Nutrition Data

What is the coverage for different recognition targets?

Here is a quick overview of Nutrition-AI taxonomy of recognized and supported foods. We recognize visual food groups and map them to unique nutrition references (nutrition alternatives). A visual food group represents a visually distinct group of foods that can be uniquely identified using computer vision: green tea, glass of milk, red pureed soup are typical groups that are visually unique. Nutrition alternatives are mapped to visually recognized food groups. For example, glass of milk is linked (mapped) to soy milk, coconut milk, whole milk, etc. Each of those are linked to nutrition references from USDA or other databases. For example soy milk could be linked to Standard Reference soy milk, but also to branded soy milk versions such as 365 Unsweetened Vanilla Soy Milk.

Here is a quick summary of the nutrition database and recognition targets we support right now and are working to expand:

• Visually recognized labels: over 4100

• Visual labels are mapped to over 23,000 unique nutrition references

• Via text search: over 30,000 (only nutrition references, excluding visual labels) or 37,000 (including visual labels)

• Via barcodes: over 1.5M Barcodes (including via OpenFood.org database)

• Accessible via OCR (scanning of the front side of product): over 5300

If at any point you need help from the Passio team, please reach out to us at [e-mail]