Deep Learning and Optimization Approaches in Dual-Discriminator Spiking Generative Adversarial Network Based Classification and Segmentation for Predicting Pathogenesis of Foot Ulcers in Patients with Diabetes: A Review

Abstract

Diabetic foot ulcers (DFUs) represent one of the most severe complications of diabetes mellitus, often leading to infection, amputation, and increased mortality. Early detection and accurate prediction of ulcer pathogenesis remain critical challenges in clinical practice due to the complex interplay of physiological, vascular, and neuropathic factors. Recent advancements in artificial intelligence, particularly deep learning, have significantly improved automated diagnosis and prognostic analysis using medical imaging and multimodal data. This review explores the integration of dual-discriminator spiking generative adversarial networks (DDS-GANs) with optimization techniques for enhanced classification and segmentation of diabetic foot ulcers. The proposed paradigm leverages spiking neural dynamics to mimic biological neuron behavior, improving temporal feature representation while reducing computational overhead. Dual-discriminator architectures enhance generative stability and improve feature discrimination, leading to superior segmentation accuracy and robust classification performance. Furthermore, optimization strategies such as metaheuristic algorithms and hyperparameter tuning are analyzed for their role in improving convergence and generalization. This paper systematically reviews existing literature, highlights emerging trends, and provides a comprehensive analysis of deep learning-based DFU prediction systems. The findings indicate that hybrid architectures combining GANs, spiking neural networks, and optimization methods offer promising directions for accurate, scalable, and real-time clinical decision support systems in diabetic wound management.