An image-based food quality analysis framework driven by large language models

Key Points

• To develop a framework for food quality analysis leveraging large language models and visual inspection.
• Created a GUI for one-click automation of food quality assessment.
• Implemented two inference paths: text-only and vision-language.
• Models evaluated on tasks for ripeness, freshness, and authenticity.
• Framework automates the inspection process for food quality.
• Demonstrated effectiveness in assessing ripeness and authenticity.