2.4 A Comparison of Tracking Techniques

An overview of the advantages and drawbacks of the techniques described in the preceding paragraphs can be seen in table 1; we have left fields blank in cases where the appropriate information was unavailable. The dimensions of usability refer to the list , at the start of this section-we have omitted the dimension of recording compatibility because any recording in principle is compatible, given enough computer power, with body localization recordings. In some cases we indicate the accuracy or range for both the horizontal (h) and vertical (v) directions. We have also indicated whether the technique is able to measure the pupil diameter, and finally we have listed the price range for the equipment. The information in the table is mainly a combination of what Scott & Findlay (1993), Wooding (1995) and Baluja & Pomerleau (1994) report-please note that some of these figures originate from the manufacturers, and could thus be `ideal' values. Furthermore, several of the techniques are supplied commercially in different "flavours" (e.g. headband-mounted or trial frame mounted), and thus some of the figures listed are dependant upon which flavour of the technique is used. For example, the corneal/pupil reflection relationship technique normally requires the user to sit fairly still so the tracked eye is in the field of view of the tele-lens; but the same technique is also supplied in an "extended" flavour, where a second, wide-angle camera is used to locate the user's head and subsequently orient the tele-lens correctly (see figure 4). This will of course improve the head mobility figures.

Figure 4: Eye tracking using a mirror in front of the camera. Figure taken from Bolt (1984).