WizViz

spritesheet, original art

Inspiration

The inspiration for WizViz came from our love for both wizardry and computer vision (CV) games. We wanted to create an immersive experience that combines the magical elements of wizardry with the technical challenges of computer vision, resulting in a more engaging game with more interactive controls!

What it does

WizViz is a wizardry-themed computer vision game where players use their webcam to cast spells and interact with magical elements on the screen. The game leverages computer vision techniques to detect and track the player's movements, allowing them to perform various magical actions and solve puzzles.

Spells you can use:

Point: This is the attack function
Cross arms: This is the defense function
Drinking: This is the healing function
Nothing: Arms at side, this is resting
Arms in a bow-like stance: This is the special attacks function

How we built it

We built WizViz using a combination of computer vision and game development technologies. The game is developed with Unity for the frontend, providing a rich and interactive 3D environment. For the computer vision aspect, we used OpenCV and MediaPipe to implement real-time gesture recognition and tracking. The backend is powered by Python, handling data processing and integration with the computer vision models.

Challenges we ran into

One of the main challenges we faced was achieving accurate and real-time gesture recognition. Ensuring that the computer vision models could reliably detect and track the player's movements was crucial for creating a seamless gaming experience. Additionally, integrating the computer vision components with the engine required careful coordination and optimization.

Accomplishments that we're proud of

We are proud of successfully developing a working prototype of WizViz that combines wizardry with computer vision. This accomplishment demonstrates our ability to innovate and create a game that offers a unique and magical experience for players. We are also proud of the accuracy and responsiveness of our gesture recognition system (primarily on M-based chips).

What we learned

Throughout the development process, we gained valuable insights into computer vision techniques, game development, and the integration of these technologies. We learned how to implement real-time gesture recognition and track movements accurately.

What's next for WizViz

Looking ahead, we plan to add more spells and magical elements to enhance the gameplay experience. We also aim to refine the gesture recognition system to make it even more accurate and responsive. Furthermore, we will continue to improve the game's graphics and user interface to create a more immersive and visually appealing experience.

Built With

mediapipe
opencv
pygame
python

Submitted to

SpartaHack X
- Winner [Interactive Media] Best Game

Created by

I worked on the AI logic based on difficulty. If easy/medium the AI performs random actions. If on hard, the AI follows a FSM to find the best possible decision based on the current situation. I also helped with the engine and overall looks of the game.

Gabe De Almeida
I worked on all the player systems and integrating the engine with CV and game engine logic.

Agniva Ganguly
I worked on the computer vision backend system along with the engine that ran the game algorithm.

Ankur Desai
Ammar Ateya