as infection and amputation. This study presents a comparative analysis between a hybrid ensemble deep learning model and a classical machine learning approach based on Support Vector Machine (SVM) for automated DFU detection and stage classification. The proposed deep learning framework integrates EfficientNetB3, MobileNetV3Large, and ResNet50 to exploit their complementary feature extraction capabilities, enhancing classification performance. In contrast, the SVM model utilizes CLAHE-enhanced grayscale images combined with Histogram of Oriented Gradients (HOG) features to establish a computationally efficient baseline. The system follows a twostage classification approach, where the first stage identifies images as Healthy, Wound, or DFU, and the second stage further classifies DFU cases into Stage I, Stage II, Stage III, and Stage IV. Experimental results indicate that the ensemble deep learning model achieves superior accuracy, robustness, and generalization compared to the SVM model, while the latter offers faster computation and serves as an effective baseline for comparison.

Suganthi B, Shanmugavadivu N, Nandhini D, Nathiya V, Nithya S, Srilekha K,"An Intelligent Comparative Study of Ensemble Deep Learning and SVM for Accurate Diabetic Foot Ulcer Detection and Severity Analysis" International Journal of Scientific Research and Engineering Development, V9(3): Page(340-343) May-June 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.