What is it about?
Keeping a daily food journal is tedious and often inaccurate. To solve this, we created DietGlance, an automated dietary tracking system that uses smart glasses to seamlessly record your meals. As you eat, the glasses automatically capture privacy-protected images of your food and use advanced AI to identify exactly what you are consuming and in what portion sizes. We didn't stop at just identifying food. We connected the AI to a massive, reliable library of verified nutritional science. This allows DietGlance to not only break down the exact nutrients you consume but also provide highly accurate, personalized dietary advice. Through a mobile or desktop dashboard, users can review their habits and chat with an interactive AI nutritionist to help them meet personal health goals, such as managing weight or balancing their diet.
Featured Image
Photo by ooneiroslyl on Unsplash
Why is it important?
While many diet-tracking apps exist, they rely heavily on tedious manual data entry, which causes many people to eventually quit using them. Furthermore, while modern generative AI is great at identifying food, it often "hallucinates" or gives medically inaccurate nutrition advice. DietGlance is important because it solves both of these major hurdles. It completely automates the tracking process using wearable technology, and it solves the AI accuracy problem by using a technique called "Retrieval-Augmented Generation" (RAG). By forcing the AI to strictly reference verified government and scientific nutrition databases, our system provides trustworthy, evidence-based health interventions. In our 4-week study, this approach genuinely helped participants adopt healthier eating habits, reduce their sugar and fat intake, and become more mindful of their nutrition.
Perspectives
Developing DietGlance was an exciting journey of bridging the gap between cutting-edge wearable technology and practical, reliable health science. One of the most rewarding moments for our team was seeing the results of our longitudinal study—participants weren't just testing a novel gadget; they were actively making healthier choices, like reducing sugar and eating more vegetables, based on the AI's real-time guidance. This work represents a successful integration of multimodal wearable sensing, LLMs with RAG, and interactive design to foster meaningful dietary improvements.
Zhihan Jiang
Columbia University
Read the Original
This page is a summary of: DietGlance
: Dietary Monitoring and Personalized Analysis at a Glance with Knowledge-Empowered AI Assistant, ACM Transactions on Computing for Healthcare, April 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3797883.
You can read the full text:
Resources
Contributors
The following have contributed to this page







