Back to results
US 12,754,965 B2Granted
Information processing apparatus, method of controlling information processing apparatus, and storage medium
AssigneeCanon Kabushiki Kaisha
InventorKeiichi Takashima
Life of this patent
Estimating the term…
Abstract
An information processing apparatus that is connectable to the Internet, includes: an obtaining unit configured to obtain a logo image corresponding to a logo representing a particular organization from a scanned image of a document; a search unit configured to search for a web page including an image similar to the obtained logo image through the Internet; and an inference unit configured to infer an organization name corresponding to the image included in the web page by performing named entity recognition using a named entity inference model for the web page searched by the search unit.
Claims8
An information processing apparatus connectable to the Internet, the information processing apparatus comprising:
- at least one processor or circuitry configured to:
- generate a named entity inference model from a first learning character string in which a named entity label is applied to each character string and dividing the first learning character string into words;
- generate an organization information inference model from a second learning character string in which a label is applied to an organization name and dividing the second learning character string into words;
- obtain a first logo image corresponding to a logo representing a particular organization, and information about the first logo image from a scanned image of a document;
- perform a first search to search for a web page including an image similar to the first logo image through the Internet;
- infer an organization name corresponding to the image included in the searched web page by performing named entity recognition using the organization information inference model that receives the image and the information about the image obtained from the first search as an input, and that outputs an organization name associated with the image and a label of the information about the image obtained from the first search;
- perform a second search to search for a second logo image associated with the inferred organization name through the Internet;
- determine whether the second logo image found from the second search matches the first logo image obtained from the scanned image;
- store the inferred organization name as a recognition result of the first logo image obtained from the scanned image in a case where the second logo image found from the second search is determined to match the first logo image obtained from the scanned image;
- setting the recognition result of the first logo image obtained from the scanned image as empty in a case where the second logo image found from the second search is determined to not match the first logo image obtained from the scanned image;
- infer a named entity label of the information about the first logo image by performing named entity recognition using the named entity inference model that receives only the information about the first logo image as an input, and that outputs the named entity label of the information about the first logo image;
- compare first content in the label of the information about the first logo image with second content in the label of the information about the image obtained from the first search;
- and
- determine the inferred organization name of the image in which the first content matches the second content, as an organization name of the first logo image.
11 more elements
A method of controlling an information processing apparatus connectable to the Internet, the method comprising:
- generating a named entity inference model from a first learning character string in which a named entity label is applied to each character string and dividing the first learning character string into words;
- generating an organization information inference model from a second learning character string in which a label is applied to an organization name and dividing the second learning character string into words;
- obtaining a first logo image corresponding to a logo representing a particular organization, and information about the first logo image from a scanned image of a document;
- performing a first search to search for a web page including an image similar to the first logo image through the Internet;
- inferring an organization name corresponding to the image included in the searched web page by performing named entity recognition using the organization information inference model that receives the image and the information about the image obtained from the first search as an input, and that outputs an organization name associated with the image and a label of the information about the image obtained from the first search;
- performing a second search to search for a second logo image associated with the inferred organization name through the Internet;
- determining whether the second logo image found from the second search matches the first logo image obtained from the scanned image;
- storing the inferred organization name as a recognition result of the first logo image obtained from the scanned image in a case where the second logo image found from the second search is determined to match the first logo image obtained from the scanned image;
- setting the recognition result of the first logo image obtained from the scanned image as empty in a case where the second logo image found from the second search is determined to not match the first logo image obtained from the scanned image;
- inferring a named entity label of the information about the first logo image by performing named entity recognition using the named entity inference model that receives only the information about the first logo image as an input, and that outputs the named entity label of the information about the first logo image;
- comparing first content in the label of the information about the first logo image with second content in the label of the information about the image obtained from the first search;
- and
- determining the inferred organization name of the image in which the first content matches the second content, as an organization name of the first logo image.
10 more elements
A non-transitory computer readable storage medium storing a program executable by a computer to perform a method of controlling an information processing apparatus connectable to the Internet, the method comprising:
- generating a named entity inference model from a first learning character string in which a named entity label is applied to each character string and dividing the first learning character string into words;
- generating an organization information inference model from a second learning character string in which a label is applied to an organization name and dividing the second learning character string into words;
- obtaining a first logo image corresponding to a logo representing a particular organization, and information about the first logo image from a scanned image of a document;
- performing a first search to search for a web page including an image similar to the first logo image through the Internet;
- inferring an organization name corresponding to the image included in the searched web page by performing named entity recognition using the organization information inference model that receives the logo image and information about the image obtained from the first search as an input, and that outputs an organization name associated with the input image and a label of the information about the image obtained from the first search;
- performing a second search to search for a second logo image associated with the inferred organization name through the Internet;
- determining whether the second logo image found from the second search matches the first logo image obtained from the scanned image;
- storing the inferred organization name as a recognition result of the first logo image obtained from the scanned image in a case where the second logo image found from the second search is determined to match the first logo image obtained from the scanned image;
- setting the recognition result of the first logo image obtained from the scanned image as empty in a case where the second logo image found from the second search is determined to not match the first logo image obtained from the scanned image;
- inferring a named entity label of the information about the first logo image by performing named entity recognition using the named entity inference model that receives only the information about the first logo image as an input, and that outputs the named entity label of the information about the first logo image;
- comparing first content in the label of the information about the first logo image with second content in the label of the information about the image obtained from the first search;
- and
- determining the inferred organization name of the image in which the first content matches the second content, as an organization name of the first logo image.
10 more elements
Dependencies are inferred from the claim text.