# US 6,731,810 B1 · Method and apparatus for coding moving image and medium for recording program of coding moving image

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

- Number: US 6,731,810 B1
- Document ID: US6731810B1
- Status: Granted
- Kind: B1
- Application number: 09/461,872
- Filed: 1999-12-15
- Granted: 2004-05-04
- Earliest claimed: 1998-12-24
- Assignees: Hudson Soft Co., Ltd.
- Inventors: Takashi Miura; Fumihiko Itagaki; Miyuki Kawashima
- Web page: https://patentgrep.com/patent/US6731810B1

## CPC

- G06T 9/008: {Vector quantisation}
- H04N 19/649: {the transform being applied to non rectangular image segments}
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION

## Abstract

A method and an apparatus for coding a moving image, capable of obtaining high image quality and high rate of data compression, are provided. The method compares a pixel block of current image to be coded and a pixel block of preceding image within a predetermined region of preceding frame, one after another. The method seeks a specific preceding-image pixel block minimizing matching error. If the matching error relative to the specific preceding-image pixel block exceeds an acceptable value, than the method obtains one or more of orthogonal base systems for approximating AC component vector in the current-image pixel block by inter-frame adaptive orthogonal transformation having, as nest thereof, preceding image data within a predetermined region including the specific preceding-image pixel block. The method thereby codes the moving image data.

## Claims

### Claim 1

A method for coding a moving image, comprising the steps of:
comparing a pixel block of current image to be coded and a pixel bock of preceding image within a predetermined region of preceding frame, one after another;
seeking a specific preceding pixel block minimizing matching error;
if said matching error relative to said specific preceding pixel block exceeds an acceptable value, obtaining one or more of orthogonal base systems for approximating AC component vector in said pixel block of current image by inter-frame adaptive orthogonal transformation having, as nest thereof, preceding image data within a predetermined region including said specific preceding pixel block; and
thereby coding said moving image.

### Claim 2 (depends on 1)

The method for coding a moving image defined in claim 1, wherein a first base for approximating AC component vector in said pixel block of current image is produced based on an AC component vector of said specific preceding pixel block.

### Claim 3 (depends on 2)

The method for coding a moving image defined in claim 2, wherein:
assuming that said one or more orthogonal base systems for approximating AC component vector of said pixel block of current image is represented by first linear combination:
α1<V′1>+α2<V′2>+ . . . +αnk<V′nk>
of respective normalized orthogonal bases <V′q> based on AC component vectors <Uq> of nest pixel block Uq, nk in number, including said specific preceding pixel block;
said first linear combination is transformed to second linear combination:
β1<U1>+β2<U2>+ . . . +βnk<Unk>,
utilizing AC-component vectors <Uq>, the second linear combination being equivalent to said first linear combination, and
said number of said bases nk, scalar development coefficients βq(q=1 to nk), and co-ordinate (x,y) of the nest pixel block as well as sub-sampling intervals (sx,sy) at least related to AC-component vectors <Uq>(q=2 to nk) are coded.

### Claim 4 (depends on 3)

The method for coding a moving image defined in claim 3, wherein:
said pixel block of current image itself is coded if said number of bases nk exceeds a predetermined value.

### Claim 5 (depends on 1)

The method for coding a moving image defined in claim 1, wherein:
said pixel block of current image and specific preceding pixel block corresponding thereto are either divided into subordinate pixel blocks of the same dimension; and
each of said subordinate pixel blocks is subjected to inter-frame adaptive orthogonal transformation.

### Claim 6 (depends on 1)

The method for coding moving image defined in claim 1, wherein:
said pixel block of current image is coded based on movement vector indicating a specific preceding-image pixel block if matching error relative to said specific preceding pixel block does not exceed an acceptable value.

### Claim 7

An apparatus for coding a moving image, which comprises:
a first memory storing data of a pixel block of current image;
a second memory storing data of image in preceding frame containing a pixel block;
a moving vector-calculating unit for calculating a moving vector indicating a specific pixel block in said preceding frame that minimizes matching error by comparing sequentially said pixel block of current image to be coded and preceding pixel block within a predetermined region of said preceding frame; and
an inter-frame adaptive orthogonal transformation coding unit for calculating one or more orthogonal base systems for approximating AC-component vector of said pixel block of current image by inter-frame adaptive orthogonal transformation having as nest a preceding image data within a predetermined region including said specific preceding pixel block and coding said orthogonal base systems, if matching error relative to said specific preceding pixel block exceeds an acceptable value.

### Claim 8 (depends on 7)

The apparatus for coding a moving image defined in claim 7 wherein:
said pixel block of current image is coded based on a moving vector indicating the specific ii preceding pixel block, by means of said inter-frame adaptive orthogonal transformation coding unit, if the matching error relative to the specific preceding pixel block does not exceed an acceptable value.

### Claim 9 (depends on 7)

The apparatus for coding a moving image defined in claim 7, which further comprises:
a decoding unit for decoding/storing sequentially said pixel block of current image based on coding output of said inter-frame adaptive orthogonal transformation coding unit and on said image data of preceding frame in said second memory.

### Claim 10 (depends on 1)

A recording medium capable of reading out by a computer, wherein a program for executing a process defined in claim 1 in said computer is recorded.

### Claim 11 (depends on 2)

The method for coding a moving image defined in claim 2, wherein:
said current pixel block of current image and specific preceding pixel block corresponding thereto are either divided into subordinate pixel blocks of the same dimension; and
each of said subordinate pixel blocks is subjected to inter-frame adaptive orthogonal transformation.

### Claim 12 (depends on 3)

The method for coding a moving image defined in claim 3, wherein:
said pixel block of current image and specific preceding pixel block corresponding thereto are either divided into subordinate pixel blocks of the same dimension; and
each of said subordinate pixel blocks is subjected to inter-frame adaptive orthogonal transformation.

### Claim 13 (depends on 4)

The method for coding a moving image defined in claim 4, wherein:
said pixel block of current image and specific preceding pixel block corresponding thereto are either divided into subordinate pixel blocks of the same dimension; and
each of said subordinate pixel blocks is subjected to inter-frame adaptive orthogonal transformation.

### Claim 14 (depends on 8)

The apparatus for coding a moving image defined in claim 8, which further comprises:
a decoding unit for decoding/storing sequentially said pixel block of current image based on coding output of said inter-frame adaptive orthogonal transformation coding unit and on said image data of preceding frame in said second memory.

### Claim 15 (depends on 6)

A recording medium capable of reading out by a computer, wherein a program for executing a process defined in claim 6 in said computer is recorded.
