Stereoscopic picture generating apparatus
View Patent ↗An apparatus for rendering a three-dimensional model from a plurality of eye points to present a three-dimensional picture, wherein the position of a virtual camera for rendering apexes of the three-dimensional model is controlled in accordance with the amount of protrusion of an object under rendering from a display plane to create and present a non-perspective projection picture as a picture at each eye point position. Also, for this control, a weighting parameter is incorporated in each part of the three dimensional model, such that the presentation of stereoscopic sense can be finely controlled.
1. A stereoscopic picture generating apparatus comprising:
a storage device for storing information indicative of positions of a plurality of predetermined eye positions; and
a processing unit for creating a stereoscopic picture of a three-dimensional model based on the information, wherein said processing unit creates two-dimensional images that are rendered as images of the three-dimensional model from converted eye positions, wherein the two dimensional images are created by applying conversion, the conversion is applied by calculating a two-dimensional image position at an eye point e for global position coordinate values vs_k of apex data of a model M i, a calculation is performed using the projection matrix P ei preserved in the model M i,
[
ax
ay
az
]
←
P
ei
·
(
vs
j
-
p
ei
)
u
k
←
[
ax
az
+
1
ay
az
+
1
]
wherein,
f
xi
=
f
xj
1
p
ejz
+
1
,
f
yi
=
f
yj
1
p
ejz
+
1
q
xi
=
f
xi
p
ejx
,
q
xy
=
f
yi
p
ejy
P
ei
=
[
f
xi
0
q
xi
0
f
yi
q
yi
0
0
1
]
,
f_x, f_y represent vertical and horizontal values in three-dimensional plane; p_ex, p_ey, p_ez represent eye point position represented in three dimensional vector; ax, ay, az are positions in three-dimensional coordinate wherein of different parameters are used for different ones of the predetermined eye positions during the conversion, wherein each of the parameters is set to make a distance between the converted eye positions shorter than a distance between the predetermined eye positions, and to make a distance between a screen on which the two-dimensional images are rendered and the converted eye positions shorter than a distance between the screen and the predetermined eye positions, wherein said processing unit displays each of the two-dimensional images which are switched from one converted eye position to another.
2. A stereoscopic picture generating apparatus according to claim 1 , wherein:
said predetermined calculation includes finding a product of a weight registered in each part of the model and a protrusion amount of a apex from a display plane.
3. A stereoscopic picture generating apparatus according to claim 2 , wherein:
said processing unit reduces the amount of parallax between images corresponded to the plurality of predetermined eye positions in accordance with the product.
4. A stereoscopic picture generating apparatus according to claim 2 , wherein:
said processing unit makes a rate at which the parallax of each model is reduced proportional to an inverse of a depth distance from the predetermined eye positions.
5. A stereoscopic picture generating apparatus according to claim 4 , wherein:
said processing unit selects a projection parameter such that a linear portion of the model maintains the linearity even after the conversion.
6. A stereoscopic picture generating apparatus according to claim 2 , wherein:
said processing unit renders a protruding portion of the model while limiting the amount of parallax between images corresponded to the respective predetermined eye positions to a fixed value.
7. A stereoscopic picture generating apparatus according to claim 2 , wherein:
said processing unit changes a rendering method based on switching information.
8. A stereoscopic picture generating apparatus comprising: an interface for connecting to an external device; a storage device for storing a program and information; and
a CPU for executing the program,
wherein said information includes information indicative of positions of a plurality of predetermined eye positions, each corresponding to a view from one eye, and said program comprises a routine for creating a two-dimensional images that are rendered as images of the three-dimensional model from converted eye positions, wherein the two-dimensional images are created by applying conversion, wherein different parameters are used for different ones of the predetermined eye positions during the conversion, the two-dimensional images are created by applying conversion of the two-dimensional images positions at an eye point e for global position coordinate values vs k of apex data of a model M i, a calculation is performed using a projection matrix P ei preserved in the model M i,
[
ax
ay
az
]
←
P
ei
·
(
vs
j
-
p
ei
)
u
k
←
[
ax
az
+
1
ay
az
+
1
]
wherein
f
xi
=
f
xj
1
p
ejz
+
1
,
f
yi
=
f
yj
1
p
ejz
+
1
q
xi
=
f
xi
p
ejx
,
q
xy
=
f
yi
p
ejy
P
ei
=
[
f
xi
0
q
xi
0
f
yi
q
yi
0
0
1
]
,
f_x, f_y represent vertical and horizontal values in three-dimensional plane; p_ex, p_ey, p_ez represent eye point position represented in three dimensional vector; ax, ay, az are positions in three-dimensional coordinate wherein each of the parameters is set to make a distance between the converted eye positions shorter than a distance between the predetermined eye positions, and to make a distance between a screen on which the two-dimensional images are rendered and the converted eye positions shorter than a distance between the screen and the predetermined eye positions, wherein said program includes a routine for displaying each of the two-dimensional images which are switched from one converted eye position to another.