Patents by Inventor Hidetomo Sohma
Hidetomo Sohma has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11763582Abstract: Provided is an information processing apparatus that applies correction using a character recognition error pattern to a character recognition result of a document image, wherein the character recognition error pattern includes error pattern information on a character recognition result of a part where an error occurs in character recognition, correct pattern information applicable to the part where the error occurs, information on a frequency that the error occurs, and information on a state where the error occurs, and wherein the character recognition error pattern to be used in the correction is narrowed down based on the information on the frequency that the error occurs and the information on the state where the error occurs.Type: GrantFiled: August 31, 2022Date of Patent: September 19, 2023Assignee: CANON KABUSHIKI KAISHAInventor: Hidetomo Sohma
-
Publication number: 20230140357Abstract: An image processing apparatus acquires a character recognition result by performing character recognition processing on a document image, detects a character string candidate described in a predetermined format from the character recognition result, determines a likelihood of the character string candidate based on another character string existing in the vicinity of the detected character string candidate, and outputs, in a case where a plurality of character string candidates is detected, an item value based on a character string candidate having a high likelihood.Type: ApplicationFiled: October 27, 2022Publication date: May 4, 2023Inventor: Hidetomo Sohma
-
Publication number: 20230078322Abstract: Provided is an information processing apparatus that applies correction using a character recognition error pattern to a character recognition result of a document image, wherein the character recognition error pattern includes error pattern information on a character recognition result of a part where an error occurs in character recognition, correct pattern information applicable to the part where the error occurs, information on a frequency that the error occurs, and information on a state where the error occurs, and wherein the character recognition error pattern to be used in the correction is narrowed down based on the information on the frequency that the error occurs and the information on the state where the error occurs.Type: ApplicationFiled: August 31, 2022Publication date: March 16, 2023Inventor: Hidetomo Sohma
-
Patent number: 11412096Abstract: In an image forming apparatus that cooperates with a management service and a blockchain service, the management service receiving a document via a network and storing the document, and the blockchain service managing information on the document in block units, defining a relation between each of a plurality of blocks and blocks before and/or after the block, and managing the plurality of blocks using a plurality of nodes. In a case where an error occurs during printing of the document, new information to be registered in response to completion of the printing of the document is not registered in the blockchain service, the printing is stopped, and an error process is performed.Type: GrantFiled: January 12, 2021Date of Patent: August 9, 2022Assignee: Canon Kabushiki KaishaInventor: Hidetomo Sohma
-
Publication number: 20210243329Abstract: In an image forming apparatus that cooperates with a management service and a blockchain service, the management service receiving a document via a network and storing the document, and the blockchain service managing information on the document in block units, defining a relation between each of a plurality of blocks and blocks before and/or after the block, and managing the plurality of blocks using a plurality of nodes. In a case where an error occurs during printing of the document, new information to be registered in response to completion of the printing of the document is not registered in the blockchain service, the printing is stopped, and an error process is performed.Type: ApplicationFiled: January 12, 2021Publication date: August 5, 2021Inventor: Hidetomo Sohma
-
Patent number: 10628476Abstract: An information processing apparatus includes: an analysis unit configured to analyze a text; an obtaining unit configured to obtain term expressions from the text based on a result of the analysis; a classifying structuring unit configured to classify the term expressions based on a usage type of the term expressions; and a presentation unit configured to present a result of the classification based on a unified presentation sequence.Type: GrantFiled: February 10, 2016Date of Patent: April 21, 2020Assignee: Canon Kabushiki KaishaInventor: Hidetomo Sohma
-
Publication number: 20160239564Abstract: An information processing apparatus includes: an analysis unit configured to analyze a text; an obtaining unit configured to obtain term expressions from the text based on a result of the analysis; a classifying structuring unit configured to classify the term expressions based on a usage type of the term expressions; and a presentation unit configured to present a result of the classification based on a unified presentation sequence.Type: ApplicationFiled: February 10, 2016Publication date: August 18, 2016Inventor: Hidetomo Sohma
-
Patent number: 8971585Abstract: A face image is detected for each frame at a predetermined interval in moving image data, and the face image is traced using a frame in which the face image is detected and frames subsequent to the frame. A face sequence including an interval in which the face can be traced and motion velocity vectors of the face indicating a change in the position of the face image in the interval is generated based on the tracing result. Further, camera operation information about when the moving image data is acquired is generated from the frame image of the moving image data. When there is an overlap in the plurality of intervals in which the face images are traced, the face being tracked by the camera is determined using the face sequence and the camera operation information of each of the plurality of face images. The face determined to be tracked is then determined to be a key object.Type: GrantFiled: January 16, 2013Date of Patent: March 3, 2015Assignee: Canon Kabushiki KaishaInventors: Hiroshi Tojo, Hidetomo Sohma, Satoshi Yashiro
-
Patent number: 8954845Abstract: A region division portion extracts an “object”, an “anchor expression accompanying the object” and a “text including the anchor expression” from image data based on a paper document and an electronic document. A link processing portion generates link information that associates, in two ways, the “object”, the “anchor expression included in the text” or the “text including the anchor expression” with each other. Then, a format conversion portion converts the link information into electronic document data including two-way link information. When this electronic document data is displayed by an application and one of the “object” and the “anchor expression included in the text” is selected, the other can be displayed according to the link information.Type: GrantFiled: March 29, 2011Date of Patent: February 10, 2015Assignee: Canon Kabushiki KaishaInventors: Reiji Misawa, Ryo Kosaka, Tomotoshi Kanatsu, Hidetomo Sohma
-
Patent number: 8866900Abstract: An information processing apparatus for extracting a more appropriate representative frame image from moving image data that includes a plurality of frames of image data arranged in a time series includes: an input unit configured to input moving image data; a detecting unit configured to detect a frame image, which includes an image similar to a prescribed image pattern; a tracking unit configured to detect a frame image, which includes an image similar to the image included in the detected frame image; a storage unit configured to store the successive frame images in association with time information; a splitting unit configured to split the moving image data into a plurality of time intervals with starting time and end time of each of one or more image sequences; and an extracting unit configured to extract a representative frame image.Type: GrantFiled: November 5, 2009Date of Patent: October 21, 2014Assignee: Canon Kabushiki KaishaInventors: Hidetomo Sohma, Hiroshi Tojo, Satoshi Yashiro
-
Patent number: 8644621Abstract: A image processing apparatus extracts a plurality of feature points, comprising a local feature amount, from an inputted image, and a region information that relates to the feature point, identifies at least one of the plurality of feature points, associates the feature point thus identified with a feature point with a high relational degree between the feature point thus identified and a region that is denoted by the region information of the feature point thus identified, and registers the feature point thus identified and associated as an index of the inputted image.Type: GrantFiled: May 6, 2008Date of Patent: February 4, 2014Assignee: Canon Kabushiki KaishaInventors: Hidetomo Sohma, Tomomi Takata, Koichi Magai, Hirotaka Shiiyama
-
Patent number: 8614838Abstract: According to the present invention, it is possible to create electronic document data capable of highlighting an object detected through a search so that a user can easily recognize it. An image processing apparatus extracts an object from an input image and extracts metadata related to the object. The image processing apparatus, when determines to describe with a shape in accordance with the shape of the object, creates a vector path description of frame described with a shape in accordance with the shape of the object. Then, the image processing apparatus creates an electronic document including data of the input image and the vector path description of frame with which the metadata is associated. When a keyword search is performed on the created electronic document, highlight display is performed in accordance with the vector path description of frame with which metadata that matches the keyword is associated.Type: GrantFiled: December 20, 2010Date of Patent: December 24, 2013Assignee: Canon Kabushiki KaishaInventors: Ryo Kosaka, Reiji Misawa, Tomotoshi Kanatsu, Hidetomo Sohma
-
Patent number: 8548240Abstract: An apparatus comprises: unit configured to divide input document data into a body region, a caption region, and an object region; unit configured to acquire text information included in each of the body region and the caption region; unit configured to search the text information in the body region for an anchor term, to extract an anchor term from the text information in the caption region, and to generate a bi-directional link between a portion corresponding to the anchor term in the body region and a portion of the object region to which the caption region is appended; and unit configured to convert the input document data into digital document data in which the portion corresponding to the anchor term in the body region and the portion corresponding to the object region to which the caption region is appended are bi-directionally linked based on the link.Type: GrantFiled: June 16, 2011Date of Patent: October 1, 2013Assignee: Canon Kabushiki KaishaInventors: Tomotoshi Kanatsu, Hidetomo Sohma, Reiji Misawa, Ryo Kosaka
-
Patent number: 8514462Abstract: An image processing apparatus comprises: a character information acquisition unit configured to acquire character information included in each of a body region and a caption region; an accumulation unit configured to divide the character information acquired from the body region into predetermined set units and to accumulate the character information and position information of the divided set unit in a memory; an anchor term extraction unit configured to extract an anchor term from the character information acquired from the caption region; an anchor term search unit configured to search, based on the character information accumulated in the memory for each set unit, for the set unit including the anchor term extracted; a link information generation unit configured to generate link generation information that associates the set unit found by the anchor term search unit with the object region to which the caption region including the anchor term is appended.Type: GrantFiled: June 29, 2011Date of Patent: August 20, 2013Assignee: Canon Kabushiki KaishaInventors: Hidetomo Sohma, Tomotoshi Kanatsu, Reiji Misawa, Ryo Kosaka
-
Patent number: 8515177Abstract: An image processing apparatus executes smoothing processing (reduction conversion) of an input image to acquire a smoothed image (reduced image), acquires a normalization parameter for normalization from the smoothed image, and normalizes pixel values of the input image based on the normalization parameter.Type: GrantFiled: September 2, 2009Date of Patent: August 20, 2013Assignee: Canon Kabushiki KaishaInventors: Masahiro Matsushita, Hirotaka Shiiyama, Hidetomo Sohma, Koichi Magai
-
Patent number: 8472757Abstract: An image search apparatus, which determines a similarity between an input query image and a registered comparison destination image and searches an image similar to the query image, extracts a plurality of corresponding pairs of feature points in two images based on feature points selected from the both images. A coordinate transformation coefficient to execute a coordinate transformation process is decided so that coordinates of the feature point of one of the two images match coordinates of the feature point of the other image, in relation to each pair. Only if an amount of transformation of coordinates satisfies the constraint conditions designated in advance, the coordinate transformation process using the coordinate transformation coefficient is executed in relation to the plurality of pairs of feature points, and coordinates of the feature points after the transformation of one image are compared with coordinates of the corresponding feature points of the other image.Type: GrantFiled: April 27, 2010Date of Patent: June 25, 2013Assignee: Canon Kabushiki KaishaInventors: Hirotaka Shiiyama, Hidetomo Sohma, Koichi Magai, Masahiro Matsushita
-
Patent number: 8416982Abstract: An image processing method performs reduction processing on an input image to acquire a reduced image, extracts a feature point from a group of images including the input image and one or more of the reduced images, determines as a matched feature point the feature point extracted from a matching position in each of two or more images in the group of images, calculates a local feature quantity of the matched feature point determined by the determination unit, and registers the calculated local feature quantity as a local feature quantity of the input image.Type: GrantFiled: September 24, 2009Date of Patent: April 9, 2013Assignee: Canon Kabushiki KaishaInventors: Koichi Magai, Hirotaka Shiiyama, Hidetomo Sohma, Masahiro Matsushita
-
Patent number: 8411960Abstract: An image processing apparatus extracts an object area (e.g., character, picture, line drawing, and table) from an input image and acquires a metadata to be associated with the object. The image processing apparatus generates a transparent graphics description for an object area having an attribute that requires generation of the transparent graphics description, and generates an electronic document while associating the transparent graphics description with the metadata. As transparent graphics description, an arbitrary shape of graphics can be used. Accordingly, the image processing apparatus can generate electronic document data suitable for a highlight expression, which is easy for users to recognize in a search operation using a keyword to search an object included in an electronic document.Type: GrantFiled: February 25, 2010Date of Patent: April 2, 2013Assignee: Canon Kabushiki KaishaInventors: Reiji Misawa, Ryo Kosaka, Tomotoshi Kanatsu, Hidetomo Sohma
-
Patent number: 8412730Abstract: An image search apparatus extracts a symmetric partial image based on acquired pairs of feature points, and divides the partial image at a symmetry axis of the partial image to decide two partial areas, as a first area and a second area based on the tendency of image features in the partial image. In determining the similarity between the query image and a comparison destination image, a coordinate transformation coefficient for a coordinate transformation process is set based on pairs of feature points extracted from the partial area decided as the first area of the two images, the coordinate transformation process using the coordinate transformation coefficient is applied to the pairs of feature points extracted from areas including areas other than the first area of the two images, and coordinates of the pairs of feature points after the coordinate transformation process are compared.Type: GrantFiled: May 12, 2010Date of Patent: April 2, 2013Assignee: Canon Kabushiki KaishaInventors: Hirotaka Shiiyama, Hidetomo Sohma, Koichi Magai, Masahiro Matsushita
-
Patent number: 8385656Abstract: An image processing apparatus reduces two input images to be compared by the predetermined number of times to generate two image groups, extracts a plurality of feature points and a local feature amount of each feature point from these image groups, and determines a combination of feature points in which local feature amounts are similar to each other, between the image groups. Then, the image processing apparatus determines a relation of a reasonable combination, assigns high weights to the reasonable combination, and calculates a similarity degree between the two input images.Type: GrantFiled: October 21, 2009Date of Patent: February 26, 2013Assignee: Canon Kabushiki KaishaInventors: Hidetomo Sohma, Hirotaka Shiiyama, Masahiro Matsushita, Koichi Magai