Matrix Compression Arrangements

• US 20080273588A1
• Filed: 06/05/2006
• Published: 11/06/2008
• Est. Priority Date: 06/03/2005
• Status: Active Grant

• Current Assignee: Commonwealth of Australia
• Original Assignee: Commonwealth of Australia
• Inventors: Wang, Jimmy Xiaoji
• Granted Patent: US 8,358,858 B2
• Time in Patent Office: Days
• US Class Current: 375/240
• CPC Class Codes: H03M 7/30 Compression speech analysis-synthesis for redundancy reduction G10L19/00; for image communication H04N; Expansion; Suppression of unnecessary data, e.g. redundancy reduction

Abstract

This invention relates to utilising image compression and reconstruction algorithms for transmission and reception of data that can be represented as a numerical matrix or a sequence of numerical matrices in bandwidth limited channels. The method and algorithms described herein may reduce data size to such an extent that data that was considered too large to be practically transmitted in a useful time frame can be transmitted in real or near real time. The algorithms can be applied to data communications for a variety of sensor, electronic and information systems.

Claims

1. A method of compression of digital data representative of a first numerical matrix, said representation containing scalable data, the method consisting of the steps of:

a) creating a representation of a scale associated with said first numerical matrix, and

b) compressing said digital data representative of said first numerical matrix and associating a representation of said scale with said compressed data.

2. A method of compressing digital data representative of a first numerical matrix according to claim 1 wherein creating a representation of a scale includes determining the maximum and minimum values of said digital data representative of said first numerical matrix and creating a representation of those values which are the values themselves or a ratio of those values.

3. A method according to claim 1 wherein the method consists of the further step of:

predetermining the quality factor for the compression so as to allow the transmission of data through a predetermined data transmission bandwidth and predetermining the quality of the data reconstructed with the predetermined quality factor.

4. A method according to claim 1 wherein said first numerical matrix consists of a plurality of data points the method consists of the further step of:

determining the quality of the reconstructed data measured by the mean error and standard deviation in a set of data points of the difference of an data point from the first numerical matrix compared with a reconstructed data point.

5. A method according to claim 1 consisting of the further step of:

determining the largest quality factor for compression of the first numerical matrix based on the largest data transmission bandwidth available for transmission of data representing the compressed first numerical matrix.

6. A method according to claim 1 wherein said first numerical matrix consists of a plurality of data points the method consisting of the further step of:

determining the largest or smallest quality factor for compression of the first numerical matrix based on predetermined upper bounds of mean error and standard deviation of the value of a data point in a reconstructed first numerical matrix.

7. A method according to claim 6 wherein the method consists of the further step of:

determining the size of the compressed data; and

determining the bandwidth of the data transmission required for transmission of data representing the compressed first numerical matrix.

8. A method according to claim 1, wherein JPEG2000 is used to compress digital data.

9. A method according to claim 1, wherein data to be compressed consists of a sequence of sub-data sets, where each sub-data set can be represented as a numerical matrix;

each sub-data set is transmitted to a remote system or user in real-time or near real-time.

10. A method of reconstructing digital data representative of a first numerical matrix containing scalable data, the method consisting of the steps of:

a) receiving compressed data representative of said first numerical matrix;

b) receiving a representation of a scale associated with said first numerical matrix representation,

c) decompressing said data representative of said first numerical matrix; and

d) mapping said decompressed data with said representation of said scale to reconstruct data representative of said first numerical matrix.

11. A method according to claim 10 includes the further step of:

applying said values themselves or the ratio of those values to each decompressed data value to provide data representative of a version of said first numerical matrix wherein N is a matrix with real number entries, and x and y are the maximum and minimum entries of N respectively which represent the first numerical matrix, A is a matrix of the same size that represents the compressed first numerical matrix file of N and a and b are the maximum and minimum entries of A respectively, the matrix P, the reconstruction of matrix N from the compressed numerical matrix A, of the same size can is constructed so that its i-th row and j-th column entry are;

$P<br><br>\left(i,j\right)=\frac{\left(A<br><br>\left(i,j\right)-b\right)<br><br>\left(x-y\right)}{a-b}+y.$

such that P has the property of max{P(i,j), all i and j}=x, min{P(i, j), all i and j}=y to create a numerical matrix substantially the same as the original.

12. A method according to claim 11 wherein the digital data representative of a first numerical matrix consist of successive matrices of a sensor or electronic system that create data sets representative of characteristics detected by the sensors or electronic systems.

13. A method according to claim 10 consisting of the further step of:

determining the quality of the reconstructed data measured by the mean error and standard deviation in a set of data points in the first numerical matrix of the difference of an original data point compared with a reconstructed data point.