Hyundai Motor Company

Ewha University-Industry Collaboration Foundation

Kia Corporation

Kia Motors Corporation

Korea Electronics Technology Institute

Je Won Kang

Wha Pyeong Lim

Seung Wook Park

Sang Hyo Park

Na Young Kim

Jung Kyung Lee

Yong Hwan Kim

Ho Jin Jung

Hyun Soo Ahn

Young Bin Min

Ji Eon Kim

Chae Hwa Yoo

Min Joo Kang

Eun Yeo

A video decoding method, a video encoding method, and a method of transmitting a bitstream containing encoded video data are disclosed. The methods include determining first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generating a first reference block from the first reference picture referenced by the first motion vector and generating a second reference block from the second reference picture referenced by the second motion vector; and generating a prediction block of the target block using the first and second reference blocks. The methods also include generating a residual block of the target block and reconstructing the target block based on the prediction block of the target block and the residual block of the target block.

Method and apparatus for encoding and decoding video using inter-prediction

A video coding method and an apparatus for refining predicted signals in intra prediction are disclosed. The video coding method and apparatus generate refined prediction signals approximating original video signals from predicted signals of intra prediction using a variable and fixed coefficient-based deep learning model to reduce the amount of data for residual signals, which are to be encoded.

Method and apparatus for video coding for improving predicted signals of intra prediction

A video encoding method, video decoding method, and an apparatus are disclosed. The methods and the apparatus are configured to: determine first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generate a first reference block from the first reference picture referenced by the first motion vector and generate a second reference block from the second reference picture referenced by the second motion vector; and generate a prediction block of the target block using the first and second reference blocks. The predictor includes a first coding tool configured to generate the prediction block of the target block by performing a bi-directional optical flow process using the first and second reference blocks.

A method and an apparatus for video coding using a deep learning-based in-loop filter for inter-prediction are disclosed. The video coding method and the apparatus utilize a deep learning-based in-loop filter for inter-prediction of a predictive frame (P-frame) and a bipredictive frame (B-frame) in order to mitigate various levels of image distortion according to a QP (quantization parameter) value present in the P-frame and the B-frame.

Method and apparatus for video coding using deep learning based in-loop filter for inter prediction

A video decoding apparatus, a video encoding apparatus, and an apparatus for transmitting a bitstream containing encoded video data are disclosed. The apparatuses are configured to: determine first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generate a first reference block from the first reference picture referenced by the first motion vector and generate a second reference block from the second reference picture referenced by the second motion vector; generate a prediction block of the target block using the first and second reference blocks; generate a residual block based on the prediction block of the target block; and encode the residual block. The apparatuses are configured to generate the prediction block of the target block by performing a bi-directional optical flow process using the first and second reference blocks.

A video decoding method, a video encoding method, and a method for transmitting a bitstream containing encoded video data are disclosed. The methods are configured to: determine first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generate a first reference block from the first reference picture referenced by the first motion vector and generate a second reference block from the second reference picture referenced by the second motion vector; and generate a prediction block of the target block using the first and second reference blocks. The methods are also configured to generate the prediction block for a subblock of size 4×4 split from the target block by performing a bi-directional optical flow process using the first and second reference blocks.

A video encoding or decoding apparatus and method perform existing inter prediction on a current block to generate a motion vector and predicted samples. The video encoding or decoding apparatus and method generate enhanced predicted samples for the current block using a deep learning-based video prediction network (VPN) on the basis of the motion vector, the reference samples, the predicted samples, and the like to improve encoding efficiency.

Video encoding and decoding using deep learning based inter prediction

Disclosed is a video encoding and decoding technology using a differential modulation technique for a residual signal in order to enhance compression performance in a transform skip mode. The differential modulation technique is applicable to both luma and chroma components as long as the chroma format is not monochrome (i.e. as long as chroma_format_idc=0 is not true).

Video encoding and decoding using differential modulation

A frame packing method in an MPEG immersive video format is provided. A frame packing method in association with an immersive-video encoding and decoding method efficiently arranges, on a single screen, texture information and depth information of a basic view and an additional view to increase the coding efficiency.

Method for frame packing in a MPEG immersive video format

Disclosed is a video encoding/decoding method for performing inter-prediction on a current picture by resampling and referring to luma signals and chroma signals of a reference picture to improve encoding efficiency in video encoding and decoding for a frame having luma signals and chroma signals in various sample formats in one video sequence.

Video encoding and decoding based on resampling chroma signals

A video decoding method is configured to reconstruct a current block that is encoded using an intra prediction. The method includes decoding, from a bitstream, direction information indicating a direction of differential coding applied to a residual block of the current block, generating the residual block from residual signals by reconstructing and using the information on the residual signals from the bitstream and using the direction information, inversely quantizing the modified residual block, generating a prediction block for the current block through intra prediction, and reconstructing the current block by adding the prediction block and the inverse quantized residual block.

Image decoding device using differential coding

An inter prediction method allows a variable coefficient deep learning model to adaptively learn characteristics of a video; transmits a variable coefficient deep learning model parameter generated from the learning from an image encoding device to an image decoding device; and refers to a virtual reference frame generated by the variable coefficient deep learning model.

Inter prediction method based on variable coefficient deep learning

Disclosed are a method and an apparatus for efficiently applying a transform skip mode to a transform block. According to an embodiment of the present invention, a video decoding apparatus determines, based on transform skip information associated with a current block, the transform skip information being decoded from a bitstream, whether a subblock for which transform is skipped and one or more subblocks to which transform is applied are included in the current block. When determining that the current block includes the subblock for which the transform is skipped and the one or more subblocks to which the transform is applied are included in the current block, the video decoding apparatus decodes a first residual subblock using a decoding procedure not involving inverse transform for the subblock for which the transform is skipped, and decodes one or more second residual subblocks using a decoding procedure involving the inverse transform for the one or more subblocks to which the transform is applied. The video decoding apparatus adds the first residual subblock and the one or more second residual subblocks to a prediction block to reconstruct the current block. Representative drawing: FIGS.A-D

Method and device for efficiently applying transform skip mode to data block

Disclosed is a method for video encoding and decoding using adaptive color transform on a residual signal. In particular, the maximum memory size required for adaptive color transform is limited by: controlling whether to allow color space conversion on a sequence level according to a syntax element indicating the maximum conversion size allowed for a sequence of pictures; or controlling whether to apply color space conversion on a block level according to the size of a coding block.

Video encoding and decoding using adaptive color transform

A video decoding apparatus predicts a target block in a current picture to be decoded. The apparatus comprises a predictor configured to: determine first and second reference pictures and first and second motion vectors for bi-prediction by decoding a bitstream; generate a first reference block from the first reference picture referenced by the first motion vector and generate a second reference block from the second reference picture referenced by the second motion vector; and generate a prediction block of the target block using the first and second reference blocks. The predictor includes a first coding tool configured to generate the prediction block of the target block by performing a bi-directional optical flow process using the first and second reference blocks.

Growing community of inventors

Seoul, South Korea

Je Won Kang