# US 2021/0344919 A1 · Method and apparatus for performing low complexity computation in transform kernel for video compression

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0344919 A1
- Document ID: US20210344919A1
- Status: Published application
- Kind: A1
- Application number: 17/343,274
- Filed: 2021-06-09
- Published: 2021-11-04
- Earliest claimed: 2017-06-28
- Assignees: LG Lelctronics Inc.
- Applicants: LG Lelctronics Inc.
- Inventors: Bumshik LEE; Moonmo KOO; Jaehyun LIM
- Web page: https://patentgrep.com/patent/US20210344919A1

## CPC

- H04N 19/12: Selection from among a plurality of transforms or standards, e.g. selection between discrete cosine transform [DCT] and sub-band transform or selection between H.263 and H.264
- H04N 19/147: according to rate distortion criteria (rate-distortion as a criterion for motion estimation H04N19/567)
- H04N 19/159: Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
- H04N 19/176: the region being a block, e.g. a macroblock
- H04N 19/61: in combination with predictive coding
- H04N 19/625: using discrete cosine transform [DCT]
- H04N 19/70: characterised by syntax aspects related to video coding, e.g. related to compression standards
- H04N 19/593: involving spatial prediction techniques
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

The present disclosure relates to a method for processing a video signal by using a linear relationship between transform types, comprising the steps of: determining a transform group on the basis of a prediction mode of a current block; parsing a transform combination index from a video signal, wherein the transform combination index indicates anyone of a plurality of transform combinations in the transform group; deriving a transform combination corresponding to the transform combination index, wherein the transform combination consists of a horizontal transform and a vertical transform, and includes DST-7 or DCT-8; performing inverse transform on the current block on the basis of the transform combination; and restoring the video signal by using the inverse transformed current block.

## Claims

### Claim at position 1

1-12. (canceled)

### Claim 13

A method for decoding a video signal, comprising:
obtaining a transform combination index specifying one of a plurality of transform combinations based on a prediction mode of a current block;
deriving a transform combination corresponding to the transform combination index, wherein the transform combination is composed of a horizontal transform and a vertical transform, and the horizontal transform and the vertical transform include DST-7 or DCT-8;
performing an inverse-transform on coefficients of the current block based on the transform combination to generate an inverse-transformed block; and
generating a reconstructed block of the current block based on the inverse-transformed block.

### Claim 14 (depends on 13)

The method of claim 13, wherein, if the horizontal transform or the vertical transform is DST-7, the performing the inverse transform comprises:
applying Winograd FFT algorithm when an input N is a prime number, and applying Prime Factor FFT algorithm when the input N is not a prime number.

### Claim 15 (depends on 13)

The method of claim 13, wherein, if the horizontal transform or the vertical transform is DCT-8, the performing the inverse transform comprises:
performing multiplications of N-times on an input having length N;
performing a DST-6 operation of (2N+1) length by using Winograd FFT; and
outputting DST-8 inverse-transform kernel by performing addition operation of (N−1) times.

### Claim 16 (depends on 13)

The method of claim 13, wherein, if the vertical transform is DST-7 or DCT-8, the performing the inverse transform comprises:
applying the DST-7 or DCT-8 inverse-transform on each of columns; and
identifying whether the horizontal transform is DST-7 or DCT-8.

### Claim 17 (depends on 16)

The method of claim 16, wherein, if the horizontal transform is DST-7 or DCT-8 according to the identification, the performing the inverse transform comprises applying the DST-7 or DCT-8 inverse-transform on each of rows.

### Claim 18 (depends on 13)

The method of claim 13, wherein the transform combination index is transmitted with a transform unit.

### Claim 19

A method for encoding a video signal, comprising:
generating a prediction block of a current block;
generating a residual block of the current block based on the prediction block;
applying a plurality of transform combinations to the residual block based on a prediction mode applied to the current block when generating the prediction block;
determining a transform combination to be applied to the residual block among the plurality of transform combinations, wherein the transform combination comprises a horizontal transform and a vertical transform and the horizontal transform and the vertical transform include DST-7 or DCT-8;
performing a transform on the residual block based on the horizontal transform and the vertical transform of the transform combination; and
generating a transform combination index specifying the transform combination.

### Claim 20 (depends on 19)

The method of claim 19, wherein, if the horizontal transform or the vertical transform is DST-7, the performing the transform comprises:
applying Winograd FFT algorithm when an input N is a prime number, and applying Prime Factor FFT algorithm when the input N is not a prime number.

### Claim 21 (depends on 19)

The method of claim 19, wherein, if the horizontal transform or the vertical transform is DCT-8, the performing the transform comprises:
performing multiplications of N-times on an input having length N;
performing a DST-6 operation of (2N+1) length by using Winograd FFT; and
outputting DST-8 transform kernel by performing addition operation of (N−1) times.

### Claim 22 (depends on 19)

The method of claim 19, wherein, if the vertical transform is DST-7 or DCT-8, the performing the transform comprises:
applying the DST-7 or DCT-8 transform on each of columns; and
identifying whether the horizontal transform is DST-7 or DCT-8.

### Claim 23 (depends on 22)

The method of claim 22, wherein, if the horizontal transform is DST-7 or DCT-8 according to the identification, the performing the transform comprises applying the DST-7 or DCT-8 transform on each of rows.

### Claim 24 (depends on 19)

The method of claim 19, wherein the transform combination index is generated in a transform unit.

### Claim 25 (depends on 19)

A non-transitory storage medium for storing a bitstream generated according to the method of claim 19.
