# US 10,121,250 B2 · Image orientation detection > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,121,250 B2 - Document ID: US10121250B2 - Status: Granted - Kind: B2 - Application number: 15/366,478 - Filed: 2016-12-01 - Published: 2018-06-07 - Granted: 2018-11-06 - Earliest claimed: 2016-12-01 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: Stephen Chu; Min Gong; Chang Sheng Li; Dong Sheng Li; Jun Chi Yan; Weipeng Zhang - Web page: https://patentgrep.com/patent/US10121250B2 ## CPC - G06T 7/73: using feature-based methods - G06T 7/11: Region-based segmentation - G06V 10/242: {by image rotation, e.g. by 90 degrees} - G06T 2207/10004: Still image; Photographic image - G06T 2207/20081: Training; Learning - G09G 2340/0492: Change of orientation of the displayed image, e.g. upside-down, mirrored - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G09G: ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION (arrangements for transferring data between digital computers and displays G06F3/14; static indicating arrangements comprising an association of a number of separate sources or light control cells G09F9/00; static indicating arrangements comprising integral associations of a number of light sources H01J, H01K, H10H, H05B33/12; scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission, details thereof H04N1/00) ## Abstract Embodiments of the present disclosure relate to a new approach for image orientation detection. In the computer-implemented method, at least one character area in an image is sampled. The orientation of the sampled character area is determined, and the orientation of the image is determined based on the determined orientation of the sampled character area. ## Claims ### Claim 1 A device comprising: a processing unit; a screen; and a memory coupled to the processing unit and storing instructions thereon, the instructions, when executed by the processing unit, perform a method comprising: sampling at least one character area in an image by: detecting at least one text region in the image; and sampling the at least one character area in the at least one text region by: obtaining, within the detected text region, an initial predefined character area; estimating orientation and corresponding confidence score of the obtained initial predefined character area; adjusting the initial predefined character area; re-estimating the orientation and the corresponding confidence score; and in response to the corresponding confidence score being greater than a predefined threshold, determining an adjusted character area to be used in sampling of the at least one character areas in the at least one text region; determining orientation of the sampled character area; and determining orientation of the image based on the determined orientation of the sampled character area; and displaying the image on the screen based on the determined image orientation. ### Claim 2 (depends on 1) The device of claim 1, wherein, the displaying the image on the screen based on the determined image orientation comprises: in response to the determined image orientation being different from an orientation of screen, rotating the image according to the determined image orientation. ### Claim 3 (depends on 1) The device of claim 1, wherein: detecting at least one text region in the image comprises detecting the text region in part of the image displayed on the screen; determining orientation of the image based on the determined orientation of the sampled character area comprises determining orientation of the part of the image; and displaying the image based on the determined image orientation comprises displaying the image according to the orientation of the part of the image. ### Claim 4 (depends on 1) The device of claim 1, wherein sampling at least one character area in the image comprises sampling the at least one character area in the image randomly or evenly. ### Claim 5 (depends on 1) The device of claim 1, wherein determining orientation of the sampled character area comprises determining the orientation of the sampled character area by a character orientation estimation model. ### Claim 6 (depends on 5) The device of claim 5, wherein the memory further comprises instructions for: generating a train set for the at least one character area, wherein the train set includes a plurality of pairs which comprise a raw image of a character and corresponding orientation information; and training the character orientation estimation model based on the train set. ### Claim 7 A computer-implemented method comprising: sampling at least one character area in an image by: detecting at least one text region in the image; and sampling the at least one character area in the at least one text region by: obtaining, within the detected text region, an initial predefined character area; estimating orientation and corresponding confidence score of the obtained initial predefined character area; adjusting the initial predefined character area; re-estimating the orientation and the corresponding confidence score; and in response to the corresponding confidence score being greater than a predefined threshold, determining an adjusted character area to be used in sampling of the at least one character areas in the at least one text region; determining orientation of the sampled character area; determining orientation of the image based on the determined orientation of the sampled character area; and displaying the image on a screen based on the determined image orientation. ### Claim 8 (depends on 7) The method of claim 7, wherein, displaying the image based on the determined image orientation comprises: in response to the determined image orientation being different from an orientation of the screen, rotating the image according to the determined image orientation. ### Claim 9 (depends on 7) The method of claim 7, wherein, detecting at least one text region in the image comprises detecting the text region in part of the image displayed on the screen; determining orientation of the image based on the determined orientation of the sampled character area comprises determining orientation of the part of the image; and displaying the image based on the determined image orientation comprises displaying the image according to the orientation of the part of the image. ### Claim 10 (depends on 7) The method of claim 7, wherein sampling at least one character area in the image comprises sampling the character areas in the image randomly or evenly. ### Claim 11 (depends on 7) The method of claim 7, wherein determining orientation of the sampled character area comprises determining the orientation of the sampled character area by a character orientation estimation model. ### Claim 12 (depends on 11) The method of claim 11, further comprising: generating a train set for the at least one character area, wherein the train set includes a plurality of pairs which comprise raw image of a character and corresponding orientation information; and training the character orientation estimation model based on the train set. ### Claim 13 A computer program product being tangibly stored on a non-transitory machine-readable medium and comprising machine-executable instructions, the instructions, when executed on an electronic device, causing the electronic device to: sample at least one character area in an image by: detecting at least one text region in the image; and sampling the at least one character area in the at least one text region by: obtaining, within the detected text region, an initial predefined character area: estimating orientation and corresponding confidence score of the obtained initial predefined character area; adjusting the initial predefined character area; re-estimating the orientation and the corresponding confidence score; and in response to the corresponding confidence score being greater than a predefined threshold, determining an adjusted character area to be used in sampling of the at least one character areas in the at least one text region; determine orientation of the sampled character area; determine orientation of the image based on the determined orientation of the sampled character area; and display the image on a screen based on the determined image orientation.