AI-Assisted Generation of Semantic Metadata for Personal Images and Videos

Problem Statement

"Personal photo and video collections often lack meaningful contextual metadata. While existing metadata (e.g., timestamps, camera settings, GPS) captures technical information, it fails to describe the content, context, or emotional significance of visual media. For digital legacy management, such richer metadata could help preserve memories, relationships, and personal meaning. Conduct an exploratory user study to identify desired metadata types. Analyze responses to define metadata categories and prioritize those most valuable to users. Develop a prototype using computer vision or generative AI models to automatically generate the identified metadata for personal photos and videos. "