Apps and Algorithms

What is an application?

An application is a set of algorithms that process visual data (selfies, skin images, etc.). You can learn more about what each application does by clicking on its picture in the “Applications” menu.

How do I use an application?

To use any application, you first need to attach it to a dataset. You can do it in the dataset creation menu.

There is another option. Go to the “Applications” menu, click on the application you’re interested in, click on the “Attach application” button, and then choose which dataset to attach it to. Click “Start” and simply wait for the processing to finish. If you want to know how to create a dataset, look here.

What are the criteria for a selfie so that the algorithms can analyze it correctly?

The answer to this question is LIQA—our visual assistant that will give all the instructions you need to take the perfect selfie. You can use LIQA to collect data while conducting a study. For more information on how it works, go here.

If you’re uploading pictures from your desktop, however, it is most likely that they weren’t taken with LIQA. And even though our algorithms can detect and analyze the face even if the picture is low-quality and upside down, to get the best results, we recommend using high-quality images with the face clearly visible.

Be sure to remove any makeup and don’t smile or make any facial expressions, as this may affect such metrics as wrinkles, sagging, and lacrimal grooves. The face-to-camera distance has to be far enough to ensure your face fits completely into the frame but close enough to make your pores and small wrinkles visible. The resolution also has to be sufficient (but not higher than 2000 x 2000 px); otherwise, such metrics as hydration or pores might not be applicable.

After you upload your picture, our algorithms will assess its quality. If the quality of your photo is low, you can still run applications and get a report on it. However, please be aware that its accuracy may be altered.

What metrics do you use to assess the condition of the skin and hair?

You can find the list of metrics we use, along with a description of each one, here.

How do product recommendations work?

We have a matrix of categories and subcategories (in total, approx. 50) that you need to fill in with your products (name, action, image, description). Our recommendation engine can recommend basic care and special care skincare product categories based on the image analysis and survey. For now, this application is only available via API.

Last updated