Abstract: Methods and apparatus to measure brand exposure in media streams are disclosed. An example apparatus disclosed herein includes a brand identifier detector to compare first data associated with a first scene of a media stream with second data associated with a reference scene including a first brand identifier to detect the first brand identifier in the first scene of the media stream. The example apparatus also includes a measure and tracking module to combine respective locations of the first brand identifier in respective frames of a first sequence of image frames forming the first scene to determine a weighted location for the first brand identifier. The example apparatus further includes a report generator to report appearance data corresponding to the first brand identifier, the appearance data including the weighted location for the first brand identifier.

Abstract: Example methods, apparatus, systems to perform context-based image recognition are disclosed. Disclosed example methods include comparing a first and second reference region of a reference image with a first and second region of a first captured image to determine a first and second match value. In response to determining the first and second match values satisfy a first and second threshold: associating the first captured image with a product associated with the first reference image, the product being associated with a plurality of second reference images stored in the database. Decrementing index values associated with ones of the second reference images. Removing a reference record associated with one of the second reference images from the database based on a corresponding one of the second index values satisfying a first condition, the removal of the reference record reducing the storage space required to store reference records in the database.

Abstract: Methods and apparatus to extract text from imaged documents are disclosed. Example methods include segmenting an image of a document into localized sub-images corresponding to individual characters in the document. The example methods further include grouping respective ones of the sub-images into a cluster based on a visual correlation of the respective ones of the sub-images to a reference sub-image. The visual correlation between the reference sub-image and the respective ones of the sub-images grouped into the cluster exceeding a correlation threshold. The example methods also include identifying a designated character for the cluster based on the sub-images grouped into the cluster. The example methods further include associating the designated character with locations in the image of the document associated with the respective ones of the sub-images grouped into the cluster.

Abstract: Hybrid active and passive people metering for audience measurement is disclosed. Example audience identification methods disclosed herein include detecting a trigger signal output from a media device meter. Disclosed example methods also include, in response to determining that an image captured by an image sensor does not correspond to any one of a plurality of recognized audiences, passing the trigger signal to an input of a people meter to cause the people meter to output a prompt for entry of audience identification information. Disclosed example methods further include, in response to determining that the image corresponds to a first one of the plurality of recognized audiences, blocking the trigger signal from being passed to the input of the people meter to prevent the people meter from outputting the prompt.

Abstract: Example methods, apparatus, systems to perform context-based image recognition are disclosed. Disclosed example methods include comparing a first and second reference region of a reference image with a first and second region of a first captured image to determine a first and second match value. In response to determining the first and second match values satisfy a first and second threshold: associating the first captured image with a product associated with the first reference image, the product being associated with a plurality of second reference images stored in the database. Decrementing index values associated with ones of the second reference images. Removing a reference record associated with one of the second reference images from the database based on a corresponding one of the second index values satisfying a first condition, the removal of the reference record reducing the storage space required to store reference records in the database.

Abstract: Methods and apparatus to extract text from imaged documents are disclosed. Example methods include segmenting an image of a document into localized sub-images corresponding to individual characters in the document. The example methods further include grouping respective ones of the sub-images into a cluster based on a visual correlation of the respective ones of the sub-images to a reference sub-image. The visual correlation between the reference sub-image and the respective ones of the sub-images grouped into the cluster exceeding a correlation threshold. The example methods also include identifying a designated character for the cluster based on the sub-images grouped into the cluster. The example methods further include associating the designated character with locations in the image of the document associated with the respective ones of the sub-images grouped into the cluster.

Abstract: Example methods, apparatus, systems to perform context-based image recognition for consumer market research are disclosed. Disclosed example methods include comparing a first subset of feature points in a first region of interest of a reference image with a first subset of feature points of a first image residing within a first region corresponding spatially to the first region of interest and located at a first location in the first image to determine a first match value. Disclosed example methods also include, in response to determining the first match value satisfies a first threshold, comparing a second subset of feature points in a second region of interest of the reference image with a second subset of the feature points of the first image residing within a second region of the first image corresponding spatially to the second region of interest to determine a second match value.

Abstract: Methods and apparatus to detect differences between images are disclosed. Example methods disclosed herein include determining whether a difference between a block of pixels of a first image represented by a first signature and a block of pixels of a reference image represented by a second signature matching the first signature is resolvable using a transformation operation. Example methods disclosed herein also include, in response to determining the difference is unresolvable using the transformation operation, including the block of pixels of the first image in a difference region of the first image. Example methods disclosed herein further include generating a third signature different from the difference region of the first image to represent the difference region of the first image, and storing the third signature with version information associated with first image.

Abstract: Methods and apparatus to measure brand exposure in media streams are disclosed. An example apparatus disclosed herein includes a brand identifier detector to compare first data associated with a first scene of a media stream with second data associated with a reference scene including a first brand identifier to detect the first brand identifier in the first scene of the media stream. The example apparatus also includes a measure and tracking module to combine respective locations of the first brand identifier in respective frames of a first sequence of image frames forming the first scene to determine a weighted location for the first brand identifier. The example apparatus further includes a report generator to report appearance data corresponding to the first brand identifier, the appearance data including the weighted location for the first brand identifier.

Abstract: Example methods, apparatus, systems to perform context-based image recognition for consumer market research are disclosed. Disclosed example methods include comparing a first subset of feature points in a first region of interest of a reference image with a first subset of feature points of a first image residing within a first region corresponding spatially to the first region of interest and located at a first location in the first image to determine a first match value. Disclosed example methods also include, in response to determining the first match value satisfies a first threshold, comparing a second subset of feature points in a second region of interest of the reference image with a second subset of the feature points of the first image residing within a second region of the first image corresponding spatially to the second region of interest to determine a second match value.

Abstract: Methods and apparatus to specify regions of interest in video frames are disclosed. Example disclosed methods to mark a region in a graphical presentation include selecting a first point located at a substantially central position within the region, selecting a plurality of second points to define a boundary of the region, and comparing a plurality of stored templates with the selected first and second points to identify a first one of the stored templates to represent the region.

Abstract: Disclosed example methods to determine a time offset between a first media signature having a first sequence of first data pairs and a second media signature having a second sequence of second data pairs include comparing ones of the first data pairs in the first sequence with ones of the second data pairs in the second sequence to determine original time differences between respective combinations of the ones of the first data pairs and the ones of the second data pairs. Such example methods also include scaling and truncating the original time differences to determine scaled time differences between respective combinations of the ones of the first data pairs and the ones of the second data pairs. Such example methods further include determining a histogram comprising bins representative of the scaled time differences, and selecting a histogram bin to correspond to the time offset between the first and second signatures.

Abstract: Hybrid active and passive people metering for audience measurement is disclosed. Example audience identification methods disclosed herein include detecting a trigger signal output from a media device meter. Disclosed example methods also include, in response to determining that an image captured by an image sensor does not correspond to any one of a plurality of recognized audiences, passing the trigger signal to an input of a people meter to cause the people meter to output a prompt for entry of audience identification information. Disclosed example methods further include, in response to determining that the image corresponds to a first one of the plurality of recognized audiences, blocking the trigger signal from being passed to the input of the people meter to prevent the people meter from outputting the prompt.

Abstract: Methods and apparatus to measure brand exposure in media streams are disclosed. An example method to determine brand exposures included in media content disclosed herein comprises determining whether a scene detected from a media stream corresponding to the media content matches a reference scene, identifying an expected region of interest in the detected scene based on information describing the reference scene when the detected scene is determined to match the reference scene and the reference scene is not specified to be a scene of no interest, and determining whether a reference brand identifier associated with the reference scene is included in the expected region of interest identified in the detected scene.

Abstract: Image recognition methods, apparatus and articles or manufacture to support shelf auditing for consumer research are disclosed herein. Example methods disclosed herein include comparing a first image signature associated with an input image with a plurality of reference signatures associated with a plurality of reference images to identify a first reference image matching the input image. Such disclosed example methods also include identifying a first group of items depicted in the input image as corresponding to a first group of reference items registered with the first reference image. Such disclosed example methods further include determining a first region of the input image that differs from a corresponding first region of the first reference image, and processing the first region of the input image based on a template to identify a second item depicted in the input image.

Abstract: Scene-based people metering for audience measurement is disclosed. Example methods disclosed herein include grouping successive image frames depicting a location of a media presentation to form a sequence of scenes, respective scenes including respective groups of the image frames. Such example methods also include grouping matching scenes into respective scene clusters having respective sizes, the scene clusters being represented by respective key frames. Such example methods further include assigning respective ranks to the key frames of the respective scene clusters, the respective ranks being determined based on the respective sizes of the scene clusters. Such example methods additionally include processing the key frames in accordance with the respective ranks to monitor an audience of the media presentation.

Abstract: Hybrid active and passive people metering for audience measurement is disclosed. An example method disclosed herein to perform people metering for audience measurement comprises obtaining an image sequence depicting a scene in which an audience is expected to be present, comparing an image sequence signature representative of the image sequence with a set of reference signatures to determine a comparison result, and controlling audience prompting performed by a people meter based on the comparison result.

Abstract: Methods and apparatus to detect differences between images are disclosed. Example methods disclosed herein include determining whether a difference between a block of pixels of a first image represented by a first signature and a block of pixels of a reference image represented by a second signature matching the first signature is resolvable using a transformation operation. Example methods disclosed herein also include, in response to determining the difference is unresolvable using the transformation operation, including the block of pixels of the first image in a difference region of the first image. Example methods disclosed herein further include generating a third signature different from the difference region of the first image to represent the difference region of the first image, and storing the third signature with version information associated with first image.

Abstract: Methods and apparatus to detect differences between images are disclosed. Example methods disclosed herein to recognize different image versions include generating a first signature representative of a sample image, and obtaining a reference image associated with a second signature determined to substantially match the first image signature. Such example methods also include generating a third signature representative of a first region of the sample image corresponding spatially to a first region of the reference image associated with a first version of the reference image, the third signature being different from the first signature. Such example methods further include comparing the generated third signature and a fourth signature representative of the first region of the reference image to determine whether the sample image corresponds to the first version of the reference image.

Abstract: Methods and apparatus to methods and apparatus to identify images in print advertisements are disclosed. An example method comprises computing a first image feature vector for a first presented image, comparing the first image feature vector to a second image feature vector, and when the first image feature vector matches the second image feature vector, storing printed-media information associated with the first presented image in a database record associated with the second image feature vector.