Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1995-11-24
2001-09-04
Rao, Andy (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C348S424200
Reexamination Certificate
active
06285714
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to an image signal interpolating apparatus, and more particularly to an image signal interpolating apparatus preferably applied to improve the resolution of an image by interpolating picture elements which had been thinned out by subsampling.
2. Description of Related Art
Conventionally, as the method for band compression or information reduction when image signals are recorded or transmitted, methods of thinning out picture elements of an original image at a certain interval have been widely used. As an example of such method, multiple sub-nyquist sampling encoding has been known.
As an example of subsampling, offset subsampling has been well known. In the two dimensional offset subsampling of shown in
FIG. 6
, sampling interval (Tx, Ty) in the horizontal direction (x-direction) and vertical direction (y-direction) is set to be double a picture element interval (Hx, Hy) of each original signal, and subsampling (×) proceeds alternately. In the case of offset subsampling, sampling points (◯) which are positioned adjacent vertically are offset each other by a half sampling interval (TX/2). As the result, the space frequency component in horizontal and vertical directions of the transmission band of the image signal after offset subsampling is widened to the space frequency in slant direction as shown in
FIG. 7
, and consequently thin-off processing is possible without remarkable visual picture deterioration.
When the offset subsampled image signal is displayed on a monitor or printed out, as shown in
FIG. 8
, picture elements between each sampling point must be interpolated with adjacent picture elements. Such interpolation processing allows frequency component in the slant area shown in
FIG. 7
to pass, and prevent the frequency component in area which contains loopback point A from passing and functions as a space filter, thus this interpolation processing is recognized as a back-end filter in the sampling theory.
Offset subsampling is very effective method when a front-end filter is correctly used, but when a front-end filter is not sufficient because of, for example, constraint due to hardware or a front-end filter is not sufficient because widening of the transmission band, the deterioration of image due to aliasing noise is a problem.
As a method to reduce the aliasing noise, adaptive interpolation has been proposed. In this method, when subsampled image signal is interpolated, closely correlated direction is detected around a picture element to be interpolated, and depending on the detected result a plurality of different interpolation means are selectively used for interpolation.
In the adaptive interpolation, the interpolation accuracy depends seriously on the accuracy of direction detection of closely correlated direction and performance of interpolation means. Therefore, when the performance of individual interpolation means is insufficient for proper interpolation, or closely correlated direction is erroneously judged, not only the original signal component decreases but also aliasing noise increases. These disadvantages are a serious problem of this adaptive interpolation.
SUMMARY OF THE INVENTION
The present invention is accomplished to solve these problems, and the present invention provides an image signal interpolation apparatus having a simple structure for determining an interpolation picture element value near the real value without aliasing noise.
To solve these problems, an image signal interpolating apparatus (
1
) of the present invention for receiving thinned-out transmission image data (D
1
) and interpolating the thinned-out picture element comprises; flatness detecting means (
3
) for detecting the flatness near an interpolation-addressed picture element using picture elements (a to d) near the interpolation-addressed picture element (× symbol in FIG.
2
), picture element selecting means (
4
) for selecting a surrounding picture elements (a to d, or a to l) from surrounding picture elements (a to l) depending on the resultant detected flatness (D
4
) obtained by the flatness detecting means (
3
), classifying means (
5
and
6
) for classifying the interpolation-addressed picture element depending on the level distribution pattern of the surrounding picture elements (a to d or a to l) selected by the picture element selecting means (
4
), predictive coefficient generating means (
7
) for generating a predictive coefficients (D
8
) corresponding to the class classified by the classifying means (
5
and
6
), interpolation data calculating means (
8
) for calculating interpolation data (D
2
) corresponding to the interpolation-addressed picture element by predictive operation using the predictive coefficient (D
8
) and transmission image data (D
1
).
In the present invention, the picture element selecting means (
4
) selects the first number of surrounding picture elements (a to d) when the resultant detected flatness (D
4
) suggests the small flatness, and selects the second number (more than the first number) of surrounding picture element (a to l) when the resultant detected flatness (D
4
) suggests the large. flatness, while, the classifying means comprising data compressing means compresses the selected surrounding picture elements (a to d) at the first compression ratio when the resultant detected flatness (D
4
) suggests a small flatness, and compresses the selected surrounding picture elements (a to l) at the second compression ratio (larger than the first compression ratio) when the resultant detected flatness (D
4
) suggests a large flatness.
The interpolation data (D
2
) are obtained using predictive coefficients (D
8
) which correspond to the class so as to obtain near the real interpolation data (D
2
) without aliasing noise. In addition, the detection of flatness near an interpolation-addressed picture element and selection of picture elements to be used for classification depending on the detection result (D
4
) favors the correct classification of an interpolation-addressed picture element with the least number of class. This mechanism allows the structure of predictive coefficient generating means to be simple.
REFERENCES:
patent: 4858013 (1989-08-01), Matsuda
patent: 5032910 (1991-07-01), Cok
patent: 5131057 (1992-07-01), Walowit et al.
patent: 5184218 (1993-02-01), Gerdes
patent: 5363213 (1994-11-01), Coward et al.
patent: 5517245 (1996-05-01), Kondo et al.
Kawaguchi Kunio
Kondo Tetsujiro
Frommer William S.
Frommer Lawrence & Haug LLP.
Rao Andy
Sony Corporation
LandOfFree
Image signal interpolating apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image signal interpolating apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image signal interpolating apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2470094