With the development of a new AI tool, you may be able to snap a photo of your meal, and immediately know your food’s calorie count, fat content, and nutritional value.This new technology was presented at the 6th IEEE International Conference on Mobile Computing and Sustainable Informatics. It was developed by NYU Tandon School of Engineering researchers. It could become a key tool for the millions of people who need help in managing their weight and many other diet-affected health conditions such as diabetes. This AI tool uses advanced algorithms to recognize food items from their images and calculates their nutritional content, including calories, protein, carbohydrates and fat. https://www.sciencedaily.com/releases/2025/03/250318141833.htm
NYU’s Fire Research Group, including the papers’ lead authors Prabodh Panindre and Sunil Kumar, have studied firefighter health challenges for a long time. A number of studies have shown that approximately 75-90% of firefighters are overweight and therefore have increased cardiovascular and other health risks that threaten their operational readiness as firefighters. These findings helped motivate the researchers to develop their AI-powered food-tracking system.
Dr. Panindre, an Associate Research Professor at the NYU Tandon School of Engineering says, “Traditional methods of tracking food intake rely heavily on self-reporting, which is notoriously unreliable. Our system removes human error from the equation.”
Others have studied and attempted to develop a food recognition AI tool but its development has been more difficult than one would have thought it might have been. Prior attempts faced three basic challenges that this NYU team of researchers was able to finally overcome.Dr. Kumar, Professor of Mechanical Engineering explained, “The sheer visual diversity of food is staggering. Unlike manufactured objects with standardized appearances, the same dish can look dramatically different based on who prepared it. A burger from one restaurant bears little resemblance to one from another place, and homemade versions add another layer of complexity.”
Earlier systems also had difficulty with estimating portion sizes — an important factor for nutritional calculation. The NYU team used advanced image processing to measure the exact area each food occupies on a plate. The system correlates the area occupied by each food item with density and macronutrient data to convert 2D images into accurate assessments of nutritional values. This integration of volumetric computations with their AI model allows for a fairly precise analysis answering a longstanding unmet challenge enabling automated dietary tracking.
The third major hurdle in this AI tools development is called “computational efficiency”. Previous models required too much processing power to be practical for real-time use. This team overcame this by using a recognition technology called YOLOv8 with ONNX Runtime allowing a person to be able to use their phone’s web browser to analyze meals.
Dr Panindre says,”One of our goals was to ensure the system works across diverse cuisines and food presentations” . “We wanted it to be as accurate with a hot dog — 280 calories according to our system — as it is with baklava, a Middle Eastern pastry that our system identifies as having 310 calories and 18 grams of fat.” They tested a pizza slice, and their AUI tool calculated 317 calories, 10 grams of protein, 40 grams of carbohydrates, and 13 grams of fat- and these nutritional values that closely match available reference standards. It also performed well when analyzing idli sambhar, a South Indian specialty consisting of steamed rice cakes with lentil stew. The calories, 7 grams of protein, fat and carbohydrate also matched well with standard references. In their work they overcame some of the technical challenges by combining similar food categories, removing food types where there were not enough examples,to use as a reference . As a result these techniques helped them to establish a dataset consisting of 214 food categories and over 95,000 food items.
This research team claims that their AI food nutritional computation tool accurately identifies and calculates nutritional values of foods approximately 80% of the time. They have deployed this tool as a web application that works on mobile devices, making it accessible to anyone with a smartphone. The researchers describe their current system as a “proof-of-concept” that will be refined and expanded for other healthcare applications in the near future.
TRADITIONAL REFERENCES PUBLISHED BY THE NIH AND DEPT OF AGRICULTURE
https://ods.od.nih.gov/HealthInformation/nutrient recommendations.aspx
Nutrient Recommendations: Dietary Reference Intakes (DRI)
Here you will find the Food and Nutrition Board of the National Academies of Sciences Engineering, and Medicine. The Food and Nutrition Board is the official national body which establishes principles and guidelines of adequate dietary intake; and renders authoritative judgments on the relationships among food intake, nutrition, and health.
DRI is the general term for a set of reference values used to plan and assess nutrient intakes of healthy people. These values, which vary by age and sex, include:
They then give the well known RDA (Recommended Daily Allowance ) Values for foods.
- Recommended Dietary Allowance (RDA): Average daily level of intake sufficient to meet the nutrient requirements of nearly all (97–98%) healthy individuals; often used to plan nutritionally adequate diets for individuals.
This is the USDA’s comprehensive source of food composition datasFOUNDATION FOODS Analytical data/metadata on commodity and minimally processed food samples
BRANDED DATA . for commercially available foods known by their brand names From label data collected through a public-private partnership.
ata Cent
https://fdc.nal.usda.gov/food-search?type=
Here you will find 7 pages of over 300 food items listed alphabetically from apples to yogurt.
