Computer Vision Driven Recommendation System for London Explorers
Khadidja Mekiri, Kashif Hussain Talpur, Shakeel Ahmad · 2024
This paper presents an innovative system that uses computer vision and machine learning to improve tourists' experiences in London, UK. The system conducts real-time facial analysis using advanced algorithms like You Only Look Once (YOLO) for face detection and FaceNet for age, gender, and ethnicity prediction to automatically infer demographics without manual input. It leverages survey data and infers attributes to generate personalized landmark recommendations. The system is hybrid that uses demographic, collaborative filtering and content based. We achieve a precision of 79% and a root mean square error 1.26. The novel combination of computer vision profiling and recommendation techniques shows promise to transform urban tourism by automating personalization without explicit profile inputs.