Proposal Gesture Recognition Algorithm Combining CNN for Health Monitoring

Today, artificial intelligence (AI) has become a trend that is studied by many research. It simulates the process of learning human thinking for computers or supercomputers used to handle certain types of work such as controlling a house, image recognition research, processing data of disease to give treatment regimens, processing data for self-learning, answering questions of patient and customers. Besides, the interaction between humans and computers is more diverse. Using hand gestures is considered an effective method to communicate with each other for health monitoring system. The proposal algorithm consists of three steps, namely capturing frame from videos, extracting their characteristic, and classifying actions. In proposal algorithm, we select several gestures such as baby waking up, walking, and falling for detecting and warning to doctor. The results show that proposal algorithm improves the accuracy up to 90 percent. However, disadvantage of algorithm is to requires long processing time and large memory.