DIGITALINSIGHTS INC.

A method and an apparatus are disclosed for video coding using mixed cross-component prediction. In the disclosed embodiments, a video decoding device generates a linear model (LM) prediction block of a current chroma block by performing cross-component prediction according to an LM mode, and generates a non-LM prediction block of the current chroma block by performing prediction according to a non-LM mode. The video decoding device generates a final chroma prediction block of the current chroma block by combining the LM prediction block and the non-LM prediction block based on weights.

A method for video coding uses various block partitioning structures. A method performed by a video decoding apparatus includes decoding, from a bitstream, a width and a height of the current block, and a split flag indicating whether the current block is to be partitioned. The method also includes checking for a value of the split flag. If the split flag is false, the method further includes checking whether the aspect ratio satisfies a preset first condition and whether the width and the height satisfy a preset second condition. If the aspect ratio satisfies the first condition and if the width and the height satisfy the second condition, the method further includes decoding from the bitstream a sequential split flag indicating an application of a sequential quad-tree structure. If the sequential split flag is true, the method further includes partitioning the current block into lower blocks by using the sequential quad-tree structure.

In a method and a device for encoding or decoding an image according to the present invention, motion information for bidirectional prediction of a current block may be derived on the basis of an inter mode previously defined in the device for encoding or decoding an image, and inter prediction may be performed on the current block on the basis of the motion information, wherein the motion information for bidirectional prediction is adjusted to be motion information for unidirectional prediction according to the predefined inter mode.

A method and an apparatus are disclosed for video coding utilizing prediction-transform skipping. In the disclosed embodiments, a video decoding device decodes a prediction skipping flag and a quantized transform block of the current block. The video decoding device inversely quantizes the quantized transform block to generate an inversely quantized transform block of the current block. When the prediction skipping flag is true, the video decoding device inversely transforms the inversely quantized transform block to generate a reconstructed block and stores the reconstructed block to filter the current picture and predict a next block.

A method and an apparatus for video coding using template matching-based intra prediction are disclosed. The video coding method and the video coding apparatus select an intra predictor by defining previously reconstructed pixels spatially adjacent to the current block as a template and performing a template matching-based search in the previously reconstructed region neighboring the current block.

A method and an apparatus for video coding using a matrix-based cross-component prediction are disclosed. The video coding method and apparatus predict a chroma component of a current block, by using a deep learning-based matrix operation, from a chroma component spatially adjacent to a chroma block of the current block and from a luma component spatially adjacent to a luma block corresponding to the chroma block.

A video decoding method performed by a video decoding apparatus includes decoding, from a bitstream, information associated with a neural-network-based filter, the information including a ratio between a size of an input picture to the neural-network-based filter and a size of an output picture from the neural-network-based filter. The method further includes generating a reconstructed picture by decoding the bitstream, and inputting the reconstructed picture into the neural-network-based filter to output a picture having a resolution different from that of the reconstructed picture.

Disclosed herein are an image decoding method, an image decoding apparatus, an image encoding method, an image encoding apparatus and a computer-readable recording medium. An image decoding method according to the present invention may include: decoding information on a prediction method of a current block; based on the information on the prediction method, determining a prediction method of the current block; and generating a prediction block of the current block by performing prediction for the current block based on the determined prediction method. Herein, the information on the prediction method of the current block may include at least one of information on subblock merge, information on MMVD, information on multi hypothesis prediction, and information on partition prediction.

In a method and a device for encoding or decoding an image according to the present invention, motion information for bidirectional prediction of a current block may be derived on the basis of an inter mode previously defined in the device for encoding or decoding an image, and inter prediction may be performed on the current block on the basis of the motion information, wherein the motion information for bidirectional prediction is adjusted to be motion information for unidirectional prediction according to the predefined inter mode.

An image encoding/decoding method and apparatus may reconstruct a current picture based on at least one of intra prediction and inter prediction, specify a block boundary to which a deblocking filter is applied in the reconstructed current picture, and apply a deblocking filter to a block boundary based on a filter type pre-defined in an encoding apparatus.