Method and apparatus for the analysis and correction of the image gradation in image originals

• US 5,748,802 A
• Filed: 01/23/1995
• Issued: 05/05/1998
• Est. Priority Date: 04/06/1992
• Status: Expired due to Fees

• Current Assignee: Heidelberger Druckmaschinen Ag
• Original Assignee: Linotype-Hell AG (Heidelberger Druckmaschinen Ag)
• Inventors: Winkelman, Kurt-Helfried
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US08/376,377
• Time in Patent Office: 1,198 Days
• Field of Search: 382/169, 382/171, 382/172, 382/270, 382/271, 382/274
• US Class Current: 382/271
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06T 5/40 using histogram techniques
• CPC Class Codes: G06T 5/92 based on global image properties
• CPC Class Codes: H04N 1/4074 using histograms
• CPC Class Codes: H04N 1/6027 Correction or control of colour gradation or colour contrast H04N1/6058 takes precedence

Abstract

A method and apparatus for the analysis and correction of the image gradation of an image original to be reproduced by evaluating image values acquired by point-by-point and line-by-line, optoelectronic scanning with an input device in apparatus and systems for image processing. The image original is geometrically subdivided into a plurality of sub-images. The frequency distribution of the image values or, respectively, of the luminance components of the color values in a corresponding sub-image is separately identified as a sub-image histogram. The sub-image histograms of the individual sub-images are evaluated and the sub-images relevant for the image gradation are identified by means of the evaluation. An aggregate histogram that corresponds to the frequency distribution of the image values or, respectively, of the luminance component of the color values in the relevant sub-images is calculated from the sub-image histograms of the relevant sub-images. Correction values for the correction of the image gradation characteristic of the image original are subsequently calculated from the aggregate histogram according to the method of histogram modification.

Claims

1. A method for analysis and correction of an image gradation characteristic of an image original, comprising the steps of:

generating image values relating to the image original by point-by-point and line-by-line, optoelectronic scanning with an image processing input device;

geometrically subdividing the image original into a plurality of sub-images, each sub-image having corresponding image values;

for each sub-image, separately identifying frequency distributions of the image values in the corresponding sub-images as a sub-image histogram for every sub-image;

evaluating the sub-image histograms of the individual sub-images and identifying sub-images critical for image gradation since they are identified as being high in structure wherein relatively large image value changes occur as compared to other of the sub-images which are lower, in structure wherein relatively smaller image value changes occur;

generating an aggregate histogram comprising a summation of frequency distributions of the image values in the critical sub-images from the sub-image histograms of the critical sub-images; and

generating a correction curve function from the aggregate histogram for correction of the image gradation characteristic of the image original for contrast correction.

2. The method of claim 1 wherein the image values are color values each of which comprises a luminance component, and the frequency distributions are generated for the luminance component.

3. The method of claim 1, wherein the evaluation of the sub-image histograms for the identification of the critical sub-images occurs by use of statistical histogram parameters.

12. The method of claim 1, wherein for each luminance value which is an image value of a black/white image or luminance component of a color image the functionally corresponding frequency distribution values in the individual sub-image histograms of the identified critical sub-images of the image original are added, the added frequency distribution values calculated for each luminance value defining a new histogram in the form of said aggregate histogram.

24. The method of claim 1 wherein said image values used for analysis of the image gradation of the image original are generated by scanning in a pre-scan the image original with a resolution which is coarser than the resolution normally used for scanning in a fine scanning of the image original for reproduction purposes.

25. The method of claim 1, wherein the input device generates image values in a first color space, and wherein the method comprises the additional steps of:

transforming the image values of the first color space into image values of a second color space that is independent of the first color space; and

analyzing the image gradation for calculating setting values based on the transformed image values of the second color space.

4. A method for analysis and correction of an image gradation characteristic of an image original, comprising the steps of:

generating image values relating to the image original by point-by-point and line-by-line, optoelectronic scanning with an image processing input device;

geometrically subdividing the image original into a plurality of sub-images, each sub-image having corresponding image values;

for each sub-image, separately identifying frequency distributions of the image values in the corresponding sub-images as a sub-image histogram for every sub-image;

evaluating the sub-image histograms of the individual sub-images by use of statistical histogram parameters and identifying critical sub-images for image gradation based on the evaluation;

for identifying the sub-images which are critical, calculating a histogram parameter "scatter" (SDev) and a histogram parameter "relative area proportion of the most frequent image values" (FLAnt) for each of the sub-image histograms and utilizing them for the evaluation of the sub-image histograms, the histogram parameter "scatter" (SDev) and the histogram parameter "relative area proportion of the most frequent image values" (FLAnt) being measures of an amount of details present in a given sub-image;

generating an aggregate histogram comprising a summation of frequency distributions of the image values in the critical sub-images from the sub-image histograms of the critical sub-images; and

generating a correction curve function from the aggregate histogram for correction of the image gradation characteristic of the image original for contrast correction.

5. The method of claim 4, wherein the identification of the critical sub-images is undertaken by comparing the histogram parameters "scatter" (SDev) and "relative area proportion of the most frequent image values" (FLAnt) with respective thresholds (SwSDev) and (SWFLAnt) selected for the image original.

6. The method of claim 5, wherein a sub-image is classified as a critical sub-image when the value of its respective histogram parameter "scatter" (SDev) is higher than the threshold (SwSDev) and the value of the histogram parameter "relative area proportion of the most frequent image values" (FLAnt) of the sub-image is lower than the threshold (SwFLAnt).

7. The method of claim 4, wherein the threshold (SwSDev) for the histogram parameter "scatter" (SDev) is calculated from a frequency distribution of the values of the histogram parameter "scatter" (SDev) of all sub-images.

8. The method of claim 4, wherein the histogram parameter "scatter" (SDev) and the histogram parameter "relative area proportion of the most frequent image values" (FLAnt) of the sub-images are calculated by a statistical evaluation of the corresponding sub-image histograms.

9. The method of claim 4, wherein the histogram parameter "scatter" (SDev) for a sub-image is calculated by means of the following steps:

calculating a plurality (N) of image values for the sub-image from the sub-image histogram (Hi(i), according to the equation ##EQU19## whereby (H(i)) corresponds to the number of picture elements having the image value (i) in the sub-image;

calculating a mean of the frequency distribution from the sub-image histogram (Hi(i), i=1 . . . M) according to the equation;

##EQU20## calculating a variance (Var) according to the equation ##EQU21## and identifying the histogram parameter of "scatter" (SDev) according to the equation ##EQU22##

10. The method of claim 4, wherein the histogram parameter "relative area proportion of the most frequent image values" (FLAnt) for a sub-image is calculated according to the following steps:

calculating a plurality N of image values for the sub-image from corresponding sub-image histogram values (Hi(i), i=1, . . . , M) according to the equation ##EQU23## whereby (H(i)) corresponds to the plurality of picture elements having the image value (i) in the sub-image;

resorting histogram values (Hs(i)) of the corresponding sub-image histogram (Hi) in descending sequence of the frequency to form a new frequency distribution of histogram values (Hs(i));

defining a plurality (n) of histogram values (Hs(i)) to be accumulated; and

calculating the histogram parameter "relative area proportion of the most frequency image values" (FLAnt) according to the equation ##EQU24##

11. The method of claim 5, wherein a classification of said identified critical sub-images is undertaken by means of the following steps:

selecting the thresholds (SwSDev) and (SWFLAnt) for the histogram parameters of "scatter" (SDev) and of "relative area proportion of the most frequency image values" (FLAnt);

calculating the histogram parameters of "scatter" (SDev) and of "relative area proportion of the most frequent image values" (FLAnt) for all sub-images; and

evaluating the calculated histogram parameters of "scatter" (SDev) and "relative area proportion of the most frequent image values" (FLAnt) according to a classification pattern.

13. A method for analysis and correction of an image gradation characteristic of an image original, comprising the steps of:

generating image values relating to the image original by point-by-point and line-by-line, optoelectronic scanning with an image processing input device, the image values being color values each of which comprises a luminance component;

geometrically subdividing the image original into a plurality of sub-images, each sub-image having corresponding image values;

for each sub-image, separately identifying frequency distributions for the luminance component of the image values in the corresponding sub-images as a sub-image histogram for every sub-image;

evaluating the sub-image histograms of the individual sub-images and identifying sub-images critical for image gradation since they are identified as being high in structure wherein relatively large image value changes occur as compared to other of the sub-images which are lower in structure wherein relatively small image value changes occur;

generating an aggregate histogram comprising a summation of frequency distributions of the image values in the critical sub-images from the sub-image histograms of the critical sub-images, and wherein for each luminance value which is an image value of a black/white image or luminance component of a color image, functionally corresponding frequency distribution values and the individual sub-image histograms of the identified critical sub-images of the image original are added, the added frequency distribution values calculated for each luminance value defining a new histogram in the form of said aggregate histogram; and

generating a correction curve function G=f(L) from the aggregate histogram for correction of the image gradation characteristic of the image original for contrast correction, the correction curve G=f(L) being calculated according to a histogram modification method by accumulating histogram values (H(i)) of the aggregate histogram of the image-critical sub-images according to the following equation;

##EQU25## and wherein the accumulation is undertaken between a minimum value (LMin) and a value (L) of a luminance value range of the image original.

14. The method of claim 13, wherein the accumulation is undertaken between a minimum value (LMin) and a maximum value (Lmax) of the luminance of the image original.

15. The method of claim 13, wherein the correction curve G=f(L) is smoothed with a low-pass filtering.

16. The method of claim 15, wherein the smoothing of the correction curve G=f(L) is implemented according to a "sliding mean" method, such that each value of the smoothed correction curve G=f(L) is calculated as a weighted sum of neighboring values of the unsmoothed correction curve.

17. The method of claim 16, wherein a degree of correction of said image gradation characteristic is made selectable by means of a correction factor (k) in order to provide for variable contrast correction.

18. The method of claim 17, wherein:

a histogram gradation factor (HG) that corresponds to a maximum degree of correction (100%) is calculated from an image value histogram of the image original utilizing a histogram modification method;

a linear gradation factor (LG) that corresponds to a minimum degree of correction (0%) is generated; and

a correction gradation factor (KG) for the variable contrast correction is formed by adding together portions of the histogram gradation factor (HG) and of the linear gradation factor (LG).

19. The method of claim 18, wherein the formation of the correction gradation factor (KG) occurs according to the following equation, whereby the correction factor (k) lies in the value range from 0.0 through 1.0:

space="preserve" listing-type="equation">KG=k*HG+(1-k)*LG.

20. The method of claim 17, wherein the correction factor (k) is dependent on properties of the image original.

21. The method of claim 18, wherein an RMS actual value (RMS_grd) describing a maximum degree of correction is calculated from the correction curve G=f(L) as the mean quadratic deviation of the histogram gradation (HG) from the linear gradation (LG), according to the following equation:

##EQU26## with;

δ

_i =deviation of a corrected image value (histogram gradation (HG)) from an uncorrected image value i (linear gradation (LG));

N=plurality of deviations (δ

_i).

22. The method of claim 21, comprising the steps of:

calculating statistical histogram parameters of "skewness" (skew) and "Kurtosis" (Kurt) from the aggregate histogram of said critical sub-images;

comparing said calculated statistical histogram parameters to prescribed skewness and Kurtois threshold values, and

deriving an RSM value (Rmsi) as a rated value for contrast correction by evaluating results of said comparison according to a classification pattern.

23. The method of claim 22, wherein a correction factor (k) for a contrast correction is formed as a quotient from said derived RMS rated value (Rmsi) and an actual value (RMS_grad).

26. A method for analysis and correction of image gradation of an image original in apparatus and systems for image processing, comprising the steps of:

geometrically subdividing the image original to be analyzed into a plurality of sub-images;

acquiring image values or color values by point-by-point and line-by-line, optoelectronic scanning of the image original with an input apparatus;

identifying a frequency distribution of the image values or of luminance components of the color values in the corresponding sub-image for every sub-image so as to form a corresponding sub-image histogram;

evaluating the sub-image histograms of the individual sub-images, and identifying the sub-images that are image-critical for the image gradation since they are identified as being high in structure wherein relatively large image value changes occur as compared to other of the sub-images which are lower in structure wherein relatively small image value chances occur;

calculating an aggregate histogram from the sub-image histograms of the image-critical sub-images, said aggregate histogram corresponding to a frequency distribution of the image values or of the luminance components of the color values in the image-critical sub-images; and

employing the calculated aggregate histogram for calculating a correction curve (G=f(L)) for contrast correction of the image original.