Contactless human-computer interaction using a Deep Neural Network Pipeline for real-time video interpretation and classification

Abstract

Nowadays, all applications are developed with the user’s comfort in mind. Regardless of the application’s objective, it should be as simple as possible so that it is easily accepted by its users. With the evolution of technology, simplicity has evolved and has become intrinsically related to the automation of tasks. Therefore, many researchers have focused their investigations on the interaction between humans and computing devices. However, this interaction is usually still carried out via a keyboard and/or a mouse. We present an essemble of deep neural networks for the detection and interpretation of gestural movement, in various environments. Its purpose is to introduce a new form of interaction between the human and computing devices in order to evolve this paradigm. The use case focused on detecting the movement of the user’s hands in real time and automatically interpreting the movement.