Evaluasi Jarak Deteksi Antara Gestur Tangan Dan Kamera Webcam Dengan Metode Mediapipe

Abstract

In recent years, there has been a growing focus on gesture recognition and human-machine interaction, with hand gestures acting as controllers in applications such as virtual reality and interactive software. Google’s MediaPipe framework has played a significant role in advancing hand detection technology, enabling real-time recognition and tracking of hand gestures. This system leverages image processing and pattern recognition techniques. A key challenge lies in accurately measuring the distance between the user’s hand and the detection device, which is essential for ensuring a seamless and intuitive user experience. Evaluation of MediaPipe shows that it can effectively detect hand gestures at distances ranging from 20 cm to 290 cm. However, beyond 290 cm and up to 400 cm, some hand gestures become difficult to detect. Notably, when the distance exceeds 450 cm, all hand gestures are no longer detectable by the webcam. This assessment highlights the importance of understanding the limitations caused by environmental factors, lighting conditions, and hand positioning, which affect the accuracy of distance detection. This evaluation offers valuable insights into MediaPipe’s capabilities, which can guide decisions for applications that require consistent and precise hand gesture recognition.