JP3780587B2 - Color data converter - Google Patents

Description

【００４１】
指定値決定手段１０２は、入力指定値と出力指定値を決定するもので、動画像キャプチャボードと、グラフィカルユーザインタフェースを搭載したパーソナルコンピュータにより構成される。
【００４２】
近傍値演算手段１０５は、入力画素を色空間上の座標値と考え、その近傍の座標値を演算し、入力指定値特定手段１０６に出力するものであり、加算器とバッファにより構成される。
【００４３】
入力指定値特定手段１０６は、近傍値演算手段１０５で得られた座標値と、入力指定値記録手段１０３に記録された全ての入力指定値との、三次元色空間における距離を演算する加算器および乗算器と、そのうちで最も距離の近い入力指定値を選択する比較器により構成される。
【００４４】
出力指定値選択手段１０７は、入力指定値特定手段１０６で選択された入力指定値に対応する出力指定値を、出力指定値記録手段１０３から選択し、入力された座標値に対応する値として出力値演算手段１０８に出力するセレクタにより構成される。
【００４５】
出力値演算手段１０８は、出力値選択手段１０７からの出力を積和して出力するもので、バッファ、乗算器、および加算器により構成される。
【００４６】
出力手段１０９は、入力画素に対応した出力値演算手段１０８から出力されたキー信号を、画像情報として表示、あるいは記録するもので、Ｄ／Ａ変換器、ビデオセレクタ、モニタ、ビデオデッキにより構成される。
【００４７】
以上のように構成された本装置の動作について説明する。
まず、入力画像は、画素毎に三次元色空間上の座標値として近傍値演算手段１０５に入力される。ここで、画素は、輝度、色差の３属性をもつＣＣＩＲ６０１勧告のＹｕｖ色空間座標を持つとする。
【００４８】
次に、近傍値演算手段１０５によって、図２のように、近傍値演算手段に入力された値を中心にした、Ｙｕｖ各軸に沿ってδ離れた近傍６点の座標が演算され、入力指定値特定手段１０６に入力される。
【００４９】
一方で、指定値決定手段１０２では、ユーザからの指示により、入力手段１０１からの１フレームが静止画像として取りこまれ、表示される。次に、ユーザインタフェースを用いて、その画像中の複数の色が、キー信号として選択したいあるいは選択したくない色として抜きだされ、入力指定値として入力指定値記録手段１０３に出力される。また、それぞれの入力指定値は、選択信号であるか（２５５）ないか（０）を決定し、その値を出力指定値として、出力指定値記録手段１０４に出力される。
【００５０】
入力指定値記録手段１０３と出力指定値記録手段１０４には、（表１）のようなキー信号を作成する場合に代表となる色の三次元色空間上の座標値と、その色を選択するか（２５５）否か（０）を示すデータを記録するもので、あらかじめ準備して記録しておいてもよいし、ユーザが指定値決定手段１０２で、そのつど決定して記録してもよい。
【００５１】
図３は、入力指定値特定手段１０６と、出力指定値選択手段１０７の動作を説明するためのフロー図である。入力指定値特定手段１０６では、入力値と入力指定値記録手段１０３に記録されている全ての入力指定値との三次元色空間における距離が算出される。さらにその値を順次比較することによって、入力値との色空間上での距離が最小となる入力指定値が決定され、出力指定値選択手段１０７に通知される。その通知を受けた出力指定値選択手段１０７では、その入力指定値に対応する出力指定値が出力指定値記録手段１０４から読み出され、入力値に対応する出力値として出力される。
【００５２】
図４（ａ）は、三次元色空間の分割の様子を示す図であり、Ｙを一定とするｕｖ平面における断面図である。表示の便宜上、Ｙが一定となる断面を表示したが、実際の色空間の分割の様子は三次元的なものとなる。図中の白丸は（表１）の入力指定値に該当し、×印は入力値に該当する。色空間上の任意の点は距離が最も近い入力指定値の領域に属するように分割される。このような手法による空間の分割図は一般にボロノイ図（ＶｏｒｏｎｏｉＤｉａｇｒａｍｓ）と呼ばれる。入力指定値特定手段１０６と、出力指定値選択手段１０７の動作によって、境界線に囲まれた分割領域に属する入力値が与えられると、その領域中の白丸に相当する出力値が出力される。つまり、出力指定値選択手段１０７によって出力値は、０または２５５の値に変換されたことになる。
【００５３】
しかし、キー信号は、後処理で画像の合成や色彩調整などの画像編集効果の係り具合を決定するデータとして用いられることを考えると、信号自身に階調を持つことが望ましい。そこで、出力値演算手段１０７では、入力値とその近傍値に対応する６つの出力値が、（数１）の演算で積和され、入力値に対応する８階調のキー信号として出力される。ここで、ｋｉは出力選択手段の出力値、ｑｉは重み係数、ｋ＾は出力である。
【００５４】
【数１】

【００５５】
この動作を言い換えれば、図４（ｂ）のように三次元色空間上の領域の境界を近傍の点を用いて平滑化する処理がおこなわれことになる。そして、入力画素に対応した出力値演算手段１０８から出力されたキー信号は、出力手段１０９により、画像情報として表示、あるいは記録される。
【００５６】
このようにして、入力画素は、本装置によって、色空間上の距離が最も近い指定値の影響を最大限に受けた階調を持つキー信号となって出力される。また、このキー信号は、画像表示空間位置情報を全く使用しないので、画像の輪郭情報を一様に失うことがない、色情報に応じた階調情報を持つことができる。
【００５７】
従って、本装置により、処理の対象となる画像の内容に応じて、作業者が得たい信号を、階調を持つキー信号として出力することができる。
【００５８】
なお、本実施形態において入力信号系の定義された色空間をＹＵＶ色空間としているが、ＲＧＢ、ＹＣｂＣｒ、ＹＩＱ、ＣＭＹ、ＣＭＹＫ、Ｌ＊ａ＊ｂ＊、Ｌ＊ｕ＊ｖ＊、ＨＳＶ、ＨＬＳなど画像処理に一般に用いられる他の色空間を用いてもよく、また処理画像に応じて独自に定義した色空間を用いてもよい。
【００５９】
なお、本実施形態において入力指定値特定手段１０６で比較される色空間における距離は、三次元色空間における距離を用いたが、距離の二乗の値を用いてもよい。距離の二乗の値を用いた場合は、距離計算において平方根を求める処理が不要となるので処理コストの大幅な削減が可能となる。
【００６０】
なお、近傍値演算手段１０５と、出力値演算手段１０８は、６つのデータを蓄積するためのバッファを持つ構成でもよく、バッファから６つの出力値を順次出力することによって、入力指定値特定手段１０６と出力指定値選択手段１０７を複数持たない構成にすることもできる。
【００６１】
また、本装置の近傍値演算手段１０２において、近傍値として６点を出力するとしたが、周辺８点や、１２点や、あるいは演算コストが許される限りそれ以上のデータを出力するように設定できるユーザインタフェースを備えていてもよく、点数を増やすことでより多くの階調を持ったキー信号の出力が期待できる。
【００６２】
さらに、近傍値演算手段１０５は、距離範囲δをユーザが設定できるようなインタフェースを備えていてもよく、δを変更することでキー信号の中間階調をもつデータの割合を変更することができる。
【００６３】
なお、本装置においては、出力指定値としてキー信号を設定したが、（表２）のように出力指定値に色空間上の座標値を設定してもよい。この場合、出力画像として、中間調を持つポスタリゼーション効果を施した画像を得ることができる。
【００６４】
【表２】

【００６５】
さらに、出力指定値に色空間上の座標値を設定する場合、出力指定値の値として入力指定値をそのまま設定しなくてもよい。入力指定値と出力指定値を異なる値に設定したり異なる色空間で定義される値に設定しておくことにより、ポスタリゼーション効果の実施後に得られる画像の配色を自由に設定することが可能となり、出力画像の定義される色空間も、システムの要請に合わせ自由に選択することが可能となる。
【００６６】
なお、本装置では、入力指定値記録手段１０３、出力指定値記録手段１０４、近傍値演算手段１０５、入力指定値特定手段１０６、出力指定値選択手段１０７、出力値演算手段１０８は、専用ハードウェアを用いるとしているが、すべてパーソナルコンピュータと、そのメモリにロードされるソフトウェアで実現しても良い。
【００６７】
（実施の形態２）
以下に、静止画像からキーとなる信号情報を抜き出す装置である本発明の第２の実施の形態について詳しく説明する。
【００６８】
図５は、本発明の第２の実施の形態における装置の機能ブロックを示す図である。図５において、５０１は指定値決定手段、５０２は入力指定値記録手段、５０３は出力指定値記録手段、５０４は入力指定値特定手段、５０５は出力値演算手段、５０６は入力手段、５０７は出力手段である。
【００６９】
指定値決定手段５０１は、ユーザが、入力画像あるいはカラーパレットなどから入力指定値を抽出し、それに対応する出力指定値を決定するためのもので、マウスなどのポインティングデバイスとグラフィカルユーザインタフェースによって構成される。
【００７０】
（表３）は、入力指定値記録手段５０２と出力指定値記録手段５０３に記録されるデータの例である。入力指定値記録手段５０２は、キー信号を作成する場合に代表となる色のＣＩＥＬ＊ａ＊ｂ＊色空間座標値を入力指定値として記録するものである。出力指定値記録手段５０３は、実施の形態１のものと同様なもので、ここでは詳しい説明は省略する。
【００７１】
【表３】

【００７２】
入力指定値特定手段５０４は、入力指定値記録手段５０２に記録された全ての入力指定値との、三次元色空間における距離を演算し、そのうちで最も距離の近い入力指定値と二番目に近い入力指定値を選択するものである。
【００７３】
出力値演算手段５０５は、入力指定値特定手段５０４からの二つの入力指定値と、それぞれに対応する出力指定値を用いて、入力された色彩データが選択信号であるか否かを示すキー信号を演算し、出力するものである。
【００７４】
入力手段５０６は、入力静止画像を画素単位で入力指定値特定手段５０４に出力するもので、出力手段は５０７は、出力値演算手段５０５から出力されたキー信号を、画像情報として表示、あるいは記録するものである。以上の機能ブロックは、すべてパーソナルコンピュータと、そのメモリにロードされるプログラムによって構成される。
【００７５】
このように構成された本実施の形態における装置の動作について説明する。
まず、ユーザは、指定値決定手段５０１のマウスを使って、モニタに表示されている入力画像あるいはカラーパレットなどから入力指定値を抽出し、対応する出力指定値を決定する。指定されたデータは、入力指定値記録手段５０２と出力指定値記録手段５０３に記録される。
【００７６】
入力指定値と出力指定値は、入力指定値の色空間座標がＣＩＥＬ＊ａ＊ｂ＊であること以外は、第一の実施の形態で説明したものと同様であるので説明は省略する。
【００７７】
モニタに表示されている入力画像は、画素毎に三次元色空間上の座標値として入力指定値特定手段５０４に入力される。ここで、画素は、輝度、色差の３属性をもつＣＩＥＬ＊ａ＊ｂ＊色空間座標を持つとする。
【００７８】
図６（ａ）に、入力指定値特定手段５０４の動作を説明するためのフロー図を示す。ここで、図６（ｂ）は、図６（ａ）におけるブロックＡの内部構成である。入力指定値特定手段５０４では、その入力値と入力指定値記録部５０２に記録されている全ての入力指定値との三次元色空間における距離（＝√Ｌ＊２＋ａ＊２＋ｂ＊２）が算出される。さらにその値を順次比較することによって、その画素との距離が最小となる入力指定値と、二番目に距離が小さい入力指定値が決定され、対応する出力指定値とともに出力値演算手段５０５に出力される。
【００７９】
図７は、三次元色空間の分割の様子を示す図であり、Ｌ＊を一定とするａ＊ｂ＊平面における断面図である。表示の便宜上、Ｌ＊が一定となる断面を表示したが、実際の色空間の分割の様子は三次元的なものとなる。図中の×印ｐは、ある入力値の色空間上における座標値で、白丸１はその点に最も近い入力指定値、白丸２はその点に二番目に近い入力指定値である。
【００８０】
出力値演算手段５０５では、白丸１、白丸２に対応する出力指定値が出力指定値記録手段５０３から読み出され、（数２）、（数３）、（数４）の演算式で得られたデータが、階調をもつキー信号として出力される。ここで、式のＬ１２、Ｌ１ｐ、Ｌ２ｐは、図に示す各点間の距離、δは任意の値であり、境界面からのδの距離の領域の出力信号に階調を持たせることができる。
【００８１】
図８（ａ）は、ボロノイ分割によって作られた領域の出力をそのままキー信号とする場合のグラフで、図８（ｂ）は、（数２）、（数３）、（数４）によって得られる出力を表すグラフである。ここで、図の横軸は白丸１からの距離で、縦軸はキー信号出力である。
【００８２】
【数２】

【００８３】
【数３】

【００８４】
【数４】

【００８５】
このように、白丸１、白丸２に対応する出力指定値と、白丸１、白丸２にから構成される最も近い境界面からの距離情報から出力信号が導かれる。この動作を言い換えれば、三次元色空間上の領域の境界近傍の出力に階調を持たせる処理がおこなわれことになる。
【００８６】
このようにして、入力画素は、本装置によって、色空間上の距離が最も近い指定値の影響を最大限に受けた階調を持つキー信号となって出力される。また、このキー信号は、画像表示空間位置情報を全く使用しないので、画像の輪郭情報を一様に失うことがない、色情報に応じた階調情報を持つことができる。
【００８７】
従って、本実施の形態における装置により、処理の対象となる画像の内容に応じて、作業者が得たい信号を、階調を持つキー信号として出力することができる。
【００８８】
なお、本実施形態において入力信号系の定義された色空間をＬ＊ａ＊ｂ＊色空間としているが、ＲＧＢ、ＹＣｂＣｒ、ＹＩＱ、ＣＭＹ、ＣＭＹＫ、ＹＵＶ、Ｌ＊ｕ＊ｖ＊、ＨＳＶ、ＨＬＳなど画像処理に一般に用いられる他の色空間を用いてもよく、また処理画像に応じて独自に定義した色空間を用いてもよい。なお、本実施形態において入力指定値特定手段５０４で比較される色空間における距離は、三次元色空間における距離を用いたが、距離の二乗の値を用いてもよい。距離の二乗の値を用いた場合は、距離計算において平方根を求める処理が不要となるので処理コストの大幅な削減が可能となる。
【００８９】
なお、出力値演算手段５０５は、距離範囲δをユーザが設定できるようなインタフェースを備えていてもよく、δを変更することでキー信号の中間階調をもつデータの割合を変更することができる。
【００９０】
なお、本装置においては、出力指定値としてキー信号を設定したが、（表４）のように出力指定値に色空間上の座標値を設定してもよい。この場合、出力画像として、中間調を持つポスタリゼーション効果を施した画像を得ることができる。
【００９１】
【表４】

【００９２】
さらに、出力指定値に色空間上の座標値を設定する場合、出力指定値の値として入力指定値をそのまま設定しなくてもよい。入力指定値と出力指定値を異なる値に設定したり異なる色空間で定義される値に設定しておくことにより、ポスタリゼーション効果の実施後に得られる画像の配色を自由に設定することが可能となり、出力画像の定義される色空間も、システムの要請に合わせ自由に選択することが可能となる。
【００９３】
なお、本発明の実施の形態は、パーソナルコンピュータとそのメモリにロードされるプログラムにより構成されるとしたが、図５の機能ブロックの一部あるいはすべてを専用ハードウェアに置き換えてもよい。
【００９４】
（実施の形態３）
以下に、動画像からキーとなる信号情報を抜き出す装置である本発明の第３の実施の形態について図面を参照しながら説明する。図９は、本発明の第３の実施の形態における装置の構成を示す図である。
【００９５】
図９において、９０１は指定値決定手段、９０２は入力指定値記録手段、９０３は出力指定値記録手段、９０４は入力指定値特定手段、９０５は出力指定値選択手段、９０９は入力手段、９１０は出力手段で、以上は第１の実施の形態のものと同様なものであるので説明は省略する。
【００９６】
９０６はＬＵＴアドレス生成手段、９０７はＬＵＴ記録手段、９０８はＬＵＴ演算手段である。ＬＵＴアドレス生成手段９０６は、三次元色空間座標値を各軸について８等分にサンプリングした７２９点の座標値をルックアップテーブル（ＬＵＴ）座標値として入力指定値特定手段９０４に出力するものでカウンタにより構成される。ＬＵＴ記録手段９０７は、出力指定値選択手段９０５から出力される７２９個のデータをＬＵＴデータとして記録するものである。ＬＵＴ演算手段９０８は、入力画像の画素データに対応する出力値を、ＬＵＴ記録手段９０７に記録された複数のＬＵＴデータを用いた補間演算により出力するものである。
【００９７】
以上のように構成された本装置の動作について説明する。
まず、指定値決定手段９０１では、ユーザからの指示により、入力手段９０９からの１フレームが静止画像として取りこまれ、表示される。次に、ユーザインタフェースを用いて、その画像中の複数の色が、キー信号として選択したいあるいは選択したくない色として抜きだされ、入力指定値として入力指定値記録手段９０２に出力される。また、それぞれの入力指定値は、選択信号であるか（２５５）ないか（０）を決定し、その値を出力指定値として、出力指定値記録手段９０３に出力される。
【００９８】
次に、ＬＵＴアドレス生成手段９０６から、三次元色空間座標値を各軸について８等分にサンプリングした７２９点の座標値が、ルックアップテーブル（ＬＵＴ）座標値として入力指定値特定手段９０４に出力される。入力指定値特定手段９０４と出力指定値選択手段９０５は、第一の実施の形態と同様の動作で信号を出力するので、詳しい説明は省略する。
【００９９】
ＬＵＴ記録手段９０７には、出力指定値選択手段９０５からの出力が、三次元色空間座標値を各軸について８等分にサンプリングした７２９点の座標値に対応したキー信号として記録される。
【０１００】
入力画像は、画素毎に三次元色空間上の座標値としてＬＵＴ演算手段９０８に入力される。ここで、画素は、輝度、色差の３属性をもつＣＣＩＲ９０１勧告のＹｕｖ色空間座標を持つとする。
【０１０１】
次に、ＬＵＴ演算手段９０８によって、入力座標値に対応する出力が補間演算され、出力手段９１０に出力され、出力手段９１０により画像情報として表示あるいは記録される。入力手段９０９と出力手段９１０は、第一の実施の形態における装置と同様なので詳しい説明は省略する。
【０１０２】
ここでＬＵＴを用いた補間演算について詳しく説明する。ＬＵＴ記録手段９０７は、三次元色空間を各軸８等分した５１２個の立方体の頂点に対応する７２９個の格子点の座標値に対応する出力値を持つ。ＬＵＴ参照手段は、ＬＵＴの格子点の値のうち入力値を囲む頂点に対応する近傍６点のデータを用いたＰＲＩＳＭ補間演算（Ｆａｓｔｃｏｌｏｒｐｒｏｃｅｓｓｏｒｗｉｔｈｐｒｏｇｒａｍｍａｂｌｅｉｎｔｅｒｐｏｌａｔｉｏｎｂｙｓｍａｌｌｍｅｍｏｒｙ（ＰＲＩＳＭ），ＪｏｕｒｎａｌｏｆＥｌｅｃｔｒｏｎｉｃＩｍａｇｉｎｇ，Ｖｏｌ．２（３）（１９９３）参照）をおこない、出力値を決定する。
【０１０３】
図１０（ａ）は、補間演算を説明するための概念図である。実際の色空間は三次元であるが、説明のために、ここでは横軸を色空間座標、縦軸を出力値とする。また、図１０（ｂ）のようにＬＵＴを実際の出力レンジより大きい値に設定し、補間結果を実際の出力レンジでクリッピングすることにより補間における中間階調をもつデータの範囲を調整することができる。
【０１０４】
このようにして、入力画素は、本装置によって、色空間上の距離が最も近い指定値の影響を最大限に受けた階調を持つキー信号となって出力される。また、このキー信号は、画像表示空間位置情報を全く使用しないので、画像の輪郭情報を一様に失うことがない、色情報に応じた階調情報を持つことができる。
【０１０５】
また、第１および第２の実施の形態と異なり、入力指定値特定手段９０４は、入力画像に対して処理をおこなわない前演算処理なので、指定値の数によらない一定の演算時間で、入力画像に対する出力値を演算するという効果も期待できる。従って、本装置により、処理の対象となる画像の内容に応じて、作業者が得たい信号を、階調を持つキー信号として出力することができる。
【０１０６】
なお、本実施形態において入力信号系の定義された色空間をＹＵＶ色空間としているが、ＲＧＢ、ＹＣｂＣｒ、ＹＩＱ、ＣＭＹ、ＣＭＹＫ、Ｌ＊ａ＊ｂ＊、Ｌ＊ｕ＊ｖ＊、ＨＳＶ、ＨＬＳなど画像処理に一般に用いられる他の色空間を用いてもよく、また処理画像に応じて独自に定義した色空間を用いてもよい。
【０１０７】
なお、本実施の形態において、ＬＵＴは７２９個の値を持つものとしたが、その数は出力誤差が許される範囲内でこれより少なくてもよく、ＬＵＴ演算手段９０８で実用的に処理可能な範囲内でこれより多くても同様の効果が期待できる。また、本実施例のＬＵＴ演算手段９０８における補間計算において、入力画素値を囲む近傍６点の値を用いたが、近傍８点などを用いた他の補間演算をおこなうハードウェアを用いてもよい。
【０１０８】
なお、本装置においては、出力指定値としてキー信号を設定したが、（表２）のように出力指定値に色空間上の座標値を設定してもよい。この場合、出力画像として、中間調を持つポスタリゼーション効果を施した画像を得ることができる。
【０１０９】
さらに、出力指定値に色空間上の座標値を設定する場合、出力指定値の値として入力指定値をそのまま設定しなくてもよい。入力指定値と出力指定値を異なる値に設定したり異なる色空間で定義される値に設定しておくことにより、ポスタリゼーション効果の実施後に得られる画像の配色を自由に設定することが可能となり、出力画像の定義される色空間も、システムの要請に合わせ自由に選択することが可能となる。
【０１１０】
なお、本装置では、入力指定値記録手段９０２、出力指定値記録手段９０３、ＬＵＴアドレス生成手段９０６、入力指定値特定手段９０４、出力指定値選択手段９０５は、専用ハードウェアを用いるとしているが、すべてパーソナルコンピュータと、そのメモリにロードされるソフトウェアで実現しても良い。
【０１１１】
【発明の効果】
以上のように、本発明によれば、入力画像は、色空間上の距離が最も近い指定値の影響を最大限に受けた階調を持つキー信号となって出力される。また、このキー信号の階調情報は、画像表示空間における位置情報とは無関係なので、画像の輪郭情報を一様に失うことがないという利点を持つ。
【０１１２】
つまり、処理の対象となる画像の内容に応じて、作業者が得たい信号を、階調を持つキー信号として出力することができる。
【０１１３】
なお、出力指定値に色空間上の座標値を設定した場合、出力画像として、作業者の得たい自由な配色のポスタリゼーション効果を施した出力画像を得ることができる。
【０１１４】
また、ＬＵＴ（ルックアップテーブル）を用いた補間演算により、指定色の数によらない一定の演算時間で、入力画像に対する出力値を演算するという効果も期待できる。
【図面の簡単な説明】
【図１】第１の実施の形態の装置構成図
【図２】第１の実施の形態の近傍値演算手段の動作を説明するための図
【図３】第１の実施の形態の入力指定値特定手段と出力指定値選択手段の動作を説明するための図
【図４】第１の実施の形態の三次元色空間の分割図
【図５】第２の実施の形態の機能ブロック構成図
【図６】第２の実施の形態の入力指定値特定手段の動作を説明するための図
【図７】第２の実施の形態の三次元色空間の分割図
【図８】第２の実施の形態の出力演算を説明するためのグラフ
【図９】第３の実施の形態の装置構成図
【図１０】第３の実施の形態の補間演算を説明するためのグラフ
【図１１】第１の従来技術における装置の構成図
【図１２】第１の従来技術における装置の選択信号生成部の動作を説明するためのグラフ
【図１３】第１の従来技術における装置の乗算部の動作を説明するためのグラフ
【図１４】第２の従来技術における装置の構成図
【図１５】第２の従来技術における装置の信号変換部の動作を説明するためのグラフ
【図１６】第２の従来技術における装置の三次元色空間の分割図
【図１７】第２の従来技術における装置の三次元色空間を説明するための図
【符号の説明】
１０１ 入力手段
１０２ 指定値決定手段
１０３ 入力指定値記録手段
１０４ 出力指定値記録手段
１０５ 近傍値演算手段
１０６ 入力指定値特定手段
１０７ 出力指定値選択手段
１０８ 出力値演算手段
１０９ 出力手段
５０１ 指定値決定手段
５０２ 入力指定値記録手段
５０３ 出力指定値記録手段
５０４ 入力指定値特定手段
５０５ 出力値演算手段
５０６ 入力手段
５０７ 出力手段
９０１ 指定値決定手段
９０２ 入力指定値記録手段
９０３ 出力指定値記録手段
９０４ 入力指定値特定手段
９０５ 出力指定値選択手段
９０６ ＬＵＴアドレス生成手段
９０７ ＬＵＴ記録手段
９０８ ＬＵＴ演算手段
９０９ 入力手段
９１０ 出力手段
１１０１ 極座標変換部
１１０２ 選択信号生成部
１１０３ 選択信号生成部
１１０４ 選択信号生成部
１１０５ 乗算部
１４０１ 信号変換部
１４０２ 信号変換部
１４０３ 信号変換部[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an apparatus for performing a specific process on a color image, and in particular, an apparatus for extracting a specific area of a color image or a special effect (posterization effect) on an image by reducing the gradation of a natural image. Relates to the device.
[0002]
[Prior art]
(First prior art)
First, as a first conventional technique, a method for selecting an area using color information will be described with reference to the drawings.
[0003]
Area selection is called mask data generation in still image processing, and key signal generation in moving image processing, and is essential for processing a specific area independently from other areas in subsequent processing. It is a process.
[0004]
Both mask data and key signals are represented by values that have a maximum value in the selected part and a minimum value in the non-selected part, and have certain types of image data that have a format that conforms to the signal of the image to be processed. It is.
[0005]
Here, a case will be described in which a signal for selecting only a monochromatic background region, that is, a so-called key signal is generated from a television image of a person or the like by using color information in the image. In the actual video production site, this key signal is frequently generated as a preceding process in a device that replaces only the background of a captured video with another video, a so-called chroma key composition device.
[0006]
FIG. 11 is a block diagram showing a configuration of a conventional color data conversion apparatus that generates a key signal. If the three-dimensional signal components constituting the pixels of the input video are Y, U and V, the U and V signals which are the color difference components of the original signal in the polar coordinate conversion unit 1101 are the saturation component C (Chroma) and the hue component H ( (HueAngle). The three components Y, C, and H are converted into selection signals key-y, key-c, and key-h in the selection signal generation units 1102 to 1104 corresponding to the respective components, and then combined in the multiplication unit 1105 to be finally obtained. Is output as a key signal key.
[0007]
FIG. 12 is a diagram for explaining the operation of the selection signal generators 1102 to 1104 in FIG. 11, and shows the input / output relationship of signals. This relationship is a rectangular graph defined by a selection center value (CenterValue) and a selection width (Width) preset for each component. Therefore, the Y, C, and H components of the input signal are converted according to the graph shown in FIG. 12, and become selection signals represented by binary values (0/255).
[0008]
The selection signal for each component is subjected to a logical product in the multiplication unit 1105 and output as a key signal (Key). 13A shows a C component selection signal (Key-c), FIG. 13B shows an H component selection signal (Key-h), and FIG. 13C shows a combined key signal (Key). The distribution in the UV input color difference plane is shown.
[0009]
Note that FIG. 13 shows a cross-sectional view in which the Y component in the divided color space is constant, that is, only the UV input color difference plane, but in practice, a specific region of the specific YUV space is selected three-dimensionally. Is done.
[0010]
In this manner, the key signal is generated based on the selection center value and the selection width preset for each color component by the first conventional technique.
[0011]
(Second prior art)
Next, as a second conventional technique, a special effect by reducing the number of gradations of color components in image processing, that is, a so-called posterization effect implementation method will be described with reference to the drawings.
[0012]
Posterization transforms an original natural image into an image expressed with a small number of colors that can be used in poster designs, etc., by reducing the gradation component of a natural image with multiple gradations to a specified number of gradations. Refers to image processing. The number that can be taken as the value of each component constituting the color is called the number of gradations, and the number that can be taken as the type of color expressed as a combination of the components is called the number of colors.
[0013]
Designers who produce advertisement images and others create works that they imagine by combining various natural effects, including posterization effects, with natural images. Among these special effects, the posterization effect is used widely in various scenes as a means for easily and effectively generating images expressed with a small number of colors such as cell images used for animation from material images. Yes.
[0014]
FIG. 14 is a block diagram showing a configuration of a conventional color data conversion apparatus for posterization. The components of the three-dimensional values (here, R, G, and B components) constituting the pixels of the input image are output with their gradations reduced by independent signal converters 1401 to 1403, respectively.
[0015]
FIG. 15 is a diagram for explaining the signal conversion operation in the signal conversion units 1401 to 1403. For example, when the number of gradations on the output side is 3, as shown in FIG. 15, a multi-gradation input signal is converted into three gradation values by the signal converters 1401 to 1403.
[0016]
Here, since the values constituting the pixels of the input image have a three-dimensional component, when each component is expressed in two gradations, eight colors (2 to the third power) are expressed in three gradations. An output image with 27 colors (3 to the third power) is obtained. FIG. 16 shows an example of color space division when each component is expressed by three gradations. FIG. 16 shows only a cross-sectional view in which the B component in the divided color space is constant, that is, only the RG plane, but the color space is actually divided three-dimensionally.
[0017]
In this way, in the second conventional technique, the color space is fixedly divided, and the color of the input pixel belonging to the same divided region is changed to the same color representing the region, thereby changing the color of the input pixel. The number is decreasing.
[0018]
In the above description, each component of the pixel is represented by a value based on (R, G, B). However, as a three-dimensional value component used for posterization in the conventional method, processing of a television video signal, etc. In some cases, values such as (Y *, U *, V *) and (L *, a *, b *) defined in the uniform perceptual color space used for printing images and the like are used.
[0019]
[Problems to be solved by the invention]
However, in the key signal generation using the color information according to the first prior art, the three-dimensional signal constituting the pixel of the input video is converted into the selection signal independently for each component and then synthesized as the key signal. Therefore, there is a problem that the shape in the color space of the region selected as the key signal is limited to a certain shape along the axis of each component.
[0020]
For example, in the signal conversion as shown in FIG. 11, the color region to be finally selected is a fan shape (actually in the Y direction) as shown in FIG. Therefore, the distribution in the color space of the color to be selected is shaped like a region indicated by diagonal lines as shown in FIG. 17 with respect to the input image to be actually processed. If so, it is no longer possible to cut out such a color region. In other words, when the color to be selected is distributed in a direction inclined with respect to the defined axis of any component of the three-dimensional signal, it is difficult to accurately cut out the region.
[0021]
Further, the selection center value and selection width of the cut-out area to be input are also subject to the skill of the operator, and a method that can easily obtain an optimum value is desired.
[0022]
In many cases, the key signal is used as weight data for determining the degree of image editing effects such as image synthesis and color adjustment in post-processing. When the key signal is binary, a pseudo contour may occur in the edited image, so it is desired that the key signal itself has a gradation.
[0023]
In the posterization effect implementation method according to the second prior art, the gradation division method and the color scheme of the image after the effect are applied depend on each component of the three-dimensional values constituting the pixel and the designated number of gradations. It is determined uniquely. That is, when the input signal is reduced to three gradations, it is similarly fixedly divided into three as shown in FIG. 15, and these division methods are uniformly determined irrespective of the image to be implemented. Is. Therefore, there is a problem that an optimal posterization effect according to the content and type of the image cannot be performed.
[0024]
As for the color arrangement, the color space is divided evenly and each area is painted with each representative color, and there is no room for reflecting the operator's intention.
[0025]
Therefore, the present invention solves the above-described conventional problems, and in performing area selection using color information, a simple operation allows a free shape in the color space in which each component of the input signal is defined. It is an object of the present invention to provide a color data conversion apparatus that can cut out the above-mentioned area and can further have an intermediate gradation in the selection signal output.
[0026]
It is another object of the present invention to provide a color data conversion device that can use any number of colors and color schemes by free color division with a simple operation in implementing the posterization effect.
[0027]
[Means for Solving the Problems]
In order to solve this problem, a first color data conversion device according to the present invention is a color data conversion device that converts color data of pixels constituting a color image, and represents a representative when a key signal is created. Input designation value recording means for recording the coordinate value of the color in the three-dimensional color space, and output designation value recording for recording information for determining whether or not to select each input designation value as a key signal Means, input value determining means for determining the input specified value and output specified value, input means for outputting the input image for each pixel, and calculating the coordinate value in the vicinity considering the input pixel as a coordinate value in the color space Selected by the input specified value specifying means, the input specified value specifying means for calculating the distance between the input value and all input specified values in the three-dimensional color space, and specifying the input specified value closest to the distance Input specification value Output designation value selection means for outputting the corresponding output designation value, output value calculation means for multiplying and outputting the outputs from the output value selection means, and displaying the key signal output from the output value calculation means as image information Or the output means to record is provided, It is characterized by the above-mentioned.
[0028]
A second color data conversion apparatus according to the present invention is a color data conversion apparatus that converts color data of pixels constituting a color image, and is similar to the first color data conversion apparatus. In addition to the value recording means, the output designated value recording means, the designated value determining means, the input means, and the output means, the three-dimensional color of the color data of the pixels constituting the color image and the input designated values recorded in the input value designating means The input specified value that minimizes the distance by calculating and comparing the distance in the space, the input specified value specifying means for specifying the input specified value having the second smallest distance, and the input specified value specifying means Output calculating means for determining output using output specified values corresponding to two input specified values.
[0029]
A third color data conversion apparatus according to the present invention is a color data conversion apparatus that converts color data of pixels constituting a color image, and is the same as the first color data conversion apparatus. In addition to the value recording means, the output designated value recording means, the designated value determining means, the input means, and the output means, the coordinates of the color data obtained by equally sampling the three-dimensional color space for each axis are look-up table (LUT) coordinate values. The LUT address generation means that outputs the input specified value and the input specified value that minimizes the distance by calculating and comparing the distance in the three-dimensional color space between the color data output from the LUT address generating means and each input specified value. An input specified value specifying means to be specified and an output specified value corresponding to the input specified value specified by the input specified value specifying means are read from the output specified value recording means and output. Force value selection means, LUT recording means for recording data output from the output value selection means as LUT, and interpolation using a plurality of data of the LUT recording means with color data of pixels constituting a color image as coordinate values LUT calculation means for determining an output value by calculation is provided.
[0030]
Further, the fourth color data conversion apparatus according to the present invention is the same as the first color data conversion apparatus, that is, input designation value recording means, output designation value recording means, designation value determining means, input means, and output means. In addition to the input specified value specifying means, output specified value selecting means, and output value calculating means, the input pixel is considered as a coordinate value in the color space, and its neighboring coordinate values are calculated, and the number of color data to be output is determined. And a neighborhood value calculation means having a user interface.
[0031]
Further, the fifth color data conversion apparatus according to the present invention is the same as the first color data conversion apparatus, that is, input designation value recording means, output designation value recording means, designation value determining means, input means, and output means. In addition to the input specified value specifying means, the output specified value selecting means, and the output value calculating means, the input pixels are considered as coordinate values in the color space, the coordinate values in the vicinity thereof are calculated, and the colors of the pixels constituting the color image It has a neighborhood value calculation means having a user interface for determining the distance between the data and the output color data.
[0032]
Further, a sixth color data conversion apparatus according to the present invention is the same as the first color data conversion apparatus, that is, an input designated value recording means, an output designated value recording means, a designated value determining means, an input means, and an output means. In addition to the input specified value specifying means, the output specified value selecting means, and the neighborhood value calculating means, for the output from the output specified value determining means, the color data and the color image which are the inputs of the output specified value determining means are configured It is characterized by having output calculation means for multiplying and adding weights corresponding to the distance from the color data of the pixel to be processed.
[0033]
A seventh color data conversion apparatus according to the present invention is a color data conversion apparatus that converts color data of pixels constituting a color image, and is similar to the second color data conversion apparatus. In addition to the value recording means, the output specified value recording means, the specified value determining means, the input means, the output means, and the input specified value specifying means, the output specified values corresponding to the two input specified values specified by the input specified value specifying means and And an output calculation means for determining an output using distance information between a surface having the same distance from two input designation values and color data of pixels constituting a color image.
[0034]
An eighth color data conversion apparatus according to the present invention is a color data conversion apparatus that converts color data of pixels constituting a color image, and is similar to the second color data conversion apparatus. In addition to the value recording means, the output designated value recording means, the designated value determining means, the input means, the output means, and the input designated value specifying means, the surface having the same distance from the two input designated values and the color data of the pixels constituting the color image, If the distance is more than a certain threshold, there is an output calculation means for determining the output by a weight function such that the output is equal to the output specified value corresponding to the input specified value at which the distance is minimum. Features.
[0035]
DETAILED DESCRIPTION OF THE INVENTION
(Embodiment 1)
The first embodiment of the present invention, which is an apparatus for extracting signal information as a key from a moving image, will be described in detail below.
[0036]
FIG. 1 is a diagram illustrating a configuration of a color data conversion apparatus according to the first embodiment. In FIG. 1, 101 is an input means, 102 is a designated value determining means, 103 is an input designated value recording means, 104 is an output designated value recording means, 105 is a neighborhood value calculating means, 106 is an input designated value specifying means, and 107 is an output. Designated value selection means 108, output value calculation means 108, and 109 output means.
[0037]
The input unit 101 outputs an input moving image in units of pixels to the neighborhood value calculation unit 105, and includes a video camera, a video deck, a selector for selecting signals from the two devices, and an A / D converter. The
[0038]
Table 1 is an example of data recorded in the input specified value recording unit 103 and the output specified value recording unit 104. The input designation value recording means 103 records the coordinate value in the three-dimensional color space of a representative color when creating a key signal as an input designation value, and is composed of a rewritable memory.
[0039]
The output designation value recording means 104 records information for determining whether or not to select each input designation value as a key signal as an output designation value, and is constituted by a rewritable memory.
[0040]
[Table 1]

[0041]
The specified value determining means 102 determines an input specified value and an output specified value, and includes a moving image capture board and a personal computer equipped with a graphical user interface.
[0042]
The neighborhood value calculation means 105 considers an input pixel as a coordinate value in the color space, calculates a coordinate value in the vicinity of the input pixel, and outputs it to the input designation value specifying means 106, and includes an adder and a buffer.
[0043]
The input specified value specifying means 106 is an adder that calculates the distance in the three-dimensional color space between the coordinate value obtained by the neighborhood value calculating means 105 and all the input specified values recorded in the input specified value recording means 103. And a multiplier, and a comparator for selecting an input designation value having the closest distance.
[0044]
The output designation value selection means 107 selects an output designation value corresponding to the input designation value selected by the input designation value identification means 106 from the output designation value recording means 103 and outputs it as a value corresponding to the input coordinate value. The selector is output to the value calculation means 108.
[0045]
The output value calculation means 108 is a product-sum of outputs from the output value selection means 107, and is constituted by a buffer, a multiplier, and an adder.
[0046]
The output unit 109 displays or records the key signal output from the output value calculation unit 108 corresponding to the input pixel as image information, and includes a D / A converter, a video selector, a monitor, and a video deck. The
[0047]
The operation of the present apparatus configured as described above will be described.
First, the input image is input to the neighborhood value calculation means 105 as a coordinate value in the three-dimensional color space for each pixel. Here, it is assumed that the pixel has Yuv color space coordinates of CCIR601 recommendation having three attributes of luminance and color difference.
[0048]
Next, as shown in FIG. 2, the coordinates of six points near the point separated by δ along each axis of the Yuv are calculated by the neighborhood value computing unit 105 as shown in FIG. Input to the value specifying means 106.
[0049]
On the other hand, the designated value determination unit 102 captures and displays one frame from the input unit 101 as a still image according to an instruction from the user. Next, using the user interface, a plurality of colors in the image are extracted as colors that the user wants to select or do not want to select as key signals, and are output to the input specified value recording means 103 as input specified values. Also, each input designation value is determined to be a selection signal (255) or not (0), and the value is output to the output designation value recording means 104 as an output designation value.
[0050]
In the input specified value recording means 103 and the output specified value recording means 104, the coordinate value in the three-dimensional color space of a representative color when generating a key signal as shown in (Table 1) and its color are selected. (255) or (0) is recorded and may be prepared and recorded in advance, or may be determined and recorded by the user using the designated value determining means 102 each time. .
[0051]
FIG. 3 is a flowchart for explaining the operation of the input specified value specifying means 106 and the output specified value selecting means 107. The input designation value specifying unit 106 calculates the distance in the three-dimensional color space between the input value and all the input designation values recorded in the input designation value recording unit 103. Further, by sequentially comparing the values, an input designation value that minimizes the distance in the color space from the input value is determined and notified to the output designation value selection means 107. In response to the notification, the output designation value selection unit 107 reads the output designation value corresponding to the input designation value from the output designation value recording unit 104 and outputs it as an output value corresponding to the input value.
[0052]
FIG. 4A is a diagram showing how the three-dimensional color space is divided, and is a cross-sectional view in the uv plane where Y is constant. For the sake of display, a cross section in which Y is constant is displayed, but the actual color space division is three-dimensional. White circles in the figure correspond to the input designation values in (Table 1), and x marks correspond to the input values. Arbitrary points on the color space are divided so as to belong to the region of the input designation value that is closest in distance. A space division diagram by such a method is generally called a Voronoi Diagram. When an input value belonging to the divided area surrounded by the boundary line is given by the operation of the input specified value specifying means 106 and the output specified value selecting means 107, an output value corresponding to a white circle in the area is output. That is, the output value is converted into a value of 0 or 255 by the output designated value selection means 107.
[0053]
However, considering that the key signal is used as data for determining the degree of image editing effects such as image synthesis and color adjustment in post-processing, it is desirable that the signal itself has a gradation. Therefore, in the output value calculation means 107, the six output values corresponding to the input value and its neighboring values are multiplied and summed by the calculation of (Equation 1), and output as an 8-gradation key signal corresponding to the input value. . Here, ki is an output value of the output selection means, qi is a weighting factor, and k ^ is an output.
[0054]
[Expression 1]

[0055]
In other words, as shown in FIG. 4B, the process of smoothing the boundary of the region in the three-dimensional color space using the nearby points is performed. Then, the key signal output from the output value calculation means 108 corresponding to the input pixel is displayed or recorded as image information by the output means 109.
[0056]
In this way, the input pixel is output as a key signal having a gradation having the maximum influence of the designated value having the closest distance in the color space by this apparatus. Further, since this key signal does not use any image display space position information, it can have gradation information corresponding to the color information without losing the contour information of the image uniformly.
[0057]
Therefore, this apparatus can output a signal that the operator wants to obtain as a key signal having gradation according to the content of the image to be processed.
[0058]
In this embodiment, the color space defined for the input signal system is a YUV color space. However, RGB, YCbCr, YIQ, CMY, CMYK, L * a * b *, L * u * v *, HSV, HLS Other color spaces that are generally used for image processing may be used, or a color space that is uniquely defined according to the processed image may be used.
[0059]
In this embodiment, the distance in the color space compared by the input designation value specifying unit 106 is the distance in the three-dimensional color space, but a square value of the distance may be used. When the value of the square of the distance is used, processing for obtaining the square root in the distance calculation is not required, so that the processing cost can be greatly reduced.
[0060]
The neighborhood value calculation means 105 and the output value calculation means 108 may have a configuration for storing six data. The input specified value specifying means 106 can be obtained by sequentially outputting six output values from the buffer. It is also possible to adopt a configuration without a plurality of output designated value selection means 107.
[0061]
Also, the neighborhood value calculation means 102 of this apparatus outputs 6 points as neighborhood values, but it can be set to output 8 points, 12 points, or more data as long as the calculation cost is allowed. A user interface may be provided, and output of a key signal having more gradations can be expected by increasing the number of points.
[0062]
Further, the neighborhood value calculation means 105 may be provided with an interface that allows the user to set the distance range δ. By changing δ, the ratio of the data having the intermediate gradation of the key signal can be changed. .
[0063]
In this apparatus, the key signal is set as the output designation value. However, a coordinate value in the color space may be set as the output designation value as shown in (Table 2). In this case, an image having a posterization effect having a halftone can be obtained as an output image.
[0064]
[Table 2]

[0065]
Further, when a coordinate value in the color space is set as the output specified value, the input specified value may not be set as it is as the value of the output specified value. By setting the input specified value and output specified value to different values or values defined in different color spaces, it becomes possible to freely set the color scheme of the image obtained after the implementation of the posterization effect, The color space in which the output image is defined can also be freely selected according to the system requirements.
[0066]
In this apparatus, the input specified value recording means 103, the output specified value recording means 104, the neighborhood value calculating means 105, the input specified value specifying means 106, the output specified value selecting means 107, and the output value calculating means 108 are dedicated hardware. However, it may be realized by a personal computer and software loaded in the memory.
[0067]
(Embodiment 2)
Hereinafter, a second embodiment of the present invention, which is an apparatus for extracting signal information as a key from a still image, will be described in detail.
[0068]
FIG. 5 is a diagram showing functional blocks of the device according to the second embodiment of the present invention. In FIG. 5, 501 is a designated value determining means, 502 is an input designated value recording means, 503 is an output designated value recording means, 504 is an input designated value specifying means, 505 is an output value calculating means, 506 is an input means, and 507 is an output. Means.
[0069]
The designated value determining means 501 is used by a user to extract an input designated value from an input image or a color palette and determine an output designated value corresponding to the input designated value, and is configured by a pointing device such as a mouse and a graphical user interface. The
[0070]
(Table 3) is an example of data recorded in the input specified value recording unit 502 and the output specified value recording unit 503. The input designated value recording means 502 records the CIE L * a * b * color space coordinate values of the representative colors when creating the key signal as the input designated value. The output specified value recording means 503 is the same as that of the first embodiment, and detailed description thereof is omitted here.
[0071]
[Table 3]

[0072]
The input designation value specifying unit 504 calculates the distance in the three-dimensional color space with all the input designation values recorded in the input designation value recording unit 502, and among them, the input designation value closest to the distance is the second closest. This is to select an input specification value.
[0073]
The output value calculation means 505 uses the two input designation values from the input designation value specifying means 504 and the corresponding output designation values to indicate whether or not the input color data is a selection signal. Is calculated and output.
[0074]
The input unit 506 outputs the input still image to the input designation value specifying unit 504 in units of pixels. The output unit 507 displays or records the key signal output from the output value calculation unit 505 as image information. To do. The above functional blocks are all constituted by a personal computer and a program loaded in the memory.
[0075]
The operation of the apparatus according to this embodiment configured as described above will be described.
First, the user uses the mouse of the specified value determining means 501 to extract an input specified value from an input image or a color palette displayed on the monitor, and determines a corresponding output specified value. The designated data is recorded in the input designated value recording unit 502 and the output designated value recording unit 503.
[0076]
The input specified value and the output specified value are the same as those described in the first embodiment except that the color space coordinates of the input specified value are CIE L * a * b *, and thus description thereof is omitted.
[0077]
The input image displayed on the monitor is input to the input designation value specifying unit 504 as a coordinate value in the three-dimensional color space for each pixel. Here, it is assumed that the pixel has CIEL * a * b * color space coordinates having three attributes of luminance and color difference.
[0078]
FIG. 6A is a flowchart for explaining the operation of the input designation value specifying unit 504. Here, FIG. 6B shows the internal configuration of the block A in FIG. The designated input value specifying unit 504 calculates the distance (= √L * 2 + a * 2 + b * 2) in the three-dimensional color space between the input value and all the designated input values recorded in the designated input value recording unit 502. The Further, by sequentially comparing the values, an input designation value that minimizes the distance to the pixel and an input designation value that has the second smallest distance are determined and output to the output value calculation means 505 together with the corresponding output designation value. Is done.
[0079]
FIG. 7 is a diagram showing how the three-dimensional color space is divided, and is a cross-sectional view in the a * b * plane in which L * is constant. For the convenience of display, a cross section in which L * is constant is displayed, but the actual color space division is three-dimensional. The x mark p in the figure is a coordinate value of a certain input value in the color space, the white circle 1 is the input designation value closest to the point, and the white circle 2 is the input designation value closest to the point.
[0080]
In the output value calculation means 505, the output designation values corresponding to the white circle 1 and the white circle 2 are read from the output designation value recording means 503 and obtained by the arithmetic expressions of (Equation 2), (Equation 3), and (Equation 4). The data is output as a key signal having a gradation. Here, L12, L1p, and L2p in the equation are the distances between the points shown in the figure, and δ is an arbitrary value, and the output signal in the region at the distance δ from the boundary surface can be given a gradation. .
[0081]
FIG. 8A is a graph when the output of the region created by Voronoi division is used as a key signal as it is, and FIG. 8B is obtained by (Equation 2), (Equation 3), and (Equation 4). It is a graph showing the output to be output. Here, the horizontal axis of the figure is the distance from the white circle 1, and the vertical axis is the key signal output.
[0082]
[Expression 2]

[0083]
[Equation 3]

[0084]
[Expression 4]

[0085]
As described above, the output signal is derived from the output designation values corresponding to the white circle 1 and the white circle 2 and the distance information from the nearest boundary surface constituted by the white circle 1 and the white circle 2. In other words, the process of giving gradation to the output near the boundary of the region in the three-dimensional color space is performed.
[0086]
In this way, the input pixel is output as a key signal having a gradation having the maximum influence of the designated value having the closest distance in the color space by this apparatus. Further, since this key signal does not use any image display space position information, it can have gradation information corresponding to the color information without losing the contour information of the image uniformly.
[0087]
Therefore, the apparatus according to the present embodiment can output a signal that the operator wants to obtain as a key signal having gradation according to the content of the image to be processed.
[0088]
In this embodiment, the defined color space of the input signal system is the L * a * b * color space. However, RGB, YCbCr, YIQ, CMY, CMYK, YUV, L * u * v *, HSV, HLS Other color spaces that are generally used for image processing may be used, or a color space that is uniquely defined according to the processed image may be used. In this embodiment, the distance in the color space compared by the input designation value specifying unit 504 is the distance in the three-dimensional color space, but a square value of the distance may be used. When the value of the square of the distance is used, processing for obtaining the square root in the distance calculation is not required, so that the processing cost can be greatly reduced.
[0089]
The output value calculating means 505 may be provided with an interface that allows the user to set the distance range δ. By changing δ, the ratio of data having an intermediate gray level of the key signal can be changed. .
[0090]
In this apparatus, the key signal is set as the output designation value, but a coordinate value in the color space may be set as the output designation value as shown in (Table 4). In this case, an image having a posterization effect having a halftone can be obtained as an output image.
[0091]
[Table 4]

[0092]
Further, when a coordinate value in the color space is set as the output specified value, the input specified value may not be set as it is as the value of the output specified value. By setting the input specified value and output specified value to different values or values defined in different color spaces, it becomes possible to freely set the color scheme of the image obtained after the implementation of the posterization effect, The color space in which the output image is defined can also be freely selected according to the system requirements.
[0093]
Although the embodiment of the present invention is configured by a personal computer and a program loaded in its memory, a part or all of the functional blocks in FIG. 5 may be replaced with dedicated hardware.
[0094]
(Embodiment 3)
Hereinafter, a third embodiment of the present invention, which is a device for extracting signal information as a key from a moving image, will be described with reference to the drawings. FIG. 9 is a diagram showing a configuration of an apparatus according to the third embodiment of the present invention.
[0095]
In FIG. 9, 901 is a designated value determining means, 902 is an input designated value recording means, 903 is an output designated value recording means, 904 is an input designated value specifying means, 905 is an output designated value selecting means, 909 is an input means, and 910 is Since the output means is the same as that of the first embodiment, description thereof will be omitted.
[0096]
Reference numeral 906 denotes LUT address generation means, 907 denotes LUT recording means, and 908 denotes LUT calculation means. The LUT address generation means 906 outputs the coordinate values of 729 points obtained by sampling the three-dimensional color space coordinate values into eight equal parts for each axis as a look-up table (LUT) coordinate value to the input designated value specifying means 904, and is a counter. Consists of. The LUT recording unit 907 records 729 data output from the output designated value selection unit 905 as LUT data. The LUT calculation unit 908 outputs an output value corresponding to the pixel data of the input image by interpolation calculation using a plurality of LUT data recorded in the LUT recording unit 907.
[0097]
The operation of the present apparatus configured as described above will be described.
First, in the designated value determining unit 901, one frame from the input unit 909 is captured as a still image and displayed in accordance with an instruction from the user. Next, using the user interface, a plurality of colors in the image are extracted as colors that the user wants to select or do not want to select as key signals, and are output to the input specified value recording unit 902 as input specified values. Also, each input designation value is determined to be a selection signal (255) or not (0), and that value is output to the output designation value recording means 903 as an output designation value.
[0098]
Next, from the LUT address generation unit 906, the coordinate values of 729 points obtained by sampling the three-dimensional color space coordinate values into eight equal parts for each axis are output to the input specified value specifying unit 904 as lookup table (LUT) coordinate values. Is done. Since the input specified value specifying unit 904 and the output specified value selecting unit 905 output signals by the same operation as in the first embodiment, detailed description is omitted.
[0099]
In the LUT recording unit 907, the output from the output designation value selection unit 905 is recorded as a key signal corresponding to the coordinate values of 729 points obtained by sampling the three-dimensional color space coordinate values equally into 8 for each axis.
[0100]
The input image is input to the LUT calculation unit 908 as a coordinate value in the three-dimensional color space for each pixel. Here, it is assumed that the pixel has Yuv color space coordinates of CCIR901 recommendation having three attributes of luminance and color difference.
[0101]
Next, an output corresponding to the input coordinate value is interpolated by the LUT calculating means 908, output to the output means 910, and displayed or recorded as image information by the output means 910. Since the input unit 909 and the output unit 910 are the same as those in the first embodiment, detailed description thereof is omitted.
[0102]
Here, the interpolation calculation using the LUT will be described in detail. The LUT recording unit 907 has output values corresponding to the coordinate values of 729 lattice points corresponding to the vertices of 512 cubes obtained by equally dividing the three-dimensional color space into eight axes. The LUT reference means is a PRISM interpolation calculation using data of six neighboring points corresponding to the vertices surrounding the input value among the values of the grid points of the LUT (refer to Fastcolor processor interpolation programmable memory (PRISM), JournalofElectricalImag. 93). ) To determine the output value.
[0103]
FIG. 10A is a conceptual diagram for explaining the interpolation calculation. Although the actual color space is three-dimensional, for the sake of explanation, the horizontal axis is the color space coordinate and the vertical axis is the output value. Further, as shown in FIG. 10B, the range of data having intermediate gradations in the interpolation can be adjusted by setting the LUT to a value larger than the actual output range and clipping the interpolation result in the actual output range. it can.
[0104]
In this way, the input pixel is output as a key signal having a gradation having the maximum influence of the designated value having the closest distance in the color space by this apparatus. Further, since this key signal does not use any image display space position information, it can have gradation information corresponding to the color information without losing the contour information of the image uniformly.
[0105]
Unlike the first and second embodiments, the input designated value specifying unit 904 is a pre-computation process that does not perform processing on the input image. The effect of calculating the output value for the image can also be expected. Therefore, this apparatus can output a signal that the operator wants to obtain as a key signal having gradation according to the content of the image to be processed.
[0106]
In this embodiment, the color space defined for the input signal system is a YUV color space. However, RGB, YCbCr, YIQ, CMY, CMYK, L * a * b *, L * u * v *, HSV, HLS Other color spaces that are generally used for image processing may be used, or a color space that is uniquely defined according to the processed image may be used.
[0107]
In the present embodiment, the LUT has 729 values. However, the number may be smaller than that within an allowable range of output error, and can be practically processed by the LUT calculation means 908. A similar effect can be expected even if the number is larger than this within the range. In addition, in the interpolation calculation in the LUT calculation unit 908 of the present embodiment, the values of the six neighboring points surrounding the input pixel value are used. However, hardware that performs other interpolation calculations using the eight neighboring points may be used. .
[0108]
In this apparatus, the key signal is set as the output designation value. However, a coordinate value in the color space may be set as the output designation value as shown in (Table 2). In this case, an image having a posterization effect having a halftone can be obtained as an output image.
[0109]
Further, when a coordinate value in the color space is set as the output specified value, the input specified value may not be set as it is as the value of the output specified value. By setting the input specified value and output specified value to different values or values defined in different color spaces, it becomes possible to freely set the color scheme of the image obtained after the implementation of the posterization effect, The color space in which the output image is defined can also be freely selected according to the system requirements.
[0110]
In this apparatus, the input specified value recording unit 902, the output specified value recording unit 903, the LUT address generating unit 906, the input specified value specifying unit 904, and the output specified value selecting unit 905 use dedicated hardware. All may be realized by a personal computer and software loaded in its memory.
[0111]
【The invention's effect】
As described above, according to the present invention, the input image is output as a key signal having a gradation that is maximally influenced by the designated value having the closest distance in the color space. Further, since the gradation information of the key signal is not related to the position information in the image display space, there is an advantage that the contour information of the image is not lost uniformly.
[0112]
That is, a signal that the operator wants to obtain can be output as a key signal having gradation according to the content of the image to be processed.
[0113]
When a coordinate value in the color space is set as the output designation value, an output image having a posterization effect with a free color scheme desired by the operator can be obtained as the output image.
[0114]
In addition, an effect of calculating an output value for an input image in a fixed calculation time that does not depend on the number of designated colors can be expected by interpolation using a LUT (look-up table).
[Brief description of the drawings]
FIG. 1 is an apparatus configuration diagram of a first embodiment.
FIG. 2 is a diagram for explaining the operation of the neighborhood value calculation means of the first embodiment;
FIG. 3 is a diagram for explaining operations of an input designated value specifying unit and an output designated value selecting unit according to the first embodiment;
FIG. 4 is a division diagram of a three-dimensional color space according to the first embodiment.
FIG. 5 is a functional block configuration diagram of the second embodiment;
FIG. 6 is a diagram for explaining an operation of an input designation value specifying unit according to the second embodiment;
FIG. 7 is a division diagram of a three-dimensional color space according to the second embodiment.
FIG. 8 is a graph for explaining an output calculation according to the second embodiment;
FIG. 9 is an apparatus configuration diagram of the third embodiment.
FIG. 10 is a graph for explaining the interpolation calculation according to the third embodiment;
FIG. 11 is a block diagram of an apparatus according to the first prior art.
FIG. 12 is a graph for explaining the operation of the selection signal generation unit of the device according to the first conventional technique;
FIG. 13 is a graph for explaining the operation of the multiplication unit of the device according to the first prior art;
FIG. 14 is a block diagram of an apparatus according to the second prior art.
FIG. 15 is a graph for explaining the operation of the signal conversion unit of the apparatus according to the second prior art.
FIG. 16 is a division diagram of the three-dimensional color space of the apparatus according to the second prior art.
FIG. 17 is a diagram for explaining a three-dimensional color space of an apparatus according to the second prior art.
[Explanation of symbols]
101 Input means
102 Specified value determination means
103 Input designation value recording means
104 Output specified value recording means
105 Neighborhood value calculation means
106 Input specified value specifying means
107 Output designation value selection means
108 Output value calculation means
109 Output means
501 Specified value determination means
502 Input designation value recording means
503 Output specified value recording means
504 Input specification value specifying means
505 Output value calculation means
506 Input means
507 Output means
901 Specified value determination means
902 Input designation value recording means
903 Output specified value recording means
904 Input designation value specifying means
905 Output specified value selection means
906 LUT address generation means
907 LUT recording means
908 LUT calculation means
909 input means
910 Output means
1101 Polar coordinate converter
1102 Selection signal generator
1103 Selection signal generator
1104 Selection signal generator
1105 Multiplier
1401 Signal converter
1402 Signal converter
1403 Signal converter

Claims (4)

A color data conversion device for converting color data of pixels constituting a color image, wherein an input designated value is a reference color for dividing a three-dimensional color space used for specifying the color into predetermined regions Input designation value recording means recording as:
An output designation value recording means for recording an output designation value that is a value associated with each of the input designation values and that is to be finally output;
Given the color data of the pixels constituting the color image, neighborhood value calculation means for outputting color data at a certain distance from the color in a three-dimensional color space;
Calculating a distance in a three-dimensional color space between each of the input specified values recorded in the input value specifying means for each of the color data of the pixels constituting the color image and the color data as the output of the neighborhood value calculating means; An input specification value specifying means for specifying an input specification value that minimizes the distance by comparing;
An output designation value selecting means for reading out and outputting an output designation value corresponding to the input designation value identified by the input designation value identification means from the output designation value recording means;
Output calculation means for multiplying and outputting the output specified by the output designated value selection means for each of the color data of the pixels constituting the color image and the color data output from the neighborhood value calculation means. A color data converter characterized by the above.   2. The color data conversion apparatus according to claim 1, wherein the neighborhood value calculation means includes a user interface for determining the number of color data to be output.   2. The color data conversion apparatus according to claim 1, wherein the neighborhood value calculating means includes a user interface for determining a distance between color data of pixels constituting a color image and output color data.   The output calculation means assigns a weight corresponding to the distance between the color data input to the output designated value determining means and the color data of the pixels constituting the color image to the output from the output designated value determining means. The color data conversion apparatus according to claim 1, wherein product summation is performed.

Images