Instead of writing the texts directly into the components we could use json files for our texts. One file per language or so? We can start with english & german and ask users to contribute translations

fix the issue with the certificates so that we can tackle multiplayer

When user is in forming expressions view wrong detections can result into skipping of tasks. We should ask the user to keep the expression for a view seconds.

Make pong usable on mobile devices. Maybe only in landscape mode?

Write it in TypeScript

Sometimes the webcam component breaks when switching between sites

We should provide the users with more information than we do today.

• Emotions: What emotions we are addressing and definitions or descriptions of these emotions.
• How we help: Our approach to the problem. E.g. different phases during training and playing
• How we respect privacy
• Who we are
• That we are open for help (from developers) or donations (patreon?)

What do you guys think about a memory game?

We can use open datasets to create cards of faces that show the same emotion. This way our users can try to match cards based on the emotion of the faces on the cards.

Different difficulties or versions are possible:

• two faces with the same emotion
• one face with emotion and name of emotion
• face with emotion and emoji

memory game example https://github.com/sandraisrael/Memory-Game-fend
data https://www.kaggle.com/c/challenges-in-representation-learning-facial-expression-recognition-challenge/data

The speed of our app would improve when we use webworkers for model queries

We could support individual high scores and some success statistics to the user.

This of course needs user accounts and a database in the backend. We can continue respecting privacy by only asking for username, mail and password. I wouldn't recommend cookies because then high scores are locked to the device.

Maybe this feature should be only available for supporters in order to get the servers paid?

do we want this project to be tested?

We should work on a mobile version of this app soon. Maybe we can switch our way of designing this app into a "mobile first approach" in which we think about a desktop UI afterwards, since chances are high that users will try to use this app on their mobile devices.

We could enhance our facial expression feature once we have more data of different facial expressions. This data can be provided (or donated) by users to help us improving our service