Tech Stack -
Python, Mediapipe, OpenCV, pyautogui
Algorithm -
Landmark detection
- The version 1 of this project used
color detection
technology to detect the position of the hand.
- The project can be viewed
HERE
.
- The problem with that approach was that user had to have something of green color in their hands.
- The version 2 of this program uses
hand tracking technology
to detect hands in the real time
.
Mediapipe
, a library buit by google, provides ready-to-use real time hand tracking and landmark estimation
feature.
- Although, using this model affects the FPS a little bit, but overall it was found to be a better solution.
pyautogui
was used to control the mouse pointer.