Computer vision for tourism information systems: Research on object recognition and image classification for mobile applications.

This article focuses on the development of an intelligent signpost navigation system for mobile tourism applications that integrates computer vision techniques to enhance real-time landmark recognition and location guidance. The system employs an advanced object recognition and image classification model called the Intelligent Ant Colony-tuned Bottleneck Residual Network (IntACO-Bottleneck-ResNet), which combines a modified ResNet architecture with an optimized Ant Colony Optimization algorithm for efficient pathfinding. Using a curated and annotated dataset of 110 images from various tourist sites under diverse environmental conditions, the model achieved superior performance metrics—97% accuracy, 96% precision, 95% recall, and a 94% F1-score—outperforming comparable methods. The research highlights the system's potential to improve user experience and location-based services in dynamic tourism contexts, while noting limitations related to dataset scope and computational demands on mobile devices, suggesting future work to enhance generalizability and efficiency.