Image coder-decoder using a matrix transform with weighted contribution of several points of the image to the formation of one point of the transform

• US 4,055,756 A
• Filed: 09/30/1976
• Issued: 10/25/1977
• Est. Priority Date: 02/03/1975
• Status: Expired due to Term

• Current Assignee: Societe Anonyme de Telecommunications (Safran SA)
• Original Assignee: Societe Anonyme de Telecommunications (Safran SA)
• Inventors: Stouls, Francois-Xavier Antoine, Jolivet, Jean-Claude
• Primary Examiner(s): Malzahn, David H.
• Application Number:
  US05/728,189
• Time in Patent Office: 390 Days
• Field of Search: 235/156, 235/152, 235/164
• US Class Current: 708/400
• CPC Class Codes: G06T 1/0007 Image acquisition
• CPC Class Codes: H04N 19/86 involving reduction of coding artifacts, e.g. of blockiness

Abstract

A matrix transform system for coding and decoding images. The system comprises means for sampling the lines and the columns of an image and forming with these samples a sample square matrix and storing the same. This sample square matrix is split into component square matrices of order 3N × 3N having a central part of order N × N. Each of these component matrices is multiplied by a first rectangular coefficient matrix of order N × 3N having a square central part in which the coefficients are equal to unity and two square lateral parts in which the coefficients are selectively equal to zero and a predetermined factor smaller than unity, which gives an intermediate matrix of order N × 3N. This intermediate matrix is multiplied by a second rectangular coefficient matrix of order 3N × N which is the transpose of the first rectangular coefficient matrix, which gives an output matrix of order N × N which is the transform of the central part of the component matrix. If N = 2, the output matrix is of order 2 × 2; one of its terms depends on four samples of the image, two of its terms depend on 12 samples of the image and the last of its terms depends on 36 samples of the image.

Claims

1. A matrix transform system for coding images with weighted contribution of several points of the image to each points of the coded image and for decoding coded images with weighted contribution of several points of the coded image to each point of the decoded image, said system comprising:

a. means for sampling the lines of an image to be coded and forming with said image samples a square matrix associated with the image;

b. means for splitting said square matrix associated with the image into a plurality of first input component matrices of order 3N ×

3N having a central part of order N ×

N ;

c. means for multiplying each of said first input component matrices by a first rectangular coefficient matrix of order N ×

3N having a square central part in which the coefficients are equal to positive or negative unity and two square lateral parts in which the coefficients are selectively equal to zero and ±

α

where α

is a predetermined factor smaller than unity, and forming first intermediate matrices of order N ×

3

d. means for multiplying each of said first intermediate matrices by a second rectangular coefficient matrix of order 3N ×

N which is the transpose of said first rectangular coefficient matrix, and forming first output matrices of order N ×

N, each of said first output matrices being the transform of the central part of a first input component matrix;

e. means for forming with said first output matrices a square matrix associated with the coded image;

f. means for splitting said square matrix associated with the coded image into a plurality of second input component matrices of order 3N ×

3N having a central part of order N ×

N;

g. means for multiplying each of said second input component matrices by a third rectangular coefficient matrix of order N ×

3N having a square central part in which the coefficients are equal to positive or negative unity and two square lateral parts in which the coefficients are selectively equal to zero and (a) ±

α

where α

is said predetermined factor smaller than unity and forming second intermediate matrices of order N ×

3N;

h. means for multiplying each of said second intermediate matrices by a fourth rectangular coefficient matrix of order 3N ×

N which is the transpose of said third rectangular coefficient matrix and forming second output matrices of order N ×

N, each of said second output matrices being the transform of the central part of a second input component matrix; and

i. means for forming with said second output matrices a square matrix associated with the decoded image.

2. A matrix transform system for coding and decoding images according to claim 1 in which N is a power of two and the central part of the first, second, third and fourth rectangular coefficient matrices is a Hadamard matrix.

3. A matrix transform system for coding and decoding images according to claim 1 in which N is a power of two and the central part of the first, second, third and fourth rectangular coefficient matrices is a Haar matrix.

4. A matrix transform system for coding images according to claim 1, in which N = 2, said means for multiplying each of the first input component matrices by a first rectangular coefficient matrix of order 2 ×

6 having a square central part in which the coefficients are equal to positive or negative unity and two square lateral parts in which the coefficients are selectively equal to zero and ±

α

where α

is a predetermined factor smaller than unity comprises;

a. means for splitting the samples of each line of a first input component matrix into a central group of two samples and two lateral groups of two samples;

b. a first adder and subtractor circuit for forming the sum and the difference of the samples of the central sample group;

c. a second adder and subtractor circuit for forming the sums of the samples of the lateral sample groups and forming the difference of said sample sums;

d. a first multiplier and adder circuit for multiplying said difference of said sample sums by α

to obtain a first multiplication result and for forming the sum of said first multiplication result and of said difference of the samples of the central sample group;

the signals provided by said first adder circuit and by said first multiplier and adder circuit forming the coefficients of first intermediate matrices and said means for multiplying each of the first intermediate matrices by a second rectangular coefficient matrix of order 6 ×

2 having a square central part in which the coefficients are equal to positive or negative unity and two square lateral parts in which the coefficients are selectively equal to zero and ±

α

comprises;

e. means for splitting the coefficients or each line of a first intermediate matrix into a central group of two coefficients and two lateral groups of two coefficients;

f. a third adder and subtractor circuit for forming the sum and the difference of the intermediate coefficients of the central intermediate coefficient group;

g. a fourth adder and subtractor circuit for forming the sums of the coefficients of the lateral coefficient groups and forming the difference of said coefficient sums;

h. a second multiplier and adder circuit for multiplying said difference of said coefficient sums by α

to obtain a second multiplication result and for forming the sum of said second multiplication result and of said difference of the intermediate coefficients of the central intermediate coefficient group;

the signals provided by said third adder circuit and by said second multiplier and adder circuit forming the coded image samples.

5. A matrix transform system for decoding coded images according to claim 1, in which N = 2, the means for multiplying each of the second input component matrices by a third rectangular coefficient matrix or order 2 ×

6 having a square central part in which the coefficients are equal to positive or negative unity and two square lateral parts in which the coefficients are selectively equal to zero and ±

α

where α

is a predetermined factor smaller than unity comprises;

a. means for splitting the samples of each line of a second input component matrix into a central group of two samples and two lateral groups of two samples;

b. a first subtractor circuit for forming the difference of one sample of a lateral group and of one sample of the other lateral group;

c. a first multiplier and summing circuit for multiplying said difference of samples of lateral groups by α

to obtain first multiplication result and for forming algebric summation of said first multiplication result and of both said samples of the central group;

the signals provided by said first multiplier and summing circuit forming the coefficients of a second intermediate matrix and said means for multiplying each of the second intermediate matrices by a fourth rectangular coefficient matrix of order 6 ×

2 having a square central part in which the coefficients are equal to positive or negative unity and two square lateral parts in which the coefficients are selectively equal to zero and ±

α

comprises;

d. means for splitting the coefficients of each line of a second intermediate matrix into a central group of two coefficients and two lateral groups of two coefficients;

e. a second subtractor circuit for forming the difference of one coefficient of a lateral coefficient group and of one coefficient of the other lateral coefficient group;

f. a second multiplier and summing circuit for multiplying said difference of coefficients or lateral coefficient groups by α

to obtain second multiplication result and for forming algebric summation of said second multiplication result and of both said coefficients of the central coefficient group;

the signals provided by said second multiplier and summing circuit forming the decoded image samples.