Sk Telecom Company, Ltd.

Sungwon Lim

Byeungwoo Jeon

Yunglyul Lee

Joohee Moon

Haekwang Kim

Jeongyeon Lim

Jongki Han

Juock Lee

Joong Gunn Park

Mincheol Park

A video encoding apparatus includes: a predictor to generate a predicted block of a current block corresponding to a predicted motion vector and a differential motion vector; and an entrophy encoder to encode an index for use in identifying the predicted motion vector of the current block and the differential motion vector in a bitstream, where the predicted motion vector of the current block corresponds to a motion vector identified by the index in an inter prediction mode, and where the index corresponds to at least one of representative blocks determined using neighboring blocks of the current block, in consideration of priorities of block positions.

Method and apparatus for predictive encoding/decoding of motion vector

A video encoding method includes: when an inter prediction mode corresponds to a merge mode, checking for whether predetermined candidate adding conditions are satisfied; adding into a merge candidate set including merge candidates corresponding to neighboring blocks of a current block, a combined bi-predictive motion parameter, the combined bi-predictive motion parameter generated by combining a list 0 motion parameter of a first available candidate block and a list 1 motion parameter of a second available candidate block, when the predetermined candidate adding conditions are satisfied; and predicting the current block in the merge mode by using a motion parameter corresponding to a merge candidate selected from the merge candidate set, to generate a predicted block for the current block, wherein the merge candidate set corresponds to a maximum number of the merge candidates, the maximum number of the merge candidates corresponding to information on the number of the merge candidates.

Method and apparatus for video encoding/decoding using improved merge

A video encoding method for reducing encoding error by using a loop filter is provided, the method including: generating a reconstructed image from encoded data of an original image; determining one or more regions on which the loop filter is applied in the reconstructed image; transforming the original image and the reconstructed image which correspond to each region of the determined one or more regions from a spatial domain into a frequency domain; determining loop filter coefficients for said each region by comparing the transformed original image and the transformed reconstructed image; loop-filtering the reconstructed image based on the loop filter coefficients for said each region; and encoding information on the loop filter coefficients for said each region.

Method and apparatus for encoding/decoding image by using adaptive loop filter on frequency domain using conversion

A video decoding apparatus for predicting a current block to be decoded, includes: a bitstream decoder to extract an index for use in identifying a predicted motion vector of the current block and a differential motion vector from a bitstream; and a predictor to generate a predicted block of the current block based on the predicted motion vector and the extracted differential motion vector, wherein the predicted motion vector of the current block is determined to be a motion vector identified by the extracted index in a motion estimation mode of an inter prediction mode, and wherein the extracted index corresponds to one or more representative blocks determined using neighboring blocks of the current block, in consideration of priorities of block positions.

A video encoding apparatus for predicting a motion vector of a current block to be encoded is to determine one or more representative blocks from neighboring blocks of the current block according to priorities of the neighboring blocks; and encode the motion vector of the current block by using, as a predicted motion vector of the current block, a first motion vector selected from motion vectors of the determined one or more representative blocks and a second motion vector of at least one block within a reference picture.

A video decoding method includes: extracting a merge flag and a merge index from a bitstream; when the merge flag indicates a merge mode, checking for whether predetermined candidate adding conditions are satisfied, and adding into a merge candidate set including the merge candidates corresponding to neighboring blocks of a current block, a combined bi-predictive motion parameter, the combined bi-predictive motion parameter generated by combining a listmotion parameter of a first available candidate block and a listmotion parameter of a second available candidate block, when the predetermined candidate adding conditions are satisfied; and generating a predicted block of the current block by using a motion parameter corresponding to a merge candidate selected from the merge candidate set, the merge candidate selected by the extracted merge index.

A video encoding apparatus includes an inter predictor to check for availability of merge candidates configured from neighboring blocks of a current block, add into a merge candidate set, as one of the merge candidates, a combined bi-predictive motion parameter with respect to an unavailable block, the combined bi-predictive motion parameter generated by combining a listmotion parameter of a first available andidate block and a listmotion parameter of a second available candidate block, when the unavailable block is checked to be present among the merge candidates, and predict the current block by using a motion parameter corresponding to a merge candidate selected from the merge candidates including the combined bi-predictive motion parameter, to thereby generate a predicted block for the current block.

A video encoding apparatus includes an inter predictor to check for whether predetermined candidate adding conditions are satisfied, add into a merge candidate set including merge candidates corresponding to neighboring blocks of a current block, a combined bi-predictive motion parameter, the combined bi-predictive motion parameter generated by combining a listmotion parameter of a first available candidate block and a listmotion parameter of a second available candidate block, when the predetermined candidate adding conditions are satisfied, and predict the current block by using a motion parameter corresponding to a merge candidate selected from the merge candidate set, to thereby generate a predicted block for the current block. Herein, the merge candidate set corresponds to a maximum number of the merge candidates, the maximum number of the merge candidates corresponding to information on the number of the merge candidates included in a bistream.

Growing community of inventors

Seoul, South Korea

Sungwon Lim