The growing use of internet of things (IoT) technologies in higher education is transforming how institutions manage infrastructure, deliver teaching, and engage with students. While these advancements offer considerable benefits, they also introduce significant security risks. Common threats include weak access controls, insufficient data protection, outdated software, exposure to denial-of-service (DoS) attacks, and lack of physical safeguards for connected devices. This study provides a comprehensive review of these vulnerabilities within academic environments and proposes a security framework adapted to the specific operational and technical realities of universities. Unlike generic approaches, this research focuses on the unique challenges of higher education, such as decentralized information technology (IT) structures, limited resources, and diverse user groups. The main contribution lies in identifying and evaluating security measures that are both effective and applicable in academic contexts. These include encryption methods, identity verification techniques, secure update mechanisms,and intelligent systems for detecting abnormal behavior. The analysis is supported by case examples from real institutions, illustrating both successes and limitations of current practices. This work aims to guide educational institutions in improving the resilience of their IoT systems. It also outlines areas for future research, particularly in the development of lightweight and scalable security solutions suited to the evolving needs of smart learning environments.

Recognition problems, including object detection, scene understanding, and fine-grained categorisation, are popular subjects in computer vision. However, it is challenging to model spatial coherence and contextual dependencies in response to changes in configurations. Human Vs computers' ability in perception-although convolutional neural networks (CNNs) do well in the extraction of features, they have high dependence on local receptive fields and are not able to capture long-range spatial relationships and high-order interactions. To alleviate the shortcomings of the current approaches, we present an enhanced hybrid CNNs two dimensional hidden Markov model (2D-HMM) framework that combines 2D-HMM, Markov random fields (MRF) and variational autoencoders (VAEs) into a single model. The model employs 2D-HMMs for pairwise spatial modelling, MRFs for higher order context, and VAEs for stable latent representation learning. Tested on the MNIST and CIFAR-10 benchmark datasets, our approach consistently outperforms the state-of-the-art performance by 98.2% and 89.5%, respectively, with high robustness to noise and occlusion. Results from ablation studies further show that MRFs improve recall by 1.6% and VAEs improve precision by 1.3%, suggesting that they complement each other sufficiently with respect to overall testing performance. This work unifies deep learning and probabilistic graphical models, leading to more interpretable, scalable, and accurate recognition systems.