A healthy pregnancy diet is crucial for supporting the development of your baby. However, when you find out that you’re pregnant and your doctors are telling you to suddenly change your eating habits, it’s not immediately clear how to go about this. 

Pregnant women are aware that they must gain weight during pregnancy, however a 2017 study showed that around 47% of pregnant women gain more than their recommended weight gain target (as per the Institute of Medicine guidelines), while 23% of pregnant women gain less than the recommended amount. Excessive weight gain during pregnancy is associated with an increased risk of adverse pregnancy complications. On the other hand, a pregnancy diet that doesn’t provide the necessary nutrients for the fetus may result in preterm birth and low birthweight.

Researchers agree that dietary interventions and better nutritional planning during pregnancy can help achieve a healthy pregnancy for both the mother and baby. However, existing methods for diet tracking rely on manually reporting your food intake.

We wanted to simplify this tedious process to create a better experience for users and encourage better pregnancy outcomes. We quickly realized that our team's expertise in applied artificial intelligence, quantitative biology, UI/UX design and mobile app development could be put to good use here and our concept for nutrition tracking via image recognition was born!

Nobody enjoys the process of having to write down everything you eat, but many of us enjoy taking pictures of our food! So we thought, could we make nutrition tracking as simple as taking a picture?

Recent developments in the field of artificial intelligence, particularly in the areas of neural networks and computer vision (the kind of tech used to build driverless cars) have made it possible for computers to easily recognize patterns from images. We’ve spent the past few months applying these techniques to develop a deep learning model that recognizes a food from an image taken from your phone camera. 

Furthermore, as soon as the computer recognizes your meal from a camera, it retrieves the nutritional content from a comprehensive database containing data on over 10,000 foods. This information is written straight to HealthKit, so that your Apple Health App is always up to date with your nutritional health.

As with all forms of machine learning and artificial intelligence, these models get better over time. So over the coming months expect to see improved versions of this feature released on a rolling basis.

Time to try it out for yourself!

Open the Velmio app and create a new Diet log

Point your camera at some food and take a picture.

Velmio will try to guess the food and offer some suggestions. Make your choice and press the next button.

Now Velmio will show you nutritional information for your food. You can adjust the portion size here too and save your record. It’s that easy!


How about some tasty chicken wings?

Looks like Velmio knows what chicken wings are too.

Here we can see the nutritional information.

And here we can adjust the serving size for the most accurate nutritional data.

As you can see, we’re very excited about this feature and are looking forward to sharing improvements and new functionalities in the coming months.

Nutrition tracking doesn’t stop here. Many women are prescribed nutrition supplements during pregnancy, so to capture your full nutritional profile we’ve developed features to help you accurately record your supplement intake. Check out our article here for details on how to use this feature.

If you are interested in learning more about healthy eating during pregnancy, check out our resources on what to eat and weight gain guidelines.