Clone this repository locally:

git clone [email protected]:Sam-Sweere-Vantage-AI/cv-hackathon.git

Go into the repository directory:

cd cv-hackathon

Create a virtual environment:

python3 -m venv venv

(Developed on Python 3.10.4 but should work with any python 3.X)

Activate the environment:

source venv/bin/activate

Install the dependencies:

pip install -r requirements.txt

The game itself consists out of a simple dodge game. Dodge the obstacles to stay alive as long as possible! However, you cannot control the Vantage player with your mouse or keyboard. Only with your webcam!

The team/player that is able to reach the furthest level wins!

You can build your webcam controller in the camera_controller.py file. This file contains the CameraController class. Every tick of the game, the get_location of this class is called, given the latest input image. Using this input image, you have to control the player by returning the x and y coordinates of the player. You can use any technique you want as long as the final x and y coordinates are based on the image.

Finally, you can also return an image that will be drawn in the game's background. By default, this is the input image. But for visualization and debugging, it can be practical to draw/change this image.

Run python3 run_game.py

For fairness, stay out of the dodge_game files.