MBARI's Video Annotation and Reference System (VARS)

VARS is MBARI's platform for managing scientific video and image assets. Originally built in 2004 to annotate video from MBARI's remotely operated vehicles, it has evolved into a full media management system (M3) used by researchers to capture, annotate, query, and analyze underwater video and imagery.

MBARI provides the vars-quickstart-public project so that external teams can stand up their own VARS instance using Docker.

Capabilities

• Time-indexed annotation of video observations
• Video asset and metadata management
• Taxonomic knowledge base for organisms, equipment, and geology
• Frame extraction and image archiving
• Automated image analysis using machine learning
• Cross-service querying and bulk data retrieval

Documentation

• Getting Started — prerequisites and setup overview
• Architecture — how the system components fit together
• Services — backend microservices reference
• Applications — client applications reference