Share on Pinterest
Numerous developers are creating apps that can provide nutrition info just from a photo — but are they precise enough to aid in diabetes care?

Keeping track of everything you eat may be remarkably useful when it comes to improving your habits and health, but it’s also tedious and time-consuming.

Most food tracking apps require you to search every item or ingredient in the meal or snack you eat in order to provide data about calories, fat, protein, fiber, carbohydrates, vitamins, and minerals.

To avoid the tedium of food logging, many app developers around the globe have set their sights on creating apps that can predict the nutritional information of a meal merely from a photo.

Is this really possible? Can people with diabetes really get reliable nutrition info simply by snapping a photo? Or will they end up needing to repeatedly search and correct the app’s estimates?

Let’s take a closer look.

“It is technologically impossible to estimate carbs or calories by photo,” explains Mike Ushakov, co-founder of UnderMyFork, the first company to develop a blood sugar-specific food photo app.

“Even if you use your eyes, a much more complicated device than your iPhone camera, you cannot differentiate porridge with sugar from porridge without sugar just looking at it.”

He provides additional examples: A photo could never tell the difference between a smoothie containing unsweetened almond milk instead of 2 percent cow’s milk — and the calories, carbs, fat, and protein between the two are wildly different.

A photo could never tell if your peanut butter and jelly sandwich contains sugar-free or traditional jelly. Or if the rice on your plate is actually made of cauliflower! Inevitably, any app relying on photos to produce actual nutritional data will require some degree of manual selection and searching to ensure it’s accurate.

“Our app uses quite a different approach,” explains Ushakov, a young entrepreneur heading up the Eastern European-based startup UnderMyFork.

Their iPhone app combines meal photos with CGM (continuous glucose monitor) data so users can see how their food affects blood glucose levels and time in range. It currently interfaces with the Dexcom CGM via Apple Health and several blood glucose meter brands. The company is working towards integrating data from a variety of CGMs.

To use it, you are prompted to log food photos and insulin doses, and blood sugar levels if using a monitor that doesn’t share data automatically. The app then utilizes that info to create an ongoing graph of your blood sugar levels, identifying if they are within or outside of your goal blood sugar range.

“Your meals are classified by postprandial (post-meal) time-in-range. In other words: meals are assessed based on your blood sugar levels within a few hours after you eat a specific meal.

“We let you see which of your meals are ‘green’ (i.e. you stay in range), and which meals are ‘red’ (the meals that drive you out of range),” explains Ushakov.

“By doing so, we aim to improve user’s time in glucose range, so the next time you eat you would choose the ‘green’ meals over the ‘red’ ones. This assumption is already confirmed by some of our early users.”

So to be clear, UnderMyFork doesn’t promise to provide exact carb counts based on a photo of your meal plate. Rather, the goal is to help people with diabetes gain more awareness of how their food choices impact their blood sugar levels.

Share on Pinterest
Screen shot of the UnderMyFork app

However, those of us who live with it know that real-life with diabetes is far more complicated than “simply food + insulin = blood sugar.”

There are of course numerous variables that greatly affect blood sugar levels:

  • Inaccurate insulin dose (estimating too little or too much)
  • Exercise (and time of day, insulin-on-board, etc.)
  • Lack of exercise (if you normally exercise after dinner, but didn’t that day)
  • Stress, good or bad (arguments, presentation at work, an exam, etc.)
  • Sleep deprivation (can create short-term insulin resistance that day)
  • Menstrual cycle (the start of your period, for example, often brings blood sugar spikes)

“We do understand that there are a lot of variables,” explains Ushakov. “And our view is that we will add these variables step-by-step, layer-by-layer, to let you better see what exactly caused your out-of-range events.”

Ushakov says their most recent step in development is adding when and how much insulin was dosed with each meal. This could be significantly helpful in bringing attention to which meals were properly dosed for, and which need more insulin to stay in range.

“I also personally think that taking a photo allows you to better recall all the context of the meal — including the variables that are not easy to formalize inside of the app, like stress or an argument for example.”

Ushakov says while they do intend to make the app useful for people with type 2 diabetes, their first focus is developing further to help manage type 1 diabetes.

There are some Pros and Cons around the efficacy of this app motivating users to choose “green” meals over “red” meals.

As noted, there are tons of variables that impact blood sugar, both before and after you eat.Just because a meal sends your blood sugar out-of-range doesn’t automatically mean you shouldn’t choose to eat that meal again. In reality, managing diabetes with insulin means we are constantly estimating insulin doses based on estimates around macronutrients (carbohydrate, fat, protein).

For example, you could eat a breakfast of eggs with kale, onions, carrots, bell pepper, and mushrooms, and still end up with a high blood sugar after breakfast. Does that mean this meal is ‘red’ and you shouldn’t eat it again?

No, instead it means you either didn’t get enough insulin with that meal or there was another variable at play. For example, maybe you are getting sick that day, which can also send your blood sugars soaring.

These types of variables are a daily experience in the juggling act of type 1 diabetes management.

The flip side of this is the main benefit of the app: helping a PWD (person with diabetes) see for example that they are consistently out of range after a particular meal, like breakfast, which means they need more insulin with that meal on a regular basis.

As UnderMyFork moves into serving people with T2 diabetes, they will also be grappling with the fact that the patients who are most unaware and most in need of this education around food are likely not checking their blood sugars frequently (if at all). This app is only effective if you are using a CGM or checking your blood sugar regularly.

There are a number of other apps available today that allow users to take photos of their food, although most are not designed specifically for diabetes. Notable apps include:

Nutrino’s FoodPrint: This app asks users to log photos of meals for the sole purpose of helping identify what you ate so it can generate nutrition data. It does not generate data based on photos alone and is much more time-intensive, but rather requires that you log every specific ingredient or item in order to generate any real nutritional information. It does allow you to track insulin, medications, and blood sugar readings.

  • There are options to pay for more features, but the free version of this app gives you plenty of tools.
  • The premium version costs $7.99/monthly or $39.99/year.
  • Read more about this app in our in-depth Nutrino review

Calorie Mama: This AI-driven app bills itself as “a smart camera app that uses deep learning to track nutrition from food images.” It also doesn’t generate nutrition data from the photo alone. Instead, it relies on the photo to easily identify what you ate, and then you are still required to search and select specific foods and ingredients to generate any real nutritional data.

  • Calorie Mama’s “free” version offers plenty of tools if you don’t want to pay.
  • The full-featured premium costs $9.99/month or $29.99/year.

Foodvisor: This app touts that from a photo, it can estimate the serving size and provide a detailed nutrition report in just seconds. It creates a food diary and prompts you to log physical activity as well so it can calculate your calorie intake vs. calories burned.

We were able to give it a try, and when we entered a meal of eggs and sauteed veggies, the photo only identified the calories in the veggies. It did not identify the presence of eggs at all, so that would have to be manually searched and entered. While it did predict nutrition information from the photo alone, it was not nearly as user-friendly as UnderMyFork.

  • Foodvisor first implies that your only option is the 7-day free trial with an agreed one-time payment of $59.99 after the 7-day trial is over.
  • Only by trying to get out of that page does it become clear that you can take photos to give it a try for free.
  • Every time you open the app, it hassles you for money.

YouFood: This food journaling photo app is aimed at weight loss. It prompts users to snap a photo of meals, while also logging food, drinks, water, and exercise. It then provides daily “reflections” to help you understand your habits. And it provides a “social accountability” feature that it claims is the No.1 most effective weight-loss method.

  • Unfortunately, you cannot gain access to the “free 7-day trial” without providing payment information and requiring you to manually cancel your subscription after the 7 days are up.
  • That may be a frustrating experience for potential customers.

Snaq: This Switzerland-based startup says its app offers “reliable food recognition, image-based portion calculation and a well-structured nutrition database” built on their proprietary patent-pending nutrition analysis technology. Its CEO Aurelian Briner has a partner living with type 1 diabetes, and the company is working with the Diabetes Center Berne to help optimize the app for diabetes use, with various goal-setting functions.

  • This Android app is currently only available in select regions of Europe while they are working on it, but it’s definitely one to keep your eye on.

I think the apps provide a reasonable ‘rough’ estimate for those who don’t understand how to estimate portions or carb count properly,” says Gary Scheiner, MS, CDE, and director of Integrated Diabetes Services. “But those who depend on a reasonably accurate carb count in order to calculate the proper insulin dose, there is no substitute for nutrition education with a qualified professional.”

Realistically, any of these apps could serve as one source of support and insight into your overall diabetes management, but it’s unlikely they’ll ever be intelligent enough to tell you exactly how much insulin to dose.

As noted, blood sugar levels are the result of more than just food and insulin.

That being said, it’s never too late to gain a deeper understanding of your relationship with food, your eating habits, the choices that might send your blood sugar out-of-range most often, and a greater awareness of how much real food versus processed food you’re consuming every day.

If you haven’t taken a closer look at these aspects of nutrition in your life, it might be worth your while to visit the app store and download a few until you find one that’s right for you!