| Glossary |
List of Common Interaction Metaphors
|
Virtual Hand
|
Description
:- SELECTION/MANIPULATION technique.
One
to one mapping between physical and virtual hand.
Inputs
:-
1.
Tracked hand position
2.
Discrete input (button) to indicate selection.
Notes
:-
Disadvantage
that must be close to object to select/manipulate it. But a very natural
interaction metaphor.
Possible
Events generated:-
Implementation :-
SELECTION - Simple traversal of scene graph to
find collision between object and tracked hand.
MANIPULATION
- Attach object to hand
in scene graph, so it inherits hands transformations, then on release
reattach object to world.
|
Ray Casting
|
Description
:- SELECTION/MANIPULATION technique.
Ray
cast from tracked hand, nearest intersected object is potentially
selected.
Inputs
:-
1.
Tracked hand position and direction vector.
2. Discrete input (button) to indicate selection.
Notes
:-
Proven to perform
well, selection only needs user to control 2 DOF not 3.
Possible
Events generated:-
Implementation :-
SELECTION - Simple
traversal of scene graph to find collision between object and ray.
MANIPULATION
- Attach object to hand
ray in scene graph at the distance at time of selection, so it inherits
hands transformations, then on release reattach object to world.
|
Occlusion
Selection "Sticky Finger"
|
Description
:- SELECTION technique.
Ray cast from eye
through hand, nearest intersected object is potentially selected. Effect
is that you place hand over object on image plane to select.
Inputs
:-
1.
Tracked hand position.
2.
Tracked Head Position (eye)
3. Discrete input (button) to indicate selection.
Notes
:-
Possible
Events generated:-
Implementation :-
|
Go-Go
arm extension
|
Description
:- SELECTION/MANIPULATION technique.
Like virtual hand, but
users reach is greatly extended by a non linear mapping of real hand-torso
distance and virtual hand-torso distance. For close objects mapping is
one-one, and becomes non linear after some threshold is crossed.
Inputs
:-
Notes
:-
Possible
Events generated:-
Implementation :-
|
HOMER
(Hand centered Object Manipulation Extending Ray casting)
|
Description
:- MANIPULATION technique.
Uses ray casting to
select, then moves virtual hand to object for manipulation. Any subsequent
move of orientation of tracked hand changes orientation of object (like
virtual hand) but any move in position of hand moves object position
according to linear mapping of physical hand-torso distance, to virtual
hand-torso distance.
Inputs
:-
1.
Tracked hand position and orientation.
2.
Tracked head or torso position.
3. Discrete input (button) to indicate
manipulation start stop.
Notes
:-
Possible
Events generated :-
Implementation :-
|
Scaled
World Grab
|
Description
:- MANIPULATION technique.
On selection user is
scaled (or equivalently the world is scaled) so that the virtual hand
touches the selected object. In a mono system and/or when the user
doesn’t move, the user doesn’t notice the scaling, i.e. the image
remains the same before and after scaling. However when the user moves he
realises he is a giant (or a dwarf?).
Inputs
:-
1. Tracked hand position and
orientation.
2. Tracked head position.
3. Discrete input (button) to indicate manipulation
start stop.
Notes
:-
Possible
Events generated :-
Implementation :-
Simply? need to scale
world by the ratio
(eye - hand
dist)/(eye-object dist).
|
World
In Miniature (WIM)
|
Description
:- SELECTION/MANIPULATION/TRAVEL technique.
User selects and
manipulates objects in a miniature handheld copy of the world. Because WIM
is always near user, we can use simple virtual hand for selection and
manipulation. Can also use for navigation by including a representation of
the user in the WIM and moving this about.
Inputs
:-
1. Tracked hand position and
orientation.
2. Discrete input (button) to turn
WIM on/off.
3. Discrete input (button) to indicate
manipulation/selection start stop.
Notes
:-
Possible
Events generated :-
Implementation :-
You need a copy of the
scene graph, that has functionality of the normal full scale environment.
|
Pointing
|
Description
:- TRAVEL technique.
User specifies
direction of travel vector with some tracked device or prop, usually
attached to the hand, like a wand or glove.
Inputs
:-
1.
Direction vector (position not needed?).
2. Button or joystick to start/stop movement (if
not using continuous automatic movement).
Notes
:-
Possible
Events generated :-
Implementation :-
Very simple, translate
viewpoint by (direction vector * velocity)
|
Gaze-Directed
steering (not eye tracking)
|
Description
:- TRAVEL technique.
Users head is tracked
and direction head is pointing determines direction of travel
Inputs
:-
1)
Head direction vector (position not needed?).
2) Button to start/stop movement (if not using
continuous automatic movement).
Notes
:-
Movement maybe
constrained to ground plane or not. Disadvantage that user can’t look
around whilst moving.
Possible
Events generated :-
Implementation :-
Very simple, translate
viewpoint by direction vector * velocity
|
Map
Based Travel
|
Description
:- TRAVEL technique.
User moves “user
icon” on a map representation of VE, then system smoothly animates user
to selected position (or can instantly transport user, but can cause
disorientation, in games have a toggle for smooth animation, so for new
environments user will be smoothly animated, but for familiar environments
can just jump to selected point).
Inputs
:-
Needs user to be able
to manipulate objects on a map. So might have a ray pointer to select and
move user icon. But could also use any of selection manipulation
techniques. For this reason inputs depend on the choice of
selection/manipulation metaphor.
Notes
:-
Possible
Events generated :-
Implementation :-
Quite complex, need a
map representation of world, and then need to know mapping between map
user icon position and the Vworld (i.e. scale factor, origin of world in
map, etc). Also might need to calculate a path that avoids objects in
world.
|
Target
Selection
|
Description
:- TRAVEL technique.
Bit like map based
travel, User selects a (visible) object, then system smoothly animates
user to object position.
Inputs
:-
Needs user to be able
to select objects in world, so again can use any of selection techniques.
For this reason inputs depend on the choice of selection metaphor. Might
just select targets from a list/menu.
Notes
:-
Possible
Events generated :-
Implementation :-
Fairly simple once
selection metaphor is implemented. Again might need to calculate an
automatic path that avoids objects in world.
|
Grabbing
the air
|
Description
:- TRAVEL technique.
User stretches out arm
and grabs, then any subsequent hand gestures move the world around the
user. (like pulling on a rope, used in the game "Black and
White")
Inputs
:-
1.
Needs tracked hand (usually using pinch gloves).
2. Button to grab and release (or pinch with
gloves).
Notes
:-
Can be physically
tiring, because needs large arm/hand movements.
Possible
Events generated :-
Implementation :-
Translate world
according to vector from initial grab point to current hand position, do
this every frame until grab is released.
|
Liangs Cone
|
Description
:- SELECTION/MANIPULATION technique.
Like
Ray Casting, but a instead of a ray a cone is used. The nearest object to
the user which is inside the cone may be selected.
See
"Geometric Modeling using 6 DOF input devices" Jiandong Liang,
Mark Green.
Inputs
:-
1.
Tracked hand position and direction vector.
2. Discrete input (button) to indicate selection.
Notes
:-
Was
designed to allow easy pickup of very small objects, which might be
difficult using ray casting. Can have trouble selecting partilly
occluded objects as only nearest object to user in cone may be selected.
Possible
Events generated :-
Implementation :-
As
with ray casting but all the objects must be tested to see if they
fall within a conic volume defined by the position and direction of the
users wand. The nearest object in the cone to the user is potentially
selected.
|
Steeds Cone
|
Description
:- SELECTION/MANIPULATION technique.
A
variant of Liangs cone. When button is pressed and held all objects within
the current cone are highlighted, forming a potential set of
selectable objects P. The user then moves the cone, and any objects in P
which become outside the cone are removed from P. Once the user has a
single object left in P it may be selected by realesing the button.
Realese of the button before a single object remians in P means no
objects are selected.
Inputs
:-
1.
Tracked hand position and direction vector.
2. Discrete input (button) to indicate selection.
Notes
:-
May
have an advantage over Liangs cone in fact that can selected occluded
objects.
Possible
Events generated :-
Implementation :-
As
with Liangs cone, but all objects in cone are stored, forming a set of
potentially selected objects.
|
