Unsupervised asymmetry detection
AssigneeInternational Business Machines Corporation
InventorsSharon Alpert, Miri Erihov, Pavel Kisilev
Life of this patent
Estimating the term…
Abstract
Asymmetries are detected in one or more images by partitioning each image to create a set of patches. Salient patches are identified, and an independent displacement for each patch is identified. The techniques used to identify the salient patches and the displacement for each patch are combined in a function to generate a score for each patch. The scores can be used to identify possible asymmetries.
Claims18
A method for identifying asymmetries between a plurality of images, the method comprising:
- identifying a first image and a second image, wherein the first image and the second image comprise two discrete images or two portions of a same image wherein the two portions are separated according to an approximate axis of symmetry;
- partitioning each image to create a set of patches, wherein each patch comprises a plurality of pixels;
- identifying salient patches of each image, the salient patches identified according to a relative degree of difference with respect to an average patch;
- identifying a respective patch displacement between the first image and the second image for at least the salient patches such that each displacement is independent;
- presenting a score for at least each salient patch, wherein the score comprises a respective saliency and the respective patch displacement for at least each salient patch.
A system for identifying asymmetries between a plurality of images, the system comprising:
- a memory storing processor-executable instructions;
- and
- a processor communicatively coupled to the memory and configured to execute the processor-executable instructions to perform a method comprising:
- identifying a first image and a second image, wherein the first image and the second image comprise two discrete images or two portions of a same image wherein the two portions are separated according to an approximate axis of symmetry;
- partitioning each image to create a set of patches, wherein each patch comprises a plurality of pixels;
- identifying salient patches of each image, the salient patches identified according to a relative degree of difference with respect to an average patch;
- identifying a respective patch displacement between the first image and the second image for at least the salient patches such that each displacement is independent;
- presenting a score for at least each salient patch, wherein the score comprises a respective saliency and the respective patch displacement for at least each salient patch.
A computer program product for identifying asymmetries between a plurality of images, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions executable by a processor to cause the processor to:
- identify a first image and a second image, wherein the first image and the second image comprise two discrete images or two portions of a same image wherein the two portions are separated according to an approximate axis of symmetry;
- partition each image to create a set of patches, wherein each patch comprises a plurality of pixels;
- identify salient patches of each image, the salient patches identified according to a relative degree of difference with respect to an average patch;
- identify a respective patch displacement between the first image and the second image for at least the salient patches such that each displacement is independent;
- present a score for at least each salient patch, wherein the score comprises a respective saliency and the respective patch displacement for at least each salient patch.
Dependencies are inferred from the claim text.