-
Main menu
-
Settings
-
Poses!
Inspiration
We were inspired by the game HORSE, where a person makes a basketball shot and the next person has to recreate it. The graffiti theme was selected to make the game seem colorful and exciting, like the poses the players will make.
What it does
Poses is a movement game where a player sets 3 poses by holding a position for 2 seconds. After setting the 3 poses, the next player has to recreate the poses, and if they fail, they receive a letter of POSES. Pose setting swaps every turn, depending on if the copying player fails or succeeds. The poses are set through the devices webcam and compared using pose-tracking technology.
How we built it
React, Next.js, TailwindCSS, and TypeScript to design the frontend Python, OpenCV, and FastAPI for camera and pose tracking MediaPipe and Numpy to set the poses SpacetimeDB and TypeScript for the game backend ElevenLabs for Announcer Voices
Challenges we ran into
- Using ElevenLabs to create announcer voices
- Organizing team roles and dealing with technical issues
- Connecting the frontend computer vision system to the game backend
Accomplishments that we're proud of
- Creating a vision-based game within the time limits
- Using SpacetimeDB to implement the game rules and track game state
- Creating a pose detection and comparison system
What we learned
- How to use AI to develop effeciently
- How to Combine a python computer vision backend with a typescript and react frontend
What's next for Poses
- Multiplayer across devices
- Website deployment
- Better UI
- Additional game modes and customization options
Built With
- fastapi
- mediapipe
- next
- numpy
- opencv
- python
- react
- spacetimedb
- tailwindcss
- typescript
Log in or sign up for Devpost to join the conversation.