camera.py

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

# Load the pre-trained model
model_dict = pickle.load(open('model.p', 'rb'))
model = model_dict['model']

# Initialize MediaPipe Hands
mp_hands = mp.solutions.hands
hands = mp_hands.Hands(min_detection_confidence=0.8, min_tracking_confidence=0.8)
mp_drawing = mp.solutions.drawing_utils

class VideoCamera(object):
    def __init__(self):
        self.video = cv2.VideoCapture(0)
        self.stopped = False  # Flag to control video streaming

    def __del__(self):
        self.video.release()

    def stop_stream(self):
        self.stopped = True

    def get_frame(self):
        while not self.stopped:
            ret, frame = self.video.read()
            if not ret:
                break

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

            H, W, _ = frame.shape
            frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
            frame_rgb = cv2.flip(frame_rgb, 1)

            results = hands.process(frame_rgb)
            if results.multi_hand_landmarks:
                for hand_landmarks in results.multi_hand_landmarks:
                    mp_drawing.draw_landmarks(
                        frame_rgb,
                        hand_landmarks,
                        mp_hands.HAND_CONNECTIONS
                    )

                    for landmark in hand_landmarks.landmark:
                        x_.append(landmark.x)
                        y_.append(landmark.y)
                        data_aux.append(landmark.x)
                        data_aux.append(landmark.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
                prediction = model.predict([np.array(data_aux)[:42]])[0]

                cv2.rectangle(frame_rgb, (x1, y1 - 10), (x2, y2), (255, 99, 173), 6)
                cv2.putText(frame_rgb, prediction, (x1, y1), cv2.FONT_HERSHEY_DUPLEX, 1, (255, 0, 0), 2, cv2.LINE_AA)

            ret, jpeg = cv2.imencode('.jpg', frame_rgb)
            frame = jpeg.tobytes()
            yield (b'--frame\r\n'
                   b'Content-Type: image/jpeg\r\n\r\n' + frame + b'\r\n\r\n')