Shoutai
This was a project done for a grad-level Web Programming & Engineering class. It is an automatedmaze game that allows users to try multiple levels of maze difficulty, saves their scores in a database, a viewable leaderboard, it also features voice-based input using TensorflowJS. Thanks to Travis Nevins and Wei Zhang for their contributions!
Shout AI!
Problem statement
A game where you write or shout at your computer to command your AI minion to perform a task and get the best score.
How do you solve the problem?
We use tensorflowjs combined with speach and handwriting recongztion models. Svelte as our frontend technology along with tailwind/bootstrap css, phaser 2d/pixiJS. Along with using firebase tools.
List of all the features that your team will implement. Your team will be held accountable for the features you list!
- speach recongization
- hand writing recongization
- AI movement
- frontend UI
- leader board
What data will your prototype be using and collecting
- voice data
- User timing
- user play data
- google analytics
Any algorithms or special techniques that will be necessary
- TensorflowJS
- Deep learning algorithms for speach and writing recongization
Wireframe
Initial designs:
credit to: Wei Zhang
credit to: Emmet Hayes
technology
- firebase/google cloud
- svelte, tensorflowjs, PixiJS/phaser 2D
- backend: either flask/fastapi (py) ktor(kotlin) pocketbase (go)
- database: surrealDB/MYsql/firestore
font links
logo like
* https://fonts.google.com/specimen/Audiowide?category=Sans+Serif,Display,Monospace
for fun font
* bit game like https://fonts.google.com/specimen/Silkscreen?category=Sans+Serif,Display,Monospace
