JP3965713B2 - Color image binarization method by error diffusion method - Google Patents

Description

【００３２】
なお、前述したごとく、画素の１行目の最初の画素の２値化誤差ｅ（０，０）は、ステップＳ２０４にて「１０」が設定されているので、従来と異なり、Ｉ′は元の画素濃度Ｉよりも「１０」増加されることになる。次にステップＳ２４０にて求められた補正濃度Ｉ′と閾値Ｔとが比較される（Ｓ２６０）。ここで閾値Ｔとしては、例えば「１２８」が設定されている。
【００３３】
Ｉ′＜Ｔであれば（Ｓ２６０で「ＹＥＳ」）、出力濃度Ｏとして「０」（オフ）に設定され（Ｓ２７０）、Ｉ′≧Ｔであれば（Ｓ２６０で「ＮＯ」）、出力濃度Ｏとして「１」（オン）が設定される（Ｓ２８０）。この出力濃度Ｏの値は、出力画像データ記憶部１８にプレーンＣの２値化画像データとして、ステップＳ２７０またはステップＳ２８０の２値化毎に順次蓄積される。
【００３４】
次に、補正濃度Ｉ′と出力濃度Ｏとに基づいて２値化誤差値Ｅを次式２のごとく算出する（Ｓ２９０）。
【００３５】
【数２】

【００３６】
次に、予め設定した誤差分配マトリックスＢｍａｔ（）に基づいて、前記誤差値Ｅを、次式３に示すごとく、２値化が未処理の周辺画素の誤差バッファｅに分配する（Ｓ３００）。
【００３７】
【数３】

【００３８】
なお、「＋＝」は既に誤差バッファｅ内に存在する値と加算処理して同じ誤差バッファｅに格納することを示す演算子である。Ｂｍａｔ（）の具体例は例えば図３に示す通りであり、ｉ，ｊは注目画素位置を「＊」とすると、図３に示すごとくの値をとる変数である。
【００３９】
次に、主走査方向（ｘ方向）の２値化処理、すなわち１ライン分の処理が終了したか否かを判定する（Ｓ３１０）。終了していなければ（Ｓ３１０で「ＮＯ」）、注目画素の主走査方向の位置ｘを１つ増加させて（Ｓ３２０）、再度ステップＳ２３０から処理を繰り返す。
【００４０】
１ライン分の２値化処理を終了したと判定した場合（Ｓ３１０で「ＹＥＳ」）には、１プレーン分の全画素の２値化処理が終了したか否かを判定する（Ｓ３３０）。１プレーンの全画素の処理が終了していなければ（Ｓ３３０で「ＮＯ」）、注目画素の副走査方向の位置ｙを１つ増加させて（Ｓ３４０）、再度ステップＳ２２０から処理を繰り返す。
【００４１】
全画素の２値化が終了していれば（Ｓ３３０で「ＹＥＳ」）、次に全てのプレーンについて処理が終了しているか否かが判定される（Ｓ３５０）。最初のプレーンＣのみが終了しただけであるので（Ｓ３５０で「ＮＯ」）、次にマゼンタＭのプレーン（プレーンＭ）の処理が設定される（Ｓ３６０）。
【００４２】
そして、プレーンＭのための、最初の一行の画素に対する誤差バッファｅ（ｘ，０）を設定する（Ｓ２０４）。ここでは、前回プレーンＣに設定した誤差バッファｅ（ｘ，０）＝１０，１０，１０，１０，０，０，０，０，１０，１０，１０，１０，０，０，０，０，……とは異なるパターンで誤差初期値が設定される。例えば、プレーンＣとは、パターンをずらして、誤差バッファｅ（ｘ，０）＝０，０，０，０，１０，１０，１０，１０，０，０，０，０，１０，１０，１０，１０，……でも良い。あるいはプレーンＣとはパターンが異なれば良いので、誤差バッファｅ（ｘ，０）は、すべて０でも良い。
【００４３】
このように初期誤差バッファｅ（ｘ，ｙ）を設定して、プレーンＭについても、前述したごとくステップＳ２１０〜Ｓ３４０の処理を行なって２値化し、出力画像データ記憶部１８にプレーンＭの２値化画像データとして蓄積する。次に、プレーンＭについて２値化処理が終了すれば（Ｓ３５０で「ＮＯ」）、次にイエローのプレーン（プレーンＹ）が設定されて（Ｓ３６０）、プレーンＹの２値化処理が行われる。このプレーンＹについても、最初の一行の画素に対する誤差バッファｅ（ｘ，０）を設定する（Ｓ２０４）。ここでも、プレーンＣ，Ｍとは、別個のパターンを設定した方が良いが、３原色ＣＭＹの内で、イエローは境界線の発生には影響しにくく、そのドットの配置を考慮しなくても問題が生じにくいので、ＣＭＹのいずれかのパターンと同一でも良い。また、特に値を設定せずに、誤差バッファｅ（ｘ，０）は、すべて０でも良い。
【００４４】
こうして、プレーンＹについても２値化されて、その２値化データが出力画像データ記憶部１８にプレーンＹの２値化画像データとして蓄積されると、全てのプレーンについて２値化処理が終了し（Ｓ３５０で「ＹＥＳ」）、誤差拡散処理は終了する。
【００４５】
このときには、出力画像データ記憶部１８内には、ステップＳ２７０またはステップＳ２８０にて設定された出力濃度Ｏにて各画素が各プレーンＣ，Ｍ，Ｙ毎に２値化された擬似中間調画像データが形成されている。上述したごとく、本実施の形態１では、色成分としての３原色ＣＭＹのプレーン毎に画素の一行目において、主走査方向に異なるパターンで値が変化するように、誤差バッファｅの誤差初期値を設定している。このことにより、少なくとも、シアンとマゼンタとが中間調カラー画像において、同一濃度値であっても同一画素位置にシアンのインクとマゼンタのインクとが置かれる状態が乱されて、従来技術で述べた境界線が目立たなくあるいは全く見えなくなるという効果を生じる。
【００４６】
なお、誤差バッファｅ（ｘ，０）としては、「１０，１０，１０，１０，０，０，０，０，１０，１０，１０，１０，０，０，０，０，……」といような「１０」と「０」とからなる繰り返しのパターン以外に、例えば、「１０」と「−１０」との繰り返し、例えば、誤差バッファｅ（ｘ，０）＝１０，１０，１０，１０，−１０，−１０，−１０，−１０，１０，１０，１０，１０，−１０，−１０，−１０，−１０，……というように、Σｅ（ｘ，０）がほぼ「０」となるように配置したパターンの誤差初期値でも良い。
【００４７】
また、誤差バッファｅ（ｘ，０）の主走査方向のパターンを、各プレーンＣＭＹで変化させるのではなく、単に、画素の１行目の誤差バッファｅ（ｘ，０）の値のみを変更しても良い。例えば、プレーンＣについては、１行目の誤差バッファｅ（ｘ，０）は、すべて「１０」とし、プレーンＭについては、１行目の誤差バッファｅ（ｘ，０）は、すべて「０」とし、プレーンＹについては、１行目の誤差バッファｅ（ｘ，０）は、すべて「５」（「１０」あるいは「０」でも良い。）としても良い。また、プラスの値でなくてもマイナスの値でも良い。
【００４８】
このように、主走査方向の誤差バッファｅ（ｘ，０）のパターンを変えなくとも、値のみでも、従来技術で述べた境界線を目立たなくあるいは全く見えなくすることができる。［実施の形態２］
本実施の形態２が実施の形態１と異なるのは、誤差拡散処理であり、他は同一である。実施の形態２の誤差拡散処理を図４のフローチャートに示す。
【００４９】
処理が開始されると、まず、処理対象の原色プレーンとして、プレーンＣが設定される（Ｓ４０２）。次に、２値化処理する画素の位置を判別するための変数ｘ，ｙを０に初期化する（Ｓ４１０，Ｓ４２０）。
【００５０】
画素位置（ｘ，ｙ）に対応する入力画像の入力濃度Ｉ（０≦Ｉ≦２５５）を読み取る（Ｓ４３０）。次に、この注目画素に対応する２値化誤差値ｅ（ｘ，ｙ）を誤差バッファ１６から読み取り、次式４のごとく２値化誤差値ｅ（ｘ，ｙ）にて画素濃度Ｉを補正して、補正濃度Ｉ′を求める（Ｓ４４０）。
【００５１】
【数４】

【００５２】
なお、本実施の形態２では、１行目の画素について実施の形態１のような特別な誤差初期値の設定は行っていない。全て誤差初期値は「０」である。次に、閾値生成処理（Ｓ４５０）が行われる。この処理を図５のフローチャートに示す。
【００５３】
まず、先に処理されたプレーンの同一位置の画素について「１」（オン）が有るか否かが判定される（Ｓ４５１）。現在、最初のプレーンＣが処理されているとすると、前には２値化処理はなされていないのでオンされていることはなく（Ｓ４５１で「ＮＯ」）、プレーンＣについては、すべて閾値Ｔとして初期閾値Ｔ０が設定される（Ｓ４５２）。ここで、例えば、初期閾値Ｔ０としては、「１２８」が設定されている。
【００５４】
こうして閾値生成処理が終了すると、次にステップＳ４４０にて求められた補正濃度Ｉ′とステップＳ４５０で求められた閾値Ｔとが比較される（Ｓ４６０）。Ｉ′＜Ｔであれば（Ｓ４６０で「ＹＥＳ」）、出力濃度Ｏとして「０」（オフ）に設定され（Ｓ４７０）、Ｉ′≧Ｔであれば（Ｓ４６０で「ＮＯ」）、出力濃度Ｏとして「１」（オン）が設定される（Ｓ４８０）。この出力濃度Ｏの値は、出力画像データ記憶部１８にプレーンＣの２値化画像データとして順次蓄積される。
【００５５】
以下、ステップＳ４９０〜Ｓ５６０の処理は、実施の形態１のテップＳ２９０〜Ｓ３６０と同一の処理が行われる。そして、ステップＳ５６０にてマゼンタＭのプレーン（プレーンＭ）の処理が設定されると、前述したステップＳ４１０〜Ｓ４４０の処理を行い、次にステップＳ４５０の閾値生成処理が行われる。
【００５６】
この閾値生成処理のステップＳ４５１において、先に処理されたプレーンＣについてオンと設定された画素位置の場合には（Ｓ４５１で「ＹＥＳ」）、次に同一画素位置にオンが２つ設定されているか否かが判定される（Ｓ４５３）。今回の処理はプレーンＭの処理であるので、先の処理はプレーンＣのみであることから、ステップＳ４５３では「ＹＥＳ」と判定されることはない。したがって、ステップＳ４５３では「ＮＯ」と判定されて、閾値Ｔとして初期閾値Ｔ０＋αが設定される（Ｓ４５５）。αとしては例えば、「５０」である。
【００５７】
このように、プレーンＭの各画素の２値化処理に用いられる閾値Ｔとしては、プレーンＣにてオフに設定されている画素位置であれば（Ｓ４５１で「ＮＯ」）、Ｔ＝Ｔ０に設定され（Ｓ４５２）、プレーンＣにてオンに設定されている画素位置であれば（Ｓ４５１で「ＹＥＳ」、Ｓ４５３で「ＮＯ」）、Ｔ＝Ｔ０＋αに設定される（Ｓ４５５）。
【００５８】
そして、以下、前述したステップＳ４６０〜Ｓ５６０の処理が行われ、出力画像データ記憶部１８にプレーンＭの２値化画像データを蓄積し、プレーンＹを処理対象に設定する。プレーンＹについても同様に、前述したステップＳ４１０〜Ｓ４４０の処理を行い、次にステップＳ４５０の閾値生成処理が行われる。
【００５９】
この閾値生成処理のステップＳ４５１において、先に処理されたプレーンＣ，Ｍのいずれかについてオンと設定された画素位置の場合には（Ｓ４５１で「ＹＥＳ」）、次に同一画素位置にオンが２つ設定されているか否かが判定される（Ｓ４５３）。プレーンＣ，Ｍで共にオンが設定されていれば（Ｓ４５３で「ＹＥＳ」）、閾値Ｔとして初期閾値Ｔ０＋２αが設定され（Ｓ４５４）、プレーンＣ，Ｍのいずれか一方のみでオンが設定されていれば（Ｓ４５３で「ＮＯ」）、閾値Ｔとして初期閾値Ｔ０＋αが設定される（Ｓ４５５）。
【００６０】
したがって、プレーンＹの各画素の２値化処理に用いられる閾値Ｔとしては、Ｔ０，Ｔ０＋α，Ｔ０＋２αの３種類のいずれかが設定されることになる。なお、プレーンＹについては、実施の形態１でも述べたごとく、境界線発生への影響は少ないので、ステップＳ４５０の処理では、他のプレーンＣ，Ｍの２値化状態にかかわらず全て閾値Ｔ＝Ｔ０に設定しても良い。
【００６１】
そして、以下、前述したステップＳ４６０〜Ｓ５６０の処理が行われ、出力画像データ記憶部１８にプレーンＹの２値化画像データを蓄積し、全ての処理を終了する（Ｓ５５０で「ＹＥＳ」）。本実施の形態２は、前述のごとく、同一画素において、先に２値化処理された原色プレーンについてオンの２値化がなされた場合、この後に行われる同一画素の他の原色プレーンの２値化については、閾値Ｔを高くすることにより、オンの２値化がなされる確率を低下させている。
【００６２】
このように、同一画素において、先に２値化された原色プレーンがオンとなった場合は、すなわち、その原色のインクが置かれることが決定した位置では、他の原色のインクについては、同一画素位置には置かれにくくなる。したがって、インクの置かれる状態を乱すことができ、従来技術で述べた境界線を目立たなくあるいは全く見えなくすることができる。
【００６３】
なお、同一画素において、先に２値化処理された原色プレーンについてオンの２値化がなされなかった場合、すなわち、オフであった場合は、この後に行われる同一画素の他の原色プレーンの２値化については、オンの２値化がなされる確率を維持していたが、更にオンの２値化がなされる確率を上昇させる処理としても良い。例えば、先に２値化された原色プレーンについてオンの２値化がなされなかった場合、この後に行われる同一画素の他の原色プレーンの２値化については、２値化を判断する閾値Ｔを低くすることにより、オンの２値化がなされる確率を上昇させる。このような構成を加えることにより、一層、境界線を目立たなくあるいは全く見えなくすることができる。
【００６４】
また、図５に示した閾値生成処理の代りに、先のプレーンにてオフの処理がなされていた場合に、オンの２値化がなされる確率を上昇させる処理、例えば、閾値Ｔを低くする処理を行っても良い。［その他］前記実施の形態１，２における２値化は、注目画素を２値化した場合に２値化の誤差を未だ２値化していない周辺の画素の濃度に分配する方法による誤差拡散法であったが、２値化時に周辺画素から２値化誤差の分配を受けるタイプの誤差拡散法、いわゆる平均誤差最小法であっても良い。
【００６５】
前記実施の形態１に、平均誤差最小法を適用する場合には、最初に２値化処理される主走査方向の画素行には、それ以前の行に画素がないので予めプレーン毎に異なる値あるいは異なるパターンで誤差値を生じている仮想的画素を想定して、その仮想的画素からの誤差分配を誤差初期値として処理すれば良い。
【００６６】
前記実施の形態１，２では、色成分としては、シアン、マゼンタおよびイエローの３原色を用いたが、これ以外に、ブラックを含めても良い。この場合には、特に、シアン、マゼンタおよびブラックの内の少なくとも１つの色成分が、他の色成分とは異なる誤差初期値の設定がなされることが好ましい。また、このような原色を用いるのではなく、淡色や中間色を色成分としても良い。
【００６７】
また、実施の形態１と実施の形態２との構成を組み合わせても良く、このようにすれば、一層、境界線を目立たなく、あるいは見えなくする効果が高くなる。また、実施の形態２においては、１色目の２値化処理の後、２色目および３色目は共に、オンへの２値化の確率を小さくしたが、２色目および３色目についてはいずれか一方のみ、オンへの確率を小さくするのみでも良い。特に、前述したごとく、イエローは境界線の発生には影響しにくいので、イエローの２値化の場合はオンへの２値化の確率を変更しなくても良い。
【図面の簡単な説明】
【図１】 実施の形態１としての中間調画像データ２値化装置の概略構成を表すブロック図である。
【図２】 実施の形態１における誤差拡散処理のフローチャートである。
【図３】 実施の形態１における誤差分配マトリックスの構成説明図である。
【図４】 実施の形態２における誤差拡散処理のフローチャートである。
【図５】 実施の形態２における閾値生成処理のフローチャートである。
【図６】 従来の２値化処理において生じる現象の説明図である。
【符号の説明】
２…中間調画像データ２値化装置 １２…ＣＰＵ１３…プログラム記憶部 １４…閾値記憶部１５…誤差分配マトリックス記憶部 １６…誤差バッファ１７…入力画像データ記憶部 １８…出力画像データ記憶部１９…作業用メモリ ２０…バス ２１…キーボード２２…ディスプレイ ２３…外部記憶装置 ２４…カラープリンタ[0001]
BACKGROUND OF THE INVENTION
The present invention binarizes a halftone color image by comparing the pixel density with a threshold for a plurality of different color components included in the halftone color image, and errors generated in the binarization are corrected to the surrounding pixels. The present invention relates to a binarization method of a color image by an error diffusion method to be reflected in binarization.
[0002]
[Prior art]
An error diffusion method is known as a process for binarizing a halftone image and converting it into a pseudo halftone image. In this error diffusion method, when a certain pixel is binarized, an error caused by binarization is distributed to the density of surrounding pixels that have not yet been binarized (in a narrow sense, error diffusion method / document: Robert W. Floyd and Louis Steinberg, "An Adaptive Algorithm for Spatial Greyscale", Proceeding of the SID Vol.17 / 2, 1976, etc.) or 2 from the already binarized pixels that exist in the vicinity when binarizing A method of receiving a predetermined percentage of error that occurred during valuation (also called the mean error minimum method. Reference: JFJarvis, CNJudice, and WHNinke, "A Survey of Techniques for the Display of Continuous Tone Pictures on Bilevel Displays", Computer Graphicsand Image Processing.5, 13-40 (1976) etc. are well known. In this specification, unless otherwise specified, the error diffusion method is used in a broad sense.
[0003]
When this method is used for a color image, the color image can be binarized as a pseudo halftone image by performing error diffusion processing for each color component constituting the color image. Therefore, even in a printer that can only place color ink in units of dots, a halftone color image can be printed as a color image approximate to a halftone as a whole.
[0004]
[Problems to be solved by the invention]
However, when the color image is binarized in this way, the following problem may occur. That is, as shown in FIG. 6A, in a halftone color image in which the density level range of each color component is 0 to 255, each pixel is cyan (hereinafter referred to as C) = 100, magenta (hereinafter referred to as M). Each pixel is cyan (hereinafter referred to as C) = 50 and magenta (hereinafter referred to as M) in a uniform color arrangement area A1 of yellow (hereinafter referred to as Y) = 0. ) = 100, yellow (hereinafter referred to as Y) = 0 Consider a case where there is a uniform color arrangement area A2.
[0005]
When such a halftone color image is binarized by the error diffusion method and printed by a printer, a pseudo halftone color image as shown in FIG. 6B is formed. In this image, FIG. The boundary line L that did not exist in the halftone image shown in a) appears so as to hang down from the binarized color arrangement area B2 obtained by binarizing the color arrangement area A2.
[0006]
This is because, in the binarized color arrangement area B21 outside the boundary line L in the binarized color arrangement area B1 obtained by the error diffusion method, the pixels where cyan ink is placed and the magenta ink are arranged. The pixel to be placed is the same pixel, and the color development is expressed as a subtractive color mixture of cyan and magenta. On the other hand, in the binarized color scheme region B22 inside the boundary line L, the error from the binarized color scheme region B2 This is because the pixel where cyan ink is placed and the pixel where magenta ink is placed are shifted due to the influence, and color development is expressed as an additive color mixture of cyan and magenta. This is because the boundary line L is generated when the color looks different in spite of the area to be removed.
[0007]
Thus, the boundary line L appears in the color arrangement region that should be uniform, and the quality of the image is deteriorated. SUMMARY OF THE INVENTION An object of the present invention is to provide an error diffusion method for preventing a boundary line appearing in a uniform color arrangement region and obtaining a high-quality pseudo halftone color image.
[0008]
[Means for Solving the Problems and Effects of the Invention]
The binarization method of a color image by the error diffusion method of the present invention binarizes a halftone color image by comparing the pixel density with a threshold value for a plurality of different color components included in the halftone color image. This is a binarization method of a color image by an error diffusion method in which an error generated in the binarization is reflected in the binarization of surrounding pixels, and is reflected in the pixel row in the main scanning direction to be binarized first. The initial error value is set differently from that of the other color components for at least one color component. The color components include cyan, magenta, yellow, and black. Among the cyan, magenta, and black, The error initial value of at least one color component and the error initial value of the other color components are set to different values, and the different setting in the error initial value changes in a different pattern in the main scanning direction. Is a constant, the change pattern in the main scanning direction of the values, the sum of the error initial value of one pixel row is, as is almost zero, the value is characterized in that it is a pattern that moves up and down.
[0009]
Since the pixel row in the main scanning direction which is first binarized by the error diffusion method is not binarized in the previous pixel row, there is no error or a slight error from adjacent pixels. However, when the binarization process is performed on the first pixel row, settings different from those of the other color components are performed for at least one color component.
[0010]
The different setting in the initial error value is, for example, setting the initial error value to be different. As a result, in the example described in the prior art, if either the cyan or magenta error initial value of the first pixel row is different from the other color components, cyan and magenta are halftone color images. In this case, even if the density values are the same, the positions where the cyan ink and the magenta ink are placed at the same pixel position are disturbed, and the boundary line becomes inconspicuous or completely invisible as in the prior art. Such setting of the error initial value includes a method of setting a sufficiently large positive value or a sufficiently small negative value as the error initial value so that the value sufficiently varies depending on the color component.
[0011]
[0012]
[0013]
[0014]
[0015]
[0016]
[0017]
[0018]
[0019]
[0020]
DETAILED DESCRIPTION OF THE INVENTION
[Embodiment 1]
FIG. 1 is a block diagram showing a schematic configuration of a halftone image data binarization apparatus 2 that realizes a color image binarization method by an error diffusion method to which some of the above-described inventions are applied.
[0021]
The halftone image data binarization apparatus 2 is mainly composed of a computer, and includes a CPU 12, a program storage unit 13 comprising a ROM, a threshold storage unit 14 comprising a RAM, an error distribution matrix storage unit 15 comprising a RAM, and a RAM. An error buffer 16, an input image data storage unit 17 comprising a RAM, an output image data storage unit 18 comprising a RAM, and a working memory 19 comprising a RAM are connected via a bus 20 to send control signals and data signals. It can be exchanged.
[0022]
In addition, the halftone image data binarization apparatus 2 also includes an input / output device such as a keyboard 21 and a display 22 necessary as a computer, an external storage device 23 such as a hard disk and a floppy disk drive, and the like via a bus 20. A color printer 24 is connected.
[0023]
The program storage unit 13 stores a basic program necessary for a computer, an error diffusion processing program described later, and other processing programs, and is executed by the CPU 12 as necessary. The program is executed by reading the program from a storage medium such as a floppy disk, a magneto-optical disk, or a CD-ROM in which the various programs are stored via the external storage device 23 as needed. May be.
[0024]
The threshold storage unit 14 stores a threshold used for the error diffusion method. The error distribution matrix storage unit 15 distributes the error between the output density value calculated by the error diffusion method and the original density value to the peripheral pixels in the error buffer 16 and the distribution target peripheral pixels. The rate is stored as an error distribution matrix.
[0025]
The error buffer 16 stores an error distributed to each pixel that is an error distribution target. The input image data storage unit 17 stores halftone image data introduced from the external storage device 23 and the like for each of the three primary colors CMY as color components. The density range of each primary color is 0-255. The output image data storage unit 18 stores pseudo halftone image data obtained by binarizing halftone image data stored in the input image data storage unit 17 by error diffusion processing. The pseudo halftone image data is displayed on the display 22 or recorded by the color printer 24 as necessary.
[0026]
Next, error diffusion processing executed by the CPU 12 will be described with reference to the flowchart of FIG. This process is performed to binarize the halftone image stored in the input image data storage unit 17 to create a pseudo halftone image. When the error diffusion process is started, the following procedure is executed.
[0027]
First, a plane C (cyan plane) is set as a primary color plane to be processed (S202). Then, among the error buffers e (x, y) for the plane C stored in the error buffer 16, the error buffer e (x, 0) for the first row of pixels is set (S204).
[0028]
For example, error buffer e (x, 0) = 10, 10, 10, 10, 0, 0, 0, 0, 10, 10, 10, 10, 0, 0, 0, 0,. An error value array pattern in which “10” is four and “0” is four is repeated. In the second and subsequent rows of pixels, ie, y ≧ 1, all error buffers e (x, y) = 0.
[0029]
Next, variables x and y for determining the position of the pixel to be binarized are initialized to 0 (S210, S220). Note that the pixel indicated by the variables x and y is referred to as a target pixel. The input density I (0 ≦ I ≦ 255) of the input image corresponding to the pixel position (x, y) is read (S230).
[0030]
Next, the binarization error value e (x, y) corresponding to the target pixel is read from the error buffer 16, and the pixel density I is corrected with the binarization error value e (x, y) as shown in the following equation 1. Then, the correction density I ′ is obtained (S240).
[0031]
[Expression 1]

[0032]
As described above, since the binarization error e (0, 0) of the first pixel in the first row of pixels is set to “10” in step S204, unlike the conventional case, I ′ is the original value. The pixel density I is increased by “10”. Next, the correction density I ′ obtained in step S240 is compared with the threshold value T (S260). For example, “128” is set as the threshold T.
[0033]
If I ′ <T (“YES” in S260), the output density O is set to “0” (off) (S270), and if I ′ ≧ T (“NO” in S260), the output density O "1" (ON) is set as (S280). The value of the output density O is sequentially stored in the output image data storage unit 18 as binarized image data of the plane C for each binarization in step S270 or step S280.
[0034]
Next, based on the correction density I ′ and the output density O, a binarization error value E is calculated as in the following equation 2 (S290).
[0035]
[Expression 2]

[0036]
Next, based on the preset error distribution matrix Bmat (), the error value E is distributed to the error buffer e of peripheral pixels that have not been binarized as shown in the following equation (3) (S300).
[0037]
[Equation 3]

[0038]
Note that “+ =” is an operator indicating that the value already added in the error buffer e is added and stored in the same error buffer e. A specific example of Bmat () is as shown in FIG. 3, for example, and i and j are variables that take values as shown in FIG. 3 when the target pixel position is “*”.
[0039]
Next, it is determined whether or not the binarization process in the main scanning direction (x direction), that is, the process for one line is completed (S310). If not completed (“NO” in S310), the position x of the target pixel in the main scanning direction is increased by 1 (S320), and the process is repeated from step S230.
[0040]
If it is determined that the binarization process for one line has been completed (“YES” in S310), it is determined whether the binarization process for all pixels for one plane has been completed (S330). If processing of all the pixels of one plane has not been completed (“NO” in S330), the position y in the sub-scanning direction of the target pixel is increased by 1 (S340), and the processing is repeated from step S220 again.
[0041]
If the binarization of all the pixels has been completed (“YES” in S330), it is next determined whether or not the processing has been completed for all the planes (S350). Since only the first plane C has been completed (“NO” in S350), the process for the magenta M plane (plane M) is set (S360).
[0042]
Then, an error buffer e (x, 0) for the first row of pixels for the plane M is set (S204). Here, the error buffer e (x, 0) previously set for the plane C = 10, 10, 10, 10, 0, 0, 0, 0, 10, 10, 10, 10, 0, 0, 0, 0, The initial error is set in a pattern different from. For example, with respect to the plane C, the error buffer e (x, 0) = 0, 0, 0, 0, 10, 10, 10, 10, 0, 0, 0, 0, 10, 10, 10 is shifted from the pattern. , 10, .... Alternatively, since the pattern may be different from that of the plane C, the error buffers e (x, 0) may be all zero.
[0043]
Thus, the initial error buffer e (x, y) is set, and the plane M is also binarized by performing the processing of steps S210 to S340 as described above, and the binary of the plane M is stored in the output image data storage unit 18. Stored as converted image data. Next, if the binarization process is completed for the plane M (“NO” in S350), then a yellow plane (plane Y) is set (S360), and the plane Y binarization process is performed. Also for this plane Y, an error buffer e (x, 0) for the pixels in the first row is set (S204). Again, it is better to set separate patterns for the planes C and M. Of the three primary colors CMY, yellow is less affected by the generation of the boundary line, and the arrangement of the dots need not be considered. Since a problem does not easily occur, it may be the same as any one pattern of CMY. Further, all the error buffers e (x, 0) may be 0 without setting any value.
[0044]
In this way, when the plane Y is also binarized and the binarized data is accumulated as the binarized image data of the plane Y in the output image data storage unit 18, the binarization processing is completed for all the planes. (“YES” in S350), the error diffusion process ends.
[0045]
At this time, in the output image data storage unit 18, pseudo halftone image data in which each pixel is binarized for each plane C, M, Y at the output density O set in step S270 or step S280. Is formed. As described above, in the first embodiment, the error initial value of the error buffer e is set so that the value changes in a different pattern in the main scanning direction in the first row of pixels for each of the three primary color CMY planes as color components. It is set. This disturbs the state in which cyan ink and magenta ink are placed at the same pixel position even if the cyan and magenta halftone color images have the same density value, as described in the prior art. This produces an effect that the boundary line is inconspicuous or completely invisible.
[0046]
The error buffer e (x, 0) is “10, 10, 10, 10, 0, 0, 0, 0, 10, 10, 10, 10, 0, 0, 0, 0,. In addition to the repeated pattern of “10” and “0”, for example, the repetition of “10” and “−10”, for example, error buffer e (x, 0) = 10, 10, 10, 10 , −10, −10, −10, −10, 10, 10, 10, 10, −10, −10, −10, −10, and so on, Σe (x, 0) is substantially “0”. The initial error value of the pattern arranged so as to be
[0047]
Further, the pattern in the main scanning direction of the error buffer e (x, 0) is not changed for each plane CMY, but only the value of the error buffer e (x, 0) in the first row of pixels is changed. May be. For example, for the plane C, the error buffers e (x, 0) in the first row are all “10”, and for the plane M, the error buffers e (x, 0) in the first row are all “0”. For the plane Y, the error buffers e (x, 0) in the first row may be all “5” (may be “10” or “0”). Further, a negative value may be used instead of a positive value.
[0048]
In this way, even if the pattern of the error buffer e (x, 0) in the main scanning direction is not changed, the boundary line described in the prior art can be made inconspicuous or completely invisible even with only the value. [Embodiment 2]
The second embodiment differs from the first embodiment in error diffusion processing, and the others are the same. The error diffusion process of the second embodiment is shown in the flowchart of FIG.
[0049]
When the process is started, first, the plane C is set as the primary color plane to be processed (S402). Next, variables x and y for determining the position of the pixel to be binarized are initialized to 0 (S410, S420).
[0050]
The input density I (0 ≦ I ≦ 255) of the input image corresponding to the pixel position (x, y) is read (S430). Next, the binarization error value e (x, y) corresponding to the target pixel is read from the error buffer 16, and the pixel density I is corrected by the binarization error value e (x, y) as shown in the following equation 4. Then, the correction density I ′ is obtained (S440).
[0051]
[Expression 4]

[0052]
In the second embodiment, a special error initial value is not set for the pixels in the first row as in the first embodiment. All the error initial values are “0”. Next, threshold value generation processing (S450) is performed. This process is shown in the flowchart of FIG.
[0053]
First, it is determined whether or not “1” (ON) exists for the pixel at the same position in the previously processed plane (S451). Assuming that the first plane C is currently being processed, the binarization process has not been performed before, so it is not turned on (“NO” in S451). An initial threshold value T0 is set (S452). Here, for example, “128” is set as the initial threshold value T0.
[0054]
When the threshold value generation process is completed in this way, the corrected density I ′ obtained in step S440 is compared with the threshold value T obtained in step S450 (S460). If I ′ <T (“YES” in S460), the output density O is set to “0” (off) (S470), and if I ′ ≧ T (“NO” in S460), the output density O "1" (ON) is set as (S480). The value of the output density O is sequentially accumulated as binary image data of the plane C in the output image data storage unit 18.
[0055]
Hereinafter, the processes in steps S490 to S560 are the same as those in steps S290 to S360 of the first embodiment. When the process for the magenta M plane (plane M) is set in step S560, the processes in steps S410 to S440 described above are performed, and then the threshold value generation process in step S450 is performed.
[0056]
In step S451 of this threshold value generation process, if the pixel position is set to ON for the previously processed plane C ("YES" in S451), then whether two ONs are set to the same pixel position? It is determined whether or not (S453). Since the current process is the process of plane M, the previous process is only plane C, and therefore “YES” is not determined in step S453. Therefore, it is determined as “NO” in step S453, and the initial threshold value T0 + α is set as the threshold value T (S455). For example, α is “50”.
[0057]
As described above, the threshold T used for the binarization processing of each pixel of the plane M is set to T = T0 if the pixel position is set to be off in the plane C (“NO” in S451). If the pixel position is set to ON in the plane C (“YES” in S451, “NO” in S453), T = T0 + α is set (S455).
[0058]
Thereafter, the processing in steps S460 to S560 described above is performed, the binarized image data of the plane M is accumulated in the output image data storage unit 18, and the plane Y is set as a processing target. Similarly, the process of steps S410 to S440 described above is performed for the plane Y, and then the threshold value generation process of step S450 is performed.
[0059]
In step S451 of this threshold value generation process, if the pixel position is set to ON for either of the previously processed planes C and M ("YES" in S451), then the same pixel position is set to ON. It is determined whether or not two are set (S453). If both planes C and M are set to ON (“YES” in S453), the initial threshold value T0 + 2α is set as the threshold T (S454), and only one of planes C and M is set to ON. If “NO” in S453, the initial threshold value T0 + α is set as the threshold value T (S455).
[0060]
Accordingly, any one of three types T0, T0 + α, and T0 + 2α is set as the threshold value T used for the binarization processing of each pixel of the plane Y. Note that the plane Y has little influence on the generation of the boundary line as described in the first embodiment. Therefore, in the process of step S450, the threshold value T = is set regardless of the binarized states of the other planes C and M. You may set to T0.
[0061]
Thereafter, the processes of steps S460 to S560 described above are performed, the binarized image data of the plane Y is accumulated in the output image data storage unit 18, and all the processes are ended (“YES” in S550). In the second embodiment, as described above, in the same pixel, when the binarization of the primary color plane previously binarized is performed, the binarization of the other primary color plane of the same pixel performed after this is performed. As for conversion, by increasing the threshold value T, the probability of binarization of ON is reduced.
[0062]
Thus, in the same pixel, when the previously binarized primary color plane is turned on, that is, at the position where the primary color ink is determined to be placed, the other primary color inks are the same. It becomes difficult to be placed at the pixel position. Therefore, the state where the ink is placed can be disturbed, and the boundary line described in the prior art can be made inconspicuous or completely invisible.
[0063]
In the same pixel, when the binarization of the primary color plane previously binarized is not performed, that is, when it is off, 2 of the other primary color planes of the same pixel to be performed thereafter As for the binarization, the probability that the binarization of ON is performed is maintained. However, the probability that the binarization of ON is further performed may be increased. For example, when the binarization of the primary color plane previously binarized has not been turned on, the threshold T for judging binarization is set for the binarization of other primary color planes of the same pixel performed after this. Lowering increases the probability of binarization of ON. By adding such a configuration, the boundary line can be made inconspicuous or completely invisible.
[0064]
Further, instead of the threshold value generation process shown in FIG. 5, when the off process has been performed in the previous plane, a process for increasing the probability that the binarization will be performed, for example, the threshold value T is lowered. Processing may be performed. [Others] The binarization in the first and second embodiments is an error diffusion method based on a method of distributing the binarization error to the density of surrounding pixels that have not been binarized when the pixel of interest is binarized. However, an error diffusion method in which binarization error is distributed from surrounding pixels at the time of binarization, a so-called average error minimum method may be used.
[0065]
When the average error minimum method is applied to the first embodiment, pixel rows in the main scanning direction that are first binarized do not have pixels in the previous rows, and therefore different values for each plane in advance. Alternatively, assuming a virtual pixel in which an error value is generated in a different pattern, error distribution from the virtual pixel may be processed as an error initial value.
[0066]
In the first and second embodiments, three primary colors of cyan, magenta, and yellow are used as color components, but black may be included in addition to this. In this case, in particular, it is preferable that at least one color component of cyan, magenta, and black is set with an error initial value different from the other color components. Further, instead of using such primary colors, light colors or intermediate colors may be used as color components.
[0067]
Further, the configurations of the first embodiment and the second embodiment may be combined. In this way, the effect of making the boundary line inconspicuous or invisible is further enhanced. In the second embodiment, after the binarization process for the first color, the probability of binarization to ON is reduced for both the second color and the third color, but one of the second color and the third color is selected. Only the probability of turning on may be reduced. In particular, as described above, yellow hardly affects the occurrence of a boundary line. Therefore, in the case of yellow binarization, it is not necessary to change the probability of binarization to ON.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a schematic configuration of a halftone image data binarization apparatus as a first embodiment.
2 is a flowchart of error diffusion processing in Embodiment 1. FIG.
FIG. 3 is a configuration explanatory diagram of an error distribution matrix according to the first embodiment.
4 is a flowchart of error diffusion processing in Embodiment 2. FIG.
5 is a flowchart of threshold value generation processing according to Embodiment 2. FIG.
FIG. 6 is an explanatory diagram of a phenomenon that occurs in a conventional binarization process.
[Explanation of symbols]
2 ... Halftone image data binarization device 12 ... CPU 13 ... Program storage unit 14 ... Threshold storage unit 15 ... Error distribution matrix storage unit 16 ... Error buffer 17 ... Input image data storage unit 18 ... Output image data storage unit 19 ... Work Memory 20 ... Bus 21 ... Keyboard 22 ... Display 23 ... External storage device 24 ... Color printer

Claims (2)

For a plurality of different color components included in a halftone color image, the halftone color image is binarized by comparing the pixel density with a threshold value, and errors generated in the binarization are binarized for surrounding pixels. A color image binarization method by an error diffusion method to be reflected in
The error initial value reflected in the pixel row in the main scanning direction to be binarized first is set differently from the other color components for at least one color component,
The color components include cyan, magenta, yellow, and black, and the error initial value of at least one color component of cyan, magenta, and black is different from the error initial value of the other color components. Set,
The different setting in the initial error value is a setting in which the value changes in a different pattern in the main scanning direction,
The change pattern of values in the main scanning direction is a pattern in which values move up and down so that the sum of initial error values for one pixel row is substantially zero, and a color image obtained by an error diffusion method is characterized in that Binarization method.   2. The method of binarizing a color image by an error diffusion method according to claim 1, wherein an error initial value different from that of the other color components is set for at least one color component of cyan and magenta.

Images