Hand gesture control accuracy through increased epoch and batch size

Abstract. The conventional method of mechanism control which requires physical contact is being replaced with remote control, especially the advent Internet things (IoT). Facial recognition used to identify and authenticate faces from images or videos. It has many applications, including granting restricting access a facility, secured areas, preventing unauthorized users, selective access. accuracy such system very important depends on how was modeled adopted. In this study, facial gesture were collated in python neural network, then optimized using TensorFlow. algorithm compiled unto raspberry pi for testing developed automatic gate. An effective achieving epoch presented work. Five (0, 5, 10, 15, 20) batch number respectively achieve system. model trained five gestures; fist, palm, thumb up, thumb, last finger. achieved through maximum at 99.97 when both set 20. However, no set, below 10.2, whereas there 98% introduced. This depicts importance image recognition. also discovered that higher greater accuracy, but processing would require high unit.