OpenCV & Mediapipe Hand Tracking Module

A hand tracking module for OpenCV and Mediapipe. Contains various functions such as counting fingers, finding positions of landmarks (return a list), checking if the hand is grabbing, checking which direction the hand is dragging/swipping, etc.

Functions

Functions:
findHands():            Finds and plots the skeletal structure on image.                    [Returns an Image]
findPosition():         Finds the position of each hand landmark.                           [Returns a list]
fingerCount():          Counts how many fingers are held up. Works best with upright hand.  [Returns a list]
checkGrab():            Checks if the hand makes a grab gesture.                            [Returns a boolean]
checkGrabAlt():         Alternative method for checkGrab(). [Returns a boolean]
swipeDirection():       Checks which way the user swipes their hand while grabbing. [Returns a string]
center_of_mass():       Finds the center of mass of the hand, indicated by a red dot.       [Returns a list]
palm_center():          Finds the center of the palm, indicated by a blue dot.              [Returns a list]
dist():                 Finds the distance between two points.                              [Returns a float]

How to use:

• create detector object and then call .findHands() and .findPosition() at beginning of each OpenCV iteration.
• Call the functions as needed, they can be called in the loop.
• Example:

# Declare stuff

detector = hm.HandDetector(detectionCon = 0.75, trackCon = 0.6)

while cap.isOpened():

    success, frame = cap.read()

    # You can change the display window size by resizing frame here.

    img = detector.findHands(frame) # Draws hands onto img

    detector.findPosition(frame, 2)

    # do stuff:
    # If you want to draw stuff or add text, draw it onto img, not frame.


    cv.imshow('Video', img)         # Display img.

cap.release()
cv.destroyAllWindows()

Libraries:

opencv
mediapipe
numpy
math
time

Other things

The module is not perfect, support for multiple hands in the frame at the same time is still being developed. It will hopefully be updated regularly.

Thanks!