Abstract
The objective of the research is to address increasing privacy and safety issues related to hidden cameras in personal spaces such as bedrooms, bathrooms, or dressing areas, where such cameras can wirelessly transmit video signals clandestinely. The aim is to develop and evaluate an Artificial Intelligence (AI)-Enabled Hidden Camera Localization (AHCL) platform capable of identifying and locating hidden video streams through analysis of real-time network traffic. The methodology involves packet capturing, statistical analysis, and deep learning-based classifiers to detect anomalous streaming traffic in captured packets. The research generated a dataset comprising 60,412 packets, labeled as either 'normal' or video streaming, which was used to train and evaluate several models, including Support Vector Machines (SVM), Denoising Autoencoders, and ensemble deep learning models. The experimental results indicate that the ensemble model achieved the highest performance, with a detection accuracy of up to 98.27%, demonstrating good generalization and robustness across different network environments and over multiple days. The findings show that the AHCL platform is highly reliable in detecting hidden camera traffic from benign traffic. The practical contribution of this research is significant, providing users with an intelligent and affordable system for real-time privacy protection that can be deployed in residential or commercial settings, thereby enhancing trust and safety in a connected environment.