Digital Image Boundary Detection
AssigneeAdobe Inc.
InventorsBrian Lynn Price, Peng Zhou, Scott David Cohen, Gregg Darryl Wilensky
Life of this patent
Estimating the term…
Abstract
In implementations of object boundary generation, a computing device implements a boundary system to receive a mask defining a contour of an object depicted in a digital image, the mask having a lower resolution than the digital image. The boundary system maps a curve to the contour of the object and extracts strips of pixels from the digital image which are normal to points of the curve. A sample of the digital image is generated using the extracted strips of pixels which is input to a machine learning model. The machine learning model outputs a representation of a boundary of the object by processing the sample of the digital image.
Claims20
In a digital medium environment, a method implemented by a computing device, the method comprising:
- receiving, by a sampling module, image data describing a digital image depicting an object;
- generating, by the sampling module, a sample image that includes strips of pixels extracted from the digital image based on a contour of the object;
- detecting, by a machine learning module, a boundary of the object by processing the sample image using a machine learning model trained to detect boundaries of objects using training data describing training digital images and ground truth boundaries of objects depicted in the training digital images;
- and
- outputting, by a mapping module, a representation of the boundary of the object for display in a user interface.
One or more non-transitory computer-readable storage media comprising instructions stored thereon that, responsive to execution by a computing device, causes the computing device to perform operations including:
- determining a contour of an object depicted in a digital image by applying a mask to a downsampled version of the digital image;
- mapping a curve to the contour of the object, the curve having a resolution corresponding to a resolution of the digital image;
- generating a sample image based on the curve and pixels of the digital image;
- detecting a boundary of the object by processing the sample image using a machine learning model trained to detect boundaries of objects using training data describing training digital images and ground truth boundaries of objects depicted in the training digital images;
- and
- outputting a representation of the boundary of the object for display in a user interface.
A system comprising:
- means for determining a contour of an object depicted in a digital image;
- means for generating a sample image that includes strips of pixels extracted from the digital image based on a contour of the object;
- means for detecting a boundary of the object by processing the sample image using a machine learning model trained to detect boundaries of objects using training data describing training digital images and ground truth boundaries of objects depicted in the training digital images;
- and
- means for outputting a representation of the boundary of the object for display in a user interface.
Dependencies are inferred from the claim text.