Gesture centric interaction: evaluating hand and head gestures in touchless cursor control.

Abstract

Touchless interfaces have gained considerable importance in the modern era, particularly due to their user-friendly and hygienic nature of interaction. This article presents the designing of two touchless cursor control systems based on hand gestures and head movements utilising the MediaPipe framework to extract the key landmarks of the hand and face utilising a laptop camera. The index finger's landmark points are tracked and converted to corresponding screen coordinates for cursor control. Similarly, yaw and pitch angles of head movements are computed in the head movement-based cursor design. A comprehensive performance evaluation of the two proposed systems based on a two-dimensional (2D) Fitts' law experiment revealed superior performance for the hand gesture-controlled cursor compared to the head movement-controlled cursor with throughputs of 0.59 bps and 0.53 bps respectively. Participants also favoured hand gesture-based cursor control over head movement-based cursor control in terms of overall experience and task difficulty.