A Unified, Interpretable, and Scalable Deep Learning Framework Integrating Foundation Models, Self-Supervised Learning, and Neuromorphic Computing for Robust Multi-Class Image Recognition
Authors: Raghavendra Rao Ankam , Burra Ramanuja Srinivas , G Maheswara Rao, S.Mallikarjunaiah
DOI : 10.46335/IJIES.2026.1.1.3
Abstract – Recent developments in deep learning have substantially advanced image recognition performance; however, achieving scalable, interpretable, and reliable results in large multi-class classification problems remains challenging. This research proposes a unified deep learning framework that combines a strong feature extraction backbone with self-supervised representation learning and neuromorphic-inspired computational mechanisms to address these challenges. The framework is designed to improve feature robustness, ensure stable convergence during training, and maintain computational efficiency. The effectiveness of the proposed framework is evaluated on the CIFAR-100 benchmark dataset, which consists of 100 object categories and represents a demanding multi-class recognition task. Experimental results demonstrate consistent improvements in both training and validation accuracy across epochs.
