SYSTEMS AND METHODS FOR TREATING OCCLUSIONS IN 2-D TO 3-D IMAGE CONVERSION
The present invention is directed to systems and methods for processing 2-D to 3-D images. The system and method includes a procedure for optimizing occlusion and/or texturing by creating tolerances in which such texturing need not occur.
1 . A method for processing occlusions in 2-D to 3-D image conversion, said method comprises: forming a inter-object occlusion image (O) from at least two objects;
generating both a right eye occlusion image (OR) and a left eye occlusion image (OL) based on respective right eye and left eye views of said objects; and
replacing said occlusion image (O) with an image formed by an intersection of O, OR and OL.
2 . The method of claim 1 , wherein the image (O) is a boolean image.
3 . The method of claim 1 , further comprising:
generating a difference between O and OR to form a visible right occlusion;
generating a difference between O and OL to form a visible left occlusion; and
generating a union between OR and OL to form a total stereo occlusion.
4 . The method of claim 1 , wherein OR and OL are formed with respect to an original camera of the 2D image.
5 . The method of claim 1 wherein said forming comprises:
graphics application program (API) using shadow mapping.
6 . The method of claim 5 wherein a depth image (S) of said object is pre-computed.
7 . The method of claim 6 further comprising:
calculating said O directly from a single pass render of a selected one of said objects in isolation, said calculating utilizing said shadow mapping.
8 . The method of claim 1 wherein said forming comprises:
applying a tolerance value.
9 . The method of claim 8 wherein said tolerance value is applied manually.
10 . The method of claim 8 wherein said tolerance value is variable.
11 . The method of claim 1 wherein said forming complies with the equation:
O[p]:=if M[p] then S[p]˜< (or ≠) T[p] else false
where M is an object mask, S is a depth image of an entire scene, and T is a second depth image of said object in isolation.
12 . The method of claim 11 wherein said tolerance value is applied as a condition to said equation, said condition being:
ti abs ( S[p]−T[p] )>tolerance.
13 . The method of claim 1 further comprising:
determining, based on said replaced occluded image, portions of an image to be rendered.
14 . The method of claim 13 further comprising:
texturing only said portions of to be rendered.
15 . Code for controlling a processor to establish occlusions in 2-D to 3-D image conversion, said code comprising:
control sequences for forming a boolean inter-object occlusion image (O) from at least two objects;
control sequences for generating both a right eye occlusion image (OR) and a left eye occlusion image (OL) based on respective right eye and left eye views of said objects; and
control sequences for replacing said occlusion image (O) with an image formed by the intersection of O, OR and OL.
16 . The code of claim 15 wherein said control sequences for forming comprises:
control sequences for controlling shadow mapping.
17 . The code of claim 16 wherein a depth image (S) of said object is pre-computed.
18 . The code of claim 15 further comprising:
control sequences for calculating said O directly from a single pass render of a selected one of said objects in isolation, said calculating utilizing said shadow mapping.
19 . The code of claim 15 wherein said forming comprises:
control sequences for controlling a tolerance value.
20 . The code of claim 19 wherein said tolerance value is supplied manually.
21 . The code of claim 16 wherein said tolerance value is variable.
22 . The code of claim 15 wherein said forming complies with the equation:
O[p]:=if M[p] then S[p]˜?=T[p] else false
where M is an object mask, S is a depth image of an entire scene, and T is a second depth image of said object in isolation.
23 . The code of claim 22 wherein said tolerance value is applied as a condition to said equation, said condition being:
ti abs ( S[p]−T[p] )>tolerance.
24 . A method of minimizing processing resources during a 2-D to 3-D image conversion, said method comprising:
forming an inter-object occlusion image (O) of an object;
generating both a right eye occlusion image (OR) and a left eye occlusion image (OL) based on respective right eye and left eye views of said object; and
determining a portion of an image to be rendered into 3-D by replacing said occlusion image (O) with an image formed by an intersection of O, OR and OL.
25 . The method of claim 24 wherein a depth image (S) of said object is pre-computed.
26 . The method of claim 25 further comprising:
calculating said O directly from a single pass render of said object in isolation,
27 . The method of claim 24 wherein said forming comprises:
applying a tolerance value.
28 . The method of claim 27 wherein said tolerance value is applied manually.
29 . The method of claim 27 wherein said forming complies with the equation:
O[p]:=if M[p] then S[p]˜?=T[p] else false
where M is an object mask, S is a depth image of an entire scene, and T is a second depth image of said object in isolation.
30 . The method of claim 29 wherein said tolerance value is applied as a condition to said equation, said condition being:
abs ( S[p]−T[p] )>tolerance.