Patents by Inventor Sung-Jea Ko
Sung-Jea Ko 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).
-
Publication number: 20240119571Abstract: Disclosed is a system for improving the detection performance of objects of interest in blurry images. According to an embodiment of the disclosed system for improving the detection performance of objects of interest in blurry images, the system includes a teacher network including a first base subnet, a first feature pyramid subnet and a first detection subnet; and a student network including a second base subnet, a second feature pyramid subnet, a deblurring subnet and a second detection subnet, wherein the teacher network provides hints to the student network.Type: ApplicationFiled: July 19, 2023Publication date: April 11, 2024Applicant: KOREA UNIVERSITY RESEARCH AND BUSINESS FOUNDATIONInventors: Sung Jea KO, Sung Jin CHO
-
Patent number: 11783460Abstract: The present disclosure relates to an apparatus and method for optimizing inverse tone mapping on basis of a single image to reconstruct into a high quality High Dynamic Range (HDR) image using a Low Dynamic Range (LDR) image and a recording medium for performing the method, and includes inputting a LDR image to Inverse Tone Mapping (ITM) operator, inputting a result value by the ITM operator to Tone Mapping (TM) operator, comparing a result value by the TM operator with a threshold for HDR luminance determination, and selecting the result value by the TM operator which is smaller than the threshold as an HDR image.Type: GrantFiled: January 28, 2019Date of Patent: October 10, 2023Assignee: Korea University Research and Business FoundationInventors: Sung Jea Ko, Kwanghyun Uhm, Sungjin Cho, Hyongkeun Kook, Seungwook Kim
-
Publication number: 20220051376Abstract: The present disclosure relates to an apparatus and method for optimizing inverse tone mapping on basis of a single image to reconstruct into a high quality High Dynamic Range (HDR) image using a Low Dynamic Range (LDR) image and a recording medium for performing the method, and includes inputting a LDR image to Inverse Tone Mapping (ITM) operator, inputting a result value by the ITM operator to Tone Mapping (TM) operator, comparing a result value by the TM operator with a threshold for HDR luminance determination, and selecting the result value by the TM operator which is smaller than the threshold as an HDR image.Type: ApplicationFiled: January 28, 2019Publication date: February 17, 2022Applicant: Korea University Research and Business FoundationInventors: Sung Jea KO, Kwanghyun UHM, Sungjin CHO, Hyongkeun KOOK, Seungwook KIM
-
Patent number: 11244584Abstract: The present invention relates to a technology for a sensor network, machine to machine (M2M) communication, machine type communication (MTC), and the Internet of things (IoT). The present invention can be utilized for intelligent services (smart home, smart building, smart city, smart car or connected car, health care, digital education, retail, security and safety-related services, and the like) based on the technology. The present invention relates to an efficient image processing method and device for virtual reality content, and according to one embodiment of the present invention, the image processing method for projecting an image of virtual reality content comprises the steps of: acquiring a first planar image projected by dividing a front part and a rear part of a spherical image for expressing a 360-degree image; generating a second planar image projected by sampling the first planar image on the basis of a pixel position; and encoding the second planar image.Type: GrantFiled: September 12, 2017Date of Patent: February 8, 2022Assignees: Samsung Electronics Co., Ltd., KOREA UNIVERSITY RESEARCH AND BUSINESS FOUNDATIONInventors: Jae-Yeon Song, Sung-Jea Ko, Eric Yip, Byeong-Doo Choi, Sung-Ho Lee, Sung-Tae Kim
-
Patent number: 10943109Abstract: An electronic device is provided. The electronic device includes a light emitting unit configured to alternately emit light having a preset pattern to an eye area of a user; an inputter configured to receive a plurality of images which capture an eye area of the user; and a processor configured to detect a user's gaze using a first image that is captured at the time of irradiating a preset first pattern and a second image that is captured at the time of irradiating a preset second pattern, and the processor detects a user's gaze by extracting reflection points of each of the first image and the second image and using remaining reflection points excluding reflection points at a same position from among the extracted reflection points.Type: GrantFiled: September 5, 2018Date of Patent: March 9, 2021Assignees: Samsung Electronics Co., Ltd., Korea University Research and Business FoundationInventors: Kang-won Jeon, Sung-Jea Ko, Tae-young Na, Mun-Cheon Kang, Kwang-Hyun Uhm
-
Patent number: 10817054Abstract: An electronic apparatus is disclosed. The electronic apparatus includes an inputter configured to receive a binocular image which is a captured image of both eyes of a user and a stereo image which is an image of a direction corresponding to a gaze of the user captured at locations spaced apart from each other, and a processor configured to detect a watch point of a user in the stereo image by using the binocular image, obtain a disparity map in the input stereo image, and compensate the detected watch point using the obtained disparity map.Type: GrantFiled: September 5, 2018Date of Patent: October 27, 2020Assignees: Samsung Electronics Co., Ltd., Korea University Research and Business FoundationInventors: Kang-won Jeon, Sung-Jea Ko, Tae-young Na, Mun-Cheon Kang, Sung-Ho Chae
-
Publication number: 20190236990Abstract: The present invention relates to a technology for a sensor network, machine to machine (M2M) communication, machine type communication (MTC), and the Internet of things (IoT). The present invention can be utilized for intelligent services (smart home, smart building, smart city, smart car or connected car, health care, digital education, retail, security and safety-related services, and the like) based on the technology. The present invention relates to an efficient image processing method and device for virtual reality content, and according to one embodiment of the present invention, the image processing method for projecting an image of virtual reality content comprises the steps of: acquiring a first planar image projected by dividing a front part and a rear part of a spherical image for expressing a 360-degree image; generating a second planar image projected by sampling the first planar image on the basis of a pixel position; and encoding the second planar image.Type: ApplicationFiled: September 12, 2017Publication date: August 1, 2019Inventors: Jae-Yeon SONG, Sung-Jea KO, Eric YIP, Byeong-Doo CHOI, Sung-Ho LEE, Sung-Tae KIM
-
Publication number: 20190073793Abstract: An electronic apparatus is disclosed. The electronic apparatus includes an inputter configured to receive a binocular image which is a captured image of both eyes of a user and a stereo image which is an image of a direction corresponding to a gaze of the user captured at locations spaced apart from each other, and a processor configured to detect a watch point of a user in the stereo image by using the binocular image, obtain a disparity map in the input stereo image, and compensate the detected watch point using the obtained disparity map.Type: ApplicationFiled: September 5, 2018Publication date: March 7, 2019Inventors: Kang-won JEON, Sung-Jea KO, Tae-young NA, Mun-Cheon KANG, Sung-Ho CHAE
-
Publication number: 20190073513Abstract: An electronic device is provided. The electronic device includes a light emitting unit configured to alternately emit light having a preset pattern to an eye area of a user; an inputter configured to receive a plurality of images which capture an eye area of the user; and a processor configured to detect a user's gaze using a first image that is captured at the time of irradiating a preset first pattern and a second image that is captured at the time of irradiating a preset second pattern, and the processor detects a user's gaze by extracting reflection points of each of the first image and the second image and using remaining reflection points excluding reflection points at a same position from among the extracted reflection points.Type: ApplicationFiled: September 5, 2018Publication date: March 7, 2019Inventors: Kang-won JEON, Sung-Jea KO, Tae-young NA, Mun-Cheon KANG, Kwang-Hyun UHM
-
Patent number: 9880109Abstract: A display apparatus includes a background image estimator configured to restructure a plurality of pixel signals of a panel image corresponding to a plurality of pixels arranged in an (n×m) matrix array into a row dataset and a column dataset and generate a row background image and a column background image with a Mura defect removed from the panel image using the row dataset and the column dataset through a Principal Component Analysis (PCA), a Mura image generator configured to generate a row binary image and a column binary image including a background and the Mura defect using differences between the panel image and the row background image and between the panel image and the column background image.Type: GrantFiled: January 21, 2015Date of Patent: January 30, 2018Assignees: Samsung Display Co., Ltd., Korea University Research and Business FoundationInventors: Se-Yun Kim, Sung-Jea Ko, Hoi-Sik Moon
-
Patent number: 9355471Abstract: A color balancing apparatus for multiple image synthesis and a method therefore, in which an automatic white balance unit, with respect to each of a plurality of input images that are obtained by sequentially capturing adjacent scenes, is configured to compare a neutral color representing the input image with chromaticity of standard sources that are predefined, determine a standard source which is closest to the neutral color among the standard sources as a standard illuminant, and perform a white balance on the input image based on the standard illuminant, so that a natural wide-area synthesis image is generated by use of a light source estimation scheme considering multiple light sources, without complex computation of the conventional wide-area image synthesis, are provided.Type: GrantFiled: January 9, 2012Date of Patent: May 31, 2016Assignee: KOREA UNIVERSITY RESEARCH AND BUSINESS FOUNDATIONInventors: Sung-Jea Ko, Seung-Kyun Kim, Seung-Won Jung
-
Publication number: 20160012759Abstract: A display apparatus includes a background image estimator configured to restructure a plurality of pixel signals of a panel image corresponding to a plurality of pixels arranged in an (n×m) matrix array into a row dataset and a column dataset and generate a row background image and a column background image with a Mura defect removed from the panel image using the row dataset and the column dataset through a Principal Component Analysis (PCA), a Mura image generator configured to generate a row binary image and a column binary image including a background and the Mura defect using differences between the panel image and the row background image and between the panel image and the column background image.Type: ApplicationFiled: January 21, 2015Publication date: January 14, 2016Inventors: Se-Yun Kim, Sung-Jea Ko, Hoi-Sik Moon
-
Patent number: 9025893Abstract: Provided are a method and an apparatus for encoding images using template matching and a method and an apparatus for decoding images. The method for encoding the images can determine a template of an encoding target block and can determine a matching search target image for performing a matching search with the determined template among the recovered reference images. In addition, an optimum prediction block can be determined on the basis of the determined matching search target image and the template of the encoding target block.Type: GrantFiled: July 8, 2011Date of Patent: May 5, 2015Assignee: Electronics and Telecommunications Research InstituteInventors: Ha Hyun Lee, Sung Chang Lim, Hui Yong Kim, Se Yoon Jeong, Jong Ho Kim, Jin Ho Lee, Suk Hee Cho, Jin Soo Choi, Jin Woong Kim, Chie Teuk Ahn, Sung Jea Ko, Seung Won Jung, Yeo Jin Yoon
-
Patent number: 8908760Abstract: The present invention relates to a method and apparatus for processing video signals, the method comprising: decoding a current frame of a video signal in block units to generate a restored frame, and acquiring first and second flags to acquire a filter coefficient which is adaptively determined from the video signal for each slice on the basis of flag information, or performing deblocking filtering with a predetermined filter coefficient. The present invention relates to a video signal processing method which includes an inter-frame prediction method that determines a motion vector, and a method for performing inter-frame prediction for a residual image. The method for processing video signals can increase the compression encoding efficiency by performing accurate motion prediction and residual image prediction, and can improve the quality of a restored image by removing distortion and performing a filtering for restoring an image that is close to the original image.Type: GrantFiled: March 30, 2010Date of Patent: December 9, 2014Assignees: LG Electronics Inc., Korea University Research and Business FoundationInventors: Byeong Moon Jeon, Joon Young Park, Young Hee Choi, Seung Wook Park, Jung Sun Kim, Seung-Won Jung, Sung Jea Ko, Seung-Jin Baek
-
Publication number: 20140126815Abstract: A color balancing apparatus for multiple image synthesis and a method therefore, in which an automatic white balance unit, with respect to each of a plurality of input images that are obtained by sequentially capturing adjacent scenes, is configured to compare a neutral color representing the input image with chromaticity of standard sources that are predefined, determine a standard source which is closest to the neutral color among the standard sources as a standard illuminant, and perform a white balance on the input image based on the standard illuminant, so that a natural wide-area synthesis image is generated by use of a light source estimation scheme considering multiple light sources, without complex computation of the conventional wide-area image synthesis, are provided.Type: ApplicationFiled: January 9, 2012Publication date: May 8, 2014Applicant: KOREA UNIVERSITY RESEARCH AND BUSINESS FOUNDATIONInventors: Sung-Jea Ko, Seung-Kyun Kim, Seung-Won Jung
-
Patent number: 8693549Abstract: Provided are an enhanced inter-layer reference picture designating method and apparatus and an enhanced inter-layer reference picture recovering method and apparatus capable of increasing coding efficiency by allowing different blocks in an upper spatial layer picture of a current to-be-decoded or to-be-encoded picture to refer to blocks in different lower spatial layer pictures.Type: GrantFiled: January 16, 2007Date of Patent: April 8, 2014Assignee: Electronics and Telecommunications Research InstituteInventors: Hae-Chul Choi, Jae-Gon Kim, Jin-Woo Hong, Sung-Jea Ko
-
Publication number: 20140015850Abstract: A color transformation method and apparatus are provided for a person with a color vision defect. A color vision characteristic of a user is determined by receiving at least one of a kind of color vision defect and a degree of color vision defect selected by the user via a preset user color vision characteristic input menu. A preset basic hue table of a terminal is transformed according to the color vision characteristic of the user and transforming colors of a video to be displayed by applying the transformed basic hue table.Type: ApplicationFiled: July 11, 2013Publication date: January 16, 2014Inventors: Sung-Joo AHN, Sung-Jea KO, Jung-Hoon PARK, Hang-Sik SHIN, Hyun-Ji KIM, Jae-Yun JEONG, Hyun-Guk YOO
-
Patent number: 8576739Abstract: Disclosed is an apparatus and method for estimating a channel capacity of an end-to-end terminal. The channel capacity estimating method of the end-to-end terminal includes calculating an average value of channel capacities for each size of data packets, the data packets varying in size, measuring an actual channel capacity with respect to the data packets, and deriving a linear model of a channel capacity based on the size of the data packets using the average value of the channel capacities and the actual channel capacity.Type: GrantFiled: July 2, 2009Date of Patent: November 5, 2013Assignee: Electronics and Telecommunications Research InstituteInventors: Seong-Jun Bae, Jeong Ju Yoo, Jin Woo Hong, Sung Jea Ko, Hye Soo Kim, Hyung Min Nam
-
Publication number: 20130177078Abstract: Provided are a method and apparatus for encoding/decoding video. The method and apparatus of the present invention involve generating a first prediction block for a block to be decoded, calculating a filter coefficient on the basis of the block adjacent to the first prediction block, and performing filtering on the first prediction block using the filter coefficient so as to generate a second prediction block if information relating to the performance of filtering indicates that filtering should be performed. According to the present invention, accuracy in video prediction and encoding performance are improved.Type: ApplicationFiled: September 30, 2011Publication date: July 11, 2013Applicant: Electronics and Telecommunications Research InstituteInventors: Ha Hyun Lee, Hui Yong Kim, Sung Chang Lim, Se Yoon Jeong, Jong Ho Kim, Jin Ho Lee, Suk Hee Cho, Jin Soo Choi, Jin Woong Kim, Chie Teuk Ahn, Sung Jea Ko, Yeo Jin Yoon, Keun Yung Byun
-
Patent number: 8483201Abstract: A server and a terminal enabling channel capacity estimation in a High-Speed Downlink Packet Access (HSDPA) network and a method of controlling the server and the terminal are provided. More particularly, when transmitting data between both terminals in an HSDPA network, a server end may transmit a packet pair of the same size and a client end may measure a time difference between the packet pair and thereby proceed filtering. Through this, it is possible to estimate the channel capacity. Therefore, it is possible to prevent data transmission from exceeding the channel capacity in the HSDPA to thereby prevent a packet loss and to effectively use a channel.Type: GrantFiled: December 15, 2008Date of Patent: July 9, 2013Assignee: Electronics and Telecommunications Research InstituteInventors: Seong-Jun Bae, Jeong Ju Yoo, Jin-Woo Hong, Sung Jea Ko, Hye Soo Kim, Hyeong Min Nam