JP3789243B2

JP3789243B2 - Image processing apparatus and method

Info

Publication number: JP3789243B2
Application number: JP34336698A
Authority: JP
Inventors: 幸二小林; 敏大内
Original assignee: Ricoh Co Ltd
Current assignee: Ricoh Co Ltd
Priority date: 1998-12-02
Filing date: 1998-12-02
Publication date: 2006-06-21
Anticipated expiration: 2018-12-02
Also published as: JP2000175032A

Description

【０００１】
【発明の属する技術分野】
本発明は、画像領域を文字領域とそうでない領域（非文字領域）に分離して各領域毎に画像処理を行うための画像処理装置および方法に関する。
【０００２】
【従来の技術】
一般に、デジタル複写機等において読み取られる原稿は、文字原稿、写真原稿及び網点印刷原稿の３種類に大別され、また、これらが混在する原稿も取り扱われる。このような原稿はそれぞれの画像に要求される画質評価が異なるので、各領域を識別して各領域毎に最適な画像処理を行う必要がある。しかしながら、網点上の文字や色地上の文字を識別しないと、文字ではなく網点と誤って識別して平滑化等の絵柄と同様な処理を行い、このため網点上の文字の解像度が悪くなり、読みにくくなる。
【０００３】
そこで、従来例として例えば特開平７−９５４０９号公報には、網点上の文字や色地上の文字、網点原稿、写真原稿を高画質に再生するために、像域分離処理により白地上の文字領域と非文字領域を分離し、次いで文字領域にはエッジ強調を施し、他方、非文字領域に対して平滑化を行い、次いでこの平滑化後の信号に対して注目画素のエッジ度に応じた適応エッジ強調を行う方法が提案されている。
【０００４】
また、他の従来例として例えば特開平７−２６４３９９号公報には、色成分毎の画像のエッジ情報（エッジ強度と方向）に基づいて原稿を写真画像領域と、カラー網点画像領域と文字線画像領域に分離し、また、色成分毎の領域分離と単色用の領域分離を併用して網点上の文字領域を分離して、各領域毎に画像処理を行う方法が提案されている。
【０００５】
【発明が解決しようとする課題】
しかしながら、上記の特開平７−９５４０９号公報記載の発明では、像域分離処理により網点上の文字や色地上の文字を非文字領域と判定して処理を行うので、例えばカラー複写機のようにこの処理の後段に、色補正処理や中間調処理を行う場合には、網点上の文字や色地上の文字が誤って絵柄処理されるという問題点がある。
【０００６】
また、上記の特開平７−２６４３９９号公報記載の発明では、色成分毎の画像のエッジ方向を高網点率の領域やその境界において判定することが困難であり、誤判定の確率が高いという問題点がある。また、これを防止するために分離信号の平滑化をより強力に行うと、網点上の文字を判定することが困難となる。加えて、色成分毎にエッジ情報検出手段が必要になるので、ハード量が増加してコストアップとなったり、処理時間が長くなる等の問題点がある。
【０００７】
本発明は上記従来例の問題点に鑑み、白地上の文字のみならず網点上の文字や色地上の文字を正確に判定して高画質で再生することができる画像処理装置を提供することを目的とする。
【０００８】
【課題を解決するための手段】
第１の手段は上記目的を達成するために、入力画像データから文字領域とその他の領域とに分離する像域分離処理を施す画像処理装置であって、前記入力画像データに対してエッジ強調処理を施す第１のフィルタリング手段と、前記第１のフィルタリング手段により処理された画像データに基づき、エッジか否かを判定する第１のエッジ判定手段と、前記入力画像データに対して平滑化処理を施す第２のフィルタリング手段と、前記第２のフィルタリング手段により処理された画像データに基づき、エッジか否かを判定する第２のエッジ判定手段と、前記入力画像データが網点領域か否かを判定する網点領域判定手段と、前記第１のエッジ判定手段、前記第２のエッジ判定手段、及び前記網点領域判定手段の判定結果に基づき、前記画像データにおける文字領域とその他の領域とを判定する総合判定手段とを有する画像処理装置を特徴とする。
【０００９】
第２の手段は、第１の手段において、前記総合判定手段は、前記第２のエッジ判定手段が文字領域だけではなく、前記網点領域判定手段が網点領域ではなく、かつ前記第１のエッジ判定手段が文字領域と判定した場合には、文字領域と判定し、それ以外は、その他の領域と判定することを特徴とする。
【００１０】
第３の手段は、第１の手段において、前記総合判定手段は、前記第２のエッジ判定手段が文字領域と判定した場合には文字領域と判定し、前記第２のエッジ判定手段が文字領域ではなく、かつ前記網点領域判定手段が網点領域と判定した場合には、絵柄領域と判定し、前記第２のエッジ判定手段が文字領域ではなく、前記網点領域判定手段が網点領域ではなく、かつ前記第１のエッジ判定手段が文字領域と判定した場合には、文字領域と判定し、前記第２のエッジ判定手段が文字領域ではなく、前記網点領域判定手段が網点領域ではなく、かつ前記第１のエッジ判定手段が文字領域ではないと判定した場合には、絵柄領域と判定することを特徴とする。
【００１１】
第４の手段は、第１の手段において、前記総合判定手段は、前記第２のエッジ判定手段が文字領域と判定し、かつ前記第１のエッジ判定手段が網点領域と判定した場合には、文字領域と判定し、前記第２のエッジ判定手段が文字領域ではなく、かつ前記網点領域判定手段が網点領域と判定した場合には、絵柄領域と判定し、前記網点領域判定手段が網点領域ではなく、かつ前記第１のエッジ判定手段が文字領域ではないと判定した場合には、文字領域と判定し、前記網点領域判定手段が網点領域ではなく、かつ前記第１のエッジ判定手段が文字領域ではないと判定した場合には、絵柄領域と判定することを特徴とする。
第５の手段は、入力画像データから文字領域とその他の領域とに分離する像域分離処理を施す画像処理方法であって、前記入力画像データに対してエッジ強調処理を施す第１の工程と、前記第１の工程で処理された画像データに基づき、エッジか否かを判定する第２の工程と、前記入力画像データに対して平滑化処理を施す第３の工程と、前記第３の工程で処理された画像データに基づき、エッジか否かを判定する第４の工程と、前記入力画像データが網点領域か否かを判定する第５の工程と、前記第２の工程、前記第４の工程、及び前記第５の工程の判定結果に基づき、前記入力画像データにおける文字領域とその他の領域とを判定する第６の工程とを有することを特徴とする。
第６の手段は、第５の手段において、前記第６の工程で、前記第４の工程の判定結果が文字領域だけではなく、前記第５の工程の判定結果が網点領域ではなく、かつ前記第２の工程の判定結果が文字領域である場合には、文字領域と判定し、それ以外は、その他の領域と判定することを特徴とする。
第７の手段は、第５の手段において、前記第６の工程で、前記第４の工程の判定結果が文字領域である場合には文字領域と判定し、前記第４の工程の判定結果が文字領域ではなく、かつ前記第５の工程の判定結果が網点領域である場合に、絵柄領域と判定し、前記第４の工程の判定結果が文字領域ではなく、前記第５の工程の判定結果が網点領域ではなく、かつ前記第２の工程の判定結果が文字領域である場合に、文字領域と判定し、前記第４の工程の判定結果が文字領域ではなく、前記第５の工程の判定結果が網点領域ではなく、かつ前記第２の工程の判定結果が文字領域ではない場合には、絵柄領域と判定することを特徴とする。
第８の手段は、第５の手段において、前記第６の工程で、前記第４の工程の判定結果が文字領域であり、かつ前記第２の工程の判定結果が網点領域である場合には、文字領域と判定し、前記第４の工程の判定結果が文字領域ではなく、かつ前記第５の工程の判定結果が網点領域である場合には、絵柄領域と判定し、前記第５の工程の判定結果が網点領域ではなく、かつ前記第２の工程の判定結果が文字領域ではない場合には、文字領域と判定し、前記第５の工程の判定結果が網点領域ではなく、かつ前記第２の工程の判定結果が文字領域ではない場合には、絵柄領域と判定することを特徴とする。
【００１４】
【発明の実施の形態】
１．第１の実施形態
以下、図面を参照して本発明の実施の形態を説明する。図１は本発明に係る画像処理装置の一実施形態を示すブロック図、図２は図１の像域分離処理部を詳細に示すブロック図、図３は図２のエッジ強調処理部の構成及び処理を示す説明図、図４は図２の平滑化処理部の構成及び処理を示す説明図、図５は図２のエッジ判定部を詳細に示すブロック図、図６は図５の黒画素判定部と白画素判定部の処理を示す説明図、図７は図２の総合判定部の処理を示す説明図である。
【００１５】
図１において、スキャナ１により読み取られた原稿画像データは文字用画像処理部２と、絵柄用画像処理部３と像域分離処理部４に印加される。文字用画像処理部２は一般的には、文字のエッジを強調するエッジ強調処理を行い、また、カラー画像の場合には高ＵＣＲ（下地除去）率によるＵＣＲ処理、高解像度の中間調処理を行う。絵柄用画像処理部３は一般的には、網点画像上のモアレを除去するための平滑化処理や、高階調性を重視した中間調処理を行う。なお、文字用画像処理部２と絵柄用画像処理部３の処理については、本発明に直接関係しないので、詳細な説明を省略する。文字用画像処理部２と絵柄用画像処理部３により処理された画像データは、セレクタ５により図２に詳しく示す像域分離処理部４により検出された文字／非文字検出信号に基づいて選択されてプリンタ部６に印加される。
【００１６】
図２を参照して像域分離処理部４について詳しく説明する。像域分離処理部４は概略的に、エッジ強調処理部１１及びエッジ判定部１３を有する白地上文字領域判定部（１１，１３）と、平滑化処理部１２及びエッジ判定部１４を有する網点上文字領域判定部（１２，１４）と総合判定部１５により構成されている。スキャナ１からの画像データは、白地上文字領域判定部（１１，１３）ではまず、エッジ強調処理部１１により例えば図３に示すような５×３画素のマトリクスで係数＝２４の注目画素がエッジ強調フィルタ処理されてエッジ判定部１３に出力される。また、これに並行して網点上文字領域判定部（１２，１４）ではまず、平滑化処理部１２により例えば図４に示すような５×３画素のマトリクスで係数＝２の注目画素が平滑化されてエッジ判定部１４に出力される。
【００１７】
＜白地上文字領域判定＞
図５に詳しく示すようにエッジ判定部１３、１４は同じ構成であるが、３値化部２１の閾値が異なる。白地上文字領域判定回路（１１、１３）におけるエッジ判定部１３の３値化部２１は、エッジ強調処理部１１からの入力データ（画素値Ｘ＝０〜２５５）を白画素、黒画素及びその他の画素に分類するために、閾値ＴＭ１、ＴＭ２（ＴＭ１＞ＴＭ２）により３値化する。すなわち
Ｘ≧ＴＭ１の場合→黒画素
ＴＭ１＞Ｘ≧ＴＭ２の場合→その他の画素
ＴＭ２＞Ｘの場合→白画素
のように分類してこの３値データを黒画素判定部２２と白画素判定部２３に出力する。
【００１８】
黒画素判定部２２はこの３値データに基づいて、図６（ａ）〜（ｄ）に示すように３つの黒画素が縦、横、斜めに連続するパターンと一致するか否かを判定してその判定結果を２値で判定部２４に出力する。また、白画素判定部２３はこの３値データに基づいて、図６（ｅ）〜（ｈ）に示すように３つの白画素が縦、横、斜めに連続するパターンと一致するか否かを判定してその判定結果を２値で判定部２４に出力する。判定部２４は黒画素判定部２２と白画素判定部２３の各判定結果に基づいて、５×５画素のブロックにおけるパターンの一致が両者ともに１個以上存在する場合に注目ブロックを「白地上のエッジ領域」と判定し、そうでなければ非エッジ領域と判定する。
【００１９】
ここで、白地上の文字領域を判定するためにエッジ強調処理を行う理由は、エッジ強調により文字エッジとその背景（白地）の差を強調してエッジ判定の精度を向上させるとともに、ペアライン（一定間隔の黒線）等のように線画像密集領域においてエッジ判定の精度を向上させるためである。なお、この白地上の文字領域を判定する方法は、例えば電子情報通信学会論文誌、D-II Vol. J75-D-II No. 1 pp39-47, 1992年1月"文字／絵柄（網点、写真）混在画像の像域分離方式”に記載されているが、他の方法でもよい。
【００２０】
＜網点上文字領域判定＞
網点上文字領域判定回路（１２、１４）におけるエッジ判定部１４の３値化部２１は、平滑化処理部１２からの入力データ（画素値Ｘ＝０〜２５５）を白画素、黒画素及びその他の画素に分類するために、閾値ＴＡ１、ＴＡ２（但し、ＴＭ１＞ＴＡ１＞ＴＡ２＞ＴＭ２）により３値化する。すなわち
Ｘ≧ＴＡ１の場合→黒画素
ＴＡ１＞Ｘ≧ＴＡ２の場合→その他の画素
ＴＡ２＞Ｘの場合→白画素
のように分類してこの３値データを黒画素判定部２２と白画素判定部２３に出力する。黒画素判定部２２と、白画素判定部２３と判定部２４の処理は白地上文字領域判定の場合と同一であり、これにより「網点上の文字領域」か否かが判定される。
【００２１】
ここで、網点上の文字領域を判定するために平滑化処理を行う理由は、網点部に見られる画像データの起伏を除去して、網点上の文字領域を判定可能にするためである。なお、エッジ判定部１３、１４はこれに限定されず、また、同じ構成でなくてもよい。また、網点上文字領域判定回路（１２、１４）では、平滑化処理部１２の出力に基づいて直接エッジ判定を行う代わりに、平滑化処理部１２の出力をさらにエッジ強調し、その結果に基づいてエッジ判定を行うようにしてもよい。
【００２２】
＜総合判定＞
総合判定部１５は白地上文字領域判定部（１１，１３）の判定結果ＳＭＯと網点上文字領域判定部（１２，１４）の判定結果ＡＭＯに基づいて図７に示すような総合判定を行う。すなわち
（１）網点上文字領域判定結果ＡＭＯ＝文字の場合には、白地上文字領域判定結果ＳＭＯに関係なく、総合判定結果＝文字、
（２）白地上文字領域判定結果ＳＭＯ＝文字の場合には、網点上文字領域判定結果ＡＭＯに関係なく、総合判定結果＝文字、
（３）網点上文字領域判定結果ＡＭＯ＝白地上文字領域判定結果ＳＭＯ＝非文字の場合には、総合判定結果＝絵柄
と総合判定する。
【００２３】
したがって、上記実施形態によれば、エッジ強調した信号をエッジ判定することにより白地上の文字領域を判定するとともに、平滑化した信号をエッジ判定することにより網点上の文字領域を判定し、両者の論理和（ＯＲ）により文字領域を総合判定するので、白地上の文字のみならず網点上の文字や色地上の文字を判定することができる。このため、図１に示す文字用画像処理部２と絵柄用画像処理部３の出力を誤りなく選択することができるので、白地上の文字のみならず網点上の文字や色地上の文字を正確に判定して高画質で再生することができる。
【００２４】
２．第２の実施形態
次に図８〜図１０を参照して第２の実施形態について説明する。この第２の実施形態の像域分離処理部は、図８に示すように第１の実施形態（図２参照）に対して網点領域判定部１６ａが追加され、この網点領域判定部１６はスキャナ１により読み取られた原稿画像データに基づいて網点領域か否かを判定し、その判定結果ＡＭＩを総合判定部１５ａに出力する。
【００２５】
網点領域判定部１６ａでは図９に詳しく示すように、まず、ピーク画素検出部３１により３×３画素のブロックにおける中心画素の値が周囲画素より一定レベル以上か否かを判定することによりピーク画素を検出する。次いで網点領域検出部３２により４×４画素のブロックにおいてピーク画素を含むブロックが２ブロック以上存在する場合に注目ブロックを網点候補領域とし、そうでない場合には非網点候補領域として検出する。最後に網点領域補正部３３により４×４画素を１ブロックとして、注目ブロックを中心ブロックとする３×３ブロックにおいて４ブロック以上が網点候補領域の場合に注目ブロックを網点領域（ＡＭＩ＝１）とする。なお、この網点領域判定は図９に限定されず、他の公知の方法を用いてもよい。
【００２６】
そして、総合判定部１５ａは白地上文字領域判定部（１１，１３）の判定結果ＳＭＯと、網点上文字領域判定部（１２，１４）の判定結果ＡＭＯと網点領域判定部１６の判定結果ＡＭＩに基づいて図１０に示すような総合判定を行う。すなわち
（１）網点上文字領域判定結果ＡＭＯ＝文字の場合には、白地上文字領域判定結果ＳＭＯと網点領域判定結果ＡＭＩに関係なく、総合判定結果＝文字、
（２）網点上文字領域判定結果ＡＭＯ＝非文字、且つ網点領域判定結果ＡＭＩ＝網点領域の場合には、白地上文字領域判定結果ＳＭＯに関係なく、総合判定結果＝絵柄（網点）、
（３）網点上文字領域判定結果ＡＭＯ＝非文字、網点領域判定結果ＡＭＩ＝非網点領域且つ白地上文字領域判定結果ＳＭＯ＝文字の場合には、総合判定結果＝文字、
（４）網点上文字領域判定結果ＡＭＯ＝非文字、網点領域判定結果ＡＭＩ＝非網点領域且つ白地上文字領域判定結果ＳＭＯ＝非文字の場合には、総合判定結果＝絵柄（写真）
と総合判定する。
【００２７】
したがって、第２の実施形態によれば、総合判定の優先度は、網点上の文字領域が最も高く、以下、網点（絵柄）領域、白地上の文字領域、写真（絵柄）領域の順に低くなるので、第１の実施形態の効果の他に、さらに像域分離精度を上げることができる。
【００２８】
３．第３の実施形態
次に図１１を参照して第３の実施形態について説明する。この第３の実施形態の構成は第２の実施形態（図８、図９）と同一であるが、総合判定部１５ａは白地上文字領域判定部（１１，１３）の判定結果ＳＭＯと、網点上文字領域判定部（１２，１４）の判定結果ＡＭＯと網点領域判定部１６の判定結果ＡＭＩに基づいて図１１に示すような総合判定を行う。すなわち
（１）網点上文字領域判定結果ＡＭＯ＝文字、且つ網点領域判定結果ＡＭＩ＝網点の場合には、白地上文字領域判定結果ＳＭＯに関係なく、総合判定結果＝文字、
（２）網点上文字領域判定結果ＡＭＯ＝非文字、且つ網点領域判定結果ＡＭＩ＝網点領域の場合には、白地上文字領域判定結果ＳＭＯに関係なく、総合判定結果＝絵柄（網点）、
（３）網点領域判定結果ＡＭＩ＝非網点領域、且つ白地上文字領域判定結果ＳＭＯ＝文字の場合には、網点上文字領域判定結果ＡＭＯに関係なく、総合判定結果＝文字、
（４）網点領域判定結果ＡＭＩ＝非網点領域、且つ白地上文字領域判定結果ＳＭＯ＝非文字の場合には、網点上文字領域判定結果ＡＭＯに関係なく、総合判定結果＝絵柄（写真）
と総合判定する。
【００２９】
したがって、第３の実施形態によれば、第２の実施形態と比較すると、網点領域でのみ網点上の文字領域を判定するので、第２の実施形態の効果の他に、さらに像域分離精度を上げることができる。
【００３０】
４．第４の実施形態
次に図１２〜図１４を参照して第４の実施形態について説明する。この第４の実施形態では、図１２に示すように第１の実施形態（図１参照）における像域分離処理部４の代わりに特徴量算出部７が設けられ、セレクタ５の代わりに演算部８が設けられている。他の構成は同一である。ここで、第１〜第３の実施形態では、文字用画像処理部２又は絵柄用画像処理部３の処理信号を選択、すなわち２値的に選択するように構成されているが、この第４の実施形態では、特徴量算出部７によりエッジの度合いに応じて文字度ＭＯＪＩを算出し、文字度ＭＯＪＩに応じて文字用画像処理部２処理信号ＣＨＡＲと絵柄用画像処理部３の処理信号ＩＭＡＧＥを演算部８により混合するように構成されている。
【００３１】
特徴量算出部７は図１３に詳しく示すように、エッジ強調処理部１１及びエッジ度算出部１６を有する白地上文字度算出部（１１，１６）と、平滑化処理部１２及びエッジ度算出部１７を有する網点上文字度算出部（１２，１７）と演算部１８により構成されている。エッジ強調処理部１１と平滑化処理部１２は第１の実施形態と同じ構成である。
【００３２】
白地上文字度算出におけるエッジ度算出部１６では図１４に詳しく示すように、まず、エッジ強調処理部１１からの入力データに対して、ラプラシアンフィルタ部４１により例えば図１５に示すような３×３画素のマトリクスでラプラシアンフィルタ処理を行うことによりエッジ度（量）を算出する。次いでこのエッジ度（量）をＬＵＴ（ルックアップテーブル）部４２により白地上の文字度ＳＭＯに変換する。ここで、エッジ度が大きい場合には文字度ＳＭＯも大きくなるのでＬＵＴ部４２は必ずしも必要はないが、それぞれのシステムに応じて最適な文字度ＳＭＯを得るためにはＬＵＴ部４２を設けた方がよい。
【００３３】
網点上文字度算出におけるエッジ度算出部１７は、図１４に示すエッジ度算出部１６と同じ構成であるが、ＬＵＴ部４２の変換が異なる。ここで、エッジ度算出部１６、１７におけるラプラシアンフィルタ部４１は必ずしも必要はなく、また、エッジ度算出部１６、１７は図１４に示す構成に限定されない。また、第１の実施形態においても説明したが、平滑化処理部１２の出力に基づいて直接エッジ度を算出する代わりに、平滑化処理部１２の出力をさらにエッジ強調し、その結果に基づいてエッジ度を算出するようにしてもよい。
【００３４】
図１３に示す演算部１８は、白地上文字度算出部（１１，１６）により算出された白地上の文字度ＳＭＯと網点上文字度算出部（１２，１７）により算出された網点上の文字度ＡＭＯに基づいて、以下のように数値が大きいほど大きな文字度ＭＯＪＩを算出して図１２に示す演算部８に出力する。
【００３５】
ＭＯＪＩ＝max（ＳＭＯ，ＡＭＯ）
但し、ＭＯＪＩ，ＳＭＯ，ＡＭＯ＝８ビット
図１２に示す演算部８は、文字度ＭＯＪＩに応じて文字用画像処理部２の処理信号ＣＨＡＲと絵柄用画像処理部３の処理信号ＩＭＡＧＥを以下のように混合する。
【００３６】
ＯＵＴ＝｛ＣＨＡＲ×ＭＯＪＩ＋ＩＭＡＧＥ×（２５５−ＭＯＪＩ）｝／２５５
但し、ＣＨＡＲ，ＩＭＡＧＥ＝８ビット
なお、文字度ＭＯＪＩを算出する他の方法としては、第１〜第３の実施形態における２値の像域分離信号を、文字領域と判定された場合を「２５５」、判定されなかった場合を「０」として平滑化することにより８ビットの文字度ＭＯＪＩに変換するようにしてもよい。また、他の方法として、第２、第３の実施形態における網点領域判定部１６ａ（図８参照）の判定結果を文字度ＭＯＪＩと併用するようにしてもよい。
【００３７】
したがって、この第４の実施形態によれば、エッジ強調した信号のエッジ度により白地上の文字度を算出するとともに、平滑化した信号のエッジ度により網点上の文字度を算出し、両者を総合判定して文字度を算出するので、白地上の文字のみならず網点上の文字や色地上の文字を判定することができるとともに、文字処理と絵柄処理を２値的ではなく、段階的に切り替えることができる。このため、白地上の文字のみならず網点上の文字や色地上の文字を正確に判定して高画質で再生することができる。
【００３８】
【発明の効果】
以上説明したように、本発明によれば、白地上の文字には、エッジ強調処理を行うことにより下地とエッジ部とのコントラストを高めてエッジ判定の精度を向上させ、網点領域の場合には、網点に対して平滑化処理を行ってぼかすことにより網点をエッジ判定しないようにして網点上の文字と網点の誤判定を減少させることが可能となり、これにより白地上の文字のみならず網点上の文字や色地上の文字を正確に判定して高画質で再生することができる。
【図面の簡単な説明】
【図１】本発明に係る画像処理装置の一実施形態を示すブロック図である。
【図２】図１の像域分離処理部を詳細に示すブロック図である。
【図３】図２のエッジ強調処理部の構成及び処理を示す説明図である。
【図４】図２の平滑化処理部の構成及び処理を示す説明図である。
【図５】図２のエッジ判定部を詳細に示すブロック図である。
【図６】図５の黒画素判定部と白画素判定部の処理を示す説明図である。
【図７】図２の総合判定部の処理を示す説明図である。
【図８】第２の実施形態の像域分離処理部を詳細に示すブロック図である。
【図９】図８の網点領域判定部を詳細に示すブロック図である。
【図１０】図８の総合判定部の処理を示す説明図である。
【図１１】第３の実施形態の総合判定部の処理を示す説明図である。
【図１２】第４の実施形態の画像処理装置を示すブロック図である。
【図１３】図１２の特徴量算出部を詳細に示すブロック図である。
【図１４】図１３のエッジ度算出部を詳細に示すブロック図である。
【図１５】図１４のラプラシアンフィルタ部の処理を示す説明図である。
【符号の説明】
２文字用画像処理部
３絵柄用画像処理部
４像域分離処理部
８，１８演算部
１１エッジ強調処理部
１２平滑化処理部
１３，１４エッジ判定部
１５，１５ａ総合判定部
１６ａ網点領域判定部
１６，１７エッジ度算出部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and method for performing image processing for each area by separating an image area into a character area and a non-character area (non-character area).
[0002]
[Prior art]
In general, documents read by a digital copying machine or the like are roughly classified into three types: a text document, a photo document, and a halftone print document, and a document in which these are mixed is also handled. Since such an original requires different image quality evaluations for each image, it is necessary to identify each area and perform optimum image processing for each area. However, if the characters on the halftone dots and the characters on the color ground are not identified, they are mistakenly identified as halftone dots instead of characters, and the same processing as the pattern such as smoothing is performed. It gets worse and harder to read.
[0003]
Therefore, as a conventional example, for example, Japanese Patent Application Laid-Open No. 7-95409 discloses a method for reproducing a character on a white ground by image area separation processing in order to reproduce a character on a halftone dot, a character on a color ground, a halftone dot manuscript, and a photo manuscript with high image quality. The character area is separated from the non-character area, then edge enhancement is applied to the character area, and the non-character area is smoothed. Then, the smoothed signal is subjected to the edge degree of the pixel of interest. A method for performing adaptive edge enhancement has been proposed.
[0004]
As another conventional example, for example, Japanese Patent Laid-Open No. 7-264399 discloses a document based on edge information (edge strength and direction) of an image for each color component, a photographic image region, a color halftone image region, and a character line. A method has been proposed in which image areas are separated and character areas on halftone dots are separated by using both area separation for each color component and area separation for a single color to perform image processing for each area.
[0005]
[Problems to be solved by the invention]
However, in the invention described in Japanese Patent Laid-Open No. 7-95409, a character on a halftone dot or a character on a color ground is determined as a non-character region by image region separation processing, so that, for example, a color copying machine is used. In addition, when color correction processing or halftone processing is performed after this processing, there is a problem that characters on halftone dots or characters on the color ground are erroneously subjected to pattern processing.
[0006]
In the invention described in Japanese Patent Laid-Open No. 7-264399, it is difficult to determine the edge direction of an image for each color component in a high halftone dot area or its boundary, and the probability of erroneous determination is high. There is a problem. Further, if the separation signal is more smoothly smoothed to prevent this, it becomes difficult to determine the character on the halftone dot. In addition, since edge information detection means is required for each color component, there is a problem that the amount of hardware increases, resulting in an increase in cost and a longer processing time.
[0007]
SUMMARY OF THE INVENTION In view of the problems of the above-described conventional example, the present invention provides an image processing apparatus that can accurately determine not only characters on white ground but also characters on halftone dots and characters on color ground and reproduce them with high image quality. With the goal.
[0008]
[Means for Solving the Problems]
For the first means to achieve the above object, an image processing apparatus which performs image area separation processing for separating from the input image data into a character region and another region, edge enhancement processing to the input image data a first filtering means for performing said-out first filtering means based on the image data processed by the first edge determination means for determining whether or not the edge, smoothing with respect to the input image data and second filtering means for performing processing, the-out second filtering means based on the image data processed by the second edge determination means for determining whether or not an edge, the input image data is halftone area and whether determining dot area determination means for determining said first edge determination unit,-out the second edge determination means, and based on the determination result of the halftone region judging means, the image data Wherein the image processing apparatus and a total determination means for determining a character region and other regions in.
[0009]
The second means is the first means, wherein the comprehensive judgment means is such that the second edge judgment means is not only a character area, the halftone area judgment means is not a halftone area, and the first When the edge determination means determines that the character area is detected , it is determined that the character area is determined. Otherwise, the other area is determined .
[0010]
Third means, in the first hand stage, the overall judging means, when said second edge determination unit determines that the character region is determined as a character region, the second edge determination unit characters If it is not an area and the halftone dot area determination means determines that it is a halftone dot area, it is determined as a picture area, the second edge determination means is not a character area, and the halftone dot area determination means is a halftone dot. If the first edge determination means determines that the area is not a character area, it is determined to be a character area, and the second edge determination means is not a character area, and the halftone dot area determination means is a halftone dot. If it is not an area and the first edge determining means determines that the area is not a character area, it is determined that the area is a picture area .
[0011]
Fourth means, in the first hand stage, the total determination unit, when the second edge determination means determines that the character region and the first edge determination unit determines that the halftone dot area Is determined as a character area, and when the second edge determination means is not a character area and the halftone dot area determination means determines a halftone dot area, it is determined as a picture area, and the halftone dot area determination is performed. If the means is not a halftone dot area, and the first edge determination means determines that it is not a character area, it is determined as a character area, the halftone dot area determination means is not a halftone dot area, and the first area When one edge determination means determines that the area is not a character area, it is determined that the area is a picture area .
A fifth means is an image processing method for performing image region separation processing for separating character regions and other regions from input image data, the first step of performing edge enhancement processing on the input image data; A second step of determining whether or not the edge is based on the image data processed in the first step, a third step of performing a smoothing process on the input image data, and the third step A fourth step of determining whether or not an edge is based on the image data processed in the step; a fifth step of determining whether or not the input image data is a halftone region; the second step; And a sixth step of determining a character region and other regions in the input image data based on the determination result of the fourth step and the fifth step.
The sixth means is the fifth means in which, in the sixth step, the determination result of the fourth step is not only the character region, the determination result of the fifth step is not a halftone dot region, and If the determination result of the second step is a character area, it is determined as a character area, and otherwise it is determined as another area.
A seventh means is a fifth means in which, in the sixth step, if the determination result of the fourth step is a character area, it is determined as a character area, and the determination result of the fourth step is When it is not a character region and the determination result of the fifth step is a halftone dot region, it is determined as a picture region, and the determination result of the fourth step is not a character region but the determination of the fifth step If the result is not a halftone dot region and the determination result of the second step is a character region, it is determined as a character region, and the determination result of the fourth step is not a character region, but the fifth step If the determination result is not a halftone dot region and the determination result of the second step is not a character region, it is determined as a pattern region.
The eighth means is the fifth means, wherein, in the sixth step, the determination result of the fourth step is a character area, and the determination result of the second process is a halftone dot area. Is determined to be a character region, and if the determination result of the fourth step is not a character region and the determination result of the fifth step is a halftone dot region, it is determined to be a pattern region, and the fifth region If the determination result of the step is not a halftone dot region and the determination result of the second step is not a character region, it is determined as a character region, and the determination result of the fifth step is not a halftone dot region. And when the determination result of the said 2nd process is not a character area, it determines with a pattern area, It is characterized by the above-mentioned.
[0014]
DETAILED DESCRIPTION OF THE INVENTION
1. First Embodiment Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention, FIG. 2 is a block diagram showing in detail an image area separation processing unit in FIG. 1, and FIG. 3 is a configuration of an edge enhancement processing unit in FIG. FIG. 4 is an explanatory diagram showing the configuration and processing of the smoothing processing unit in FIG. 2, FIG. 5 is a block diagram showing the edge determination unit in FIG. 2 in detail, and FIG. 6 is a black pixel determination in FIG. FIG. 7 is an explanatory diagram showing the processing of the comprehensive determination unit in FIG. 2.
[0015]
In FIG. 1, document image data read by the scanner 1 is applied to a character image processing unit 2, a design image processing unit 3, and an image area separation processing unit 4. The character image processing unit 2 generally performs edge emphasis processing for emphasizing the edges of characters, and in the case of a color image, UCR processing with a high UCR (background removal) rate and high resolution halftone processing are performed. Do. The pattern image processing unit 3 generally performs a smoothing process for removing moire on a halftone image and a halftone process that emphasizes high gradation. Note that the processing of the character image processing unit 2 and the design image processing unit 3 is not directly related to the present invention, and thus detailed description thereof is omitted. The image data processed by the character image processing unit 2 and the pattern image processing unit 3 is selected by the selector 5 based on the character / non-character detection signal detected by the image area separation processing unit 4 shown in detail in FIG. Applied to the printer unit 6.
[0016]
The image area separation processing unit 4 will be described in detail with reference to FIG. The image area separation processing unit 4 generally includes a white background character region determination unit (11, 13) having an edge enhancement processing unit 11 and an edge determination unit 13, and a halftone dot having a smoothing processing unit 12 and an edge determination unit 14. The upper character area determination unit (12, 14) and the comprehensive determination unit 15 are configured. For the image data from the scanner 1, the white ground character area determination unit (11, 13) first converts the target pixel having a coefficient of 24 into a 5 × 3 pixel matrix as shown in FIG. The enhancement filter process is performed and the result is output to the edge determination unit 13. In parallel with this, in the halftone dot character region determination unit (12, 14), first, the smoothing processing unit 12 smooths the target pixel of coefficient = 2 in a matrix of 5 × 3 pixels as shown in FIG. And output to the edge determination unit 14.
[0017]
<White ground character area judgment>
As shown in detail in FIG. 5, the edge determination units 13 and 14 have the same configuration, but the threshold value of the ternary unit 21 is different. The ternarization unit 21 of the edge determination unit 13 in the white ground character region determination circuit (11, 13) converts the input data (pixel value X = 0 to 255) from the edge enhancement processing unit 11 into white pixels, black pixels, and others. In order to classify the pixel into three pixels, ternarization is performed using threshold values TM1 and TM2 (TM1> TM2). That is, if X ≧ TM1 → black pixel TM1> X ≧ TM2 → other pixels TM2> X → classify as white pixels and classify the ternary data into black pixel determination unit 22 and white pixel determination unit 23. Output to.
[0018]
Based on this ternary data, the black pixel determination unit 22 determines whether or not three black pixels match a pattern in which the vertical, horizontal, and diagonal continuous as shown in FIGS. 6 (a) to 6 (d). The determination result is output to the determination unit 24 in binary. Further, based on the ternary data, the white pixel determination unit 23 determines whether or not three white pixels coincide with a pattern that is continuous in the vertical, horizontal, and diagonal directions as shown in FIGS. The determination is made and the determination result is output to the determination unit 24 in binary. Based on the determination results of the black pixel determination unit 22 and the white pixel determination unit 23, the determination unit 24 selects the block of interest as “white ground” when there is at least one pattern match in a 5 × 5 pixel block. Edge region ”is determined. Otherwise, it is determined as a non-edge region.
[0019]
Here, the reason for performing the edge enhancement processing to determine the character area on the white background is to enhance the edge determination accuracy by enhancing the difference between the character edge and the background (white background) by edge enhancement, This is to improve the accuracy of edge determination in a line image dense region such as a black line at regular intervals. The method for determining the character area on the white ground is, for example, the Journal of the Institute of Electronics, Information and Communication Engineers, D-II Vol. J75-D-II No. 1 pp39-47, January 1992 , Photo) “Image area separation method for mixed images”, but other methods may be used.
[0020]
<Character determination on halftone dot>
The ternarization unit 21 of the edge determination unit 14 in the halftone dot character region determination circuit (12, 14) converts the input data (pixel value X = 0 to 255) from the smoothing processing unit 12 into white pixels, black pixels, and In order to classify into other pixels, ternarization is performed by threshold values TA1 and TA2 (where TM1>TA1>TA2> TM2). That is, when X ≧ TA1 → black pixel TA1> X ≧ TA2 → other pixels TA2> X → classify as white pixels and ternary data is classified into black pixel determination unit 22 and white pixel determination unit 23. Output to. The processing of the black pixel determination unit 22, the white pixel determination unit 23, and the determination unit 24 is the same as in the case of white ground character region determination, and thereby, it is determined whether or not the character region is on a halftone dot.
[0021]
Here, the reason for performing the smoothing process in order to determine the character area on the halftone dot is to remove the undulation of the image data seen in the halftone dot portion and to make it possible to determine the character area on the halftone dot. is there. The edge determination units 13 and 14 are not limited to this, and may not have the same configuration. Further, in the halftone dot character region determination circuit (12, 14), instead of directly performing edge determination based on the output of the smoothing processing unit 12, the output of the smoothing processing unit 12 is further edge-enhanced, The edge determination may be performed based on this.
[0022]
<Comprehensive judgment>
The comprehensive determination unit 15 performs a comprehensive determination as shown in FIG. 7 based on the determination result SMO of the white ground character region determination unit (11, 13) and the determination result AMO of the halftone character region determination unit (12, 14). . That is, (1) when halftone dot character region determination result AMO = character, regardless of white ground character region determination result SMO, comprehensive determination result = character,
(2) When white ground character area determination result SMO = character, regardless of halftone dot character area determination result AMO, overall determination result = character,
(3) When halftone dot character region determination result AMO = white ground character region determination result SMO = non-character, comprehensive determination result = pattern is comprehensively determined.
[0023]
Therefore, according to the above embodiment, the character area on the white background is determined by performing edge determination on the edge-enhanced signal, and the character area on the halftone dot is determined by performing edge determination on the smoothed signal. Since the character area is comprehensively determined by the logical sum (OR), it is possible to determine not only the characters on the white background but also the characters on the halftone dots and the characters on the color ground. Therefore, the output of the character image processing unit 2 and the pattern image processing unit 3 shown in FIG. 1 can be selected without error, so that not only the characters on the white background but also the characters on the halftone dots and the characters on the color ground can be selected. Accurate determination and reproduction with high image quality are possible.
[0024]
2. Second Embodiment Next, a second embodiment will be described with reference to FIGS. In the image area separation processing unit of the second embodiment, as shown in FIG. 8, a halftone dot region determination unit 16a is added to the first embodiment (see FIG. 2). Determines whether or not it is a halftone dot region based on the document image data read by the scanner 1, and outputs the determination result AMI to the comprehensive determination unit 15a.
[0025]
As shown in detail in FIG. 9, the halftone dot region determination unit 16 a first determines whether the peak pixel detection unit 31 determines whether the value of the central pixel in the 3 × 3 pixel block is equal to or higher than a certain level from the surrounding pixels. Detect a pixel. Next, when there are two or more blocks including peak pixels in the 4 × 4 pixel block, the halftone dot detection unit 32 detects the block of interest as a halftone dot candidate region, otherwise, it is detected as a non-halftone dot candidate region. . Finally, when the 4 × 4 pixels are set as one block by the halftone dot region correcting unit 33 and 4 blocks or more are halftone dot candidate regions in the 3 × 3 block having the target block as the central block, the target block is converted into a halftone dot region (AMI = 1). This halftone dot region determination is not limited to that shown in FIG. 9, and other known methods may be used.
[0026]
Then, the overall determination unit 15a determines the determination result SMO of the white ground character region determination unit (11, 13), the determination result AMO of the halftone dot character region determination unit (12, 14), and the determination result of the halftone region determination unit 16. Based on the AMI, comprehensive determination as shown in FIG. 10 is performed. That is, (1) when halftone dot character area determination result AMO = character, regardless of white ground character area determination result SMO and halftone dot area determination result AMI, overall determination result = character,
(2) When halftone dot character area determination result AMO = non-character and halftone dot area determination result AMI = halftone dot area, comprehensive determination result = picture (halftone dot) regardless of white ground character area determination result SMO ),
(3) When halftone dot character region determination result AMO = non-character, halftone dot region determination result AMI = non-halftone dot region and white ground character region determination result SMO = character, overall determination result = character,
(4) When halftone dot character region determination result AMO = non-character, halftone dot region determination result AMI = non-halftone dot region and white ground character region determination result SMO = non-character, comprehensive determination result = picture (photo)
And comprehensive judgment.
[0027]
Therefore, according to the second embodiment, the priority of the comprehensive determination is the highest in the character area on the halftone dot, and in the order of the halftone dot (picture) area, the character area on the white background, and the photograph (picture) area. Therefore, in addition to the effects of the first embodiment, the image area separation accuracy can be further increased.
[0028]
3. Third Embodiment Next, a third embodiment will be described with reference to FIG. Although the configuration of the third embodiment is the same as that of the second embodiment (FIGS. 8 and 9), the overall determination unit 15a uses the determination result SMO of the white background character region determination unit (11, 13) and the network. Based on the determination result AMO of the dot character region determination unit (12, 14) and the determination result AMI of the halftone region determination unit 16, comprehensive determination as shown in FIG. 11 is performed. That is, (1) When the halftone dot character region determination result AMO = character and the halftone dot region determination result AMI = halftone dot, regardless of the white ground character region determination result SMO, the comprehensive determination result = character,
(2) When halftone dot character area determination result AMO = non-character and halftone dot area determination result AMI = halftone dot area, comprehensive determination result = picture (halftone dot) regardless of white ground character area determination result SMO ),
(3) When halftone dot area determination result AMI = non-halftone dot area and white ground character area determination result SMO = character, regardless of halftone dot character area determination result AMO, overall determination result = character,
(4) When halftone dot area determination result AMI = non-halftone dot area and white ground character area determination result SMO = non-character, regardless of halftone dot character area determination result AMO, overall determination result = picture (photograph )
And comprehensive judgment.
[0029]
Therefore, according to the third embodiment, compared with the second embodiment, the character area on the halftone dot is determined only in the halftone dot area, so that in addition to the effect of the second embodiment, the image area is further increased. Separation accuracy can be increased.
[0030]
4). Fourth Embodiment Next, a fourth embodiment will be described with reference to FIGS. In the fourth embodiment, as shown in FIG. 12, a feature amount calculation unit 7 is provided instead of the image area separation processing unit 4 in the first embodiment (see FIG. 1), and an arithmetic unit instead of the selector 5 8 is provided. Other configurations are the same. Here, in the first to third embodiments, the processing signal of the character image processing unit 2 or the design image processing unit 3 is selected, that is, selected in a binary manner. In the embodiment, the character amount MOJI is calculated by the feature amount calculation unit 7 according to the degree of the edge, and the character image processing unit 2 processing signal CHAR and the processing image IMAGE of the pattern image processing unit 3 are calculated according to the character degree MOJI. Are mixed by the calculation unit 8.
[0031]
As shown in detail in FIG. 13, the feature amount calculation unit 7 includes a white background character degree calculation unit (11, 16) having an edge enhancement processing unit 11 and an edge degree calculation unit 16, a smoothing processing unit 12, and an edge degree calculation unit. A halftone dot character degree calculation unit (12, 17) having 17 and a calculation unit 18. The edge enhancement processing unit 11 and the smoothing processing unit 12 have the same configuration as in the first embodiment.
[0032]
In the whiteness character degree calculation, as shown in detail in FIG. 14, first, the Laplacian filter unit 41 applies 3 × 3 as shown in FIG. 15, for example, to the input data from the edge enhancement processing unit 11. The edge degree (amount) is calculated by performing Laplacian filter processing with a matrix of pixels. Next, this edge degree (amount) is converted into a white ground character degree SMO by an LUT (look-up table) unit 42. Here, when the edge degree is large, the character degree SMO also becomes large, so the LUT unit 42 is not necessarily required. However, in order to obtain the optimum character degree SMO according to each system, the LUT unit 42 is provided. Is good.
[0033]
The edge degree calculation unit 17 in calculating the character degree on halftone dot has the same configuration as the edge degree calculation unit 16 shown in FIG. 14, but the conversion of the LUT unit 42 is different. Here, the Laplacian filter unit 41 in the edge degree calculation units 16 and 17 is not necessarily required, and the edge degree calculation units 16 and 17 are not limited to the configuration shown in FIG. Further, as described in the first embodiment, instead of directly calculating the edge degree based on the output of the smoothing processing unit 12, the output of the smoothing processing unit 12 is further edge-enhanced, and based on the result. The edge degree may be calculated.
[0034]
The calculation unit 18 shown in FIG. 13 includes the white ground character degree SMO calculated by the white background character degree calculation unit (11, 16) and the halftone character degree calculation unit (12, 17). Based on the character degree AMO, the greater the numerical value, the greater the character degree MOJI is calculated and output to the computing unit 8 shown in FIG.
[0035]
MOJI = max (SMO, AMO)
However, MOJI, SMO, AMO = 8 bits The arithmetic unit 8 shown in FIG. 12 generates the processing signal CHAR of the character image processing unit 2 and the processing signal IMAGE of the picture image processing unit 3 in accordance with the character degree MOJI as follows. To mix.
[0036]
OUT = {CHAR × MOJI + IMAGE × (255−MOJI)} / 255
However, CHAR, IMAGE = 8 bits As another method for calculating the character degree MOJI, the case where the binary image area separation signal in the first to third embodiments is determined to be a character area is “255”. ”, It may be converted to 8-bit character degree MOJI by smoothing the case where it is not determined as“ 0 ”. As another method, the determination result of the halftone dot region determination unit 16a (see FIG. 8) in the second and third embodiments may be used together with the character degree MOJI.
[0037]
Therefore, according to the fourth embodiment, the character degree on the white ground is calculated from the edge degree of the edge-enhanced signal, and the character degree on the halftone dot is calculated from the edge degree of the smoothed signal. Since the character degree is calculated by comprehensive determination, not only the characters on the white ground but also the characters on the halftone dots and the characters on the color ground can be determined, and the character processing and the pattern processing are not binary, but step by step. You can switch to For this reason, not only the characters on the white ground but also the characters on the halftone dots and the characters on the color ground can be accurately determined and reproduced with high image quality.
[0038]
【The invention's effect】
As described above, according to the present invention, for the characters on the white background, the edge enhancement processing is performed to increase the contrast between the background and the edge portion, thereby improving the accuracy of edge determination. Makes it possible to reduce the misjudgment of characters on halftone dots and halftone dots by performing smoothing processing on the halftone dots so that the halftone dots are not edge-determined. In addition, the characters on the halftone dots and the characters on the color ground can be accurately determined and reproduced with high image quality.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention.
FIG. 2 is a block diagram illustrating in detail an image area separation processing unit in FIG. 1;
FIG. 3 is an explanatory diagram showing a configuration and processing of an edge enhancement processing unit in FIG. 2;
4 is an explanatory diagram showing a configuration and processing of a smoothing processing unit in FIG. 2; FIG.
FIG. 5 is a block diagram illustrating in detail an edge determination unit in FIG. 2;
6 is an explanatory diagram showing processing of a black pixel determination unit and a white pixel determination unit in FIG. 5. FIG.
7 is an explanatory diagram showing processing of a comprehensive determination unit in FIG. 2; FIG.
FIG. 8 is a block diagram illustrating in detail an image area separation processing unit according to a second embodiment;
9 is a block diagram illustrating in detail a halftone dot region determination unit in FIG. 8;
10 is an explanatory diagram showing processing of a comprehensive determination unit in FIG. 8. FIG.
FIG. 11 is an explanatory diagram illustrating processing of a comprehensive determination unit according to the third embodiment.
FIG. 12 is a block diagram illustrating an image processing apparatus according to a fourth embodiment.
13 is a block diagram illustrating in detail a feature amount calculation unit in FIG. 12;
14 is a block diagram illustrating in detail an edge degree calculation unit in FIG. 13;
15 is an explanatory diagram showing processing of a Laplacian filter unit in FIG. 14;
[Explanation of symbols]
2 Image processing unit 3 Character image processing unit 4 Image area separation processing unit 8, 18 Calculation unit 11 Edge enhancement processing unit 12 Smoothing processing unit 13, 14 Edge determination unit 15, 15 a Total determination unit 16 a Halftone region determination Units 16 and 17 Edge degree calculation unit

Claims

An image processing apparatus for performing image area separation processing for separating from the input image data into a character region and another region,
A first filtering means for performing edge enhancement processing to the input image data,
Wherein-out first filtering means based on the image data processed by the first edge determination means for determining whether or not an edge,
And second filtering means for performing a smoothing process on the input image data,
The-out second filtering means based on the image data processed by the second edge determination means for determining whether or not an edge,
Halftone dot area determination means for determining whether the input image data is a halftone dot area;
The first edge determination unit,-out the second edge determination means, and based on the determination result of the halftone region judging means, and comprehensive determining means for determining a character region and another region in the input image data ,
An image processing apparatus comprising:

The comprehensive determination means includes
If the second edge determining means is not only a character area, the halftone dot area determining means is not a halftone dot area, and the first edge determining means determines a character area, it is determined as a character area. ,
The image processing apparatus according to claim 1, wherein other areas are determined as other areas .

The comprehensive determination means includes
If the second edge determination means determines that it is a character area, it is determined as a character area;
If the second edge determination means is not a character area and the halftone dot area determination means determines that it is a halftone dot area, it is determined as a pattern area;
If the second edge determination means is not a character area, the halftone dot area determination means is not a halftone dot area, and the first edge determination means determines a character area, it is determined as a character area;
If the second edge determining means is not a character area, the halftone dot area determining means is not a halftone dot area, and the first edge determining means is not a character area, it is determined as a picture area. The image processing apparatus according to claim 1, wherein:

The comprehensive determination means includes
When the second edge determination means determines that the character area and the first edge determination means determines that the area is a halftone dot area, the character area is determined.
If the second edge determination means is not a character area and the halftone dot area determination means determines that it is a halftone dot area, it is determined as a pattern area;
If the halftone dot area determination means is not a halftone dot area and the first edge determination means is not a character area, it is determined as a character area;
2. The image processing according to claim 1, wherein when the halftone dot area determination unit determines that the area is not a halftone dot area and the first edge determination section determines that the area is not a character area, the image area is determined as a picture area. apparatus.

An image processing method for performing image region separation processing for separating character regions and other regions from input image data,
  A first step of performing edge enhancement processing on the input image data;
  A second step of determining whether or not an edge is based on the image data processed in the first step;
  A third step of performing a smoothing process on the input image data;
  A fourth step of determining whether or not an edge is based on the image data processed in the third step;
  A fifth step of determining whether the input image data is a halftone dot region;
  A sixth step of determining a character region and other regions in the input image data based on the determination results of the second step, the fourth step, and the fifth step;
An image processing method comprising:

The sixth step includes
If the determination result of the fourth step is not only a character region, the determination result of the fifth step is not a halftone dot region, and the determination result of the second step is a character region, the character region And
6. The image processing apparatus according to claim 5, wherein the other areas are determined as other areas.

The sixth step includes
  If the determination result of the fourth step is a character area, it is determined as a character area;
  When the determination result of the fourth step is not a character region and the determination result of the fifth step is a halftone dot region, it is determined as a pattern region,
  When the determination result of the fourth step is not a character region, the determination result of the fifth step is not a halftone dot region, and the determination result of the second step is a character region, it is determined as a character region And
  When the determination result of the fourth step is not a character region, the determination result of the fifth step is not a halftone dot region, and the determination result of the second step is not a character region, 6. The image processing method according to claim 5, wherein the determination is performed.

The sixth step includes
  When the determination result of the fourth step is a character region and the determination result of the second step is a halftone region, it is determined as a character region,
  When the determination result of the fourth step is not a character region and the determination result of the fifth step is a halftone dot region, it is determined as a pattern region,
  When the determination result of the fifth step is not a halftone dot region and the determination result of the second step is not a character region, it is determined as a character region,
  6. The image processing according to claim 5, wherein when the determination result of the fifth step is not a halftone dot region and the determination result of the second step is not a character region, it is determined as a picture region. Method.