System and method for creating 3D models from 2D sequential image data
AssigneePush Entertainment, Inc.
InventorsTodd Simpson, Edward C. Vanwieren, David Alan Spooner, Leonard Thomas Bruton
Life of this patent
Estimating the term…
Abstract
A system and method providing for the generation of an image having enhanced three-dimensional attributes from a source image having only two-dimensional spatial domain information is disclosed. Object image discrimination through the use of machine vision techniques is utilized to identify objects within the source image. Thereafter, object images are dissected into image cells wherein the object images are supplemented to include missing information. Generation techniques are utilized to reassemble the image cells into an image having the desired enhanced three-dimensional attributes.
Claims20
A system for digitally processing a sequence of source images, said system comprising:
- means for differentiating between objects represented in said source images to produce differentiated objects;
- means for parsing said source image sequences according to said differentiated objects into image cells, wherein each said image cell substantially consists of information associated with a differentiated object;
- means for identifying selected differentiated objects that are associated with scaling transformations in a temporal domain;
- means for determining relative planes of said selected differentiated objects;
- means for producing stereoscopically enhanced image cells, wherein said means for producing stereoscopically enhanced image cells utilizes at least said relative planes to produce said stereoscopically enhanced image cells.
A method for digitally processing a sequence of source images, said method comprising:
- differentiating between objects represented in said source images to produce differentiated objects;
- parsing said source image sequences according to said differentiated objects into image cells, wherein each said image cell substantially consists of information associated with a differentiated object;
- identifying selected differentiated objects that are associated with scaling transformations in a temporal domain;
- determining relative planes of said selected differentiated objects;
- and
- producing stereoscopically enhanced image cells from said image cells utilizing at least said relative planes.
A computer readable medium containing processor executable instructions for digitally processing a sequence of source images, said computer readable medium comprising:
- code for differentiating between objects represented in said source images to produce differentiated objects;
- code for parsing said source image sequences according to said differentiated objects into image cells, wherein each said image cell substantially consists of information associated with a differentiated object;
- code for identifying selected differentiated objects that are associated with scaling transformations in a temporal domain;
- code for determining relative planes of said selected differentiated objects;
- and
- code for producing stereoscopically enhanced image cells, wherein said code for producing stereoscopically enhanced image cells utilizes at least said relative planes to produce said stereoscopically enhanced image cells.
Dependencies are inferred from the claim text.