Information Technology Reference
In-Depth Information
Fig. 3.
The overview of the head rotation measuring process
actions because it was adequately small. Meanwhile, we measured the horizontal
head rotation by using image processing, for which we used a head mounted
camera and infrared-LEDs.
The overview of the measuring process is shown in Figure 3. The inputs of the
measuring process are three-axis acceleration data and a gray scale image taken
by a head mounted camera. The gray scale image captures the IR-LED patterns
which are set between the LCD monitors composed the immersive display. The
IR-LED patterns indicate the horizontal angle. In the first step, the angle of
vertical head rotation is extracted from the three-axis acceleration data. In the
next step, alignment errors of IR-LED patterns in the gray scale image are cor-
recting by using the acceleration data. In the third step, the angle of horizontal
head rotation is estimated by reading the IR-LED patterns.
Although the method to measure head rotation is robust, the head mounted
camera made some people feel uneasy in test installations. We can measure head
rotation by using a geomagnetic sensor, infrared range sensors and only image
processing. We are developing another method to measure head rotation by using
a small sensor which can measure geomagnetism and 3-axis acceleration because
of reducing this uneasiness of users.
4.2 The Implementation to Estimate Angles of Arms
In the CEBE for general situations, we implemented the method to estimate
angles of arms by using distance data of IR-range sensors. We could easily detect
the region of a human body in an image by using the IR-range sensor because
the sensor could obtain many 3D positions of measurement objects' surface. In
addition, the IR-range sensor was robust to occlusion than an optical motion
capture system in the prototype system. We, therefore, expected that we could
use this method for investigations about cultural differences. In this section, we
briefly describe the method.
