app.py

import cv2
import gradio as gr
import mediapipe as mp

mp_drawing = mp.solutions.drawing_utils
mp_drawing_styles = mp.solutions.drawing_styles
mp_hands = mp.solutions.hands

def process_frame(img):
    if img is None:
        return None
    with mp_hands.Hands(
        model_complexity=0,
        min_detection_confidence=0.5,
        min_tracking_confidence=0.5
    ) as hands:
        img.flags.writeable = False
        image = cv2.flip(img[:, :, ::-1], 1)
        # Convert the BGR image to RGB before processing.
        results = hands.process(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
        img.flags.writeable = True
        if results.multi_hand_landmarks:
            for hand_landmarks in results.multi_hand_landmarks:
                mp_drawing.draw_landmarks(
                    image,
                    hand_landmarks,
                    mp_hands.HAND_CONNECTIONS,
                    mp_drawing_styles.get_default_hand_landmarks_style(),
                    mp_drawing_styles.get_default_hand_connections_style()
                )
        return cv2.flip(image[:, :, ::-1], 1)

with gr.Blocks(title="Realtime Keypoint Detection | Data Science Dojo", css="footer {display:none !important} .output-markdown{display:none !important}") as demo:
    with gr.Row():
        with gr.Column():
            video_input = gr.Video(label="Webcam Input", format="mp4")  # Video input for real-time processing

        with gr.Column():
            output_image = gr.Image(label="Output Image")  # Processed image output

    def update_frame(video):
        frame = cv2.VideoCapture(video).read()[1]  # Read the current frame
        if frame is not None:
            processed_frame = process_frame(frame)
            return processed_frame
        return None

    video_input.change(
        update_frame, 
        inputs=[video_input], 
        outputs=[output_image]
    )

demo.launch(debug=True)