SceneScout leverages large language models to enable blind and low vision (BLV) users to explore unfamiliar environments through Street View imagery before traveling.

• Creates an interactive conversational experience where users can ask questions about street views
• Provides detailed visual context beyond simple landmarks and turn-by-turn directions
• Uses AI agents to interpret images and answer questions about the environment
• Reduces travel uncertainty by giving BLV users access to rich environmental information

This research represents a significant advancement in assistive technology, empowering individuals with visual impairments to gain independence and confidence when navigating new places.

SceneScout: Towards AI Agent-driven Access to Street View Imagery for Blind Users