Visage Technologies
Face Tracking & Analysis

Blog

#SNAPSTAR is the innovative line of dolls that lets kids change dolls' appearance and surroundings using a mobile app. The app uses visage SDK FaceTrack to accurately track dolls' faces.

Revolutionizing childrens’ toys with face tracking

Mar 7, 2019

Snapstar dolls by Yulu

About Yulu

YULU is an international creator and manufacturer of toys and games that aim to promote an active lifestyle. The company’s goal is to help children nurture their creativity and develop problem-solving skills through active play.

Business challenge

The popularity of social media and influencers is growing, and children are also fascinated by that world. YULU wanted to give them an opportunity to create their own influencers.

#SNAPSTAR squad is a line of 6 dolls with different personalities, but the same love for fashion and travel. However, fashion influencers need to get creative when it comes to style, so YULU also needed to go one step further.

The company wanted to let kids change their dolls’ appearance and “surroundings” and take the perfect selfies, just like real influencers would. To do so, they needed an app that could provide all these options.

Solution

After checking out various face tracking providers, YULU chose to license visage|SDK FaceTrack for several reasons:

  • visage|SDK is quick and lightweight, delivering great results with a broad range of cameras and resolutions on both iOS and Android phones.
  • Users don’t have to depend on internet connection since our SDK works offline.
  • visage|SDK integration is very simple and quick thanks to the detailed development guide, which is crucial on a short deadline.
  • We offer custom development service that allows clients to quickly implement additional features.

Our face tracking technology was used to create the Snapstar Studio App.

FaceTrack immediately delivered good results with dolls, although it has been primarily optimized to track human faces. However, to get the optimal results, we needed to develop a custom algorithm that would ensure flawless tracking of dolls’ faces.

The project was done on a very short deadline. Our team started by creating a doll image dataset and annotating each image. Then, that data was used to train the new algorithm. We wanted to ensure accurate tracking even in more challenging conditions (less light, different head poses, etc.).

As a result, the app provides a fun and interactive experience. The owner can pose their doll, change its hair and makeup and choose from a wide selection of backgrounds. There are countless possible combinations children can create to tell a unique story of their little star.

Since their launch, #SNAPSTAR dolls have been featured in media worldwide and have won several “best toy” awards.

← back  
Go to top