Computer Vision News - March 2022

9 MediaPipe # Run MediaPipe Hands and plot 3d hands world landmarks. with mp_hands.Hands( static_image_mode=True, max_num_hands= 2 , min_detection_confidence= 0.7 ) as hands: for name, image in images.items(): # Convert the BGR image to RGB and process it with MediaPipe Hands. results = hands.process(cv2.cvtColor(image, cv2.COLOR_BGR2RGB)) # Draw hand world landmarks. print (f'Hand world landmarks of {name}:') if not results.multi_hand_world_landmarks: continue for hand_world_landmarks in results.multi_hand_world_landmarks: mp_drawing.plot_landmarks( hand_world_landmarks, mp_hands.HAND_CONNECTIONS, azimuth=5) We are at the end of this first chapter on MediaPipe. We gave a basic introduction to this platform, and we looked at one of their machine learning pipelines in their Visualizer software and through the interaction with Python. In the second chapter of this series, we will further explore its functions and code base, while looking at the interaction with Javascript and mobile solutions. I look forward to discovering more together, and to your feedback and questions on MediaPipe  Figure 2: Result of Hand Tracking pipeline when using parameter min_detection_confidence set to 0.5 Figure 1: Result of Hand Tracking pipeline when using parameter min_detection_confidence set to 0.7