MoroccoLens: An ML-Based Mobile Application for Monument Recognition

Abstract

This paper introduces MoroccoLens, a mobile application designed to leverage machine learning techniques for the recognition and contextualization of Moroccan monuments. By employing convolutional neural networks (CNNs) within an image recognition framework, MoroccoLens facilitates the identification of architectural landmarks from user-uploaded images, subsequently providing detailed historical, architectural, and cultural information. With a focus on achieving high recognition accuracy across diverse geographic and structural monument types, the application addresses a critical gap in accessible cultural knowledge, enhancing the experience of tourists, locals, and scholars. The system architecture integrates a ResNet-50 model, optimized for mobile deployment, alongside a robust data pipeline that enables real-time classification and retrieval of monument information. Security and privacy considerations are integrated into the system design, ensuring secure handling of user data. Through advanced image processing and classification methods, MoroccoLens presents a novel approach to the intersection of mobile technology and heritage conservation, contributing to the preservation and engagement with Moroccan architectural heritage. The findings highlight the potential for such technology to broaden cultural access and deepen public interaction with cultural assets. Future work includes expanding the monument dataset, optimizing model robustness under challenging conditions, and introducing augmented reality features to enhance user experience.