Dugout Vision

Inspiration

It was inspired by MLB challenge on transforming video Through Computer Vision.by the lack of analytics with inexpensive equipments.

What it does

Dugout Vision is a web-based application that uses HTML5 Canvas and computer vision with homeruns MLB dataset to create an video analysis of exit velocity thru ball position. Here's what it does:

1. Video Slicing: The app takes a single video and programmatically slices it into frames. As the ball moves, the app manipulates these slices to get velocity.
2. Backend (Python/OpenCV): Ball detection using color segmentation and shape analysis. Velocity calculation based on position changes.
   
3. Frontend (React): Captures video frames using canvas Sends frames to backend as base64-encoded images Receives and visualizes ball position and velocity

How we built it

We started by setting up a React application using Create React App, which provides a modern build setup with no configuration Integrating CreateJS: CreateJS, a library for working with the HTML5 Canvas element, was integrated to manage and manipulate the canvas. used opencv and computer vision

Challenges we ran into

One of the initial challenges was effectively integrating python and frontend react.

Accomplishments that we're proud of

Development of seamless experience to bring baseball video into stacast like metric by computer vision.

What we learned

here is a learning curve to openCv and ball detection

What's next for Dugout Vision

Built With

• opencv
• python
• react
• vertex