![](https://csdnimg.cn/release/download_crawler_static/15168376/bg5.jpg)
Multimed Tools Appl
Fig. 3 Selection gestures. a Activating the selection mode. b Selecting a block. c–d Viewing more
blocks on other pages
3 Body gestures
We design a set of body gestures for 3D scene navigation and virtual object control.
Our body gestures can be classified into three categories: selection, navigation, and
assembly. One of the most important principles in gesture design is to make all
gestures clearly distinguish from each other, so as to avoid the misinterpretation or
ambiguity. Another factor considered in our current system is that the user always
faces the screen, so that he can see the virtual scene on the screen.
1
Last but not
the least, the designed gestures should be intuitive, and comfortable to perform. The
navigation gestures are inspired from road traffic control gesture, while the selection
gestures and the assembly gestures mimic the mouse operations and real operating
gestures (such as how people hold object with two hands).
Selection gestures are used in the selection mode allowing the user to browse all the
available building blocks and select one of them. To activate a selection session, the
user quickly waves his hand at least two times (see Fig. 3a). In the selection mode, the
user can freely move his hand to control the cursor in the browsing window. To select
a block, the user simply pushes his hand forward (see Fig. 3b). Furthermore, the user
can also swipe to left or right to view more blocks on other pages (see Fig. 3c, d).
Navigation gestures are used to control the avatar in the virtual environment. In
our system, the avatar can walk along the directions of coordinate axes. The walking
speed of the avatar is specified by the user before starting the system. As shown in
Fig. 4 ,we define five gestures to control the avatar’s movement (go straight, turn left,
turn right, stop and finish).
Assembly gestures are used to place the block in the virtual structure and further
adjust its location/orentation. To do that, the user first performs a holding gesture to
1
This requirement would not be necessary if we configurate our system in an immersive environment.