WO2003039132A1 - Lissage de gradation - Google Patents

Lissage de gradation Download PDF

Info

Publication number
WO2003039132A1
WO2003039132A1 PCT/IL2001/001010 IL0101010W WO03039132A1 WO 2003039132 A1 WO2003039132 A1 WO 2003039132A1 IL 0101010 W IL0101010 W IL 0101010W WO 03039132 A1 WO03039132 A1 WO 03039132A1
Authority
WO
WIPO (PCT)
Prior art keywords
vignette
gray level
line
along
image
Prior art date
Application number
PCT/IL2001/001010
Other languages
English (en)
Inventor
Gideon Amir
Original Assignee
Indigo N.V.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Indigo N.V. filed Critical Indigo N.V.
Publication of WO2003039132A1 publication Critical patent/WO2003039132A1/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/40—Picture signal circuits

Definitions

  • the invention relates to the field of printing especially digital printing.
  • the invention relates to methods and apparatus of reducing banding in vignettes, caused by limited quantization gray levels.
  • BACKGROUND OF THE INVENTION Bands caused by limited quantization gray levels, is a well-known phenomenon in digital printing. It generally occurs when printing an image with slowly varying portions. In general, the number of gray levels per pixel is limited in order to reduce the amount of memory and data handling. Thus, for many images, an 8 bit per pixel representation is adequate. However, in slowly varying portions, the transition between gray levels may make the borders between regions that have only one gray level difference in a given color, obviously visible to a viewer.
  • Figs. 1 and 2 are graphs of a string of data along a line 10, wherein a vignette having a changing slope and two image features are present.
  • the scale of the Figs, is different (Fig. 1 is, in fact, drawn as an analog curve), so that the banding is not apparent in Fig. 1.
  • the vertical scale is expanded to clearly show the banding, which manifests itself as flat portions 12 along the line, separated by changes in pixel values 14 of a single gray level value. Note that lines are used to connect the gray level values only in flat areas. To the extent that the pixel values represent a true gradation in the data, the banding is probably caused by the finite gray level digitization of the data.
  • Fig. 3 A illustrates a method of printing, in which banding caused by limited number of gray levels is not ameliorated.
  • Fig. 3B One method, illustrated in Fig. 3B, of overcoming this problem is to use a larger number of gray levels. In most cases, printing 10 bits/pixel would provide a sufficient number of gray levels to reduce the problem to an extent that it would no longer be apparent. However, since most computers operate based on 8 bit words, this generally results in using 16 bits/pixel and thus twice as much memory and processing than is actually required. Thus, it is seen that there is a severe conflict between the requirement for smooth images and a minimum amount of data. Algorithms exist for reducing banding. In some of these algorithms, a low bit digitized image is analyzed for the presence of vignettes by operating on the image with a kernel and then adding information (i.e., additional gray levels) where required to make the image smoother.
  • Such algorithms utilize substantial computational and memory resources and they are generally not optimized for hardware implementation. Furthermore, by their nature, such algorithms have a sensitivity that depends on the slope of the gradient of the density change, so that they work differently with low or high gradients. In general, in order to detect very low gradients, very large kernels are required, which results in very high consumption of computing resources.
  • An aspect of the present invention is concerned with a new, simple method for determining the presence and slope of a vignette.
  • the presence of a vignette is determined in an image, prior to rendering the image for printing, whenever two steps of one gray level in one direction are separated by a constant gray level value.
  • a vignette with a slope of 1/m is considered to be present.
  • This methodology can be very efficient since only two numbers and a sign need be stored during the determination. These are the value of the present pixel, the number of pixels since the last step of 1 and whether the step was positive or negative. Thus, not only is the computation simple, but the amount of memory is minuscule. As indicated below, depending on the method used to suppress banding and/or whether separate suppression is provided in two directions, either a single line or the entire image may be stored.
  • An aspect of some embodiments of the invention is concerned with methods of suppression of banding once the vignette is detected and its slope determined.
  • the gray level values are adjusted to reduce or cancel the visual banding effect.
  • This requires the insertion of additional gray level values between the steps of one gray level. For example, if the gray levels of the original image were specified with m-bit resolution, if an extra bit is used in the adjusted image, one extra gray level may be inserted between each of the original gray levels. If two extra bits are used in specifying the adjusted image, 3 extra gray levels may be inserted between each gray level. Round off errors may exist and may result in there being residual "flat" areas in the image, which can potentially cause banding. However, since they are both smaller than the original bands and are at a higher gray level resolution, they may not be troublesome.
  • the slope of the vignette is assumed to remain constant, after it is dete ⁇ nined, until a new determination is made.
  • the gray level values after the gray level step at the end of the portion are adjusted, assuming the determined slope. If the slope does, in fact, not change, the adjustment is "exact", except for round off errors. On the other hand, if the slope increases or decreases after the step, the gray level correction will not be exact in this sense. In any event, the gray level adjustment optionally terminates at the earlier of the adjustment being equal to one gray level or the presence of a new change in the gray level. Utilizing this method, it is not necessary to save any part of the image, unless it is desired to apply the banding suppression algorithm in the transverse direction as well.
  • the gray levels are adjusted on the portion over which the slope was determined. This results in a more exact correction (in the sense given above). However, it does require saving of the values along the line, so that a posteriori correction of previous data can be made. However, if correction is made based on an a priori assumption that the slope is constant, only the pixel value and the number of pixels since the last change of gray level in the original image needs to be made.
  • An aspect of some embodiments of the invention is concerned with suppression of banding in two directions, generally in orthogonal directions.
  • Vignettes can occur in any direction. Generally, the minimum slope of a vignette need not occur along the scan lines (data axes), but may occur in any direction. If the direction of minimum slope of a vignette makes a substantial angle (and especially if this angle is nearly 90 degrees), the correction described above may not sufficiently suppress the banding. In some embodiments of the invention, the above process is repeated in a second direction, preferably, in a direction orthogonal to the direction of scan used in the methods as described above (hereinafter the "first scan direction.”).
  • the adjusted or original values in the first scan direction of a line are stored.
  • a new line (the line succeeding the first line) is analyzed and processed in the first direction, every pixel that enters the correction process, is analyzed in the second direction.
  • a detection and adjustment of the gray level values is performed in a manner similar to that described above for the first direction.
  • the distance since the last gray level step of 1 (or less if adjusted values are used) and the direction of the step as well as the previously computed slope that is corrected in the second direction (if applicable) are stored.
  • the previously processed portions of the image or the entire image is saved as it is being adjusted. This allows for a posteriori correction in both directions, using the methodology of correction described above for the first scan direction. It should be understood that although the a posteriori correction gives, in principle, better results, it may not be necessary in the second direction since, generally, the banding has been substantially reduced by applying any of the processes described above in the first scan direction.
  • a posteriori correction is used in one direction and a priori correction is used in the other direction.
  • An aspect of some embodiments of the invention is concerned with the determination of both the presence of and slope of a vignette using a single image measurement.
  • both the presence of and slope of a vignette are determined from the same measurement.
  • a method of transforming an image from a first lower gray-level resolution to a second higher gray level resolution comprising:
  • (a) and (b) comprise (d) scanning the image along the line.
  • the method includes repeating (a) to (c) for a plurality of lines in the given direction.
  • the plurality of lines covers substantially the entire image.
  • adjusting gray level values comprises interpolating gray level values along the length of the determined slow vignette at the higher gray level resolution.
  • adjusting gray level values comprises correcting gray level values along the line at positions following the position at which the presence of the slow vignette was determined, assuming that the slope of the vignette is not changed.
  • the method includes:
  • the determination of an adjusted gray level value is responsive to the vignette along the line and along the vignette in the different direction.
  • the vignette is determined by the slower of the vignette along the line and in the different direction.
  • the vignette is determined based on an average or weighted average of the vignette along the line and in the different direction.
  • (e) and (f) comprise (h) scanning the image along a line in the different direction.
  • the image scanned in (e) is at the lower gray level resolution.
  • the image scanned in (e) is at the higher resolution.
  • the method includes repeating (e) to (g) for a plurality of lines in the different direction.
  • the plurality of lines in the different direction covers substantially the entire image.
  • the lower resolution is 8 bits.
  • the higher resolution is 10 bits.
  • the higher resolution is greater than 10 bits.
  • determining the length or slope of the vignette comprises determining the number of pixels having a constant gray level along the scan line prior to a change in gray level of 1 at the lower resolution, or less. In an embodiment of the invention, determining the length or slope of the vignette comprises determining the number of pixels having a constant gray level value along the scan line, between two changes in gray level of 1 gray level at the lower resolution or less of a same sign.
  • the method includes half-toning the transformed imaged.
  • the method includes printing the image.
  • a method of page composition comprising: providing a plurality of image elements; composing a page using the elements; and transforming at least some of the elements in accordance with any of the methods of the invention for ameliorating the effects of banding caused by gray level quantization.
  • the method includes half-toning at least a portion of the page.
  • the half-toned portion includes at least some of the transformed elements.
  • the method includes printing the composed page.
  • Fig. 1 is a graph of a string of data along a line, wherein a vignette having a changing order and two image features are present;
  • Fig. 2 is an expanded version of the graph of Fig. 1, which shows the vignette more clearly and the generation of bands;
  • Fig. 3A schematically illustrates a prior art method of printing in which the effects of banding are not reduced
  • Fig. 3B schematically illustrates a prior art method of printing in which more digitization (gray) levels are used, reducing the effect of the banding
  • Fig. 4 schematically illustrates a method of printing having reduced banding in accordance with an embodiment of the invention
  • Fig. 5 A is a flow chart of a method of determining the slope of vignettes and reducing banding, in accordance with an embodiment of the invention
  • Fig. 5B is a flow chart of a method of reducing banding, in accordance with a first embodiment of the invention.
  • Fig. 5C is a flow chart of a method of reducing banding, in accordance with a second embodiment of the invention
  • Fig. 6 is a table of original pixel data values and values that are calculated and printed, in accordance with the flow chart of Fig. 5B;
  • Fig. 7 is a graph similar in format to Fig. 2, showing the original gray level and calculated values of Fig. 6. DESCRIPTION OF EXEMPLARY EMBODIMENTS
  • Fig. 4 schematically illustrates a method 100 of printing having reduced banding in accordance with an embodiment of the invention.
  • an image is first formed as a low gray level resolution image, for example an 8 bit image (102) for example, from a postscript file, using a postscript interpreter.
  • Other methods for forming the low gray- level resolution image may also be employed depending on the nature of the original data or for other reasons.
  • the page for printing is composed (104), using the low gray-level resolution image.
  • Banding suppression, as described below is performed (106) to provide an image that has a higher resolution, for example 10 bits.
  • the entire image is provided as a 10-bit image. In others, only portions for which banding suppression is necessary are so provided.
  • 16 bits of memory are used to store gray level values of each pixel of the image.
  • banding suppression provides an image, at least portions of which have a gray level resolution that is higher than 10 bits.
  • the image is then optionally half-toned (108) and printed.
  • Fig. 5A is flow chart of a method 150 of determining the slope of a vignette and suppressing banding caused by slow vignettes.
  • the image is scanned, optionally line by line, to determine the presence of slow vignettes that can cause banding.
  • a slow vignette for the purpose of the present invention, is determined to exist if two gray level changes of one gray level each, in the same direction are separated by more than one pixel, with no intermediate changes being present.
  • a search is made (as the line is scanned) for single gray level steps in the gray level value.
  • the gray level value, n after the step is saved together with the direction of the step (156).
  • a count, m since the previous single step is kept (157) and the presence of the first gray level value other than n, namely n', is noted (158).
  • a determination is made as to whether the difference between n and n' is a single gray level (162) and, if it is, whether the step is in the same direction as stored (164). If the step is not equal to 1, the method returns to act 154. Otherwise, the search resumes at act 156. If both are positive, the slope of the vignette is determined (166) as the reciprocal of the count, i.e., 1/m. It is not always necessary to determine the slope, it is generally sufficient to determine m itself.
  • Correction is then applied (200) to reduce or remove possible banding, using methods such as those described below with respect to Fig. 5B or Fig. 5C.
  • the search for a next single gray level step resumes.
  • the gray level values are adjusted to reduce or cancel the visual banding effect.
  • this is performed by the insertion of additional gray level values between the steps of one gray level. For example, if the gray levels of the original image were specified with m-bit resolution, if an extra bit is used in the adjusted image, one extra gray level may be inserted between each of the original gray levels.
  • Fig. 5B illustrates a first embodiment 200' of a method for dealing with potential banding.
  • the slope of the vignette is assumed to remain constant, after it is determined, until a new determination is made. It is therefore termed herein, an a priori correction.
  • the gray level values after the gray level step at the end of the portion are adjusted, assuming the determined slope. This involves changing the gray level of pixels subsequent to the m-th pixel (i.e., the second step in gray level values) by an amount proportional to the slope (202).
  • the adjustment is "exact", except for round off errors. If the slope increases or decreases after the step, the gray level correction will not be exact in this sense. In any event, the gray level adjustment optionally terminates at the earlier of the adjustment being equal to one gray level or the presence of a new change in the gray level. Utilizing this method, it is not necessary to save any part of the image, unless it is desired to apply the banding suppression algorithm in the transverse direction as well.
  • Fig. 5C shows an alternative methodology (200") for reducing banding.
  • at least one line of the image (the line being scanned) is saved.
  • the gray levels between the prior single gray level step and the present gray level step are corrected (250) by adding (or subtracting) a value of m"q/m to each of the pixels, where m" is the count m for a particular pixel. It is therefore termed herein, an a posteriori correction.
  • the search for a new vignette optionally continues in parallel with acts 200' or 200".
  • the values of the pixels between the steps are changed to provide as smooth a transition as possible, with as few remaining (smaller) vignettes as possible.
  • the new gray level pixel values (as determined using one of these methods) are then digitized in accordance with the increased resolution gray level values as defined by the increased number of bits used to define the image gray levels values after the correction. It is understood that for higher order vignettes (and for some very short vignettes) there may remain some low order vignette bands.
  • Vignettes can occur in any direction. Generally, the minimum slope of a vignette need not occur along the scan lines (data axes), but may occur in any direction. If the direction of minimum slope of a vignette makes a substantial angle (and especially if this angle is nearly 90 degrees), the corrections described above may not sufficiently suppress the banding.
  • the above process is repeated in a second direction, preferably, in a direction orthogonal to the direction of scan used in the methods as described above (hereinafter the "first scan direction.”).
  • the adjusted or original values in the first scan direction of a line are stored.
  • every pixel that enters the correction process is analyzed in the second direction.
  • a detection and adjustment of the gray level values is performed in a manner similar to that described above for the first direction.
  • the distance since the last gray level step of 1 (or less if adjusted values are used) and the direction of the step as well as the previously computed slope that is corrected in the second direction are stored.
  • previous portions of the image are saved as it is being adjusted. This allows for a posteriori correction in both directions, using the methodology of correction described above for the first scan direction. It should be understood that although the a posteriori correction gives, in principle, better results, it may not be necessary in the second direction since, generally, the banding has been substantially reduced by applying any of the processes described above in the first scan direction.
  • two methods can be used in the determination of the slope in the second direction. These methods are generally the same as those shown in Fig 5A, with the following differences.
  • the corrections can then be a priori or a posteriori, according to the methodologies described above.
  • the data used for second direction correction is first direction corrected data, at the higher resolution.
  • the analysis looks for a first step that is higher than 1 least significant bit (LSB) at the corrected gray level resolution and less than or equal to 1 LSB at the original resolution. Then a second step is found that is less than or equal to 1 LSB at the original resolution. Corrections between the two steps (a posteriori) or following the second step (a priori) is then made in the manner described above.
  • LSB least significant bit
  • the original data is analyzed in both the first and second directions.
  • the slope in both directions is determined.
  • the corrected value (a priori or a posteriori) is then based on the lower slope.
  • the correction can use an algorithm that takes into account the slope in both directions, for example using the average value of the correction from the two directions or a weighted average of the values.
  • Figs. 1 and 2 illustrate, in addition to the vignettes, a number of larger step values of gray-level, representative of thin lines.
  • a step returns to the base level and the vignette becomes apparent (by the presence of a number of pixels having the same gray level)
  • the process described above with respect to Figs. 4 and 5 begins again.
  • the large step is considered to be the beginning of the vignette portion (even though the step it is larger than 1) and the first single gray-level step is considered to be the end of the portion.
  • Slow vignette correction (either a priori of a posteriori) is performed as described above.
  • Fig. 6 is a table of values of input values (at the input to act 202 of method 200) and calculated (re-digitized) values (at the output of method of Fig. 5B). These values are plotted in Fig. 7. Note that in Fig. 7 (in contrast to Fig. 2), lines are used to connect all gray level values.
  • Fig. 7 in contrast to Fig. 2, lines are used to connect all gray level values.
  • the correction method need not require the use of the determination method described herein.
  • the a priori and a posteriori correction methods can be carried out in any direction, once the slope of the vignette is found by any method.
  • other methods of vignette correction can be used after the method described above for determination of the vignette has be used to find the vignette.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Image Processing (AREA)

Abstract

L'invention concerne un procédé de transformation d'une image qui passe d'une première résolution à échelle de gris inférieure à une seconde résolution à échelle de gris supérieure. Ce procédé consiste à: (a) déterminer la présence d'une vignette lente le long d'une ligne; (b) déterminer la longueur ou la pente de la vignette le long de la ligne; et enfin, (c) régler à la résolution à échelle de gris la plus élevée les valeurs de l'échelle de gris le long d'un segment de la ligne, en fonction de la longueur ou la pente déterminée de la vignette.
PCT/IL2001/001010 2001-10-08 2001-10-31 Lissage de gradation WO2003039132A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
IL0100934 2001-10-08
ILPCT/IL01/00934 2001-10-08

Publications (1)

Publication Number Publication Date
WO2003039132A1 true WO2003039132A1 (fr) 2003-05-08

Family

ID=11043101

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IL2001/001010 WO2003039132A1 (fr) 2001-10-08 2001-10-31 Lissage de gradation

Country Status (1)

Country Link
WO (1) WO2003039132A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101542580B (zh) * 2006-11-29 2012-05-09 夏普株式会社 液晶显示装置、液晶显示装置驱动方法、液晶显示装置源极驱动器、以及液晶显示装置控制器

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4511992A (en) * 1981-05-08 1985-04-16 Organisme Autonome Dote de la Personnalite Civile Agence France Presse System for reconstituting, by filtering, an analog signal from a pseudo-analog signal
US5408338A (en) * 1992-02-19 1995-04-18 Ricoh Company, Ltd. Image processing unit processing pixel data according to scan line density
EP1039416A2 (fr) * 1999-02-05 2000-09-27 Scitex Corporation Ltd. Une méthode pour la réduction des effets de troncature dans des images numériques
EP1096425A2 (fr) * 1999-10-28 2001-05-02 Hewlett-Packard Company, A Delaware Corporation Méthode et appareil de filtrage d'images ayant une précision de tonalité insuffisante

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4511992A (en) * 1981-05-08 1985-04-16 Organisme Autonome Dote de la Personnalite Civile Agence France Presse System for reconstituting, by filtering, an analog signal from a pseudo-analog signal
US5408338A (en) * 1992-02-19 1995-04-18 Ricoh Company, Ltd. Image processing unit processing pixel data according to scan line density
EP1039416A2 (fr) * 1999-02-05 2000-09-27 Scitex Corporation Ltd. Une méthode pour la réduction des effets de troncature dans des images numériques
EP1096425A2 (fr) * 1999-10-28 2001-05-02 Hewlett-Packard Company, A Delaware Corporation Méthode et appareil de filtrage d'images ayant une précision de tonalité insuffisante

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101542580B (zh) * 2006-11-29 2012-05-09 夏普株式会社 液晶显示装置、液晶显示装置驱动方法、液晶显示装置源极驱动器、以及液晶显示装置控制器

Similar Documents

Publication Publication Date Title
JP3188330B2 (ja) ２値画像ビットマップの予備補償方法
US7643180B2 (en) Image processing and display apparatus and method for tone resolution expansion of digital images
JP4100782B2 (ja) 誤差拡散値の処理方法
US5307426A (en) Image processing apparatus with improved dithering scheme
US20050285974A1 (en) Apparatus and method of smoothing video signal using pattern adaptive filtering
JPH05225332A (ja) 空間的可変濾波の方法及び装置
JPH05308515A (ja) 画像中のピクセル値を量子化する方法
US20090041369A1 (en) Image enhancement employing partial template matching
US5434931A (en) System and method for picture image processing employing transformation of picture data
JPH05219377A (ja) 画像内画素値量子化方法
JP3171993B2 (ja) 画像処理方法及び装置
US10497101B2 (en) Image processing apparatus and image processing method
US10388021B2 (en) Image processing apparatus, image forming apparatus and program
CN103793892B (zh) 图像处理装置和图像处理方法
WO2003039132A1 (fr) Lissage de gradation
US7298928B2 (en) Method for converting image from low resolution into high resolution
US6757079B1 (en) Method and apparatus for forming multi-level dither images from an input digital image
JP2006254432A (ja) デジタル画像データを処理するデジタル画像データ処理方法
JP2941288B2 (ja) 画像処理システム
JP2860039B2 (ja) 擬似中間調画像縮小装置
JP3222585B2 (ja) 中間調二値データ発生回路
JPH06187436A (ja) 画像データ処理装置
JPH10304196A (ja) 濃淡画像情報の二値化装置および二値化方法
JPH06225124A (ja) 階調画像の二値化方法および装置
JP2978363B2 (ja) 中間調画像推定装置

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ OM PH PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
122 Ep: pct application non-entry in european phase
NENP Non-entry into the national phase

Ref country code: JP