Nuvilab, established in 2018, is an award-winning “Food Vision AI” startup. It specializes in developing a unique combination of hardware, software, and cloud services aimed at accurately measuring food servings and waste in environments such as cafeterias, restaurants, buffets, and hospitals. Over the years, Nuvilab has amassed substantial data, enabling the sophisticated training of its AI software, leading to the latest AI Food Scanner 3.0.
Nuvilabs service relies on several scanning devices that sense the type and quantity of food served (or discarded) in various places. The goal is to accurately track nutrient consumption, carbon emissions, and food waste.
The data can be utilized to gauge food intake to create an accurate report about food consumption for a person or organization. Nuvilab avoids making “diet recommendations” because more context (personal information) would be required. However, it can quickly surface information that could interest nutritionists or doctors.
The scanning devices use two cameras. In less than one second, they capture a stereoscopic view that produces photo and depth data. By combining both information, the AI can categorize the type of food in the photo and evaluate “how much” of it (= volume) was served or discarded. From there, the software can cross-reference a food database and infer more nutritional details.
Different scanners are used depending on how the food is served, for example, cafeteria trays or buffet bins. That’s because the required camera field of view is different, and the buffet bin is a slightly different use case since it tends to contain a single type of food in substantial volume.
There are some alternatives to Nuvilab in the industry, but the ones I’ve heard of rely on scales to measure the volume of food. That decision tends to be more approximative but also has a negative impact on the system deployment, which is why Nuvilab opted for a vision-based approach. This also happens to be ~50% less expensive to deploy and maintain, according to the company.
Nuvilabs operates a cloud service for its customers. Data from the scanners is sent to a secure cloud (hosted on Amazon AWS) where customers can get insights from their data. New features added recently include an Inventory Manager to optimize purchases, but many other actions can be derived from the data.
The company mentioned working with Aramak in Korea. The same company is also a significant provider of catering services, including here in Silicon Valley. Having such partners could be great for the service’s global expansion as both companies can work closely at every level, including updating the food database since Amarak knows precisely what goes into the food as far as I know.
Nuvilab representatives mentioned that the company is looking at the hospitality market in North America, evaluating compliance needs in the European Union (the most stringent worldwide), and believes that the Middle East will see the most growth in the short term.
At CES 2024, Nuvilab will show its latest services and demonstrate different solutions, including nutrition intake, kitchen optimizations, dining hall use cases, and more. You can find our complete conversation in the video below: