Introduction of TRYPERFECT

Use our app

Don't know how to do your makeup ? Worried about what suits you most ? Our app TRYPERFECT can test the makeup that suits you best through the camera or your photos !

Simple use

Our app just use your camera or your photo to catch your face by using machine learning. This app is designed to help users to choose their best makeup.

Start right now !

You can download our app throgh the app store and use it for free ! We promise the use of the technology will not reveal your privacy.

The technology we use

Before testing, users will choose the style they want to present through multiple choice questions. We use AI technology to analyze people's facial features mainly through big data, and we will cooperate with beauty brands to make makeup suitable for users appear on the camera or change the form of photos.

We combined the technology of face recognition, gesture recongization and AI prediction throgh database.

We mainly use facial recognition, motion capture and big data to predict the most suitable makeup. User preferences are also under our consideration, so our app is different from ordinary beauty cameras. We are aimng at the prediction of user preferences, not a summary of user preferences.

We will add some new features, such as beauty products that recommend buying related makeup for business purpose.

How to code it

In order to creat our app, we use the ml5.js. The code of ml5.js is machine learning for the web in your web browser. Through some clever and exciting advancements, the folks building TensorFlow.js figured out that it is possible to use the web browser's built in graphics processing unit (GPU) to do calculations that would otherwise run very slowly using central processing unit (CPU). A really nice explantion of what is happening with GPUs can be found here - Why are shaders fast?. ml5 strives to make all these new developments in machine learning on the web more approachable for everyone.

As we know, Ml5.js is a library which provide access to machine learning algorithms and models in the browser. ml5.js has created an API to face-api.js that allows users to access face and face landmark detection. To go further, it would be interesting to increase the accuracy percentage of the detection by adding database through network. It is possible with TensorFlow.js and PoseNet to analyze the gestures of an individual. We mainly use those three code to creat the our app.