Extension Of Fitts’ Law In Non-contact Gesture Interaction In Machine-vision

In the field of human-computer interaction technology, the non-contact machine-visiongesture interaction has become the mainstream mode due to its features--easy to berecognized, natural and intuitive. However, this way of interaction has limitations from thepoint of view of science and technology or human physiology. Therefore, it is critical toestablish a model for the evaluation of interaction performance, which is also a guidance toimprove the user interface design. As widely applied, Fitts’ law is used frequently as theperformance evaluating model in pointing tasks. Researchers found Fitts’ law can be appliedfrom one-dimension to two or three dimension situation. Regarding interaction devices, Fitts’law is used to evaluate the performance of pin, track ball, touch screen, and mouse, etc. Butwhether it can be applied to gesture interaction is still unknown.This research firstly reviews the implementation and other related concepts of Fitts’ law.Secondly, based on the technical characteristics of the non-contact machine-vision gestureinteraction, a machine-vision imaging model is established theoretically. Thirdly, based on theanalysis of the mapping relationship of gesture interaction, this research proposes a conclutionthat Fitt’s law can be also applied to gesture interaction and an experiment was conducted toprove the conclution.This research found Fitts’ law can be used in most conditions of non-contactmachine-vision gesture interaction. Compared to the applicability for mouse and other contactpointing devises, applicability for gesture interaction is lower. Further, Fitts’ law is applicablefor8angles, in a medium or long distance（30cm-40cm）. This research also found that Fitts’law is applicable in shorter or longer distance for some angles. The best distances are30cmand40cm according to the experiment. When the ID is low, the best distance is around30cmwhile the best distance is around40cm when the ID is high. The result of the experiment alsosuggests when the gesture is finished by right hand, performance of moving down and left islow, which implies performance of gesture interaction is influenced by interaction technologyand human joint structure. Finally, hope the conclusions of this paper can provide referencefor interaction design of non-contact machine-vision gesture in the future.