Method and system for image matching in a mixed media environment
First Claim
Patent Images
1. A computer-implemented method of image matching, the method comprising:
- receiving a first image of at least part of a first media type;
distinguishing between a first text patch and an image patch in the first image;
determining a quality of the first text patch;
determining that the quality of the first text patch is unreliably processable for extracting the plurality of features from the first text patch;
generating instructions for modifying a focus of an optics device based on the quality of the first text patch;
receiving a second image of at least part of the first media type;
determining the quality of a second text patch in the second image;
validating the quality of the second text patch as being reliably processable for extracting a plurality of features from the second text patch;
extracting from the second text patch the plurality of features including a vertical grouping of a first word boundary with one or more second word boundaries present above or below overlapping the first word boundary in the second text patch;
extracting from the second text patch a geometric relationship between the plurality of features, the geometric relationship including one or more angles between the first word boundary and the one or more second word boundaries in the plurality of features;
generating a query for identifying a document corresponding to the second text patch using the plurality of features and the geometric relationship;
applying the query against a collection of documents in a database; and
in response to positively applying the query, returning an identification comprising a matching document and a location of the second text patch within the portion of the matching document.
2 Assignments
0 Petitions
Accused Products
Abstract
A Mixed Media Reality (MMR) system and associated techniques are disclosed. The MMR system provides mechanisms for forming a mixed media document that includes media of at least two types (e.g., printed paper as a first medium and digital content and/or web link as a second medium). In one particular embodiment, the MMR system provides for image matching portions of a document.
437 Citations
35 Claims
-
1. A computer-implemented method of image matching, the method comprising:
-
receiving a first image of at least part of a first media type; distinguishing between a first text patch and an image patch in the first image; determining a quality of the first text patch; determining that the quality of the first text patch is unreliably processable for extracting the plurality of features from the first text patch; generating instructions for modifying a focus of an optics device based on the quality of the first text patch; receiving a second image of at least part of the first media type; determining the quality of a second text patch in the second image; validating the quality of the second text patch as being reliably processable for extracting a plurality of features from the second text patch; extracting from the second text patch the plurality of features including a vertical grouping of a first word boundary with one or more second word boundaries present above or below overlapping the first word boundary in the second text patch; extracting from the second text patch a geometric relationship between the plurality of features, the geometric relationship including one or more angles between the first word boundary and the one or more second word boundaries in the plurality of features; generating a query for identifying a document corresponding to the second text patch using the plurality of features and the geometric relationship; applying the query against a collection of documents in a database; and in response to positively applying the query, returning an identification comprising a matching document and a location of the second text patch within the portion of the matching document. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. A system for image matching, comprising:
-
one or more processors; a memory; a quality assessment module stored on the memory and executable by the one or more processors, the quality assessment module operable to receive a first image of at least part of a first media type, to distinguish between a first text patch and an image patch in the first image, to determine a quality of the first text patch, to determine that the quality of the first text patch is unreliably processable for extracting a plurality of features from the first text patch, generate instructions for modifying a focus of an optics device based on the quality of the first text patch, receive a second image of at least part of the first media type, determine the quality of a second text patch in the second image and validate the quality of the second text patch as being reliably processable for extracting a plurality of features from the second text patch; a feature extraction module stored on the memory and executable by the one or more processors, the feature extraction module coupled to the quality assessment module and operable to extract from the second text patch the plurality of features including a vertical grouping of a first word boundary with one or more second word boundaries present above or below overlapping the first word boundary in the second text patch and a geometric relationship between the plurality of features, wherein the geometric relationship includes at least one or more angles between the first word boundary and the one or more second boundaries in the plurality of features; a classification module stored on the memory and executable by the one or more processors, the classification module coupled to the feature extraction module and operable to receive the plurality of features and the geometric relationship, to generate and send a query for identifying a document corresponding to the second text patch, and to receive an identification comprising a matching document and a location of the second text patch within the portion of the matching document; and a database operable to store a collection of documents, to receive the query from the classification module, to apply the query to the collection of documents, and to create and return the identification to the classification module. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24)
-
-
25. A computer-implemented method of associating a first media type with a second media type, the method comprising:
-
receiving a first image of at least part of the first media type; distinguishing between a first text patch and an image patch in the first image; determining a quality of the first text patch; determining that the quality of the text patch is unreliably processable for extracting the plurality of features from the first text patch; generating instructions for modifying a focus of an optics device associated with a capture device based on the quality of the first text patch; receiving a second image of at least part of the first media type; determining the quality of a second text patch in the second image; validating the quality of the second text patch as being reliably processable for extracting a plurality of features from the text patch; extracting from the second text patch the plurality of features including a vertical grouping of a first word boundary with one or more second word boundaries present above or below overlapping the first word boundary in the second text patch; extracting from the second text patch at least one geometric relationship between the plurality of features, the geometric relationship including one or more angles between the first word boundary and the one or more second word boundaries in the plurality of features; converting the plurality of features and the at least one geometric relationship into a symbolic representation in response to the validation; generating a query for identifying a document corresponding to the second text patch using the symbolic representation; applying the query against a collection of documents in a database; in response to positively matching the query, returning an identification comprising a matching document and a location of the second text patch within the portion of the matching document; and providing the identification to the capture device, the identification comprising the second media type. - View Dependent Claims (26, 27, 28, 29, 30)
-
-
31. A computer program product comprising a non-transitory computer readable medium including a computer readable program, wherein the computer readable program when executed on a computer causes the computer to:
-
receive a first image, the first image being of at least part of a first media type; distinguish between a first text patch and an image patch in the image; determine a quality of the first text patch; determine that the quality of the first text patch is unreliably processable for extracting the plurality of features from the first text patch; generate instructions for modifying a focus of an optics device based on the quality of the first text patch; receive a second image of at least part of the first media type; determine the quality of a second text patch in the second image; validate the quality of the second text patch as being reliably processable for extracting a plurality of features from the second text patch; extract from the second text patch the plurality of features including a vertical grouping of a first word boundary with one or more second word boundaries present above or below overlapping the first word boundary in the second text parch; extract from the second text patch at least one geometric relationship between the plurality of features, the at least one geometric relationship including one or more angles between the first word boundary and the one or more second word boundaries in the plurality of features; generate a query for identifying a document corresponding to the second text patch using the plurality of features and the geometric relationship; apply the query against a collection of documents stored in a database; and in response to positively applying the query, return an identification comprising a matching document and a location of the second text patch within the portion of the matching document. - View Dependent Claims (32, 33, 34, 35)
-
Specification