Patents by Inventor Ernest Yiu Cheong Wan
Ernest Yiu Cheong Wan 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: 10019810Abstract: A method of determining a depth value of a fine structure pixel in a first image of a scene using a second image of the scene is disclosed. A gradient orientation for each of a plurality of fine structure pixels in the first image is determined. Difference images are generated from the second image and a series of blurred images formed from the first image, each difference image corresponding to one of a plurality of depth values. Each of the difference images is smoothed, in accordance with the determined gradient orientations, to generate smoothed difference images having increased coherency of fine structure. For each of a plurality of fine structure pixels in the first image, one of the smoothed difference images is selected. The depth value of the fine structure pixel corresponding to the selected smoothed difference image is determined.Type: GrantFiled: November 24, 2014Date of Patent: July 10, 2018Assignee: CANON KABUSHIKI KAISHAInventors: Matthew Raphael Arnison, Ernest Yiu Cheong Wan
-
Publication number: 20180176543Abstract: A method of refining a depth map determines a depth map by estimating a depth of pixels of an associated image. The depth map is divided into connected components based on similarity of each of the pixels of the depth map. A band of pixels along a boundary of each of the connected components is selected. A confidence score is assigned to pixels within each band of pixels. The confidence score increases with the estimated depth of the pixels. The depth map is updated based on the assigned confidence scores to determine a refined depth map.Type: ApplicationFiled: December 21, 2016Publication date: June 21, 2018Inventor: Ernest Yiu Cheong Wan
-
Patent number: 10003787Abstract: A method of refining a depth map determines a depth map by estimating a depth of pixels of an associated image. The depth map is divided into connected components based on similarity of each of the pixels of the depth map. A band of pixels along a boundary of each of the connected components is selected. A confidence score is assigned to pixels within each band of pixels. The confidence score increases with the estimated depth of the pixels. The depth map is updated based on the assigned confidence scores to determine a refined depth map.Type: GrantFiled: December 21, 2016Date of Patent: June 19, 2018Assignee: Canon Kabushiki KaishaInventor: Ernest Yiu Cheong Wan
-
Patent number: 9832456Abstract: To extend the working range of depth from defocus (DFD) particularly on small depth of field (DoF) images, DFD is performed on an image pair at multiple spatial resolutions and the depth estimates are then combined. Specific implementations construct a Gaussian pyramid for each image of an image pair, perform DFD on the corresponding pair of images at each level of the two image pyramids, convert DFD depth scores to physical depth values using calibration curves generated for each level, and combine the depth values from all levels in a coarse-to-fine manner to obtain a final depth map that covers the entire depth range of the scene.Type: GrantFiled: December 17, 2015Date of Patent: November 28, 2017Assignee: Canon Kabushiki KaishaInventor: Ernest Yiu Cheong Wan
-
Patent number: 9589319Abstract: An input depth map is received having less resolution than an image. From a window region about an image pixel of the image, pixels are selected having a substantially similar characteristic value as the image pixel. A reference depth value is determined for the image pixel from depth values in the input depth map which correspond to each of the selected pixels. Weights are determined for each pixel within the window region based on a difference between a depth value corresponding to each of the selected pixels and the determined reference depth value for the image pixel. A refined depth value is determined for the image pixel from a weighted sum of the depth values corresponding to each of the selected pixels. A high resolution depth map corresponding to the image is determined from the refined depth value.Type: GrantFiled: August 25, 2015Date of Patent: March 7, 2017Assignee: Canon Kabushiki KaishaInventor: Ernest Yiu Cheong Wan
-
Publication number: 20160182893Abstract: To extend the working range of depth from defocus (DFD) particularly on small depth of field (DoF) images, DFD is performed on an image pair at multiple spatial resolutions and the depth estimates are then combined. Specific implementations construct a Gaussian pyramid for each image of an image pair, perform DFD on the corresponding pair of images at each level of the two image pyramids, convert DFD depth scores to physical depth values using calibration curves generated for each level, and combine the depth values from all levels in a coarse-to-fine manner to obtain a final depth map that covers the entire depth range of the scene.Type: ApplicationFiled: December 17, 2015Publication date: June 23, 2016Inventor: ERNEST YIU CHEONG WAN
-
Patent number: 9292928Abstract: A method of forming a refined depth map DR of an image I using a binary depth map DI of the image, said method comprising segmenting (315) the image into a superpixel image SREP, defining (330) a foreground and a background in the superpixel image SREP, to form a superpixel depth map DS, intersecting (450) the respective foreground and the background of the superpixel depth map DS with the binary depth map DI determined independently of the superpixel image SREP, to define a trimap T consisting of a foreground region, a background region and an unknown region, and forming the refined binary depth map DR of the image from the trimap T by reclassifying (355, 365) the pixels in the unknown region as either foreground or background based on a comparison (510) of the pixel values in the unknown region with pixel values in at least one of the other trimap regions.Type: GrantFiled: June 27, 2014Date of Patent: March 22, 2016Assignee: Canon Kabushiki KaishaInventor: Ernest Yiu Cheong Wan
-
Publication number: 20160063715Abstract: An input depth map is received having less resolution than an image. From a window region about an image pixel of the image, pixels are selected having a substantially similar characteristic value as the image pixel. A reference depth value is determined for the image pixel from depth values in the input depth map which correspond to each of the selected pixels. Weights are determined for each pixel within the window region based on a difference between a depth value corresponding to each of the selected pixels and the determined reference depth value for the image pixel. A refined depth value is determined for the image pixel from a weighted sum of the depth values corresponding to each of the selected pixels. A high resolution depth map corresponding to the image is determined from the refined depth value.Type: ApplicationFiled: August 25, 2015Publication date: March 3, 2016Inventor: Ernest Yiu Cheong Wan
-
Publication number: 20150146994Abstract: A method of determining a depth value of a fine structure pixel in a first image of a scene using a second image of the scene is disclosed. A gradient orientation for each of a plurality of fine structure pixels in the first image is determined. Difference images are generated from the second image and a series of blurred images formed from the first image, each difference image corresponding to one of a plurality of depth values. Each of the difference images is smoothed, in accordance with the determined gradient orientations, to generate smoothed difference images having increased coherency of fine structure. For each of a plurality of fine structure pixels in the first image, one of the smoothed difference images is selected. The depth value of the fine structure pixel corresponding to the selected smoothed difference image is determined.Type: ApplicationFiled: November 24, 2014Publication date: May 28, 2015Inventors: Matthew Raphael ARNISON, ERNEST YIU CHEONG WAN
-
Publication number: 20150003725Abstract: A method of forming a refined depth map DR of an image I using a binary depth map DI of the image, said method comprising segmenting (315) the image into a superpixel image SREP, defining (330) a foreground and a background in the superpixel image SREP, to form a superpixel depth map DS, intersecting (450) the respective foreground and the background of the superpixel depth map DS with the binary depth map DI determined independently of the superpixel image SREP, to define a trimap T consisting of a foreground region, a background region and an unknown region, and forming the refined binary depth map DR of the image from the trimap T by reclassifying (355, 365) the pixels in the unknown region as either foreground or background based on a comparison (510) of the pixel values in the unknown region with pixel values in at least one of the other trimap regions.Type: ApplicationFiled: June 27, 2014Publication date: January 1, 2015Inventor: ERNEST YIU CHEONG WAN
-
Patent number: 8819733Abstract: An assistance method for selecting a program using a display device (112) is disclosed. The method (1200) selects one or more attributes associated with a first program, the one or more attributes being selected dynamically by a processor associated with the display device (112) according to one or more predetermined criteria. The method searches for a second program associated with one or more of the selected attributes. The second program and the one or more attributes associated therewith are displayed on the display device, according to the search for the second program.Type: GrantFiled: June 22, 2012Date of Patent: August 26, 2014Assignee: Canon Kabushiki KaishaInventors: Laurence Crew, Eileen Oi-Yan Mak, Jeonren Vendrig, Ernest Yiu Cheong Wan, Barry James Drake
-
Patent number: 8578297Abstract: Disclosed is a method of browsing a set of media items on a display device. A sequence of thumbnail images corresponding to a portion of the set are initially displayed in a looping arrangement. A first thumbnail image of the sequence is displayed at a size larger than other thumbnail images of the sequence. The first thumbnail image corresponds to a focus position in the sequence. The arrangement responds to a user scrolling action to replace display of a second thumbnail image adjacent to an insertion point in the displayed sequence with a third thumbnail image corresponding to a media item of the set not displayed prior to the user action. The method then updates the focus position and insertion point and replaces a display of the first thumbnail image with display of a fourth thumbnail image corresponding to the updated focus position.Type: GrantFiled: November 26, 2008Date of Patent: November 5, 2013Assignee: Canon Kabushiki KaishaInventors: Steven David Webster, Ernest Yiu Cheong Wan
-
Patent number: 8531557Abstract: A method of performing a zoom operation on a camera is disclosed, wherein, one or more regions of interest within a captured image of a scene are determined. Camera motion direction towards one of the regions of interest is determined. A target region of interest is determined based on the determined camera motion direction. The zoom operation is performed to the target region of interest.Type: GrantFiled: April 26, 2011Date of Patent: September 10, 2013Assignee: Canon Kabushiki KaishaInventors: Ernest Yiu Cheong Wan, Xin Yu Liu
-
Patent number: 8392415Abstract: A method for clustering a plurality of content items. The method comprises, determining intervals between at least some of the content items, and then selecting a seed pair of content items using the determined intervals. The seed pair is then assigned to a current cluster, before an interval is determined between one of the content items in the current cluster and a neighboring one of the content items. The interval is then compared to a predetermined criterion in response to a successful comparison, the content item is assigned to the cluster.Type: GrantFiled: November 20, 2006Date of Patent: March 5, 2013Assignee: Canon Information Systems Research Australia Pty. Ltd.Inventor: Ernest Yiu Cheong Wan
-
Publication number: 20120266192Abstract: An assistance method for selecting a program using a display device (112) is disclosed. The method (1200) selects one or more attributes associated with a first program, the one or more attributes being selected dynamically by a processor associated with the display device (112) according to one or more predetermined criteria. The method searches for a second program associated with one or more of the selected attributes. The second program and the one or more attributes associated therewith are displayed on the display device, according to the search for the second program.Type: ApplicationFiled: June 22, 2012Publication date: October 18, 2012Applicant: CANON KABUSHIKI KAISHAInventors: LAURENCE CREW, EILEEN OI-YAN MAK, JEONREN VENDRIG, ERNEST YIU CHEONG WAN, BARRY JAMES DRAKE
-
Patent number: 8169497Abstract: A method of segmenting a compressed video containing a plurality of frames into a multi-level hierarchy of video segments. The method comprises the steps of determining differences between the size of successive compressed frames in media time. The method then groups the frames to create a hierarchy of clusters of successive frames as video segments using a clustering process. The clustering process uses the size differences to determine boundaries between clusters at each level of the hierarchy.Type: GrantFiled: October 28, 2008Date of Patent: May 1, 2012Assignee: Canon Kabushiki KaishaInventor: Ernest Yiu Cheong Wan
-
Publication number: 20110267499Abstract: A method of performing a zoom operation on a camera is disclosed, wherein, one or more regions of interest within a captured image of a scene are determined. Camera motion direction towards one of the regions of interest is determined. A target region of interest is determined based on the determined camera motion direction. The zoom operation is performed to the target region of interest.Type: ApplicationFiled: April 26, 2011Publication date: November 3, 2011Applicant: CANON KABUSHIKI KAISHAInventors: Ernest Yiu Cheong Wan, Xin Yu Liu
-
Patent number: 7984037Abstract: A method of retrieving a plurality of resource fragments from a audio-visual resource that is encoded using a hierarchical addressing model for the class of resources of which the audio-visual resource is a member is disclosed. The method comprises inputting a URI reference (1407) comprising a Universal Resource Identifier (1401) and a complex fragment identifier (1402) comprising a plurality of audio-visual resource specific location steps. Thereafter the method comprises a step of locating the audio-visual resource using the Universal Resource Name, and establishing (2023) a logical representation (2040) of the structure of the audio-visual resource. Then the method iteratively evaluates the fragment identifier location steps (1110) against the logical representation (2040) to resolve the complex fragment identifier (1402) into a set of explicit fragment references (1406) configured to explicitly address each of the plurality of resource fragments.Type: GrantFiled: July 15, 2005Date of Patent: July 19, 2011Assignee: Canon Kabushiki KaishaInventors: Ernest Yiu Cheong Wan, Myriam Elisa Lucie Amielh
-
Patent number: 7756857Abstract: A method is disclosed that can be used to reduce the size of the indexes 130 of structured-documents 150. The method uses the schema 110 of structured documents 180, in particular XML documents, to determine the existence of certain deterministic relationships 140 among the indexing components 170. Using such knowledge when indexing the data and formulating query execution plans can greatly reduce the size of the indices 170 and allow querying of such documents more efficiently.Type: GrantFiled: June 16, 2003Date of Patent: July 13, 2010Assignee: Canon Kabushiki KaishaInventor: Ernest Yiu Cheong Wan
-
Publication number: 20100138736Abstract: A method of processing a document described in a mark up language, for example XML, is disclosed. A structure and a text content of the document are separated, and then the structure is transmitted before the text content, for example, by streaming. Parsing of the received structure is commenced before all of the text content is received. Also disclosed is a method of forming a streamed presentation from at least one media object having content and description components. A presentation description is generated from at least one component description of the media object and is processed to schedule delivery of component descriptions and content of the presentation to generate elementary data streams associated with the component descriptions and content.Type: ApplicationFiled: February 1, 2010Publication date: June 3, 2010Applicant: CANON KABUSHIKI KAISHAInventor: Ernest Yiu Cheong Wan