Frontend Backend Connection

I got the frontend to send frames to the backend, then backend runs mediapipe on the frames then sends frames back to backend where they are displayed

It is pretty laggy right now, so I will add a latest-frame buffer so that I can grab the newest frames, not the older ones.

Detecting Human Body

MediaPipe

I finally got to understand how mediapipe works and how to set it up. I got it select certain body parts that are key to shooting a basketball, like the knees, elbows, wrists, shoulder, ankles, and some more.

I used circles to identify those spots on my body, and it worked. The fps was really smooth comapared to the YOLO model for tracking the ball, so Im happy about that. Next I need to focus on finding angles of the person, like the angle when they make their shot.

Server Connection

Today i got create a server connection and connect it to multiple devices on my Wi-Fi network. I used uvicorn and fastapi to connect python to javascript.

Error

Windows firewall was blocking connection for other devices, so they weren’t able to get in. I disabled all rules for python, so then it worked. But only the frontend was working, the backend did not work.

BasketLab Home Page (Will add CSS trust)

Today’s focus was more on the website, not on the actual tracking. I got the idea for my home page.

HTML

I made the basic html file today. Right now, I just created my idea and the words. I used a lot of semantic tags, maybe too much, but overall it was pretty good.

CSS

I just touched the CSS part for this home page. I worked mainly on getting the background color, fonts and the navigator bar at the top. I got my font from google. For the navigator bar, I worked on the spacing between the buttons and logo.

Added more pictures (trust me)

• still learning how to use tag
• need to add yolo AI
• need to integrate Mediapipe

Worked on

• website (basic html)
• need to get domain maybe
• might need server later, not needed though

Created Website (trust the process)

I’m still learning html and css,

but this is a good place to start

• Might put some more memes???
• Use css a lot more
• no js right now, later implement it

Tracking stuff

• add OpenCV later
• convert yolo AI model to onnx,
• Put some better graphics

YOLO Model for tracking

Tracking System

What I Did

• Tracking.py created

  • used opencv to help me track
  • created simple tracking system for a ball
  • works for any ball
  • uses camera real time

• Split main.py into two files

  • video.py
  • image.py

• added video for data

Need to improve

• work on color detection
• switch to hsv

Setup project

This is my first devlog of this project.

What I did today:

• learned how to get images and show them (example in picture)
• learned about venv and downloaded a bunch of libraries
• still working on learning opencv-python for processing

Need to work on later

• Add video processing
• Add analysis of basketball