AIFingerCounter/HandTracker.py at main · indropal/AIFingerCounter · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
import cv2
import mediapipe as mp
import time


class fingerTip():

    def __init__( self, static_image_mode = False, max_num_hands = 2, min_detection_confidence = 0.5, min_tracking_confidence = 0.5 ):
        self.static_image_mode  = static_image_mode;
        self.max_num_hands = max_num_hands;
        self.min_detection_confidence = min_detection_confidence;
        self.min_tracking_confidence = min_tracking_confidence;

        self.mpHands = mp.solutions.hands
        self.hands = self.mpHands.Hands( self.static_image_mode,
                                         self.max_num_hands,
                                         self.min_detection_confidence,
                                         self.min_tracking_confidence
                                        );

        """
            '.Hands( static_image_mode, max_num_hands, min_detection_confidence, min_tracking_confidence )' ::

            --->
                "static_image_mode" --> parameter for Detection Or Tracking -- if set to False, then based on the confidence level it will track the detected Hand
                "max_num_hands" --> maximum Number of Hands to detect & track
                "min_detection_confidence" --> min. model confidence threshold for Hand Detection
                "min_tracking_confidence" --> min. model confidence threshold for Hand Tracking -- if confidence value goes below the set threshold, then the model will initiate Detection
        """

        self.mpDraw = mp.solutions.drawing_utils; # the drawing utilities function for drawing lines and dots of the tracked/detected hand

    def findHands(self, img):

        """
            Method to detect/track Hands using MediaPipe & mark out finger tips from the 'img' - image variable
        """
        imgRGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB); # convert captured image to RGB colors
        results = self.hands.process(imgRGB) # process the captured RGB image from the webcam using Mediapipe method
        # "results.multi_hand_landmarks" --> contains results for Hand Detection / Tracking
        if(results.multi_hand_landmarks):
            # if 'Hand' is being detected & tracked then, process them.
            for handLandmarks in results.multi_hand_landmarks:
                #print(handLandmarks); # print out the detected hand landmark features.

                # obtain the details for each individual Hand Landmarks -- Each landmark (a total of 21 such Landmarks are detected) along with coordinates
                for id, landMark in enumerate(handLandmarks.landmark):
                    # print(id, landMark) # Usually the landMark coordinates are in Ratio i.e. 0-1 values so that they can be scaled accordingly with changing image size
                    height, width, channel = img.shape
                    imgX, imgY = int(landMark.x * width), int(landMark.y * height)

                    # targetting specific landmarks with their ID values -- tip of every finger & the wrist
                    if id == 0: # wrist
                        # if 'id' --> 0 landmark is detected, then color it on the image feed differently
                        cv2.circle(img, (imgX, imgY), 15, (234, 234, 101), cv2.FILLED )
                        # Landmark with 'id = 0' will have detected landmark covered with a circle with RGB(234, 234, 101)

                    if id == 4: # thumb tip
                        # if 'id' --> 4 landmark is detected, then color it on the image feed differently
                        cv2.circle(img, (imgX, imgY), 15, (255, 128, 128), cv2.FILLED )
                        # Landmark with 'id = 4' will have detected landmark covered with a circle with RGB(255, 128, 128)

                    if id == 8: # index finger tip
                        # if 'id' --> 8 landmark is detected, then color it on the image feed differently
                        cv2.circle(img, (imgX, imgY), 15, (255, 128, 255), cv2.FILLED )

                    if id == 12: # middle finger tip
                        # if 'id' --> 12 landmark is detected, then color it on the image feed differently
                        cv2.circle(img, (imgX, imgY), 15, (150, 150, 255), cv2.FILLED )

                    if id == 16: # third finger tip
                        # if 'id' --> 16 landmark is detected, then color it on the image feed differently
                        cv2.circle(img, (imgX, imgY), 15, (0, 127, 255), cv2.FILLED )

                    if id == 20: # pinky finger tip
                        # if 'id' --> 16 landmark is detected, then color it on the image feed differently
                        cv2.circle(img, (imgX, imgY), 15, (200, 255, 75), cv2.FILLED )

                self.mpDraw.draw_landmarks(img, handLandmarks, self.mpHands.HAND_CONNECTIONS); # draw the detected hand's landmark features on the image feed using mediapipe's utility functions & 'mpHands.HAND_CONNECTIONS' draws the connecting lines between the features


        return img; # return the image on which detection/tracking has been performed

    def specificHandPosition(self, img, handIdx = 0, draw = True ):

        """
            method to return the Landmarks for a particular Hand being detected
            Specific Hand is denoted by the 'handIdx'
        """

        imgRGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB); # convert captured image to RGB colors
        results = self.hands.process(imgRGB) # process the captured RGB image from the webcam using Mediapipe method
        landMarkList = []
        # "results.multi_hand_landmarks" --> contains results for Hand Detection / Tracking

        if ( results.multi_hand_landmarks ):
            myHand = results.multi_hand_landmarks[handIdx]

            for idx, landmark in enumerate( myHand.landmark ):
                # retireve required info
                height, width, channel = img.shape;
                imgX, imgY = int(landmark.x * width), int(landmark.y * height)
                landMarkList.append( { 'landmark_ID': idx, 'X': imgX, 'Y': imgY } )

                if (draw):
                    cv2.circle(img, (imgX, imgY), 7, (0, 128, 255), cv2.FILLED); # draw the detected hand's landmark features on the image feed using mediapipe's utility functions & 'mpHands.HAND_CONNECTIONS' draws the connecting lines between the features

        return landMarkList


def main():
    prevT = 0;
    currT = 0;
    cap = cv2.VideoCapture(0); # '1' passed to get video feed from webcam

    # create new fingerTip class object.
    handDetector = fingerTip()

    while True:
        success, img = cap.read()

        #img = handDetector.findHands(img);
        landMarkCoordinates = handDetector.specificHandPosition( img, 0, draw = True )  # returns list of dictionary attributes --> { 'landmark_ID': idx, 'X': imgX, 'Y': imgY }

        if ( len(landMarkCoordinates) != 0):
            for i in landMarkCoordinates:
                if i['landmark_ID'] == 0:
                    print(i)

        currT = time.time()
        framesPerSecond = 1 / (currT - prevT)
        prevT = time.time()
        cv2.putText(img, str(int(framesPerSecond)), (20, 50) , cv2.FONT_ITALIC, 1.5, (255,255,128), 3 );

        cv2.imshow("Image", img)
        cv2.waitKey(1)


if __name__ == "__main__":
    main()