經濟部中央樣準局貝工消费合作社印家 Λ 7 -—_____ ________ H? 五、發明説明（) 5·1發明領域： 本發明係有關於一種影像處理技術，特別是關於一 種利用圈文影像分割的演算法，將所分割圈文彩像加強處 理，以改善圈文影像品質之方法》 5·2發明貧景： 一鈒的單色印表機在列印灰階影像時，大多會採用 Dithering或誤差分散法（Error diffusion)的方式以使輪出 的結果艇夠呈現出灰陏的效果，但也因而造成在列印圈文 並存的影像時’往往會使得文字變的棋輞而無法辨識其内 容’尤其是具有背景之文字内容》有時為了加強文字之辨 識度’會採用二分法（Thresholding)以得到較完整的字形， 但此法則造成圈形棋輞無法辨熾，且部分文字之背景灰階 消失》 一般列印時’欲列印具有灰階的文字或彩像時，灰陏 由0至255 ’欲調整其灰階時，通常利用dithering黑點摻 雜以形成灰陏影像，缺點為文字會變模糊，因文字邊緣舆 背景黑點運算的結果，若是影像部分當然希望有灰陏的效 果’但文字則不希望造成灰障的效果而影響文字之辨識 度，因此通常會對文字部分加強處理，以得到較清晰易讀 之文字。 -2 - 本纸張尺度適财國_家標率（CNS )从麟·（ 210X297公釐） --^---„----ο裝-- (請先閲靖背面之注意事項再填疼本頁) l·訂· •年 Λ五、發明説明（） 當一張«文並存的文件經由掃描器（Scanner)得到的 影像再透過彩色印表機出來後，可能會發現不是圈形變 暗了就是文字變的棋糊了。由於一般的彩色印表機在列 印時’並未對文字或圈形分別處理，因此基本上運是無法 兼顧圈形舆文字的特性。然而對、於人類的視覺而言，文字 舆固形卻有完全不同的特性。大部分的情況下你會希望 字舆背景有著強烈的對比而使其清晰易冑，而圈形則希 望挺约柔和亮麗，有著豐富的層次舆色彩β 因此有必要提出一種改善圈文影像品質之方法，以克 服習知技術所面雄之困擾。 5·3發明8的及輟述： 本發明之目的為提供一種改善影像品質之方法特 別是含有圈形舆文字並存之影像。 (請先閲讀背面之注意事項再填寫本頁) •裝 -1τ 經濟部中央標準局ι工消费合作社印製 本發明提供一種方法改善圈文影像品質，該方法係 利用圈文分割法，將一含有圈形舆文字之原始影像分割為 圈形影像區塊舆複數個文字影像區塊，並特別將分割出來 之文字影像部分’予以加強處理。同時，處理原始影像中 之全部影像以使其灰陏或色彩對比度較明顢，最後利用將 處理後之文字影像加成至處理後之全部影像，以合 杂f -3 - 本纸浪尺度適用t國國家揉準（CNS ) A4此格(210X297公釐）Yin Jia of the Central Samples Bureau, Ministry of Economic Affairs, Shellfish Consumer Cooperative, Λ 7 ---_____ ________ H? V. Description of the Invention (5.1) Field of the Invention: The present invention relates to an image processing technology, and particularly to an image using circle text Segmentation algorithm to enhance the processing of the segmented circle color image to improve the quality of circle circle image. "5 · 2 Invention of poor scenes: When monochromatic printers print grayscale images, they mostly use The method of Dithering or Error diffusion can make the result of the rotation appear gray, but it also causes the image of the coexistence of circle texts to often make the text become a chessboard. Recognizing its content 'especially text content with a background' Sometimes in order to enhance the recognition of the text, Thresholding will be used to obtain a more complete glyph, but this rule causes the circle-shaped chess rim to be unable to discern, and part of the text The background gray level disappears. "When printing, 'When printing text or color images with gray levels, the gray scale is from 0 to 255.' When you want to adjust the gray level, you usually use dithering black dots to dope to The gray image has the disadvantage that the text will be blurred. As a result of the calculation of the black points on the edges and background of the text, of course, if the image part is expected to have a gray effect, but the text does not want to cause a gray barrier effect and affect the recognition of the text. , So usually the text part is strengthened to get clearer and easier to read text. -2-The size of this paper is suitable for the country of wealth _ house standard rate (CNS) from Lin · (210X297 mm)-^ --- „---- ο Install-(Please read the precautions on the back of Jing before Fill in this page) l · Order · • Year Ⅴ. Description of the invention () When a «co-existing document» image obtained through a scanner (Scanner) passes through a color printer, it may be found that it is not a circle deformation Darkened text is a piece of paper. As general color printers do not process text or circles separately when printing, it is basically impossible to take into account the characteristics of circles and characters. However, yes, the In terms of human vision, the solid shape of the text has completely different characteristics. In most cases, you will want the background of the text to have a strong contrast to make it clear and easy to read, while the circle shape hopes to be soft and bright and rich. Therefore, it is necessary to propose a method to improve the quality of the circle image, in order to overcome the dilemma of the conventional technology. 5.3 Invention 8 and the description: The purpose of the present invention is to provide an improved image quality Method especially images with coexisting circle text (Please read the notes on the back before filling out this page) • Printed by -1τ Printed by the Central Standards Bureau of the Ministry of Economic Affairs, Industrial Consumer Cooperative, The present invention provides a method to improve the quality of circle images. This method uses the circle segmentation method to An original image containing a circle-shaped text is divided into a circle-shaped image block and a plurality of text-image blocks, and the segmented text-image part is particularly strengthened. At the same time, all the images in the original image are processed to make it The gray or color contrast is brighter. Finally, the processed text image is added to all processed images to combine the f -3-this paper wave standard is applicable to the national standard (CNS) A4 this grid ( 210X297 mm)
彳'" /BV 經濟部中央橾準局Λ工消费合作社印装 Λ 7 Η 7___五、發明説明（) 影像。其中上述單色圈文影像之文字影像處理方法，係利 用二分法（Thresholding)以產生白底黑字的二元文字影 像；其次，處理單色圈文影像全部影像之方法，係利用誤 差分散法（Error Diffusion)以產生能表現灰階效果的二元 影像，最後以邏輯運算將上述處理後之二元文字影像輿處 理後之全部二元影像合成為一新影像。對於彩色圖文影像 則將全部影像調整Gamma值，以複數個二元文字影像為 罩幕（mask) ’分別找出各黑色文字之像素（Pjxe|s)所對應 原始影像色彩之後’加強其色彩飽和度直接窝入巳讕整 Gamma值之影像上，所得新影像之色彩對比將較原始影 像明顧。5·4 _示钃單說明： 第一 IT說明本發明之一較佳實、施例，用以改善單色明文 影像品質之方法流程方洗圈。 第二圈顧示本發明用於合成毛色圈文^影像之方法。 第三圈‘說明本發明之另一較佳寞^施例，用以改餐彩色圖 文^像品質之j法_.流程方垅圈· 第淨B 顯示本發明用於合成彩色軋文彩像之方法·5-5發明详麯說明： 本發明所提供之改善圈文影像品質之方法，係將一 本纸張尺度適用中國β家標率（CNS > A4%ST2l〇X297^ ) (請先閲讀背面之注意事項再填寫本頁) d. •裝- 訂彳 '" / BV Printed by Central Industrial and Commercial Bureau of the Ministry of Economic Affairs Λ Industrial Consumer Cooperative Λ 7 Η 7___ V. Description of Invention () Image. The text image processing method of the monochromatic circle text image mentioned above uses a binary method (Thresholding) to generate a binary text image of black characters on a white background. Secondly, the method of processing the entire image of the monochromatic circle text image uses the error dispersion method. (Error Diffusion) to generate a binary image that can represent a grayscale effect. Finally, the binary image processed by the above-mentioned processed binary image is synthesized into a new image by a logical operation. For color graphic images, adjust the gamma value of all the images, and use multiple binary text images as a mask. 'Find the original image colors corresponding to the pixels (Pjxe | s) of each black text' and strengthen the colors. Saturation is directly embedded in the image with a full Gamma value. The color contrast of the new image obtained will be better than the original image. 5 · 4 _Instruction sheet description: The first IT describes one of the preferred embodiments and embodiments of the present invention to improve the quality of monochrome plaintext image quality. The second circle shows the method of the present invention for synthesizing fur circle images. The third circle 'illustrates another preferred embodiment of the present invention, which is used to change the color and text of the meal. The method of image quality _. Process square circle · Section B shows that the present invention is used for synthesizing color text. The method of the image · 5-5 Detailed description of the invention: The method for improving the quality of the circle image provided by the present invention is to apply a paper size to the Chinese β family standard rate (CNS > A4% ST2l0X297 ^) ( (Please read the notes on the back before filling out this page) d.
J C 線 經濟部中央橾準扃員工消費合作社印» AT -— ____ H7五、發明説明（ ） 含有圖形舆文字之原始影像分割為圈形影像區塊舆複數個 文字影像區塊’再將該複數個文字影像予以加強處理之 後，同時該原始影像亦予以加強處理，最後將處理後之文 字影像舆處理後之原始影像加以合成，所得之新影像較原 始影像清晰而不失真。傳統上，利用圈文分割法處理影像 之方法，其所得之新影像通常較原始影像失真，尤其具有 背景之文字部分’該背景部分常因此而消失。故本發明提 出一種可克服習知技術所衍生之問題，並改善圈文影像品 質之方法· 所謂二分法（Thresholding)係將影像中各點依據灰隋 值取一固定雄界值（Threshold)，例如128灰隋，較該值 低者歸為零，較該值高者歸為255，如此將影像中各點二 分為灰陏值較低之黑色舆灰陏值較高之白色，此種方法對 於處理單色文字影像之效果相當好。本發明使用動態二分 法（Dynamic Threshold丨ng)，故其所取之雄界值不一定， 一般取該部分影像區域之中問灰陏值，有文字之部分可予 以加強，而非文字部分則視為雜訊而予以消除· 本發明利用於處理灰隋圖形部分之方法為誤差分散 法（Error Diffusion)。若灰隋值由〇至255，以黑點分散 而形成不同灰陏值之影像，越亮的部分其黑點越少，同理 越晴的部分表示黑點越多。假投某一像素（口丨奸丨^之灰除值 為a，而一雄界值為128,該灰陏值a舆臨界值128之差， (請先閱讀背面之注意事項再填寫本頁) -裝. -、1-τ 線 本紙張尺度逋用中國國家揉準（CNS } A4現格（2丨0X297公漦） 經濟部中央標率扃—工消费合作社印製 AT 一_—__B7五、發明説明（） 乘1¾不同的加權指數，分別加成至該點之周团各點，一影 像中必含有各種不同灰階值之像素，依序計算影像中每一 像紊點’則可得到一灰陏效果較明顧之影像β 請參考第一圈所示，說明本發明之一較佳實施例， 用以改善單色圈文影像品質之方法流程方塊圈。首先，提 供—具有不同灰»圈形舆文字之單色原姶影像，將該原始 影像分割（步供101)以形成圈形影像區境及複數個文字影 像區塊，特別分割出該複數個文字影像區塊以加薄文字辨 識處理’本發明之方法係利用上述之二分法（下hresholding) 處理該複數個文字影像區块，低於一定灰暗值之文字影像 被處理成灰陏值更低之黑色影像，形成複數個二元文字影 像區塊（步麻102)，如此可將文字部分予以加深因而加強 其辨識度舆清晰度。 其次’將原始影像利用上述之誤差分散法（Error Diffusion)予以處理（步驟103)，以使該原始影像之圈形與 文字背景部分之灰階效果更明顢，而得到一新的灰陏影像 1〇4。最後將上述之複數個二元文字影像區堍1〇2與該灰 陏影像104做一合成影像處理（步驟1〇5)，利用一邏輯運 算操作將該複數個二元文字影像區塊102加成至該灰睹影 像中，然後輸出一新的合成影像（步驟1〇6)，如此即 使文字部分含有背景，其背景舆文字之不同灰陏效果仍然 被保留’而不會像§知技術將處理後文字部分之背景予以 •6 · 本紙张尺度遑用中羁國家標率（CNS )八4規展（2丨0X297公釐） {請先閲讀背面之注意事項再填寫本頁) d -裝 訂 a Λ五、發明説明（ 消除 經濟部中央橾準局貝工消费合作社印裝 上述之合成單色圈文影像步驟係利用一邏輯操作方 式，如第二國所示，一第一影像201舆第二影像2〇2美 有不同的影像内容，利用一"0R”邏輯運算操作之後，將第 二影像202加成至第一影像2〇1得到一合成影像2〇3，其 中合成影像203係將原第一影像2〇1與第二影像2〇2之 所有影像全部呈現，其中更含有部分重疊之影像。本發明 利用此獯邏輯運算方法，將已加強處理之二元文字影像區 塊加成至處理後之全部影像，所呈現之新影像，不只圈形 影像部分較原始影像具有較明顢對比效果，文字影像部分 同樣較原始影像更清晰易辨識，尤其是含有背景之文字影 像。 接著請參考第三圈’說明本發明之另一較佳實施例， 用以改善彩色圈文影像品質之方法流程方塊圈β首先提供 一含有彩色圈形舆文字之原始影像，本發明將調整該原始 彩色影像之Gamma值（步驟305)，以提高該原始彩色影 像之亮度而得到另一彩色影像306。另一方面，將該原始 彩色影像分割（步驟301)得到複數個彩色文字影像，利用 動態二分法（Dynamic Thresholding)將該複數個彩色文字 影像處理成複數個二元文字影像302，以該複數個二元文 字影像中之各個黑色文字做為處理文字之罩幕（mask)，找 出各黑字之像紊對應於原始影像之色彩值（步驛303)之 本紙張尺度適用中國國家標率（CNS ) Α4規格（210Χ297公釐） Ψ hi-1- i-·— ^^1 HI nn In· m Ia繁 (請先閲讀背面之注^^項再填轉本頁} 訂 經濟部中央揉準局舅工消费合作社印« A7 H7五、發明説明（) 後’依各革幕像素所對應之色彩值不同而予以加強處理， 例如增加其色彩飽和度（步驟304)。 最後將上述以二元文字為革幕之各像素對應原始彩 像色彩值之後，加強其色彩鉋和度（步驟3 04)直接寫入上 述彩色影像306做一合成影像處理（步驟3〇7)，然後輪出 一新的合成彩色影像（步驟308) 如此該合成彩色影像中 不輪是彩色圈形或彩色文字或具有背景之文字部分，其背 景舆文字、圈形之不同彩色對比效果仍然被保留，甚至較 原始影像更清晰而易辨識，而不會像習知技術將處理後文 字部分之背景予以消除。 請參考第四圈’顢示本發明用於合成彩色圈文影像 之方法•圈形影像401(含文字背景影像）為經過調整原始 影像Gamma值之影像’而文字影像402部分則為透過一 文字影像罩幕’以該文字罩幕之各像素對應於其原始影像 之色彩並調整其色彩飽和度後直接寫至圖形影像4〇1上， 以得到一新的輸出影像，該輸出影像之文字背景並不會因 影侏處理而消失，甚且較原始影像更清晰，更具有色彩對 比度* 本發明以一較佳實施例說明如上，僅用於藉以繫助 了解本發明之實施，非用以限定本發明之·精神，而熟悉此 領域技藝者於領悟本發明之精神後，在不脫離本發明之精 本紙珉足度適用t國踽家搮準（CNS ) A4规格（2丨0X297公釐） (請先閲讀背面之注f項再填寫本頁) d 裝 訂 ϋ 線！ B7五、發明説明（） 專。 其定 ， 而 換域 替領 化同 變等 之其 同及 等圍 及範 飾利 潤專 動請 更申 許之 些附 作後 可視 當當 ’ 圍 内範 圍護 範保 神利 (請先閏讀背面之注意事項再填寫本頁) Q.Printed by the Central Ministry of Economic Affairs of the Ministry of Economic Affairs of the People's Republic of China on the Consumer Consumption Cooperative »AT -— ____ H7 V. Description of the Invention () The original image containing graphic text is divided into a circle image block and a plurality of text image blocks. After the text image is enhanced, the original image is also enhanced. Finally, the processed text image is synthesized with the processed original image, and the new image is clearer than the original image without distortion. Traditionally, when using circle segmentation to process an image, the new image obtained is usually more distorted than the original image, especially the text portion with the background ', which often disappears as a result. Therefore, the present invention proposes a method that can overcome the problems derived from the conventional technology and improve the quality of the circle image. The so-called dichotomy (Thresholding) is to take a fixed threshold value (Threshold) of each point in the image according to the gray value. For example, 128 gray sui, the lower value is classified as zero, and the higher value is classified as 255. In this way, each point in the image is divided into black with lower gray value and white with higher gray value. This method It is quite effective for processing monochrome text images. The present invention uses a dynamic dichotomy method (Dynamic Threshold). Therefore, the value of the male boundary is not necessarily determined. Generally, the gray value in the part of the image area is taken. The part with text can be strengthened, and the part without text is treated as Elimination for Noise · The method used by the present invention to process the gray-sei graphics part is Error Diffusion. If the gray value is from 0 to 255, the black dots are scattered to form images with different gray values. The brighter parts have fewer black dots, and the clearer parts indicate more black dots. Fake a certain pixel (the gray division value of mouth 丨 ^ ^^ is a, and the male boundary value is 128, the difference between the gray value a and the critical value 128, (Please read the precautions on the back before filling this page) -Packing.-, 1-τ line paper size standard (CNS} A4 is now standard (2 丨 0X297) 扃 Central Standards of the Ministry of Economic Affairs 扃 —Printed by Industrial and Consumer Cooperatives Description of the invention () Multiply 1¾ different weighting indexes and add them to each point of the weekly group at that point. An image must contain pixels of various grayscale values. Calculate each turbulent point in the image in sequence. An image with a gray effect is more obvious. Please refer to the first circle to illustrate a preferred embodiment of the present invention to improve the quality of the monochrome circle image quality. First, provide-with different gray »The monochrome original image of the circle-shaped text, divide the original image (step 101) to form a circle-shaped image area and a plurality of text image blocks, and specifically divide the plurality of text image blocks to thin Character recognition processing 'The method of the present invention uses the above-mentioned dichotomy (hres below holding) to process the plurality of text image blocks, and text images lower than a certain gray value are processed into black images with lower gray value to form a plurality of binary text image blocks (step hemp 102). Partially deepen it to enhance its visibility. Secondly, the original image is processed using the above-mentioned Error Diffusion method (step 103), so that the circle shape of the original image and the grayscale effect of the background portion of the text Make it clearer, and get a new gray image 104. Finally, make a composite image processing (step 105) of the above-mentioned multiple binary text image areas 102 and the gray image 104 (step 105). A logical operation adds the plurality of binary text image blocks 102 to the gray text image, and then outputs a new composite image (step 106), so that even if the text part contains a background, the background and text of the text Different gray effects are still preserved 'instead of the background of the processed text part as in § knowing technology. • This paper size uses the National Standards of China Standards (CNS) Regulation 8 (2 丨 0X297 mm). {Please read the notes on the back before filling this page) d-Binding a Λ V. Description of the invention (Elimination of the printing of the above-mentioned synthetic monochromatic circle image by the Central Laboratories of the Ministry of Economic Affairs of the Bayong Consumer Cooperative, a logical operation Method, as shown in the second country, a first image 201 and a second image 202 have different image contents. After using a "0R" logical operation operation, the second image 202 is added to the first image. In 2001, a composite image 203 is obtained, in which the composite image 203 presents all the images of the original first image 201 and the second image 202, including a partially overlapping image. The present invention utilizes this method. The logical operation method adds the enhanced binary text image block to all the processed images. The new image presented, not only the circle image part has a sharper contrast effect than the original image, but the text image part is also more The original image is clearer and easier to identify, especially the text image with background. Next, please refer to the third circle 'to describe another preferred embodiment of the present invention. A method for improving the quality of the color circle image. The square circle β first provides an original image containing the color circle text. The present invention will adjust the Gamma value of the original color image (step 305), to increase the brightness of the original color image to obtain another color image 306. On the other hand, the original color image is divided (step 301) to obtain a plurality of color text images, and the plurality of color text images are processed into a plurality of binary text images 302 using Dynamic Thresholding. Each black text in the binary text image is used as a mask to process the text. Find out that the image disorder of each black text corresponds to the color value of the original image (step 303). CNS) Α4 size (210 × 297 mm) Ψ hi-1- i- · — ^^ 1 HI nn In · m Ia Fan (Please read the note on the back ^^ before filling in this page} Order the central Ministry of Economic Affairs Printed by the Bureau ’s Consumer Cooperatives «A7 H7 V. Description of the invention () After the ', the color value of each leather screen pixel is different and strengthened, such as increasing its color saturation (step 304). Finally, the above is binary After each pixel of the leather screen corresponds to the original color image color value, strengthen its color plan (step 304) and directly write the above color image 306 for a composite image processing (step 307), and then roll out a new one. Synthetic color shadow (Step 308) As a result, in the synthesized color image, the color circle or color text or the text portion with the background is not retained. The contrast effect of the background color and the circle color is still retained, and it is even clearer and easier than the original image. Recognize without removing the background of the processed text part like conventional technology. Please refer to the fourth circle 'to show the method for synthesizing the color circle text image of the present invention. The circle image 401 (including the text background image) is The image after adjusting the Gamma value of the original image ', and the text image 402 part is written through a text image mask'. Each pixel of the text mask corresponds to the color of the original image and its color saturation is adjusted. Then it is directly written to the graphic image. 4 〇1, in order to obtain a new output image, the text background of the output image will not disappear due to shadow processing, and even clearer than the original image, more color contrast * The present invention is described in a preferred embodiment As mentioned above, it is only used to help understand the implementation of the present invention, not to limit the spirit of the present invention, and those skilled in the art will understand this book. After the spirit of the Ming, the national standard (CNS) A4 (2 丨 0X297 mm) is applicable without departing from the fine paper of the present invention. (Please read the note f on the back before filling this page) d Binding线 Line! B7 V. Description of the invention () Specialized. It is determined, and the equivalent of the replacement domain, etc., and the equivalent of the special profit of the model and the profit of the jewelry, please apply for some attachments, and you can see it as Dangdang. Scope protection and protection (please read the precautions on the back before filling this page) Q.
.1T 泉- /V- 經濟部中央標準局貝工消费合作社印裝 张紙 本 準 標 家 國 國 中 用 適 A4) 釐 π 9 2.1T Quan- / V- Printed by the Central Standards Bureau of the Ministry of Economy