Hand pose estimation and gesture detection has been challenging but has garnered a lot of recognition recently as it allows an interesting way for us to interact with our devices. From its inception, gesture recognition techniques have been used to understand human gestures and body language, thereby building a bridge between humans and machines. This enables us to replace input devices such as keyboard and mouse, and interact naturally with the machine. This can significantly improve user experience and accessibility for people with special needs. This thesis is centered on hand gesture recognition, which is both a challenging and promising topic in the area of human-computer interaction. This study explores the detection of hand gestures and poses through webcam images, and compares two different implementations: classical approach and MediaPipe.The classical approach is based on the computer vision methods, where the image frames obtained from the webcam are subjected to filtering, detection and recognition. However, MediaPipe is a multimodal framework developed by Google to create Machine Learning pipelines for handling gesture recognition. From testing and experimental results it can be concluded that MediaPipe is more precise and faster than the classical approach, and can handle more complicated motions, poses and gestures, since it incorporates machine learning technologies under the hood.     «

Hand pose estimation and gesture detection has been challenging but has garnered a lot of recognition recently as it allows an interesting way for us to interact with our devices. From its inception, gesture recognition techniques have been used to understand human gestures and body language, thereby building a bridge between humans and machines. This enables us to replace input devices such as keyboard and mouse, and interact naturally with the machine. This can significantly improve user exper...     »