US 9,286,540 B2 · Fast dense patch search and quantization

Abstract

In techniques for fast dense patch search and quantization, partition center patches are determined for partitions of example image patches. Patch groups of an image each include similar image patches and a reference image patch that represents a respective patch group. A partition center patch of the partitions is determined as a nearest neighbor to the reference image patch of a patch group. The partition center patch can be determined based on a single-nearest neighbor (1-NN) distance determination, and the determined partition center patch is allocated as the nearest neighbor to the similar image patches in the patch group. Alternatively, a group of nearby partition center patches are determined as the nearest neighbors to the reference image patch based on a k-nearest neighbor (k-NN) distance determination, and the nearest neighbor to each of the similar image patches in the patch group is determined from the nearby partition center patches.

Claims20

Dependent claims

A fast dense patch search method, comprising:

obtaining by a computing device and from storage example image patches that are taken from example images;
determining, from partitions of the example image patches and by the computing device, partition center patches for the partitions of example image patches, the example image patches in a partition being similar, one example image patch to another, a partition center patch determined to represent the example image patches of the partition;
obtaining by the computing device an input digital image;
forming patch groups by the computing device that each include similar image patches of the input digital image and a reference image patch that represents the similar image patches in a respective patch group;
determining which of the partition center patches of the partitions of the example image patches by the computing device is a nearest neighbor to the reference image patch of a patch group;
allocating the determined partition center patch by the computing device as the nearest neighbor to the similar image patches of the input digital image in the patch group;
and
outputting the allocated patch.

5 more elements

A computing device for performing a fast dense patch search, the computing device comprising:

one or more processors;
and
a memory comprising computer-readable instructions that are executable by the one or more processors to implement a fast dense patch search application to perform operations comprising:
causing an input digital image to be maintained in the memory, the input digital image including patch groups of similar image patches and a reference image patch that represents the similar image patches of the input digital image in a respective patch group;
obtaining from a database example image patches that are taken from example images;
determining, from partitions of the example image patches, partition center patches for the partitions of example image patches, the example image patches in a partition being similar, one example image patch to another, a partition center patch determined to represent the example image patches of the partition;
determining which of a partition center patches of the partitions of the example image patches is a nearest neighbor to the reference image patch of a patch group, the partition center patch included in a group of nearby partition center patches;
ascertaining, from the nearby partition center patches, the nearest neighbor to each of the similar image patches of the input digital image in the patch group;
and
outputting the ascertained patches.

7 more elements

A fast dense patch search method, comprising:

obtaining by a computing device and from storage example image patches that are taken from example images;
determining, from partitions of the example image patches and by the computing device, partition center patches for the partitions of example image patches, the example image patches in a partition being similar one example image patch to another, a partition center patch determined to represent the example image patches of the partition;
obtaining by the computing device an input digital image;
forming patch groups by the computing device that each include similar image patches of the input digital image and a reference image patch that represents the similar image patches in a respective patch group;
determining which of the partition center patches of the partitions of the example image patches by the computing device is a nearest neighbor to the reference image patch of a patch group, the partition center patch included in a group of nearby partition center patches;
ascertaining, from the nearby partition center patches and by the computing device, the nearest neighbor to each of the similar image patches of the input digital image in the patch group;
and
outputting the ascertained patches.

5 more elements

Dependencies are inferred from the claim text.

Fast dense patch search and quantization

Life of this patent

Fast dense patch search and quantization

Life of this patent

Abstract

Claims20