Patents by Inventor Sehoon Yea

Sehoon Yea 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: 10341677
    Abstract: A method for processing video signals by a decoding apparatus includes determining whether the temporal neighboring block of a current texture block is coded using inter-view prediction; and determining whether the spatial neighboring block of the current texture block is coded using inter-view prediction. When the temporal neighboring block and the spatial neighboring block are not coded using the inter-view prediction, deriving an inter-view motion vector of the current texture block as a disparity vector, and performing inter-view prediction for the current texture block using the derived inter-view motion vector of the current texture block. The disparity vector is obtained using a depth value according the current texture block.
    Type: Grant
    Filed: May 9, 2013
    Date of Patent: July 2, 2019
    Assignee: LG ELECTRONICS INC.
    Inventors: Jaewon Sung, Sehoon Yea, Moonmo Koo, Jin Heo, Taesup Kim, Jiwook Jung, Eunyong Son
  • Patent number: 10321157
    Abstract: The present invention relates to a method and an apparatus for coding a multi-view video, and a decoding apparatus according to the present invention comprises: an entropy-decoding unit for entropy-decoding a bitstream and decoding information related to residual prediction by referencing neighboring blocks of a current block; a memory for saving pictures referenced when decoding the current block; a prediction unit for inducing a prediction sample of the current block, based on the difference between sample values of a first block and a second block corresponding to the current block, from among the pictures saved to the memory, by using the residual-related information; and a filtering unit for applying filtering to a current picture restored by using the prediction sample.
    Type: Grant
    Filed: October 20, 2014
    Date of Patent: June 11, 2019
    Assignee: LG ELECTRONICS INC.
    Inventors: Jin Heo, Sunmi Yoo, Sehoon Yea
  • Publication number: 20190037232
    Abstract: The present invention provides a method for encoding a video signal by using an enhanced prediction signal, the method comprising the steps of: generating a prediction signal for a current block, wherein the prediction signal is generated by applying a prediction weight to a reconstructed pixel of the current block and a pixel of a motion-compensated previously decoded block; obtaining an optimal quantized transform coefficient for the current block through rate-distortion optimization, on the basis of the prediction signal; and performing entropy encoding of the optimal quantized transform coefficient.
    Type: Application
    Filed: February 2, 2017
    Publication date: January 31, 2019
    Inventors: Jana EHMANN, Onur G. GULERYUZ, Sehoon YEA
  • Patent number: 10171836
    Abstract: According to the present invention, an inter-view motion vector of a current coding unit is acquired using parallel processing information, and a motion vector of a current prediction unit in the current coding unit is acquired in parallel using the inter-view motion vector of the current coding unit. The present invention relates to a method and device for processing a video signal, wherein a motion vector prediction value of the current prediction unit is acquired by comparing output order information of a reference picture corresponding to the current prediction unit and output order information of a reference picture corresponding to a corresponding block. According to the present invention, through a parallel processing step, a multi-view point video image can be quickly restored, and the accuracy of the motion vector prediction value can be increased using motion information corresponding to the reference picture of the corresponding block of a different view point from the current prediction unit.
    Type: Grant
    Filed: October 4, 2013
    Date of Patent: January 1, 2019
    Assignee: LG ELECTRONICS INC.
    Inventors: Moonmo Koo, Jiwook Jung, Sehoon Yea, Jin Heo, Taesup Kim, Jaewon Sung, Eunyong Son
  • Publication number: 20180343455
    Abstract: An intra prediction method according to the present invention comprises the steps of: determining an intra prediction mode for a current block; determining whether to apply a coefficient induced intra prediction (CIIP) to the current block; deriving neighboring reference samples for the current block on the basis of whether the CIIP is applied; and generating prediction samples for the current block on the basis of the intra prediction mode and the neighboring samples. The present invention can generate modified or improved neighboring reference samples and thereby can improve the performance of intra prediction.
    Type: Application
    Filed: November 11, 2016
    Publication date: November 29, 2018
    Inventors: Hyeongmoon JANG, Bumshik LEE, Sehoon YEA
  • Patent number: 10129560
    Abstract: The present invention relates to a method and an apparatus for processing a video signal. The present invention can acquire an inter-view motion vector of a current block, acquire a first corresponding block by using the inter-view motion vector, identify a prediction method of the first corresponding block, acquire a second corresponding block when the first corresponding block is not coded by a time inter prediction, acquire a time motion vector of the second corresponding block, acquire a time motion vector of the current block by using the time motion vector of the second corresponding block, and decode the current block by using the time motion vector of the current block.
    Type: Grant
    Filed: July 18, 2014
    Date of Patent: November 13, 2018
    Assignee: LG ELECTRONICS INC.
    Inventors: Taesup Kim, Sunmi Yoo, Sehoon Yea
  • Publication number: 20180324418
    Abstract: The present invention provides a method for processing a video signal, comprising generating a prediction block for a neighbor block using a reconstructed reference pixel of the neighbor block when the neighbor block adjacent to a current block is intra-coded; determining a refined intra prediction mode of the neighbor block based on the prediction block, where the refined intra prediction mode represents a mode which minimizes an error between the prediction block and a reconstructed neighbor block; generating a prediction mode candidate list for the current block based on the refined intra prediction mode; selecting an intra prediction mode for the current block from the prediction mode candidate list; and performing a prediction for the current block based on the intra prediction mode.
    Type: Application
    Filed: October 13, 2016
    Publication date: November 8, 2018
    Inventors: Moonmo KOO, Sehoon YEA, Kyuwoon KIM, Bumshik LEE
  • Patent number: 10123007
    Abstract: The present invention relates to a method and an apparatus for coding a video signal, and more specifically, a motion vector between viewpoints is obtained by using a depth value of a depth block, which corresponds to a current texture block, and an illumination difference is compensated. By obtaining the motion vector between the viewpoints by using the depth value of the depth block, which corresponds to the current texture block, and compensating the illumination difference, the present invention can obtain an accurate prediction value of the current texture block and thus increase accuracy in inter-prediction between the viewpoints.
    Type: Grant
    Filed: January 8, 2014
    Date of Patent: November 6, 2018
    Assignee: LG ELECTRONICS INC.
    Inventors: Jin Heo, Jiwook Jung, Sehoon Yea, Moonmo Koo, Taesup Kim, Jaewon Sung, Eunyong Son
  • Publication number: 20180310000
    Abstract: An image decoding method, which is performed by a decoding apparatus, includes the steps of: deriving a frequency domain-neighboring block for a neighboring block of a current block in a spatial domain; deriving a frequency domain-prediction block for the current block by applying a prediction pattern to the frequency domain-neighboring block; generating a frequency domain-reconstructed block for the current block on the basis of the frequency domain-prediction block; and generating a reconstructed block for the current block in the spatial domain on the basis of the frequency inverse transform of the frequency domain-reconstructed block. The present invention can generate a prediction block for a current block in a frequency domain, and accordingly can more efficiently reconstruct a complex image.
    Type: Application
    Filed: September 23, 2016
    Publication date: October 25, 2018
    Inventors: Jin HEO, Eunyong SON, Junghak NAM, Seungwook PARK, Sehoon YEA
  • Publication number: 20180309996
    Abstract: The present invention provides a method for encoding a video signal comprising the steps of: obtaining activity information from the video signal, wherein the activity signal indicates information relating to edge characteristics of an image and includes edge orientation information and/or edge level information; determining conditionally non-linear transform (CNT) configuration information on the basis of the activity information; and performing CNT prediction coding on the basis of the CNT configuration information, wherein the CNT prediction coding involves performing prediction using all previously decoded pixel values.
    Type: Application
    Filed: October 13, 2016
    Publication date: October 25, 2018
    Inventors: Moonmo KOO, Sehoon YEA, Eunyong SON, Kyuwoon KIM, Jin HEO, Junghak NAM, Bumshik LEE
  • Publication number: 20180309984
    Abstract: An image decoding method, which is performed by a decoding device according to the present invention, comprises the steps of: acquiring the information on an intra-prediction mode of a current block from a received bit stream; deriving neighboring reference samples of the current block; deriving a first prediction value for a current sample of the current block on the basis of the intra-prediction mode and the neighboring reference samples; deriving a second prediction value for the current sample on the basis of adjacent samples of the current sample; and deriving a prediction value for the current sample on the basis of the first prediction value and the second prediction value. According to the present invention, the intra-prediction value can be derived on the basis of the adjacent samples of the current sample within the present block, and the intra-prediction accuracy can be improved therethrough.
    Type: Application
    Filed: September 27, 2016
    Publication date: October 25, 2018
    Inventors: Eunyong SON, Jin HEO, Seoungwook PARK, Sehoon YEA
  • Publication number: 20180302620
    Abstract: The present invention provides a method for encoding a video signal comprising the steps of: performing block division with respect to a target unit of the video signal, wherein the block division is performed on the basis of square division and/or non-square division; determining the optimal block partition on the basis of a rate distortion cost, when the target unit is divided into a plurality of non-square blocks; and performing CNT prediction coding on the non-square blocks resulting from the optimal block partition, wherein the CNT prediction coding involves performing prediction using all previously decoded pixel values.
    Type: Application
    Filed: October 17, 2016
    Publication date: October 18, 2018
    Inventors: Moonmo KOO, Bumshik LEE, Sehoon YEA, Kyuwoon KIM
  • Publication number: 20180295363
    Abstract: An image decoding method performed by a decoding apparatus according to the present invention comprises the steps of: generating a reconstruction area for a target area within a current picture; deriving a predetermined plurality of basic filters for the target area; updating at least one basic filter among the plurality of basic filters; receiving weight information for each of the basic filters; determining a merge filter for the target area on the basis of the basic filters and the weight information; and generating a modified reconstruction area by performing filtering on the target area on the basis of the filter coefficient of the determined merge filter. The present invention allows updating with filter information suitable for improving the visual quality of the target area and can thereby improve the accuracy of filtering and overall coding efficiency.
    Type: Application
    Filed: October 13, 2016
    Publication date: October 11, 2018
    Applicant: LG ELECTRONICS INC.
    Inventors: Junghak NAM, Chulkeun KIM, Sehoon YEA
  • Publication number: 20180295384
    Abstract: An intra prediction method includes: receiving intra prediction mode information through a bitstream; deriving an intra prediction mode of a current block as a first intra directional mode on the basis of the intra prediction mode information; deriving first neighboring samples of the current block; deriving a reference sample adjacent to a target sample in the current block on the basis of a prediction direction of the first intra directional mode; deriving second neighboring samples of a neighboring block to which the reference sample belongs; determining a modified prediction direction on the basis of a comparison between the reference sample and at least one of the second neighboring samples; and deriving a prediction value for the target sample on the basis of the modified prediction direction and the first neighboring samples.
    Type: Application
    Filed: September 8, 2016
    Publication date: October 11, 2018
    Applicant: LG ELECTRONICS INC.
    Inventors: Eunyong SON, Jin HEO, Seungwook PARK, Sehoon YEA
  • Publication number: 20180288438
    Abstract: The present invention provides a method for encoding a video signal on the basis of a graph-based lifting transform (GBLT), comprising the steps of: detecting an edge from an intra residual signal; generating a graph on the basis of the detected edge, wherein the graph includes a node and a weight link; acquiring a GBLT coefficient by performing the GBLT for the graph; quantizing the GBLT coefficient; and entropy-encoding the quantized GBLT coefficient, wherein the GBLT includes a partitioning step, a prediction step, and an update step.
    Type: Application
    Filed: September 29, 2016
    Publication date: October 4, 2018
    Inventors: Yung-Hsuan CHAO, Sehoon YEA, Antonio Ortega
  • Publication number: 20180288407
    Abstract: A method of encoding a video signal using a graph-based transformation includes: generating a residual block using a prediction block generated according to an intra prediction mode; obtaining at least one of a self-loop weight indicating a weight of boundary pixels in the residual block or a correlation coefficient indicating an inter-pixel correlation, on the basis of a prediction angle corresponding to the intra prediction mode; generating a graph on the basis of at least one of the self-loop weight or the correlation coefficient; determining a graph-based transformation kernel on the basis of the graph; and performing a transform for the residual block using the graph-based transformation kernel.
    Type: Application
    Filed: September 12, 2016
    Publication date: October 4, 2018
    Inventors: Bumshik LEE, Sehoon YEA, Moonmo KOO
  • Publication number: 20180288411
    Abstract: The present invention provides a method for encoding a video signal by using a single optimized graph, comprising the steps of: obtaining a residual block; generating graphs from the residual block; generating an optimized graph and an optimized transform by combining the graphs, wherein the graphs are combined on the basis of an optimization step; and performing a transform for the residual block on the basis of the optimized graph and the optimized transform.
    Type: Application
    Filed: September 29, 2016
    Publication date: October 4, 2018
    Inventors: Hilmi E. EGILMEZ, Sehoon YEA, Antonio ORTEGA, Onur G. GULERYUZ, Jana EHMANN
  • Publication number: 20180288409
    Abstract: An image decoding method according to the present invention comprises the steps of: deriving quantized transform coefficients by unit of sub-blocks in a transform unit on the basis of residual information included in a bitstream; deriving transform coefficients on the basis of the quantized transform coefficients; generating a residual sample on the basis of the transform coefficients; generating a prediction sample on the basis of an inter prediction or an intra prediction; and restoring an image on the basis of the residual sample and the prediction sample. According to the present invention, a quantity of data required for a residual signal can be reduced, and an overall coding efficiency can be improved.
    Type: Application
    Filed: September 30, 2016
    Publication date: October 4, 2018
    Inventors: Jin Heo, Eunyong Son, Jaeho Lee, Seungwook PARK, Sehoon Yea, Junghak Nam
  • Publication number: 20180270485
    Abstract: An image encoding method according to the present invention, which is performed by an encoding apparatus, comprises the steps of: generating an adaptive motion vector range (AMVR) enable flag; deriving a motion vector difference (MVD) with respect to a current block; deriving a representative MVD value of an MVD range including a value of the MVD from among a plurality of MVD ranges; generating a coded MVD corresponding to the representative MVD value; and outputting the AMVR enable flag and the coded MVD through a bit stream. According to the present invention, the amount of bits allocated to an MVD can be reduced and the overall coding efficiency can thus be increased.
    Type: Application
    Filed: March 3, 2016
    Publication date: September 20, 2018
    Inventors: Hyeongmoon JANG, Jaehyun LIM, Seungwook PARK, Sehoon YEA
  • Patent number: 10080030
    Abstract: The present invention relates to a video signal processing method and device capable of: obtaining a reference view block by using a predetermined motion vector; obtaining the depth value of a reference block which corresponds to the reference view block; obtaining an inter-view motion vector for a current block by using at least one depth value of the reference depth block; and decoding the current block by using the inter-view motion vector.
    Type: Grant
    Filed: April 11, 2014
    Date of Patent: September 18, 2018
    Assignee: LG ELECTRONICS INC.
    Inventors: Jin Heo, Taesup Kim, Jiwook Jung, Sehoon Yea