Katon: real-time hand-gesture recognition on tiny embedded systems

Abstract

As gesture recognition technology transforms human-computer interactions across various applications such as Virtual Reality (VR) and other smart wearables, the demand for real-time gesture recognition on limited-capacity hardware increases. Most of these applications need efficiency and responsiveness, which highlights the importance of research in lightweight algorithms capable of real-time processing in resource limited areas. This study evaluates various lightweight algorithms for their viability in real-time hand gesture recognition on resource-limited hardware, focusing on the Himax WE-I Development Board.