Queryable Computer Vision Pipeline

This research aims to make video data analysis more accessible by creating a scalable architecture that allows data analysts to execute insightful SQL queries without needing advanced computer vision skills. The project leverages cutting-edge algorithms for object detection, tracking, and instance segmentation to develop a framework that extracts pertinent information from video content. It focuses on determining the best level of information extraction for effective SQL queries while designing a strong database schema for efficient querying. By connecting computer vision experts with data analysts, this research intends to enhance video data analysis across various application fields, improving its accessibility and effectiveness.