Deep learning and IoT-based intelligent exhibition and visual communication design

Advancements in Internet of Things (IoT) and Deep Learning (DL) enable intelligent museum exhibitions with adaptive visual communication, overcoming limitations of static displays by enhancing interactivity, dynamic cultural content presentation, and personalized visitor experiences. Existing exhibition systems lack real-time intelligence, adaptability, and immersive Virtual Reality/Augmented Reality (VR/AR) integration for personalized visitor engagement. This research develops an IoT–DL intelligent exhibition framework to enhance interactivity, visual communication, and provide curators actionable visitor insights. The system integrates adaptive VR/AR visual communication, adjusting colors, lighting, and layouts based on visitor engagement. IoT sensors and cameras capture movement, gaze, and interaction duration, processed using Faster Region-based Convolutional Neural Network (R-CNN) for object detection and interaction identification. Engagement prediction uses proposed Bitterling Fish-Based Bidirectional Gated Recurrent Unit (BF-BGRU) model integrated with BGRU’s temporal analysis with BF optimization for faster convergence and higher accuracy. K-Means clustering segments visitor behaviors, enabling visions for exhibition optimization and enhanced interactive experiences. Experimental evaluation on a synthetic simulation-based visitor interaction dataset demonstrated that the BF-BGRU achieved 99% accuracy and 95% F1-score for engagement prediction under controlled experimental conditions, improved than traditional Machine Learning (ML) and DL models. The framework improved adaptive content presentation, interactivity, and engagement prediction within the simulated intelligent exhibition environment. The proposed IoT–DL framework provides a data-driven, adaptive, and immersive exhibition solution, improving visitor engagement, personalized experiences, and actionable insights for curators.