inference.py

import pickle
import cv2
import mediapipe as mp
import numpy as np

# Open and load model from file
model_dict = pickle.load(open('./model.p', 'rb'))
model = model_dict['model']

# Read video from camera
cam = cv2.VideoCapture(1)  # Remeber changing based on the OS, with Mac it is 1

# Drawing hands with mediapipe
mp_hands = mp.solutions.hands
mp_drawing = mp.solutions.drawing_utils
mp_drawing_styles = mp.solutions.drawing_styles

hands = mp_hands.Hands(static_image_mode=True, min_detection_confidence=0.3)

# Get labels for the machine learning models (The index represent the folder storing images for each label)
labels_dict = {0: 'A', 1: 'B', 2: 'C', 3: 'D', 4: 'E', 5: 'F', 6: 'G',
               7: 'H', 8: 'I', 9: 'K', 10: 'L', 11: 'M', 12: 'N', 13: 'O',
               14: 'P', 15: 'Q', 16: 'R', 17: 'S', 18: 'T', 19: 'U', 20: 'V',
               21: 'W', 22: 'X', 23: 'Y'}
while True:

    data_aux = []
    x_ = []
    y_ = []

    # Read frames from camera
    ret, frame = cam.read()

    # Get the dimension and shape of frame with the format x, y, z
    H, W, _ = frame.shape

    # Converting image to RGB
    frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)

    results = hands.process(frame_rgb)
    if results.multi_hand_landmarks:
        for hand_landmarks in results.multi_hand_landmarks:
            mp_drawing.draw_landmarks(
                frame,  # image to draw
                hand_landmarks,  # model output
                mp_hands.HAND_CONNECTIONS,  # hand connections
                mp_drawing_styles.get_default_hand_landmarks_style(),
                mp_drawing_styles.get_default_hand_connections_style())

        for hand_landmarks in results.multi_hand_landmarks:
            for i in range(len(hand_landmarks.landmark)):
                x = hand_landmarks.landmark[i].x
                y = hand_landmarks.landmark[i].y

                x_.append(x)
                y_.append(y)

            for i in range(len(hand_landmarks.landmark)):
                x = hand_landmarks.landmark[i].x
                y = hand_landmarks.landmark[i].y

                # Get the data closer to the value
                data_aux.append(x - min(x_))
                data_aux.append(y - min(y_))

        x1 = int(min(x_) * W) - 10
        y1 = int(min(y_) * H) - 10

        x2 = int(max(x_) * W) - 10
        y2 = int(max(y_) * H) - 10

        # Making prediction based on trained data
        prediction = model.predict([np.asarray(data_aux)])

        # Getting the predicted data from trained data
        predicted_character = labels_dict[int(prediction[0])]

        # Showing the letter or the prediction on to the windows
        cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 0, 0), 4)  # Create a box
        cv2.putText(frame, predicted_character, (x1, y1 - 10), cv2.FONT_HERSHEY_DUPLEX, 1.3, (0, 0, 0), 3,
                    cv2.LINE_AA)

    # Showing frame by frame
    cv2.imshow('frame', frame)
    cv2.waitKey(1)


cam.release()
cv2.destroyAllWindows()