Hewlett-packard Development Company, L.p.

Adobe, Inc.

Feng Tang

Suk Hwan Lim

Daniel R Tretter

Qian Lin

Nelson Liang An Chang

Jonathan Worthen Brandt

Hui Chao

Peng Wu

Nicholas P Lyons

Hao Tang

Yuli Gao

A computer device to generate an image with content and create active regions at locations or coordinates of the image including the content. The computing device associates the active regions with accessible links corresponding to the content.

Active regions of an image with accessible links

According to an example, a method for object segmentation may include receiving a digital image, performing initial segmentation on the digital image to generate a segmented digital image, and receiving refinement instructions to refine the initial segmentation. The method may further include inferring an intention of a user to correct a foreground area or a background area of the initial segmentation based on the received refinement instructions, learning a behavior of the user to further infer the intention of the user to correct the foreground area or the background area, and refining, by a processor, the initial segmentation based on the inferred intention.

Object segmentation

A reference set of image features is determined from an electronic data file specifying a reference image in a reference coordinate space. Rendering information describing a physical rendering of the reference image is ascertained. Calibration-enabling data is derived from the reference set of the image features and the ascertained rendering information. The calibration-enabling data is provided to calibrate an imaging system. The calibration-enabling data may be stored. The imaging system may capture an image of the physical rendering of the reference image in relation to a capture coordinate space. An extracted set of image features may be extracted from the captured image. Respective ones of the image features in the reference and extracted sets may be matched. The imaging system may be calibrated based on matched ones of the image features and the rendering information.

Apparatus and methods for imaging system calibration

A method for creating and using metadata supersets for matching images includes identify images from a first source of images and identify images from a second source of images. Images from the first source that match images from the second source are identified as matching images. Metadata associated with the matching images are extracted to form a superset of metadata for the matching images. The superset of metadata is analyzed to create a product.

Metadata supersets for matching images

A method for merging face clusters includes analyzing a set of digital images, grouping instances of faces within the set of digital images into a set of face clusters, each of the face clusters corresponding to a particular person, and determining a probability that a person associated with a first face cluster from the set of face clusters is the same person associated with a second face cluster of the set of face clusters. The probability is based on both a social similarity between the first face cluster and the second face cluster in addition to a facial similarity between the first face cluster and the second face cluster.

Merging face clusters

Disclosed herein are a system and a method that use a background model to determine and to segment target content from an image and replace them with different content to provide a composite image. The background model can be generated based on image data representing images of a predetermined area that does not include traversing content. The background model is compared to image data representing a set of captured images of the predetermined area. Based on the comparison, portions of an image that differs from the background model are determined as the traversing content. A target content model is used to determine the target content in the traversing content. The target content determined in the images is replaced with different content to provide a composite image.

Replacement of a person or object in an image

A method and system characterizes an image of an object. A plurality of interest points are detected within a first image and a local image feature descriptor is built for at least some of the interest points, including mapping information about the interest points according to at least circular distribution information.

Local image feature descriptors according to circular distribution information

A system and a method for near-duplicate image detection performed by a physical computing system includes applying a feature determining function to a number of images, a feature being defined by a geometric shape, comparing characteristics of said geometric shapes defining said features from at least two of said number of images, and characterizing said at least two of said number of images as a near-duplicate match if a predetermined percentage of said features of said at least two images match.

Near-duplicate image detection

Methods, devices, and systems for determining feature descriptors are provided. An example includes defining a plurality of anchor points within a patch of pixels in a particular area that includes a detected feature in an image, defining a first set of subpatches and calculating an intensity of each of the first set of subpatches, defining a second set of subpatches and calculating an intensity of each of the second set of subpatches, comparing the intensity of each of the second set of subpatches to the intensity of each of the first set of subpatches and if the intensity of a second set subpatch is higher than the intensity of a first set subpatch assign a binary value, otherwise assign an alternative binary value, and concatenating all the assigned binary values into a binary feature descriptor.

Feature descriptors

A method and system for analyzing an image made up of a plurality of pixels includes identifying a subset of the pixels, and for each identified pixel a local patch of pixels surrounding the identified pixel is defined and the pixels in the local patch are grouped into bins. The local patch is divided into a plurality of sub-patches, and for each sub-patch the method includes determining how many pixels in each sub-patch fall into each of the bins to create an intermediate vector, and the intermediate vectors are concatenated to form a feature vector and describe the local patch.

Image analysis based on pixel brightness grouping

A local image patch identified in an image is divided into respective sub-patches of respective image forming elements. For each of the respective image forming elements in the local image patch, a respective ordinal rank of the image forming element is determined, and respective contributions of the image forming element to a particular one of the respective sub-patches containing the image forming element and to one or more other ones of the respective sub-patches neighboring the particular sub-patch are ascertained. Each ordinal rank corresponds to a respective dimension of an ordinal rank feature space. For each of the respective sub-patches of the local image patch, a respective histogram of ascertained contributions of the image forming elements in the ordinal rank feature space is built. A respective feature vector representing the local image patch is generated from the respective histograms built for the sub-patches of the local image.

Ordinal and spatial local feature vector based image representation

A method, system, and computer-readable storage medium for characterizing and representing images. A plurality of feature descriptors for a plurality of images are received, where each feature descriptor encodes a respective feature in an image, and where each feature descriptor is transformed into a plurality of sub-descriptors in accordance with a specified transformation scheme. The feature descriptors correspond to an image feature descriptor space for the plurality of images, and the sub-descriptors correspond to a plurality of subspaces of the image feature descriptor space, where the plurality of subspaces span the image feature descriptor space. Each subspace of the plurality of subspaces is quantized using cluster analysis applied to the sub-descriptors, thereby determining a respective one or more feature primitives for each subspace, where the feature primitives are useable to characterize image features.

Growing community of inventors

Mountain View, CA, United States of America

Feng Tang