Facsimile and static presentation processing – Facsimile – Specific signal processing circuitry
Patent
1988-08-25
1990-02-20
Chin, Tommy P.
Facsimile and static presentation processing
Facsimile
Specific signal processing circuitry
358166, 358340, H04N 5208, H04N 521
Patent
active
049031287
DESCRIPTION:
BRIEF SUMMARY
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to the processing of digitized images and especially to systems designed to reduce the throughput of digital data needed for the storage or transmission of these images.
The encoding of color digital images at a low throughput rate requires a high data reduction ratio: for example, if a throughput rate of 10 M bits per second is sought, for a digital image source delivering data at 140 m bits per second, the data reduction rate is then 14.
2. Description of the Prior Art
A known method uses an encoder/decoder that performs an orthogonal transformation to obtain compression rates of this type. But a compression rate of this type creates faults in the images in certain circumstances.
These faults may be the following:
the boundaries of the blocks become visible in the uniform zones of the image;
visible oscillations occur at sharp, isolated transitions;
faults in restitution of data also become visible in the dark zones of the image.
An object of the invention is a method and corresponding device to reduce these faults through a filtering of blocks of images, adapted to the encoding conditions.
SUMMARY OF THE INVENTION
According to the invention, a method for the reduction of noise generated by an encoding/decoding of digital images, operating by blocks of pixels or points, comprises the processing of the characteristic value of the current point of the decoded blocks;
by determining, from a window centered on the current point of the image, a mean characteristic value Moy,
and then determining a filtering parameter p from a parameter, NR, related to the operating status of the encoder and this mean, Moy;
and by determining a filtered value, y(i,j), for the current point from its decoded characteristic value x(i,j), of the filtering parameter p, and the mean in the window, Moy: y(i,j)=p.x((i,j)+(1-p).Moy.
Another object of the invention is a device for the reduction of noise generated by an encoding/decoding device, operating by blocks of digitized pixels of N lines by M columns having an input (E) connected to the output of the decoder and designed for the implementation of the method wherein:
this input (E) is connected to the input of a line memory with a capacity equal to 2M values, the input of a column memory with a capacity equal to 2 columns of 2(N+2) values and to the input of a working memory with a capacity equal to (N+2) (M+2) values, said memory being also connected to the line and column memories,
the device further comprising
a circuit to compute the mean from values associated with the points of a window stored in the memory;
a circuit to determine the filter parameter, p, coupled firstly to the mean computing output and, secondly, to an input of characteristic data on the operating status of the encoder;
and a filtering circuit connected firstly to the output of the mean computing circuit and, secondly, to the output of the circuit for determining the filtering parameter, the output of said filtering circuit giving the sequence of filtered characteristic values.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will be better understood and other characteristics will appear from the following description, made with reference to the appended drawings, of which:
FIG. 1 shows the variations of the filtering parameter as a function of the mean in the filtering window;
FIG. 2 shows the scanning of the image by blocks;
FIGS. 3 and 4 illustrate the movement of the filtering window in a scanned block, from right to left and left to right respectively;
FIG. 5 is a block diagram of the noise reducing device according to the invention;
FIG. 6 illustrates the sequencing of the handling of the memories 10, 20 and 30 of FIG. 5.
DESCRIPTION OF THE PREFERRED EMBODIMENT
The encoding, after the orthogonal transformation of the blocks of images, amounts, very briefly, to selecting a portion, of variable size, of the spectral components of each transformed block, depending on the number of components resulting from the t
REFERENCES:
patent: 4392123 (1983-07-01), Bruggemann
patent: 4566126 (1986-01-01), Miyagawa
patent: 4729019 (1988-03-01), Rouvrris
patent: 4794454 (1988-12-01), Sugiyama
ICASSP 83 Proceddings on the IEEE International Conference on Acoustics, Speed and Signal Processing, 14-16 Apr. 1983, Boston, MA., IEEE Acoustics, Speed & Signal Processing Society, vol. 3, (US) H. C. Reeve III et al.:"Reduction of Blocking . . . " pp. 1212 & 1213.
IEEE International Conference on Communications, 13-17, Jun. 1982, Philadelphia, Conference Record, vol. 2 of 3, ICC'82, The Digital Revolution, IEEE, (US), R. C. Reininger et al.:"Soft Decision Demodulation and Transform coding of Images", pp. 4H.3.1-4H.3.6.
International Conference on Communications, ICC'80, Conference Record; (8-12 Jun. 1980, vol. 2 of 3, Seattle, WA, US) King Ngi Ngan et al.: "Lowpass Filtering in the Cosine Transform Domain", pp. 31.7.3 Left Hand Column, Line 6, p. 31.7.4 Left Hand Column.
Chin Tommy P.
Thomson Grand Public
LandOfFree
Method and device for the reduction of noise generated by a digi does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for the reduction of noise generated by a digi, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for the reduction of noise generated by a digi will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1620356