Vision-Based Finger Guessing Game in Human Machine Interaction

Abstract

This study uses motion hand gesture recognition to demonstrate a finger guessing game in human machine interaction. By analyzing body movements at the moment of signing to determine the timing of the signing gesture completion, and through imaging analysis and recognition, this game system can categorize the signs of the users, thus allowing no need of the users for the button, keyboard, mouse, or any special device for the input interface, but simply natural body movement to show the hand gestures required by the finger guessing game. Such human machine interaction provides more sense of reality and better interaction during game play between human and machines. Upon the recognition of user's hand gesture, this finger guessing game system uses the simple tabular cross-linking method to map it against the sign randomly selected by the computer, and further determine the game result. The robot used for interacting with the human is a mobile platform for an autonomic DSP-carrying(digital signal processor) closed loop positioning controller, which demonstrates a simple human machine interactive game, by showing the results of finger guessing games for the players and robot. As indicated by test result, this motion hand gesture recognition technology allows the users to interact and play games with robots in ways that are more realistic and natural, thus providing more enjoyment for people in an effective way.