A Virtual Keyboard Implementation Based on Finger Recognition

Abstract

A keyboard requires a great deal of resources and is restricted by its physical features. Additionally, discarded keyboards also inevitably contribute to environmental pollution. Consequently, the touch screen is designed to replace the physical keyboard and thus reduce these flaws. However, the internal digital keyboard on the touch screen takes up a substantial amount of space, which causes some content to be covered. Moreover, the touch screen can be dirtied by fingerprints and become worn over time by human fingernails through frequent use. Hence, it is necessary to develop a new type of environment-friendly virtual keyboard with fewer flaws.

In this thesis, the author develops a novel type of virtual keyboard that allows users to use fingers to type on a piece of paper at any fixed plane. Finger recognition is based on human skin tone and then the BWMORPH algorithm is utilized to recognize the user’s fingertip. If the user’s fingertip has remained on a key for a long time, the program will regard this key as an input. The experiments in this thesis adopt five kinds of customized paper keyboards on a wall to demonstrate the usability of the proposed virtual keyboard. Typing without touching the keyboard is fulfilled to ignore obstrutions covering the paper keyboard. The experiment results indicate that the overall recognition rate of the proposed virtual keyboard is 94.62%.

The proposed virtual keyboard can be put to use for a smartphone in the future. Furthermore, as a blueprint, it can be applied to computers after it is improved to allow ten-finger recognition. Moreover, machine learning can be potentially embedded into our virtual keyboard so as to greatly improve its finger recognition performance.