Systems and methods for converting two-dimensional images into three-dimensional images
AssigneeDD3D, Inc.
InventorsJonathan Karafin, Matthew DeJohn, Benoit Doidic
Life of this patent
Estimating the term…
Abstract
Systems and methods for converting a two-dimensional image or sequence of images into a three-dimensional image or sequence of images are presented. In embodiments, the disclosed techniques employ a suite of interactive image processing tools that allow a user to apply image pixel repositioning depth contouring effects and algorithms to efficiently create high-quality three-dimensional images.
Claims41
A method of converting a two-dimensional image into a three-dimensional image, the method comprising:
- ingesting, by a processing system, image data, said image data including at least one image, each image having one or more objects;
- segmenting, by the processing system, said one or more objects, said segmenting based on an isolation input entered by a user via said processing system;
- establishing stereoscopic depth for one or more of said segmented objects based on depth parameters entered by the user via said processing system;
- and
- compositing, by the processing system, said at least one image to remove any occluded regions subsequent to establishment of stereoscopic depth.
A user guided computer-based image processing method for converting two-dimensional images into three-dimensional images, comprising the steps of:
- ingesting one or more two-dimensional digital images by a processing system;
- creating, via said processing system, one or more multi-dimensional images from said one or more two-dimensional digital images using one or more stereoscopic techniques;
- presenting said one or more multi-dimensional images on a three-dimensional display associated with said processing system;
- identifying a first discrete portion of said one or more multi-dimensional images based on a perceived dimensional relationship between one or more pixels of said first discrete portion and one or more pixels of a second discrete portion;
- segmenting said first discrete portion of said one or more multi-dimensional images based on a perceived dimensional relationship between one or more pixels of said first discrete portion and one or more pixels of a second discrete portion;
- interpolating a plurality of occluded regions within said one or more multi-dimensional images;
- compositing one or more multi-dimensional images by combining said plurality of missing pixels and said one or more multi-dimensional images;
- depth mastering said one or more multi-dimensional images;
- and
- exporting said one or more multi-dimensional images.
A system for converting two-dimensional images into three-dimensional images, comprising:
- a memory including:
- digital data;
- and
- a first computer algorithm;
- a processor for executing a series of operations associated with the first computer algorithm;
- a display device adapted for playback of three-dimensional images;
- wherein,
- said digital data is comprised of one or more images;
- said first computer algorithm processes said plurality of digital data and creates one or more artifact-free alternate perspective images;
- said one or more artifact-free alternate perspective images are created using one or more of segmenting;
- compositing, or depth mastering;
- said display device renders, in real-time, one or more changes being made to said one or more three-dimensional images to enable a user to contemporaneously edit said digital data for improving quality of said alternate perspective images.
The system of claim 31, wherein said second computer algorithm uses one or more of object-level segmentation or pixel-level segmentation to identify and isolate said one or more discrete portions.
Dependencies are inferred from the claim text.