Back to results
US 2018/0018776 A1Published application
Content aware visual image pattern matching
AssigneeSamsung Electronics Co., Ltd.
InventorsROEE SFARADI, Omry Sendik
Life of this patent
Estimating the term…
Abstract
Exemplary embodiments of the invention as described herein generally provide for detecting the displacement of feature(s) within a visual image in cases where pattern matching fails due to the existence of aperture(s) caused for example by external condition(s) encountered in recording such an image over time. Technique(s) are disclosed for detecting the difference between displacement of a geometric feature of an object appearing within an image (e.g., an edge or smooth surface) that has an aperture and another feature (e.g., a corner) that does not since it is not symmetrically invariant.
Claims21
A method for detecting time-elapsed movement of a geometric feature of an object found within a visual image comprising:
- selecting a reference region defining a luminance value for at least a portion of an object found within a source image;
- selecting at least one other region within a target image defining another luminance value for at least a portion of the object found within the source image;
- defining a motion estimation block comprised of a plurality of image location(s) for use of a similarity measure in evaluating whether a luminance value for a selected region within the target image substantially corresponds to a luminance value for the reference region within the source image;
- forming a matrix containing an array of value(s) from the motion estimation block that define a relative distance in each geometric dimension between the source image and any one or more selected region(s) within the target image having a luminance value that substantially corresponds to a luminance value for the reference region within the source image;
- combining motion estimation block value(s) within the matrix to define at least one motion estimation vector with an absolute distance index value in each geometric dimension that is used to project movement of the object from the reference region in the source image to a corresponding region in the target image;
- and
- performing a contrast calculation using one or more projected motion estimation vector distance index value(s) to detect whether time-elapsed movement of the object has occurred in at least one geometric dimension;
- wherein an image region is comprised of an arrayed section of location(s) each having a luminance value in at least one geometric dimension.
5 more elements
A visual image formed by carrying out a procedure comprised of the following steps for detecting time-elapsed movement of a geometric feature of an object found within a visual image:
- selecting a reference region defining a luminance value for at least a portion of an object found within a source image;
- selecting at least one other region within a target image defining another luminance value for at least a portion of the object found within the source image;
- defining a motion estimation block comprised of a plurality of image location(s) for evaluating whether a luminance value for a selected region within the target image substantially corresponds to a luminance value for the reference region within the source image;
- performing a symmetry detection algorithm using at least one similarity measure to form a matrix containing an array of value(s) from the motion estimation block that define the relative distance in each geometric dimension between the source image and any one or more selected region(s) within the target image having a luminance value that substantially corresponds to a luminance value for the reference region within the source image;
- combining motion estimation block value(s) within the matrix to define a distance map containing at least one motion estimation vector with a minimum absolute distance index value in each geometric dimension that is used to project movement of the object from the reference region in the source image to a corresponding region in the target image;
- and
- performing a contrast evaluation by calculating a maximum difference between projected motion estimation vector distance index value(s) to detect whether time-elapsed movement of the object has occurred in at least one geometric dimension;
- wherein an image region is comprised of an arrayed section of location(s) each having a luminance value in at least one geometric dimension.
5 more elements
A computer device configured to execute software instructions programmed to process data for carrying out a procedure comprised of the following steps for detecting time-elapsed movement of a geometric feature of an object found within a visual image:
- selecting a reference region defining a luminance value for at least a portion of an object found within a source image;
- selecting at least one other region within a target image defining another luminance value for at least a portion of the object found within the source image;
- defining a motion estimation block comprised of a plurality of image location(s) for use of a similarity measure in evaluating whether a luminance value for a selected region within the target image substantially corresponds to a luminance value for the reference region within the source image;
- forming a matrix containing an array of value(s) from the motion estimation block that define the relative distance in each geometric dimension between the source image and any one or more selected region(s) within the target image having a luminance value that substantially corresponds to a luminance value for the reference region within the source image;
- combining motion estimation block value(s) within the matrix to define at least one motion estimation vector with an absolute distance index value in each geometric dimension that is used to project movement of the object from the reference region in the source image to a corresponding region in the target image;
- performing a contrast calculation using one or more projected motion estimation vector distance index value(s) to detect whether time-elapsed movement of the object has occurred in at least one geometric dimension;
- and
- performing a calculation for detecting the presence of noise affecting any time-elapsed movement of an object within the visual image;
- wherein an image region is comprised of an arrayed section of location(s) each having a luminance value defining an aperture in at least one geometric dimension.
6 more elements
-27. (canceled)
Dependencies are inferred from the claim text.