Time Lens: AI Powered Monument Recognition and Exploration

Abstract

India’s architectural heritage spans thousands of years and includes temples, forts, palaces, and historic structures that reflect its cultural evolution. However, many users—tourists, students, and even researchers—struggle to identify monuments or access reliable historical information from photographs taken in real-world conditions. To address this gap, TimeLens introduces an AI- powered monument recognition system designed to make heritage exploration simple, accurate, and accessible. The system allows users to upload an image of any monument, after which a multimodal vision-language model analyzes architectural style, structural elements, textures, and contextual cues to identify the monument with high confidence. When the visual input is unclear, a fallback text-based reasoning mechanism ensures robust identification. Once recognized, the system generates a structured historical narrative covering architectural features, cultural significance, and related dynasties. This content is produced in English and automatically translated into Kannada, accompanied by natural audio narration to support diverse users, including those with reading difficulties. By combining computer vision, intelligent text generation, and multilingual accessibility, TimeLens offers a modern and user- friendly approach to cultural interpretation. The project demonstrates how AI can enhance heritage preservation by making historical knowledge more approachable, interactive, and engaging for the public.

KEYWORDS: Monument Recognition, Multimodal AI, Cultural Heritage, Historical Information Retrieval, Vision-Language Model, Multilingual Narration.