Detection of Finger Height for a Multi-Touch Mouse

Abstract

An image of multiple fingers is acquired by an input system that consists of a camera with a fisheye lens, infrared LEDs, and a transparent shell. When the LEDs illuminate fingers uniformly, the distance between the fingertips and the lens can be deduced by analyzing the intensity of the fingertip regions and/or the size of the fingers in the image. Azimuth and polar angles of the fingertips are obtained from the image directly. Repeating this process provides a series of three-dimensional coordinates of multiple fingertips. Such information might be useful for realizing a multi-touch mouse with height detection capability.

Detection of touch event with a camera-based user interface

“…Absorption coefficients of deoxy-hemoglobin are similar at 700nm and 760nm. The measured emission spectra for this LED and the infrared LED used in the original system [5] are shown in Fig. 3.…”

“…4, the distributions are not quite uniform and this is a subject for future improvement. All the other components are the same with those used in the previous system [5]. …”

“…For example, the algorithm described in [5] allows us to draw a line crossing the center of each finger and the intensity profile along this line becomes available. Because the size of a finger is more or less the same among individuals, a certain minimum length can be set for the flat portion of the intensity profile.…”

“…Repeating this process provides a series of three-dimensional (3d) coordinates of multiple fingertips. In experiment, we assembled off-the-shelf components, developed an algorithm to extract the intensity information from the fingertip regions and acquired the 3d coordinates successfully [5].…”