JP3972192B2 - Image processing apparatus and image processing method - Google Patents

Description

【００９４】
標準偏差は輝度分布の広がり量に対応するものであるが、広がり量を表す意味では分散を利用してもよい。
【００９５】
以上の処理が分布検出処理に該当し、次に、このようにして求めた輝度分布の広がり量である標準偏差σに基づいて画像データの変換を行なう輝度変換処理について説明する。なお、上述したようにステップＳ１０６にて非拡大処理を実行した場合には、ステップＳ２０２にて所定のフラグを参照してそれを検知し、以下の処理を行うことなく当該画像処理を終了する。
【００９６】
輝度変換処理では、輝度分布の広がり量に基づいて分布密度の大きい範囲に多くの階調数を与えつつ分布密度の小さい範囲に少ない階調数を割り当てる。ここで、分布密度の大きい範囲に多くの階調数を与えつつ分布密度の小さい範囲に少ない階調数を割り当てるパターンについて説明する。変換前の輝度ｙ（入力）と変換先の輝度Ｙ（出力）が、
Ｙ＝ｙ …（６）
というように正比例の関係にある場合、図１５に示すように、再現可能範囲内の中央部分を基準として、変換前に割り当てられている階調範囲ｒ0と変換後に割り当てられる階調範囲Ｒ0 は一致している。しかしながら、図１６に示すように入出力の対応関係がいわゆるＳ字カーブとなると変換前に割り当てられている階調範囲ｒ0 に対して変換後に割り当てられる階調範囲Ｒ1 ，Ｒ2 は大きくなり、割り当てられた階調数が多くなったことになる。一方、入力における低輝度側と高輝度側における階調範囲ｒ0を外れた範囲についていえば、変換後に割り当てられる階調範囲は少なくなったことになる。
【００９７】
すなわち、このような対応関係こそ、分布密度の大きい範囲に多くの階調数を与えつつ分布密度の小さい範囲に少ない階調数を割り当てることを意味する。ここにおいて、この対応関係を実現する具体的な割り当て手法は各種のものが可能である。図１７は階調範囲の中心位置ｙmid から高輝度側の階調範囲上端までの領域に対してγ＜１のγ補正を施したものであり、変換前の中心位置ｙmid から上方側四分点ｙq3までの階調範囲ｒに対する変換後の階調範囲Ｒは拡大している。同様に図１８は階調範囲の中心位置ｙmid から低輝度側の階調範囲下端までの領域に対してγ＞１のγ補正を施したものであり、変換前の中心位置ｙmid から下方側四分点ｙq1までの階調範囲ｒに対する変換後の階調範囲Ｒは拡大している。
【００９８】
一方、これらの場合において、変換前の階調範囲ｒに対する変換後の階調範囲Ｒの比はγの与え方によって変化してくる。本実施形態においては、かかる比を輝度分布の広がり量である標準偏差σに基づいて制御している。すなわち、階調範囲の中心位置ｙmid を「１２８」として、この中心位置ｙmid 以下ではγ１を与えるとともに、中心位置ｙmid より大きい範囲ではγ２を与えるものとすると、
ｙ≦１２８では、
γ１＝（σstd_limit／σ）**ａ …（７）
ｙ＞１２８では、
γ２＝（σ／σstd_limit）**ａ …（８）
とし、ステップＳ２０４にてこれらのパラメータ演算を実行する。なお、上述したようにこのパラメータ演算こそ階調数割当手段を構成する。ここにおいて、σstd_limitとａは変換結果を考慮して実験的に求めて与えたパラメータであり、本実施形態においてはσstd_limitを「１２８」とするとともにａを「０．１」としている。標準偏差σは概して「１２８」よりも小さな値となるからこれらの関係式では標準偏差σが大きいと、γ２とγ１はそれぞれ「１」に近づくことになり、Ｓ字カーブの傾斜は緩やかになる。これは、広がり量が大きいときに中心位置ｙmid を中心とする階調範囲ｒに対して変換先の階調範囲Ｒはさほど広くならないことを意味しており、より具体的には画像データの輝度が広く分布しているときには輝度範囲を拡大するような変換を行わないことを意味する。これに対して、標準偏差σが小さいと、γ２とγ１はそれぞれ「１」から離れることになり、Ｓ字カーブの傾斜は急になる。これは、広がり量が小さいときに中心位置ｙmid を中心とする階調範囲ｒに対して変換先の階調範囲Ｒが広く拡大されることを意味しており、より具体的には画像データの輝度が狭い範囲にしか分布していないときには輝度範囲を拡大させる変換を行なうことを意味する。
【００９９】
この例のように、階調範囲を低輝度側と高輝度側との二つに分けるとともにそれぞれにγ補正を掛けるにあたり、γが互いに逆数となる関係を与えるようにすると、低輝度側と高輝度側との接続点で滑らかに接続することになり、良好なＳ字カーブを与えることができる。むろん、低輝度側と高輝度側とによってγ補正の極性を変えること自体がＳ字カーブの対応関係を形成し、変換前の輝度ｙに対して変換後の輝度Ｙを大きく変化させることができるようになる。
【０１００】
本実施形態においては、Ｓ字カーブの対応関係をγ補正によって成立させているが、図１９には階調範囲の中心位置ｙmid の前後においてＹ＝ｆ・ｙ＋ｇなる線形の対応関係で実現する例を示している。この例では下方側四分点ｙq1以下と上方側四分点ｙq3以上の領域で再び線形の対応関係を形成している。この例においても標準偏差σが大きいときに傾斜ｆを「１」に近づけ、標準偏差σが小さいときに傾斜ｆを「１」よりも大きくなるように対応づければよい。むろん、この場合においては対応関係の急激な変化を防止するため、図２０に示すように二つの対応直線を滑らかに接続するようにしても良い。
【０１０１】
一方、輝度分布が階調範囲の中心に収まっている場合は上述したようなγ１，γ２の設定だけで良好な輝度変換が可能となるが、図２１に示すように、輝度分布のメジアンｙMeがやや低輝度側に寄っていたり、図２２に示すように、輝度分布のメジアンｙMeがやや高輝度側に寄っていたりする場合もある。
【０１０２】
これに対し、図２３に示す例では全範囲に対してγ＜１のγ補正を施しており、この場合は変換前の低輝度側の階調範囲ｒに対する変換後の階調範囲Ｒは拡大しているし、図２４に示す例では全範囲に対してγ＞１のγ補正を施しており、この場合は変換前の高輝度側の階調範囲ｒに対する変換後の階調範囲Ｒは拡大している。
【０１０３】
従って、ステップＳ１１４にて求めたメジアンｙMeと階調範囲の中央位置である「１２８」との大小関係を比較し、図２１に示すようにメジアンｙMeが低輝度側の領域に偏っているようであれば、図２３に示すような全範囲に対してγ＜１のγ補正を施すことにより、変換前に分布密度の高かった低輝度側の階調範囲ｒが拡大することになる。また、図２２に示すようにメジアンｙMeが高輝度側の領域に偏っているようであれば、図２４に示すような全範囲に対してγ＞１のγ補正を施すことにより、変換前に分布密度の高かった高輝度側の階調範囲ｒが拡大することになる。このようにして必ずしもＳ字カーブの対応関係とすることなく階調数の割り当ての拡大や縮小は可能である。むろん、これらの場合においても、γはγ＜１とするかγ＞１とするかに応じて（７）式や（８）式に基づいて決定すればよい。
【０１０４】
さらには、図２５に示す例では、階調「０」、下方側四分点ｙq1、中心位置ｙmid 、上方側四分点ｙq3、階調「２５５」という五点を基準点としつつ、階調「０」と中心位置ｙmid と階調「２５５」に対してはＹ＝ｙとしつつ、下方側四分点ｙq1と上方側四分点ｙq3における変換点を標準偏差に基づいて決定する。そして、これらの五点を結ぶ対応関係をスプライン補間演算やニュートン補間で求めるようにしてもよい。むろん、中心位置ｙmid から下方側の三点や上方側の三点をそれぞれスプライン補間演算やニュートン補間で求めるようにしてもよい。
【０１０５】
ところで、変換前の階調範囲に対して余りにも大きな階調範囲を割り当てるとすると、却って好ましくないことになる場合もある。夕方のような薄暮の状態では最も明るい部分から暗い部分までのコントラストの幅が狭くて当然であるのに、この画像についてコントラストを大きく拡大しようとする結果、昼間の画像のように変換されてしまいかねない。このような変換は希望されないので、拡大率には制限を設けていおき、γ１，γ２ともに制限する。例えば、γ２＜０．７となっても、γ２＝０．７とし、γ１＞１．３となっても、γ１＝１．３とする。
【０１０６】
以上のようにして、本実施形態ではパラメータγ１，γ２を得ることができ、ステップＳ２０４を終了する。なお、このような階調数の割り当ての変更は、言葉を換えると量子化ビット数の割り当てを変更するともいえる。
【０１０７】
ところで、輝度の変換時に、毎回、γ補正による演算を実行するのは非合理的である。というのは、輝度ｙの取りうる範囲が「０」〜「２５５」でしかあり得ないため、予め輝度ｙが取りうる全ての値に対応して変換後の輝度Ｙを求めておくことも可能だからである。従って、ステップＳ２０６にてこの対応関係を求め、図２６に示すようなテーブルとして記憶しておく。
【０１０８】
ここにおいて、対応関係の具体的な演算は次のようにする。
【０１０９】
ｙ≦１２８では、
Ｙ＝１２８＊（ｙ／１２８）**γ１ …（９）
ｙ＞１２８では、
Ｙ＝１２８＊｛（ｙ−１２８）／１２８｝**γ２＋１２８
…（１０）
しかしながら、メジアンｙMeに基づいてγ補正の変換点を変えることも可能である。すなわち、
ｙ≦ｙMeでは、
Ｙ＝ｙMe＊（ｙ／ｙMe）**γ１ …（１１）
ｙ＞ｙMeでは、
Ｙ＝ｙMe＊｛（ｙ−ｙMe）／ｙMe｝**γ２＋ｙMe …（１２）
とすれば、メジアンｙMeを中心として高輝度側と低輝度側とで輝度変換の極性を反転させたＳ字カーブとなり分布密度の高い辺りを中心に両側に階調数を多く割り当てることができ、全体の明るさにさほど影響を与えることなく変換することができるようになるなどの効果がある。
【０１１０】
このようにして演算した変換テーブルが形成されたところで画像データを変更することが可能になる。
【０１１１】
最後に、ステップＳ２０８にて画像データの変換を行う。ここまでは輝度を変換するための対応関係を求めてきており、例えば、ＲＧＢ座標軸における成分値（Ｒp，Ｇp，Ｂp ）についての変換関係ではなかった。しかしながら、（２）式の変換式は、このＲＧＢの成分値（Ｒp，Ｇp，Ｂp ）との対応関係においても当てはめることができる。また、輝度ｙ，Ｙが階調「０」〜階調「２５５」であるのに対応してＲＧＢの各成分値（ｒ，ｇ，ｂ），（Ｒ，Ｇ，Ｂ ）も同じ範囲となっており、上述した輝度ｙ，Ｙの変換テーブルをそのまま利用すればよいといえる。
【０１１２】
従って、ステップＳ２０８では全画素の画像データ（ｒ，ｇ，ｂ）について図２６に示す変換テーブルを参照し、変換後の画像データ（Ｒ，Ｇ，Ｂ ）を得るという処理を繰り返すことになる。
【０１１３】
次に、上記構成からなる本実施形態の動作を順を追って説明する。
【０１１４】
スキャナ１１などで写真を撮像したとすると、同写真をＲＧＢの階調データで表した画像データがコンピュータ２１に取り込まれ、ＣＰＵは図５及び図６に示す画像処理のプログラムを実行して画像データのコントラストを強調する処理を実行する。
【０１１５】
まず、ステップＳ１０２では画像データを所定の誤差内となる範囲で間引き、選択した画素についての輝度ｙを求めて分布を取る。このままの分布を使用することはできず、まず、画像が白黒のような二値画像でないかステップＳ１０４にて判断するとともに、ステップＳ１０８では自然画か否かを判断する。二値画像である場合や自然画でない場合などを除き、ステップＳ１１０では画像データに枠部がないか判断し、枠部があれば除いて得られた輝度分布についてステップＳ１１４にて標準偏差σを求める。本実施形態においては、標準偏差σを求めるだけでもよいが、γ補正の極性変換点を輝度分布に合わせて変更する場合にはメジアンｙMeを求めておく。
【０１１６】
輝度分布の標準偏差σが得られたら、
ｙ≦１２８では、
γ１＝（σstd_limit／σ）**ａ …（７）
Ｙ＝１２８＊（ｙ／１２８）**γ１ …（９）
ｙ＞１２８では、
γ２（σ／σstd_limit）**ａ …（８）
Ｙ＝１２８＊｛（ｙ−１２８）／１２８｝**γ２＋１２８
…（１０）
なる関係式より、ステップＳ２０４にてパラメータγ１，γ２を求めるとともに、ステップＳ２０６では輝度ｙから輝度Ｙへの変換関係をテーブルに記憶する。そして、ステップＳ２０８にて完成した変換テーブルを参照して全画素についての画像データを変換する。
【０１１７】
むろん、上述したように二値画像や自然画でない場合においてはかかる画像処理は行われないが、本発明の画像処理が行われた場合には、写真の状態では非常にコントラストが弱かったにもかかわらず、輝度の範囲を広げるように補正することにより、明暗がはっきりして鮮明な画像を得られるようになる。
【０１１８】
なお、上述した実施形態においては、γ１やγ２を求める際のパラメータσstd_limit，ａを一定としているが、コンピュータ２１上では所定のＧＵＩを介してユーザーが選択できるようにしても良い。また、ユーザーが画像データの一部を指定して当該範囲内でのみかかるコントラストの強調処理を実行するようにすることも可能である。
【０１１９】
このように、ステップＳ１０２で間引きするなどしながら画像データの画素について輝度ｙの分布を求めた後、ステップＳ１１４にてその輝度分布の広がり量に対応する標準偏差σを求め、同標準偏差σに基づいてステップＳ２０４にてＳ字カーブの対応関係を形成するためのγ補正のパラメータ（γ１，γ２）を演算するようにしているため、ステップＳ２０８にて画像データ変換して変換元の輝度ｙにおける分布密度の高い領域には多くの階調数を割り当てるとともに分布密度の少ない領域には少ない階調数を割り当てるといった作業を自動化し、非熟練者でも容易にコントラストの強調を行うことができるようになる。
【図面の簡単な説明】
【図１】 本発明の一実施形態にかかる画像処理装置が適用される画像処理システムのブロック図である。
【図２】 同画像処理装置の具体的ハードウェア構成例を示すブロック図である。
【図３】 本発明の画像処理装置の他の適用例を示す概略ブロック図である。
【図４】 本発明の画像処理装置の他の適用例を示す概略ブロック図である。
【図５】 本発明の画像処理装置における輝度の分布検出処理部分を示すフローチャートである。
【図６】 本発明の画像処理装置における輝度変換処理部分を示すフローチャートである。
【図７】 変換元の画像における座標を示す図である。
【図８】 サンプリング周期を示す図である。
【図９】 サンプリング画素数を示す図である。
【図１０】 変換元の画像とサンプリングされる画素の関係を示す図である。
【図１１】 白黒の画像を示す図である。
【図１２】 白黒の画像の輝度分布を示す図である。
【図１３】 枠部のある画像を示す図である。
【図１４】 枠部のある画像の輝度分布を示す図である。
【図１５】 階調数の割り当てを変更しない場合の変換前の輝度と変換後の輝度との関係を示すグラフである。
【図１６】 変換前の輝度と変換後の輝度とがγ補正によってＳ字カーブの対応関係となるグラフである。
【図１７】 高輝度側の領域でγ＜１のγ補正をかけた場合のグラフである。
【図１８】 低輝度側の領域でγ＞１のγ補正をかけた場合のグラフである。
【図１９】 変換前の輝度と変換後の輝度とが線形の変換でＳ字カーブの対応関係となるグラフである。
【図２０】 同対応関係で変換特性の変換点で滑らかに変換されるようにする場合の変形例である。
【図２１】 メジアンが低輝度側に寄っている場合の輝度分布を示す図である。
【図２２】 メジアンが高輝度側に寄っている場合の輝度分布を示す図である。
【図２３】 全領域でγ＜１のγ補正をかけた場合のグラフである。
【図２４】 全領域でγ＞１のγ補正をかけた場合のグラフである。
【図２５】 特定した変換点を補間法で接続する場合のグラフである。
【図２６】 輝度分布を拡大する際の変換テーブルを示す図である。
【符号の説明】
１０…画像入力装置
１１…スキャナ
１１ｂ…スキャナ
１２…デジタルスチルカメラ
１２ａ…デジタルスチルカメラ
１２ｂ…デジタルスチルカメラ
１３ｂ…モデム
２０…画像処理装置
２１…コンピュータ
２２…ハードディスク
３０…画像出力装置
３１…プリンタ
３１ａ…プリンタ
３１ｂ…プリンタ
３２…ディスプレイ
３２ａ…ディスプレイ[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and an image processing method, and more particularly to an image processing apparatus and an image processing method for enhancing the contrast of an image in which pixels are arranged in a vertical direction and a horizontal direction.
[0002]
[Prior art]
Image data taken with a scanner or taken with a digital still camera is configured in a dot matrix. Although it is sufficient if the captured image data is in a preferable state, there are cases where it is desired to make corrections according to preferences using the characteristics of digital images. In particular, when the contrast is weak, it is difficult to correct a photo using a conventional color film. Such a photo must be re-taken, but if it is a digital image, use image correction software. The contrast can be enhanced.
[0003]
That is, conventionally, image correction software has been used to correct such digital image data, and the operator can perform various corrections by being started on the computer. For example, the contrast enhancement as described above can be performed by performing a brightness enlargement operation. More specifically, by changing the correspondence between the luminance of the conversion source and the luminance of the conversion destination from a directly proportional state to a moderately distorted S-shape, the luminance having a certain width at the conversion source is converted. It can be associated with a wider width first.
[0004]
[Problems to be solved by the invention]
In the above-described conventional image processing apparatus, although it is possible to enhance the contrast by an operator's operation, the necessary parameter values can only be given by intuition, and for those who are not used to processing image data, as expected There was a problem of operability, such as the contrast could not be emphasized, and the twilight photos became like noon.
Further, when sampling and thinning out pixels of an image in which pixels are arranged in the vertical direction and the horizontal direction, a width >> height as shown in a bitmap image having the number of pixels of (width) × (height) shown in FIG. If the sampling period is determined with the longer width, only the upper and lower lines of pixels are extracted in the vertical direction as shown in FIG. It might be.
[0005]
The present invention relates to an image processing apparatus and an image which can perform processing for enhancing the contrast of an image by eliminating the bias of extraction points of an image when sampling and thinning out pixels of an image in which pixels are arranged in a vertical direction and a horizontal direction The purpose is to provide a processing method.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, an invention according to claim 1 is an image processing apparatus for performing a process of enhancing contrast of an image in which pixels are arranged in a vertical direction and a horizontal direction. When sampling and thinning in the horizontal direction, thinning is performed so that the number of extraction based on the number of pixels in the smaller direction in the vertical direction and the horizontal direction is ensured, and for each pixel after thinning A luminance distribution of the image is obtained from a luminance equivalent value of a predetermined number of gradations, an amount representing a variation in the luminance equivalent value for each pixel is calculated from the obtained luminance distribution, and the contrast of the image is calculated based on the amount representing the variation. The processing for emphasizing is performed.
[0007]
In the invention configured as described above, when the pixels of the image are sampled and thinned out in the vertical direction and the horizontal direction, the number of pixels in the smaller direction in the vertical direction and the horizontal direction is reduced. Thinning is performed so that the number of extractions based on the number is secured. From the luminance equivalent value of the predetermined number of gradations for each pixel after thinning, the luminance distribution of the image is obtained. From the obtained luminance distribution, an amount representing the variation of the luminance equivalent value for each pixel is calculated. A process of enhancing the contrast of the image based on the amount representing the variation is performed.
Incidentally, the number of gradations may be assigned based on the amount representing the variation of the luminance equivalent value.
For example, when the luminance distribution is concentrated in a certain width, if the number of gradations is given so as to widen the width, the contrast is enhanced and the contrast becomes clear.
At this time, there is no problem even if the number of gradations assigned to the portion where the luminance distribution is not concentrated is reduced.
For this reason, according to a second aspect of the present invention, in the image processing apparatus according to the first aspect, a small number of gradations is assigned to a gradation range having a small distribution density based on an amount representing variation in the luminance equivalent value. In this configuration, a large number of gradations are assigned to a gradation range having a large distribution density, and the luminance equivalent value of each pixel is converted in accordance with the change in the assignment of the number of gradations.
[0008]
Of course, if the luminance distribution is not concentrated, it is not necessary to change the number of gradations.
[0009]
Such an operation for obtaining the amount representing the variation in the luminance equivalent value is detected based on the intuition that the operator has performed in the conventional operation, and the level corresponding to the amount representing the variation in the luminance equivalent value is determined. The operation of assigning a large number of logarithms corresponds to setting the parameter given by the operator.
[0010]
Here, various methods are possible for obtaining the amount representing the variation of the luminance equivalent value from the luminance distribution. It may be a value that corresponds to an amount that represents a known variation in luminance equivalent value, or a measure that reduces the amount of calculation to obtain an amount that represents a variation in luminance equivalent value. good. The invention according to claim 3 is the image processing apparatus according to claim 1 or 2, wherein the amount representing the variation in the luminance equivalent value is the amount representing the variation in the luminance equivalent value. Using the value corresponding to the standard deviation, when the standard deviation of the luminance distribution is large, the input / output ratio is reduced to equalize the number of gradations, and when the standard deviation is small, the input / output ratio is increased. Thus, the change in the number of gradations is enlarged.
[0011]
In the invention configured as described above, the standard deviation represents the variation amount of each element with respect to the average value. Therefore, if the variation is large, the luminance conversion is not necessary, and the input / output ratio is increased to increase the scale. There is no need for a transformation that enlarges the change in key assignment. That is, since the luminance distribution is not concentrated, it is not necessary to increase or decrease the number of gradations. However, a small standard deviation means that there is little variation, in which case the luminance distribution is concentrated, so contrast is enhanced by giving a large number of gradations to a large distribution density range. Let
[0012]
Of course, since the standard deviation is used to represent variation, it is synonymous with variance, and the calculation process does not necessarily require division by the number of samples.
[0013]
Giving more or less gradations generally corresponds to the input / output ratio. If more gradations are given, the number of gradations is larger than that given on the input side. The number of gradations is given and the gradation is expanded within the range, which is an enhancement of contrast. On the other hand, since the total number of gradations cannot be increased or decreased, if a large number of gradations is given to a certain range, it is impossible to give a large number of gradations to the remaining range as a reaction. The number of gradations smaller than the number of gradations is given, and the contrast is narrowed within the range.
[0014]
It is possible to determine how to convert the number of gradations for each gradation, but as a simpler example, the invention according to claim 4 is the invention according to claim 3. In the described image processing apparatus, the value of γ is set so that the amount of change due to γ correction is small when the luminance distribution is subjected to γ correction and the standard deviation is large, and the amount of change in γ correction is small when the standard deviation is small. The value of γ is set so as to increase.
[0015]
In the invention configured as described above, a correspondence relationship of being convex upward or convex downward is given by the method of γ correction. And a small number of gradations is assigned to the second half, and if it is convex downward, a small number of gradations is assigned to the first half and a gradation is assigned to the second half. A lot will be allocated.
The magnitude of the allocated amount can be adjusted as appropriate by the amount of change due to γ correction, and corresponds to the value of γ given.
[0016]
As one correspondence to this γ correction, the invention according to claim 5 is the image processing apparatus according to claim 4, wherein the approximate center position of the brightness distribution is obtained and the high brightness side and the low brightness side are determined based on the approximate center position. By reversing the polarity of the luminance conversion on the luminance side, the luminance conversion of the approximate S-shaped curve is performed according to the relationship between the input and the output.
[0017]
In the invention configured as described above, in order to reverse the polarity of luminance conversion by γ correction, by performing γ correction of γ <1 on the high luminance side and γ correction of γ> 1 on the low luminance side, The downwardly convex curve and the upwardly convex curve are continuous, and the so-called S-curve input / output relationship is established. In this case, a large number of gradations are assigned to the central part, so if a continuous point of two curves is brought to the approximate center of the luminance distribution, a large number of gradations will be included in the range where the luminance distribution is to be expanded. Can be assigned.
[0018]
In converting the luminance by various methods, the invention according to claim 6 is the image processing apparatus according to any one of claims 1 to 5, wherein the conversion destination is within the range that the conversion source luminance can take. Luminance is calculated and stored, and this correspondence is called up and converted at the time of conversion.
[0019]
Although it is not impossible to calculate the luminance every time based on the conversion formula, the range of values that the luminance distribution can take is determined. For this reason, if the brightness of the conversion destination is calculated and stored in advance based on the brightness of the conversion source, it is possible to perform the conversion simply by invoking the correspondence at the time of conversion.
[0020]
In converting the luminance, the image data may be included as luminance data, or may be indirect only including luminance data. Of course, if direct luminance data is included, it may be converted, and even if it is indirect luminance data, it may be converted into luminance data before performing predetermined luminance conversion. However, the luminance conversion does not have to be very accurate, and it can be said that it is sufficient to know roughly.
[0021]
In that sense, since strict accuracy is not required, the invention according to claim 7 is the image processing apparatus according to any one of claims 1 to 6, wherein the image data corresponds to the luminance. In the case of being expressed by a plurality of component values, the luminance is calculated by linear addition of the same component values.
[0022]
When the image data is represented by so-called RGB (red, green, and blue) gradation data, it can be said that each component value for red, green, and blue corresponds to the luminance. For this reason, it can be said that linear addition of the same component values sufficiently represents luminance, and can be an extremely easy conversion method.
[0023]
As the luminance for each pixel is obtained, the luminance distribution as an image is not necessarily obtained for all the pixels of the image data, and the luminance distribution is obtained by thinning out the image data corresponding to a predetermined extraction rate.
[0024]
For the purpose of obtaining the distribution, it is possible to obtain a luminance distribution with a certain degree of certainty according to the extraction rate even if thinning is performed at a predetermined extraction rate without obtaining luminance for all pixels.
[0025]
Here, the predetermined number of extractions is ensured on the short side in the vertical and horizontal ranges.
[0026]
Since the image is flat, the image data is also naturally distributed in the vertical and horizontal directions, but in determining a certain extraction rate, by ensuring a certain number of extractions at least on the short side, the extraction rate is reduced. The certainty corresponding to it is held.
[0027]
The invention according to claim 8 is characterized in that, in the image processing apparatus according to any one of claims 1 to 7, a limit is set to the degree of contrast enhancement.
[0028]
It may be natural that the contrast is narrow. For example, in the evening scenery, it is natural that the width of the luminance distribution is narrow, and if it is enlarged more than necessary, it becomes a daytime scenery. A similar example is possible in other cases, and this phenomenon is avoided by setting a limit on the expansion range of the luminance distribution.
[0029]
Furthermore, the invention according to claim 9 is the image processing apparatus according to any one of claims 1 to 8, wherein the binary image data is determined based on the luminance distribution, and if the binary image data is used. The configuration is such that no contrast enhancement is performed.
[0030]
Since it can be said that there is no substantial luminance distribution for a binary image, contrast enhancement is not performed when binary image data is determined from the luminance distribution.
[0031]
Since the binary image data can also have a certain color, it can have two luminance levels corresponding to the presence or absence of the color. Although it is possible to determine whether or not the brightness of the color, the invention according to claim 10 is reproduced in the image processing apparatus according to claim 9 in response to a case where there is no information suggesting the brightness. When the luminance distribution is concentrated at both ends within the possible range, it is determined that the binary image data is black and white.
[0032]
That is, for a black and white image, it can be determined that the luminance distribution is concentrated at both ends within the reproducible range.
[0033]
Furthermore, the invention according to claim 11 is the image processing device according to any one of claims 1 to 10, wherein the frame portion of the image data is determined based on the protruding luminance distribution and the frame portion is provided. For example, the frame data is not used for contrast enhancement.
[0034]
When an image is processed, it can frequently occur that it has a frame. If it exists as a single-color frame, only the luminance distribution corresponding to that color naturally protrudes. Therefore, if such a protruding luminance distribution is used as a criterion for enhancement determination, effective determination cannot be made. Therefore, it is determined that the frame portion is not used for contrast enhancement.
[0035]
Further, as an example, the invention according to claim 12 determines that the luminance distribution concentrated on the end within the reproducible range is the frame in the image processing apparatus according to claim 11. That is the configuration.
[0036]
A white frame or a black frame is frequently used and may be generated depending on a trimming result, and corresponds to an end portion within a reproducible range. Therefore, the luminance distribution concentrated on the end portion is determined as the frame portion.
[0037]
By the way, the invention according to claim 13 is configured such that the image processing apparatus according to any one of claims 1 to 12 does not perform contrast enhancement when the image data is not a natural image.
[0038]
The narrowness of contrast tends to be a problem for natural images such as photographs, and it can be said that there is almost no need for such things as business graphs. Conversely, modifying something like a business graph can lead to a difference from the creator's image. Therefore, the luminance distribution is expanded only in the case of such a natural image.
[0039]
As an example of determining whether or not the image is a natural image, the invention according to claim 14 is the image processing device according to claim 13, wherein the image data is determined not to be a natural image when the luminance distribution exists in a spectral form. The image processing apparatus includes a natural image determination unit that performs the operation.
[0040]
It can be said that the brightness distribution has a smooth width as a feature of the natural image. Therefore, if the luminance distribution appears in a line spectrum, it can be determined that the image is not a natural image. In the invention according to claim 14 configured as described above, the natural image determination means determines the state of the luminance distribution, determines that the image data is not a natural image when it exists in a line spectrum, and thereby the luminance distribution Will not be expanded.
[0041]
Furthermore, the invention according to claim 15 is an image processing method for performing a process of enhancing the contrast of an image in which pixels are arranged in the vertical direction and the horizontal direction, and sampling the pixels of the image in the vertical direction and the horizontal direction. When thinning out, thinning is performed so that the number of extraction based on the number of pixels in the direction of the smaller number of pixels in the vertical direction and the horizontal direction is ensured, and a predetermined number of gradations for each pixel after thinning is performed. A process of obtaining a luminance distribution of the image from the luminance equivalent value, calculating an amount representing a variation of the luminance equivalent value for each pixel from the obtained luminance distribution, and enhancing a contrast of the image based on the amount representing the variation. The configuration is to be performed.
[0042]
That is, it is not necessarily limited to a substantial apparatus, and there is no difference that the method is effective.
[0043]
By the way, such an image processing apparatus may exist independently, or may be used in a state of being incorporated in a certain device, and includes various aspects as an idea of the invention. Further, it can be changed as appropriate, such as software or hardware.
[0044]
As an example, even in a printer driver that converts image data corresponding to printing ink based on image data of a plurality of input pixels and prints it on a predetermined color printer, attention is paid to each pixel as a processing target. While moving the pixel, the luminance equivalent value of the target pixel is aggregated to obtain the luminance distribution, and the spread amount of the luminance distribution corresponding to the variation amount of each element is calculated based on the obtained luminance distribution. It is possible to adopt a configuration in which the contrast is enhanced by performing conversion processing between the input and the output so that a correspondence relationship in which a large number of gradations is assigned to a gradation range having a large distribution density based on the amount.
[0045]
In other words, the printer driver converts the input image data corresponding to the printing ink. At this time, while moving the pixel of interest with each pixel as a processing target, the luminance equivalent value of the pixel of interest is aggregated to obtain the luminance distribution. Based on the obtained luminance distribution, the spread amount of the luminance distribution corresponding to the variation amount of each element is calculated, and a large number of gradations is assigned to the gradation range having a large distribution density based on the spread amount. The input image is converted and printed so as to enhance the contrast by performing a conversion process between the input and the output so as to have a correspondence relationship.
[0046]
In the case of software for an image processing apparatus as an embodiment of the idea of the invention, it naturally exists on a recording medium on which such software is recorded and must be used. Of course, the recording medium may be a magnetic recording medium, a magneto-optical recording medium, or any recording medium that will be developed in the future. In addition, the duplication stages such as the primary duplication product and the secondary duplication product are equivalent without any question.
[0047]
Further, even when a part is software and a part is realized by hardware, the idea of the invention is not completely different, and a part is stored on a recording medium and is appropriately changed as necessary. It may be in the form of being read. Furthermore, it goes without saying that the present invention can also be applied to an image processing apparatus such as a color facsimile machine or a color copier.
[0048]
【The invention's effect】
As described above, according to the first and second aspects of the invention, since the degree of contrast enhancement is automatically determined from the amount representing the variation of the luminance equivalent value, even a non-expert can easily adjust the contrast appropriately. It is possible to enhance the contrast of the image by eliminating the bias of the extraction points of the image when sampling and thinning out the pixels of the image in which the pixels are arranged in the vertical direction and the horizontal direction. A possible image processing apparatus can be provided.
[0049]
According to the invention of claim 3, the input / output ratio is only determined based on the well-known standard deviation, and the configuration becomes easy.
[0050]
Furthermore, according to the fourth aspect of the present invention, since the conversion is performed by γ correction, the configuration becomes easy.
[0051]
Furthermore, according to the fifth aspect of the present invention, it is possible to efficiently perform distributed conversion when it is desired to enhance contrast by performing luminance conversion based on the correspondence relationship of the approximate S-curve based on the approximate center position of the luminance distribution. .
[0052]
Furthermore, according to the sixth aspect of the invention, conversion can be facilitated.
[0053]
Furthermore, according to the seventh aspect of the present invention, the luminance can be easily obtained with a necessary and sufficient degree of accuracy.
[0054]
Furthermore, according to the eighth aspect of the present invention, it is possible to prevent the contrast of the image from being changed too much by changing the atmosphere of the image.
[0055]
Furthermore, according to the ninth aspect of the present invention, it is possible to easily determine a condition that does not require contrast enhancement and prevent enhancement.
[0056]
Furthermore, according to the tenth aspect of the present invention, it is possible to efficiently determine a monochrome image having a high frequency.
[0057]
Furthermore, according to the invention concerning Claim 11, it can prevent that a process becomes inaccurate by the brightness | luminance of the frame part which tends to appear in an image.
[0058]
Furthermore, according to the twelfth aspect of the present invention, it is possible to easily determine a black and white frame portion having a high frequency.
[0059]
Furthermore, according to the invention of claim 13, it can be performed only in the case of a natural image that requires contrast enhancement. Further, according to the invention of claim 14, whether or not the image is a natural image can be determined. It can be easily determined.
[0060]
According to the fifteenth aspect of the present invention, since the degree of contrast enhancement is automatically determined from the amount representing the variation in the luminance equivalent value, even an unskilled person can easily enhance the contrast by an appropriate amount. There is also provided an image processing method capable of enhancing the contrast of an image by eliminating the bias of the extraction points of the image when sampling and thinning out the pixels of the image in which the pixels are arranged in the vertical direction and the horizontal direction. be able to.
[0061]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0062]
FIG. 1 is a block diagram illustrating an image processing system according to an embodiment of the present invention, and FIG. 2 is a block diagram illustrating a specific hardware configuration example.
[0063]
In the figure, an image input device 10 captures an image and outputs image data to the image processing device 20. The image processing device 20 performs image processing such as predetermined contrast enhancement and outputs the image data to the image output device 30. Then, the image output device 30 displays an image with enhanced contrast.
[0064]
Here, a specific example of the image input device 10 corresponds to the scanner 11, the digital still camera 12, or the video camera 14, and a specific example of the image processing device 20 corresponds to a computer system including the computer 21 and the hard disk 22. Specific examples of the output device 30 include a printer 31, a display 32, and the like. Of course, the present invention can be applied to color copiers, color facsimile machines and the like in addition to these.
[0065]
In the present image processing system, since an optimum contrast is to be given to an image having a low contrast, image data obtained by taking a photograph with a scanner 11 as the image input device 10 or a digital still camera 12 is used. The image data having a low contrast taken in step 1 or the moving image taken by the video camera 14 is a processing target and is input to the computer system as the image processing apparatus 20. Note that the calculation speed of the input image of the video camera 14 may not be in time. In such a case, it is possible to cope with this by setting the first condition that requires calculation time for each shooting scene and performing only the image conversion of each frame under the same condition setting during shooting.
[0066]
The image processing apparatus 20 includes at least a luminance distribution detection unit that extracts a luminance distribution, a luminance distribution spread amount detection unit that first detects a spread amount of the luminance distribution based on the detected luminance distribution, and the spread amount. Based on the above, a gradation number assigning unit that assigns a large number of gradations to a range with a high distribution density and assigns a small number of gradations to a range with a low distribution density, and converts image data based on the assigned luminance gradation And image data conversion means. Of course, the image processing apparatus 20 may also be a color conversion unit that corrects a color difference for each model, or a resolution conversion unit that converts a resolution corresponding to each model. . In this example, the computer 21 executes each image processing program stored in the internal ROM or the hard disk 22 while using a RAM or the like.
[0067]
The execution result of this image processing program is obtained as image data with enhanced contrast as will be described later, and is printed by the printer 31 that is the image output device 30 based on the obtained image data, or by the same image output device 30. It is displayed on a certain display 32. More specifically, this image data is RGB (green, blue, red) gradation data, and the image has dots arranged in a grid in the vertical direction (height) and the horizontal direction (width). It is configured as matrix data.
[0068]
In the present embodiment, a computer system is incorporated between image input / output devices to perform image processing. However, such a computer system is not necessarily required, and a digital still camera as shown in FIG. A system in which an image processing device for emphasizing contrast is incorporated in 12a and the converted image data is displayed on the display 32a or printed on the printer 31a may be used. As shown in FIG. 4, in the printer 31b that inputs and prints image data without using a computer system, the image data input through the scanner 11b, the digital still camera 12b, the modem 13b, or the like is automatically received. It is also possible to configure so as to enhance contrast.
[0069]
FIG. 5 shows luminance distribution detection processing corresponding to luminance distribution detection means and luminance distribution spread amount detection means in the image processing executed by the computer 21. The gradation number allocation means, image data conversion means, A luminance conversion process corresponding to is shown in FIG.
[0070]
FIG. 5 mainly corresponds to luminance distribution detection processing. First, the luminance distribution detection processing will be described.
[0071]
Before describing how to represent luminance, the pixels to be distributed will be described. As shown in step S102 of FIG. 5, a thinning process for thinning out the target pixels is executed. As shown in FIG. 7, in the case of a bitmap image, it is formed as a two-dimensional dot matrix consisting of predetermined dots in the vertical direction and predetermined dots in the horizontal direction, and all pixels can be obtained if an accurate luminance distribution is obtained. It is necessary to examine the brightness. However, this distribution extraction process is intended to determine the amount of spread of the luminance distribution, and is not necessarily accurate. Therefore, it is possible to perform thinning to such an extent that it is within a certain error range. According to the statistical error, the error with respect to the number of samples N can be expressed as 1 / (N ** (1/2)). However, ** represents a power. Therefore, in order to perform processing with an error of about 1%, N = 10000.
[0072]
Here, the bitmap screen shown in FIG. 7 has the number of pixels of (width) × (height), and the sampling period ratio is
ratio = min (width, height) / A + 1
... (1)
And This min (width, height) is the smaller of width and height, and A is a constant. Further, the sampling period ratio here indicates how many pixels are sampled, and the pixels marked with ◯ in FIG. 8 indicate the case where the sampling period ratio = 2. That is, one pixel is sampled every two pixels in the vertical and horizontal directions, and every other pixel is sampled. The number of sampling pixels in one line when A = 200 is as shown in FIG.
[0073]
As can be seen from the figure, except for the case where the sampling period ratio = 1 in which sampling is not performed, when there is a width of 200 pixels or more, the number of samples is at least 100 pixels. Therefore, in the case of 200 pixels or more in the vertical direction and the horizontal direction, (100 pixels) × (100 pixels) = (10000 pixels) is secured, and the error can be reduced to 1% or less.
[0074]
Here, the reason for using min (width, height) as a reference is as follows. For example, as shown in FIG. 10 (a), when width >> height is satisfied and the sampling period ratio is determined by the longer width, as shown in FIG. 10 (b). In addition, in the vertical direction, only two lines of the upper end and the lower end may be extracted. However, if the sampling period ratio is determined based on the smaller one as min (width, height), thinning is performed so as to include the intermediate portion in the smaller vertical direction as shown in FIG. Will be able to.
[0075]
In this example, thinning is performed with an accurate sampling period for pixels in the vertical and horizontal directions. This is suitable for processing while thinning out pixels that are sequentially input. However, when all the pixels have been input, the pixels may be selected by randomly specifying coordinates in the vertical direction or the horizontal direction. In this way, when the necessary minimum number of pixels, such as 10,000 pixels, is determined, the extraction process is repeated at random until the number of pixels reaches 10,000, and the extraction is stopped when the number of pixels reaches 10,000.
[0076]
If the pixel data for the selected pixel has luminance as its component element, the distribution can be obtained using the luminance value. However, even in the case of image data in which the luminance value is not a direct component value, the component value that indirectly represents the luminance is provided. Accordingly, the luminance value can be obtained by performing conversion from the color space where the luminance value is not a direct component value to the color space where the luminance value is a direct component value.
[0077]
Such color conversion between different color spaces is not uniquely determined by a conversion formula, but a mutual relationship is obtained for each color space having the respective component values as coordinates, and this correspondence is stored. It is necessary to perform sequential conversion with reference to the color conversion table. In relation to the table, the component value is expressed as a gradation value, and in the case of 256 gradations having a three-dimensional coordinate axis, a color conversion table of about 16.7 million (256 × 256 × 256) elements is obtained. Must have. As a result of considering the efficient use of storage resources, correspondence is not prepared for all coordinate values, but usually correspondence is prepared for appropriate discrete lattice points, and interpolation is used together. Like to do. Since this interpolation calculation is possible through several multiplications and additions, the amount of calculation processing becomes enormous.
[0078]
That is, if a full-size color conversion table is used, the processing amount is reduced, but the table size becomes an unrealistic problem, and if the table size is made a realistic size, the calculation processing amount becomes unrealistic. There are many cases.
[0079]
In view of such a situation, in the present embodiment, as used in the case of a television or the like, the following conversion formula for obtaining luminance from the three primary colors of RGB is adopted. That is, for the luminance yp at the point P, from the RGB component values (Rp, Gp, Bp),
yp = 0.30Rp + 0.59Gp + 0.11Bp (2)
And In this way, the luminance value can be obtained by only three multiplications and two additions. Then, a frequency distribution is obtained by using a variable region of an array corresponding to all gradations.
[0080]
In this embodiment, as a result of targeting the RGB color space, such a conversion formula is adopted, but since each component value indicates the brightness of the color in the background, each component is When the value is viewed alone, it has a property of linearly corresponding to the luminance. Therefore, more roughly speaking, without considering each addition ratio simply
yp = (Rp + Gp + Bp) / 3 (3)
It is not impossible to simplify it, and
yp = Gp (4)
As described above, in the equation (3), the green component value having the largest ratio can be used as the luminance value.
[0081]
In the thinning-out process, the luminance is obtained from the RGB image data at the same time for the selected pixel, and the distribution is obtained. Finally, in step S114, a standard deviation corresponding to the spread amount is obtained based on this distribution, but there are matters to consider before that.
[0082]
The first is a case where the image is a binary image such as a black and white image. The concept of contrast enhancement is inappropriate for binary images including black and white images. If there is a black and white image as shown in FIG. 11, the luminance distribution for this image is concentrated at both ends within the allocated number of gradations as shown in FIG. Basically, it concentrates on the gradation “0” and the gradation “255”.
[0083]
Therefore, when performing the black and white check in step S104, it can be determined whether or not the sum of the number of pixels of gradation “0” and gradation “255” matches the number of pixels selected by thinning. In the case of a monochrome image, the non-enlargement process is executed in step S106 in order to interrupt the process without executing the following process. In the present embodiment, the distribution extraction process and the luminance conversion process are roughly divided. Therefore, in this non-enlarging process, a flag is set so that the subsequent luminance conversion process is not executed, and the distribution extraction process is terminated.
[0084]
The binary data is not limited to black and white, and there may be binary data with color. In such a case as well, there is no need to perform the process of enhancing the contrast, and if the distribution state is examined and the distribution is concentrated only on two values (one is approximately “0”), the process is interrupted as binary data. Can be achieved.
[0085]
Second, consider whether the image is like a business graph or a natural picture like a photograph. Although a process of enhancing contrast is sometimes required for natural images, it is often preferred that contrast is not enhanced for business graphs or paintings. Accordingly, in step S108, it is checked whether or not the image is a natural image.
[0086]
In natural paintings, the number of colors including shadows is extremely large, but in certain types of paintings such as business graphs and draws, the number of colors is often limited. Therefore, if the number of colors is small, it can be determined that the image is not a natural image. In order to accurately determine the number of colors, it is necessary to determine how many of 16.7 million colors are used as described above, but this is not realistic. On the other hand, when the number of colors is extremely small as in the business graph, the probability that the colors are different and have the same luminance is low. That is, the approximate number of colors can be determined based on the luminance. If the number of colors is small, the luminance distribution is sparse, and in a business graph, it appears as a line spectrum. For this reason, in step S108, the number of luminance values whose distribution number is not “0” among the luminances of 256 gradations is counted. If it is less than “64” color (gradation), which is approximately ¼, it is determined that the image is not a natural image, and the non-enlargement process is executed in step S106 as in the case of binary data. Of course, whether or not it is equal to or less than “64” color (gradation) as the threshold can be changed as appropriate.
[0087]
It is also possible to determine whether or not the distribution is a line spectrum by the adjacent ratio of luminance values whose distribution number is not “0”. That is, it is determined whether or not there is a distribution number in the luminance values adjacent to the luminance value that is not “0”. If at least one of the two adjacent luminance values is adjacent, nothing is performed, and if both are not adjacent, the count is performed. As a result, the ratio between the number of luminance values other than “0” and the count value is used. Just judge. For example, if the number of luminance values other than “0” is “64” and the number of non-adjacent ones is “64”, it can be seen that the line spectrum is distributed.
[0088]
Further, when the image processing program is executed via the operating system, it is possible to make a determination based on the extension of the image file. Of bitmap files, especially photographic images are compressed, and an implicit extension is often used to indicate the compression method. For example, an extension “JPG” indicates that the file is compressed in the JPEG format.
Since the operating system manages the file name, if you send an inquiry to the operating system from the printer driver, the extension of the file will be answered. What is necessary is just to judge that there exists and to perform contrast emphasis. In addition, if the extension is unique to a business graph such as “XLS”, it can be determined that contrast enhancement is not performed.
[0089]
A third consideration is whether or not there is a frame around the image as shown in FIG. If such a frame portion is white or black, as shown in FIG. 14, the luminance distribution appears in a line spectrum shape at both ends in the gradation number allocation range, and both ends corresponding to the internal natural image. It also appears as a smooth luminance distribution inside
[0090]
Of course, since it is more appropriate not to take the frame portion into consideration of the luminance distribution, the sum of the number of pixels of the gradation “0” and the gradation “255” is sufficiently large in the frame check in step S108, and It is determined whether the number of pixels selected by the thinning does not match. If the result is affirmative, it is determined that there is a frame, and the frame processing is performed in step S112. In this frame processing, in order to ignore the frame, the number of pixels of gradation “0” and gradation “255” in the luminance distribution is set to “0”. Thereby, in the following process, it can handle like the thing without a frame part.
[0091]
In this example, a white or black frame portion is targeted, but there may be a case where there is a specific color frame. In such a case, a protruding line spectrum appears in the original smooth curve drawn by the luminance distribution. Accordingly, a line spectrum having a large difference between adjacent luminance values may be considered as a frame portion and not subjected to luminance distribution. In this case, since the color may be used other than the frame portion, an average of luminance values on both sides may be assigned.
[0092]
In the case where contrast enhancement is performed after considering the above consideration, in step S114, the standard deviation of the luminance distribution is obtained, and the median yMe is obtained for the luminance conversion processing in the subsequent stage. There are two ways of thinking about the standard deviation, but in the present embodiment, calculation is performed based on the following equation.
[0093]
[Expression 1]

[0094]
The standard deviation corresponds to the spread amount of the luminance distribution, but dispersion may be used to express the spread amount.
[0095]
The above processing corresponds to the distribution detection processing. Next, luminance conversion processing for converting image data based on the standard deviation σ that is the amount of spread of the luminance distribution thus obtained will be described. As described above, when the non-enlargement process is executed in step S106, the predetermined process is detected with reference to a predetermined flag in step S202, and the image process is terminated without performing the following process.
[0096]
In the luminance conversion processing, a small number of gradations is assigned to a range with a small distribution density while giving a large number of gradations to a range with a large distribution density based on the spread amount of the luminance distribution. Here, a pattern that assigns a small number of gradations to a range with a small distribution density while giving a large number of gradations to a range with a large distribution density will be described. The luminance y (input) before conversion and the luminance Y (output) of the conversion destination are
Y = y (6)
As shown in FIG. 15, the gradation range r0 assigned before the conversion and the gradation range R0 assigned after the conversion are equal to each other with the central portion in the reproducible range as a reference, as shown in FIG. I'm doing it. However, as shown in FIG. 16, when the input / output correspondence is a so-called S-shaped curve, the gradation ranges R1 and R2 assigned after conversion become larger and assigned to the gradation range r0 assigned before conversion. The number of tones increased. On the other hand, regarding the range outside the gradation range r0 on the low luminance side and the high luminance side in the input, the gradation range assigned after conversion is reduced.
[0097]
That is, this correspondence means that a large number of gradations is given to a range with a large distribution density and a small number of gradations is assigned to a range with a small distribution density. Here, various concrete allocation methods for realizing this correspondence can be used. FIG. 17 is obtained by performing γ correction of γ <1 on the region from the center position ymid of the gradation range to the upper end of the gradation range on the high luminance side, and the upper quadrant from the center position ymid before conversion. The converted gradation range R with respect to the gradation range r up to yq3 is expanded. Similarly, FIG. 18 is obtained by performing γ correction with γ> 1 on the region from the center position ymid of the gradation range to the lower end of the gradation range on the low luminance side. The converted gradation range R with respect to the gradation range r up to the dividing point yq1 is enlarged.
[0098]
On the other hand, in these cases, the ratio of the gradation range R after conversion to the gradation range r before conversion varies depending on how γ is given. In this embodiment, the ratio is controlled based on the standard deviation σ that is the amount of spread of the luminance distribution. That is, assuming that the center position ymid of the gradation range is “128”, γ1 is given below this center position ymid, and γ2 is given in a range larger than the center position ymid.
For y ≦ 128,
γ1 = (σstd_limit / σ) ** a (7)
For y> 128,
γ2 = (σ / σstd_limit) ** a (8)
In step S204, these parameter calculations are executed. As described above, this parameter calculation constitutes the gradation number assigning means. Here, σstd_limit and a are parameters obtained experimentally in consideration of the conversion result. In this embodiment, σstd_limit is set to “128” and a is set to “0.1”. Since the standard deviation σ is generally smaller than “128”, in these relational expressions, when the standard deviation σ is large, γ2 and γ1 each approach “1”, and the slope of the S-shaped curve becomes gentle. . This means that the gradation range R of the conversion destination is not so wide with respect to the gradation range r centered on the center position ymid when the spread amount is large, more specifically, the brightness of the image data. Means that the conversion for expanding the luminance range is not performed. On the other hand, when the standard deviation σ is small, γ2 and γ1 are separated from “1”, and the slope of the S-curve becomes steep. This means that when the spread amount is small, the gradation range R of the conversion destination is widened with respect to the gradation range r centered on the center position ymid. When the luminance is distributed only in a narrow range, it means that conversion for expanding the luminance range is performed.
[0099]
As shown in this example, when the gradation range is divided into two, the low luminance side and the high luminance side, and γ correction is applied to each of them, if γ is given a reciprocal relationship, the low luminance side and the high luminance side A smooth connection is made at the connection point with the luminance side, and a good S-shaped curve can be given. Of course, changing the polarity of the γ correction between the low luminance side and the high luminance side itself forms an S-shaped curve correspondence, and the luminance Y after conversion can be greatly changed with respect to the luminance y before conversion. It becomes like this.
[0100]
In the present embodiment, the correspondence relationship of the S-shaped curve is established by γ correction, but FIG. 19 shows an example in which a linear correspondence relationship of Y = f · y + g is realized before and after the center position ymid of the gradation range. Is shown. In this example, a linear correspondence is formed again in the region below the lower quadrant yq1 and above the upper quadrant yq3. In this example as well, the slope f may be made closer to “1” when the standard deviation σ is large, and the slope f may be made larger than “1” when the standard deviation σ is small. Of course, in this case, two corresponding straight lines may be smoothly connected as shown in FIG.
[0101]
On the other hand, when the luminance distribution is in the center of the gradation range, satisfactory luminance conversion is possible only by setting γ1 and γ2 as described above. However, as shown in FIG. 21, the median yMe of the luminance distribution is In some cases, the luminance distribution is slightly on the low luminance side, and as shown in FIG. 22, the median yMe of the luminance distribution is slightly on the high luminance side.
[0102]
On the other hand, in the example shown in FIG. 23, γ correction of γ <1 is applied to the entire range. In this case, the gradation range R after conversion with respect to the gradation range r on the low luminance side before conversion is enlarged. In the example shown in FIG. 24, γ correction of γ> 1 is applied to the entire range. In this case, the gradation range R after conversion with respect to the gradation range r on the high luminance side before conversion is It is expanding.
[0103]
Therefore, the magnitude relationship between the median yMe obtained in step S114 and “128”, which is the center position of the gradation range, is compared, and the median yMe seems to be biased toward the low luminance side region as shown in FIG. If so, by applying γ correction of γ <1 to the entire range as shown in FIG. 23, the gradation range r on the low luminance side having a high distribution density before conversion is expanded. Also, as shown in FIG. 22, if the median yMe seems to be biased toward the high-luminance region, the entire range as shown in FIG. The gradation range r on the high luminance side having a high distribution density is expanded. In this way, it is possible to enlarge or reduce the number of gradations without necessarily having the S-curve correspondence. Of course, in these cases, γ may be determined based on the formulas (7) and (8) depending on whether γ <1 or γ> 1.
[0104]
Further, in the example shown in FIG. 25, the gradation is “0”, the lower quadrant yq1, the center position ymid, the upper quadrant yq3, and the gradation “255”. For “0”, the center position ymid, and the gradation “255”, while Y = y, conversion points at the lower quadrant yq1 and the upper quadrant yq3 are determined based on the standard deviation. Then, a correspondence relationship connecting these five points may be obtained by spline interpolation calculation or Newton interpolation. Of course, the lower three points and the upper three points from the center position ymid may be obtained by spline interpolation calculation or Newton interpolation, respectively.
[0105]
By the way, if a tone range that is too large is assigned to the tone range before conversion, it may be undesirable. In the twilight state like the evening, the range of contrast from the brightest part to the dark part is naturally narrow, but as a result of trying to enlarge the contrast greatly for this image, it is converted like a daytime image. It might be. Since such conversion is not desired, the enlargement ratio is limited and both γ1 and γ2 are limited. For example, even when γ2 <0.7, γ2 = 0.7, and even when γ1> 1.3, γ1 = 1.3.
[0106]
As described above, in this embodiment, the parameters γ1 and γ2 can be obtained, and step S204 is ended. It can be said that such a change in the number of gradations changes the assignment of the number of quantization bits in other words.
[0107]
By the way, it is irrational to execute the calculation by γ correction every time the luminance is converted. This is because the range that the luminance y can take is only “0” to “255”, and it is possible to obtain the luminance Y after conversion corresponding to all the values that the luminance y can take in advance. That's why. Therefore, this correspondence is obtained in step S206 and stored as a table as shown in FIG.
[0108]
Here, the specific calculation of the correspondence is as follows.
[0109]
For y ≦ 128,
Y = 128 * (y / 128) ** γ1 (9)
For y> 128,
Y = 128 * {(y−128) / 128} ** γ2 + 128
(10)
However, the conversion point of γ correction can be changed based on the median yMe. That is,
If y ≦ yMe,
Y = yMe * (y / yMe) ** γ1 (11)
If y> yMe,
Y = yMe * {(y−yMe) / yMe} ** γ2 + yMe (12)
Then, it becomes an S-shaped curve in which the polarity of luminance conversion is inverted between the high luminance side and the low luminance side with the median yMe as the center, and a large number of gradations can be assigned to both sides centering around the high distribution density, There is an effect that conversion can be performed without greatly affecting the overall brightness.
[0110]
The image data can be changed when the calculated conversion table is formed.
[0111]
Finally, image data is converted in step S208. Up to this point, the correspondence relationship for converting the luminance has been obtained, and for example, it was not the conversion relationship for the component values (Rp, Gp, Bp) on the RGB coordinate axes. However, the conversion equation (2) can also be applied in correspondence with the RGB component values (Rp, Gp, Bp). In addition, the RGB component values (r, g, b), (R, G, B) are in the same range corresponding to the luminance y, Y ranging from “0” to “255”. Therefore, it can be said that the above-described conversion table of luminance y and Y may be used as it is.
[0112]
Therefore, in step S208, the conversion table shown in FIG. 26 is referred to for the image data (r, g, b) of all pixels, and the process of obtaining the converted image data (R, G, B) is repeated.
[0113]
Next, the operation of the present embodiment having the above configuration will be described step by step.
[0114]
If a photograph is taken with the scanner 11 or the like, image data representing the photograph as RGB gradation data is taken into the computer 21, and the CPU executes the image processing program shown in FIGS. 5 and 6 to execute the image data. The process of enhancing the contrast is executed.
[0115]
First, in step S102, the image data is thinned out within a predetermined error, and the luminance y for the selected pixel is obtained to obtain a distribution. The distribution as it is cannot be used. First, it is determined in step S104 whether the image is a binary image such as black and white, and in step S108, it is determined whether the image is a natural image. Except when the image is a binary image or not a natural image, in step S110, it is determined whether the image data has a frame portion, and if there is a frame portion, the standard deviation σ is determined in step S114 with respect to the luminance distribution obtained. Ask. In this embodiment, it is only necessary to obtain the standard deviation σ, but when the polarity conversion point for γ correction is changed in accordance with the luminance distribution, the median yMe is obtained.
[0116]
Once the standard deviation σ of the luminance distribution is obtained,
For y ≦ 128,
γ1 = (σstd_limit / σ) ** a (7)
Y = 128 * (y / 128) ** γ1 (9)
For y> 128,
γ2 (σ / σstd_limit) ** a (8)
Y = 128 * {(y−128) / 128} ** γ2 + 128
(10)
In step S204, parameters γ1 and γ2 are obtained, and in step S206, the conversion relationship from luminance y to luminance Y is stored in a table. Then, the image data for all the pixels is converted with reference to the conversion table completed in step S208.
[0117]
Of course, as described above, such image processing is not performed when the image is not a binary image or a natural image. However, when the image processing of the present invention is performed, the contrast is very low in the state of the photograph. Regardless, by correcting so as to widen the luminance range, a clear image with clear contrast can be obtained.
[0118]
In the above-described embodiment, the parameter σstd_limit, a for obtaining γ1 and γ2 is constant, but the computer 21 may be able to select it via a predetermined GUI. It is also possible for the user to designate a part of the image data and execute the contrast enhancement process only within the range.
[0119]
In this manner, after obtaining the luminance y distribution for the pixels of the image data while thinning out in step S102, the standard deviation σ corresponding to the spread amount of the luminance distribution is obtained in step S114, and the standard deviation σ is obtained. On the basis of this, since the γ correction parameters (γ1, γ2) for forming the correspondence relationship of the S-curve are calculated in step S204, the image data is converted in step S208 to obtain the conversion source luminance y. Automate the task of assigning a large number of gradations to areas with high distribution density and assigning a small number of gradations to areas with low distribution density so that even non-experts can easily enhance contrast. Become.
[Brief description of the drawings]
FIG. 1 is a block diagram of an image processing system to which an image processing apparatus according to an embodiment of the present invention is applied.
FIG. 2 is a block diagram illustrating a specific hardware configuration example of the image processing apparatus.
FIG. 3 is a schematic block diagram illustrating another application example of the image processing apparatus of the present invention.
FIG. 4 is a schematic block diagram illustrating another application example of the image processing apparatus of the present invention.
FIG. 5 is a flowchart showing a luminance distribution detection processing portion in the image processing apparatus of the present invention.
FIG. 6 is a flowchart showing a luminance conversion processing portion in the image processing apparatus of the present invention.
FIG. 7 is a diagram illustrating coordinates in a conversion source image.
FIG. 8 is a diagram showing a sampling period.
FIG. 9 is a diagram illustrating the number of sampling pixels.
FIG. 10 is a diagram illustrating a relationship between a conversion source image and pixels to be sampled.
FIG. 11 is a diagram illustrating a black and white image.
FIG. 12 is a diagram illustrating a luminance distribution of a black and white image.
FIG. 13 is a diagram illustrating an image with a frame portion.
FIG. 14 is a diagram illustrating a luminance distribution of an image having a frame portion.
FIG. 15 is a graph showing a relationship between luminance before conversion and luminance after conversion when the number of gradations is not changed.
FIG. 16 is a graph in which the luminance before conversion and the luminance after conversion become a correspondence relationship of the S curve by γ correction.
FIG. 17 is a graph when γ correction of γ <1 is applied in a region on the high luminance side.
FIG. 18 is a graph when γ correction of γ> 1 is applied in a low luminance side region.
FIG. 19 is a graph in which the luminance before conversion and the luminance after conversion become a correspondence relationship of the S-shaped curve by linear conversion.
FIG. 20 is a modified example in which smooth conversion is performed at the conversion point of the conversion characteristic in the same correspondence relationship.
FIG. 21 is a diagram showing a luminance distribution when the median is on the low luminance side.
FIG. 22 is a diagram showing a luminance distribution when the median is close to the high luminance side.
FIG. 23 is a graph when γ correction of γ <1 is applied in the entire region.
FIG. 24 is a graph when γ correction of γ> 1 is applied in the entire region.
FIG. 25 is a graph in a case where specified conversion points are connected by an interpolation method.
FIG. 26 is a diagram illustrating a conversion table when a luminance distribution is enlarged.
[Explanation of symbols]
10. Image input device
11 ... Scanner
11b ... Scanner
12 ... Digital still camera
12a ... Digital still camera
12b ... Digital still camera
13b Modem
20 Image processing apparatus
21 ... Computer
22 ... Hard disk
30. Image output device
31 ... Printer
31a ... Printer
31b ... Printer
32 ... Display
32a ... Display

Claims (15)

An image processing apparatus that performs processing for enhancing contrast of an image in which pixels are arranged in a vertical direction and a horizontal direction,
When sampling and thinning out the pixels of the image in the vertical direction and the horizontal direction, the number of extractions based on the number of pixels in the smaller direction in the vertical direction and the horizontal direction is ensured. The luminance distribution of the image is obtained from the luminance equivalent value of the predetermined number of gradations for each pixel after the thinning, and the amount representing the variation of the luminance equivalent value for each pixel is calculated from the obtained luminance distribution. An image processing apparatus that performs a process of enhancing the contrast of the image based on an amount representing.   The image processing apparatus according to claim 1, wherein a small number of gradations is assigned to a gradation range with a small distribution density and a large number of gradations in a gradation range with a high distribution density based on the amount representing the variation of the luminance equivalent value. An image processing apparatus characterized by assigning a number and converting a luminance equivalent value of each pixel in accordance with a change in the number of gradations.   3. The image processing apparatus according to claim 1 or 2, wherein an amount representing the variation of the luminance equivalent value uses a value corresponding to a standard deviation, and the input / output ratio is reduced when the standard deviation of the luminance distribution is large. An image processing apparatus characterized in that the gradation number assignment is made uniform, and when the standard deviation is small, the input / output ratio is increased to enlarge the gradation number assignment change.   4. The image processing apparatus according to claim 3, wherein a value of γ is set so that the amount of change due to γ correction is reduced when the luminance distribution is subjected to γ correction and the standard deviation is large, and when the standard deviation is small. An image processing apparatus, wherein the value of γ is set so that the amount of change in correction becomes large.   5. The image processing apparatus according to claim 4, wherein the approximate center position of the luminance distribution is obtained, and the input and output are performed by inverting the polarity of the luminance conversion between the high luminance side and the low luminance side with reference to the approximate center position. An image processing apparatus that performs luminance conversion of an approximately S-shaped curve in relation to   6. The image processing apparatus according to claim 1, wherein the conversion destination luminance is calculated and stored within a possible range of the conversion source luminance, and this correspondence is evoked during conversion. An image processing apparatus for converting.   The image processing apparatus according to any one of claims 1 to 6, wherein when the image data is represented by a plurality of component values corresponding to the luminance, the luminance calculation is obtained by linear addition of the same component values. An image processing apparatus.   8. The image processing apparatus according to claim 1, wherein a limit is set for a degree of contrast enhancement.   The image processing apparatus according to any one of claims 1 to 8, wherein binary image data is determined based on a luminance distribution, and contrast enhancement is not performed for binary image data. An image processing apparatus.   10. The image processing apparatus according to claim 9, wherein when the luminance distribution is concentrated at both ends within a reproducible range, the image processing apparatus determines that the binary image data is black and white.   11. The image processing apparatus according to claim 1, wherein a frame portion of the image data is determined based on the protruding luminance distribution, and if there is a frame portion, the data of the frame portion is used to enhance contrast. An image processing apparatus that is not used.   12. The image processing apparatus according to claim 11, wherein the luminance distribution concentrated at the end within a reproducible range is determined to be a frame.   13. The image processing apparatus according to claim 1, wherein contrast enhancement is not performed when the image data is not a natural image.   The image processing apparatus according to claim 13, further comprising a natural image determination unit that determines that the image data is not a natural image when the luminance distribution exists in a spectral form. An image processing method for performing processing for enhancing contrast of an image in which pixels are arranged in a vertical direction and a horizontal direction,
When sampling and thinning out the pixels of the image in the vertical direction and the horizontal direction, the number of extractions based on the number of pixels in the smaller direction in the vertical direction and the horizontal direction is ensured. The luminance distribution of the image is obtained from the luminance equivalent value of the predetermined number of gradations for each pixel after the thinning, and the amount representing the variation of the luminance equivalent value for each pixel is calculated from the obtained luminance distribution. An image processing method characterized by performing a process of enhancing the contrast of the image based on an amount representing.

Images