Emerging Technologies

Dynamic stereograms based on eye convergence for displaying multilayered images

We propose a simple method to display volumetric data such as MRI images by presenting dynamic stereograms according to eye convergence. 3D representation using binocular vision is currently quite popular; however, although this method can display a 3D surface model, it is quite difficult to grasp the internal structure of the model from a 3D representation. Furthermore, the resolution is not sufficient for many practical applications, such as diagnosis. Another currently used method is multilayered graphics, which presents cross sections of volumetric 3D data. However, although the method is practicable using 2D flat screens, it requires that the depth or index of the images be explicitly selected by the user. Therefore, an intuitive method to select cross-sectional images is required. Our idea is to use the convergence of the user’s eyes. As the convergence is semi-automatically modulated according to the distance of the image, it can be used as an index for image selection. Another key feature of our method is the use of a dynamic stereogram. A typical stereogram requires the user to adjust the convergence. In contrast, our method automatically presents a dynamic stereogram in which the disparity between images matches the user’s current convergence angle. The system is simply composed of a gaze tracker and a PC monitor. The gaze tracker measures the convergence angle. Such a system could not only be applied to the display of medical cross-sectional data but could also be used to select the various layers of an image in drawing software.

Michi Sato, The University of Electro-Communications

Hirouki Kajimoto, The University of Electro-Communications