TWI270014B - Method of image dithering process for detecting photo and character automatically - Google Patents
Method of image dithering process for detecting photo and character automatically Download PDFInfo
- Publication number
- TWI270014B TWI270014B TW92119536A TW92119536A TWI270014B TW I270014 B TWI270014 B TW I270014B TW 92119536 A TW92119536 A TW 92119536A TW 92119536 A TW92119536 A TW 92119536A TW I270014 B TWI270014 B TW I270014B
- Authority
- TW
- Taiwan
- Prior art keywords
- image
- text
- background color
- graphic
- block
- Prior art date
Links
Abstract
Description
^/UU14 五、發明說明〇) 【發明所屬之技術領域 方法本ί明係有關於-種自動偵測圖文龙Λ ^特㈣關於將文#進㈣ ☆別處理的 ',、、白文件稿或彩色文件稿之細密度:專“,提高轉換為 【先前技術】 複製或ίίϊ彩色i件(或具有灰階特性之文件)的影印 的失真程;〜複製稿的細密程度,▲少複製稿 必須將文的特性不" 為里具有灰階特性的文件在傳真時,經常轉換 在轉換時,因為文件中的資料通常包含 ==及漸層部分不失真的呈現出來,但如果將! 用半色調處理,則文字的部分會呈現破碎 口此傳統上都採取將圖像與文字分開處理的作法。 製 但在習知圖像與文字的分離 或傳真的高擬真度與品質,而 方法上,常常為了顧及複 忽略分離與圖像文字處理^/UU14 V. Inventive Note 〇) [Technical Fields of the Invention] This method is related to - Automatic detection of graphic dragons ^ (4) About the text #进(四) ☆ Do not deal with ',,, white files The fineness of the manuscript or color document: "Improve the conversion to [Previous Technology] Copy or ίίϊ color i pieces (or documents with gray-scale characteristics) of the distortion process; ~ the fineness of the copy, ▲ less copy The manuscript must not have the characteristics of the text. For files with grayscale characteristics, the fax is often converted during conversion, because the data in the file usually contains == and the gradient part is not distorted, but if it will! In the case of halftone processing, the part of the text will appear to be broken. This traditionally takes the image and the text separately. However, in the separation of the conventional image and the text or the high fidelity and quality of the fax, In terms of methods, often in order to take into account the need to ignore separation and image word processing
1270014 五、發明說明(2) 時的速度過慢之門 曰由麻鰣士 ^ 問喊,依現有的圖文债測方法而丄 ^ 疋甶硬體電路眚钼 朴α 〆女而吕,主| 泠姑彻f U 、見’其將原有的圖像先進行爭德★ , 塗抹與模糊化)饴 $ —, 衫像加工(如 使用硬體震置達成 ^像及文子辨哉分離作業;在 為加工的程序,益 買成本’且因1270014 V. Invention description (2) When the speed is too slow, the threshold is screamed by the sorcerer, according to the existing graphic debt measurement method 丄^ 疋甶 hardware circuit 眚 mo pu α α 〆 而 ,, main |泠姑彻f U , see 'the original image of the first battle for the German ★, smudge and fuzzification 饴 — $ —, shirt image processing (such as the use of hardware shock to achieve ^ image and text recognition separation work; In the process of processing, the cost of buying is 'and because of
% ί目t 1無形中增加了整體圖文偵測的處理M ; >丄 周處理方法,其取樣參考點較多,护接广 也較大,需耗費較長的轉換時間。 採樣區域 【發明内容】 以太Ϊ於上述之發明背景中,因目前所知技藝的不足,所 本龟明提出一種切割文字及圖像區塊的方式,可 象色階的分佈特性而快速分離文字及圖像,如此將: 色《周處理的區域縮小,。 、 、本發明的另一目的是配合較少的半色調處理取樣點 以減少處理時間。 " 根據以上所述之目的,本發明提供了一種自動偵測圖 文之半色調處理方法,其步驟包含選取一原稿影像的底 色、將該原稿影像以該底色為基準進行圖文分離、將分離 的圖像部分以半色調處理、將分離之文字部分以二值法處 理、以及將處理完成之該圖像部分與該文字部分合併輸 出。 口% ί目1 1 invisibly increases the processing of the overall image detection M; > 丄 Weekly processing method, which has more sampling reference points and wider protection, which requires a longer conversion time. Sampling area [Summary of the invention] In the background of the above-mentioned invention, due to the shortcomings of the prior art, the turtle proposes a method of cutting characters and image blocks, and can quickly separate texts like the distribution characteristics of color gradations. And the image, so will: Color "Zhou processing area shrinks. Another object of the present invention is to process the sampling points with less halftones to reduce processing time. According to the above, the present invention provides a halftone processing method for automatically detecting graphics, the steps of which include selecting a background color of an original image, and separating the original image based on the background color. And separating the separated image portions by halftone processing, separating the separated text portions by a binary method, and combining the processed image portions with the text portions. mouth
1270014 五、發明說明(3) 上述之本發明光學裝置中,該半色調處理為一打散 法,且該打散法之取樣公式為第五圖示之取樣點模式(標 號50)乘以六十八分之一。1270014 V. INSTRUCTION DESCRIPTION (3) In the optical device of the present invention described above, the halftone processing is a scattering method, and the sampling formula of the scattering method is the sampling point mode (reference numeral 50) of the fifth figure multiplied by six. One in eighteen.
上述之本發明光學裝置中,一較佳的實施例之步驟 有:選取一原稿影像的一第一底色、將該原稿影像以該第 一底色為基準進行圖文分離、將該原稿影像以該第一底色 進行内縮、將内縮的該原稿影像以該第一底色進行橫切、 將橫切的該原稿影像以該第一底色進行縱切,將形成複數 個獨立影像區塊、將該獨立影像區塊再選取一第二底色、 以該第二底色,進行圖形與文字的辨識、如果該獨立影像 區塊為圖形,則記錄該圖像部分、如果該獨立影像區塊為 文字,則記錄該文字部份、當無法辨識為文字或圖形時, 以該第二底色對該獨立區塊進行内縮並以第二底色取代第 一底色,則自前述内縮的部分開始重複為之、將分離的該 圖像部分以半色調處理、將分離之該文字部分以二值法處 理、以及將處理完成之該圖像部分與該文字部分合併輸 出。 【實施方式】 本發明的實施例會詳細描述如下。然而,除了詳細描 述外,本發明還可以廣泛地在其他的實施例施行,且本發In the above optical device of the present invention, the steps of a preferred embodiment include: selecting a first background color of an original image, separating the original image based on the first background color, and separating the original image. Retracting with the first background color, transversely cutting the retracted original image with the first background color, and slitting the cross-cut original image with the first background color to form a plurality of independent images Blocking, selecting a second background color for the independent image block, and identifying the graphic and the text by the second background color, if the independent image block is a graphic, recording the image portion, if the independent If the image block is a text, the text portion is recorded, and when the text or the image is not recognized, the independent block is retracted by the second background color and the first background color is replaced by the second background color. The aforementioned indented portion is repeated, the separated image portion is processed in halftone, the separated portion of the text is processed in binary, and the processed image portion is combined with the text portion. [Embodiment] An embodiment of the present invention will be described in detail below. However, the present invention can be widely applied to other embodiments in addition to the detailed description, and the present invention
第8頁 1270014 發明說明(4) " ---- 明的範圍不受限定,其以之後的專利範圍為準。 再者,為提供更清楚的描述及更易理解本發明,圖 内各部分並沒有依照其相對尺寸繪圖,某些尺寸與其他相 關尺度相比已經被誇張;不相關之細節部分也未完全繪 出,以求圖式的簡潔。 本發明之不意圖如第一圖所示,首先、先偵測原稿影 像11中所包含的文字及圖像區塊,然後實施圖文分離丨2作 業(圖文分離步驟將於第二圖中說明),接著、對分離出 的圖像做半色调處理13 (Halftone Processing,本發明 貝施例以半色调中之打散法theri ng pr〇cess丨叫做說 明)’以求得最佳的圖像灰階表現,而對分離出的文字區 塊做一值法處理15 (Line Art Processing),將處理完 成的圖像與文字部分完成圖文合併14,達成複製或傳真的 需求’且可以清楚的顯示文字與圖像的灰階。 •然而、半色調之處理方法有不同的作法,諸如打散法 (Dithering)、順序打散法(Qr(jered Dithering)及誤 差擴散法(Error Diffusi〇n)等等,本發明中的半色調 ,理方法係以打散法為實施例,在圖像的打散法中,如第 二圖所不’將具有圖文混和的文稿偵測出圖像與文字部分 後,分別對不同區塊的圖像作影像打散處理,如此、在圖 像區塊中具有灰階時(一般可以呈現〇 (最黑)到255 (最Page 8 1270014 Description of the invention (4) " The scope of the description is not limited, which is subject to the scope of the patents that follow. Further, in order to provide a clearer description and a better understanding of the present invention, the various parts of the drawings are not drawn according to their relative dimensions, and some dimensions have been exaggerated compared to other related dimensions; the irrelevant details are not fully drawn. In order to simplify the schema. The present invention is not intended to be as shown in the first figure. First, the text and image blocks contained in the original image 11 are first detected, and then the image separation operation is performed (the image separation step will be in the second figure). (Note), then, the half-tone processing 13 (Halftone Processing, the method of the present invention is called the half-tone scattering method thei ng pr〇cess 丨 called the description) is used to obtain the best image. Like the grayscale performance, the line art processing is performed on the separated text block 15 (Line Art Processing), and the processed image is merged with the text part to complete the graphic 14 to achieve the requirement of copying or faxing' and it is clear The gray scale of the display text and image. • However, the halftone processing method has different methods, such as Dithering, Qer (Jered Dithering) and Error Diffusion (Error Diffusi〇n), etc., halftone in the present invention. The method of the method is based on the scattering method. In the method of scattering the image, as shown in the second figure, the document with the text and the text is detected, and the image is separated from the text. The image is used for image scattering processing, so when there is grayscale in the image block (generally it can be 〇 (blackest) to 255 (most
第9頁 1270014 五、發明說明(5) 白)的灰度),可以有效展現圖像的灰階程度,當然、也 可以1 0 2 4個灰階程度展現圖像,這些灰階程度的設定,係 取決於設計者對圖像的要求;接著、對文字作二值法處 理,如此文字區塊僅具有0 (最黑)或2 5 5 (最白)的兩種 值,因此可以呈現出最清晰的文字部分;最後、圖像與文 字於處理完成後合併,使得完成的文稿具有清晰的文字與 高吻合度的圖像。Page 9 1270014 V. Invention Description (5) White) Grayscale can effectively display the grayscale degree of the image. Of course, the image can also be displayed in the degree of grayscale. The setting of the grayscale degree , depending on the designer's requirements for the image; then, the text is treated binary, so that the text block has only two values of 0 (blackest) or 2 5 5 (whitest), so it can be presented The clearest part of the text; finally, the image and text are merged after processing, so that the finished document has clear text and a high degree of coincidence.
而前述之影像打散法中的取樣點,本發明所提出的作 法如第五圖所示,係將鄰近像素依第五圖所顯式的取樣點 模式(標號5 0 )乘以六十八分之一的公式為之。 … 本發明提 示意,在欲偵 以該底色為準 底色對已内縮 切(步驟24 ) (步驟25 ), 26 ),以該底 不能分辨,則 文字,所以依 取另一底色, 序,而如果可 2 9) ’如果區 出之自 測圖文 ,將該 的區塊 ,此時 於是針 色分辨 可能該 照所取 回到步 以分辨 塊判別 動偵測圖文 的區塊中取 區塊的邊幅 嘗試橫切( 、原來之區 對個別區塊 是否為文字 個別區塊仍 的底色為基 驟22,再度 是圖像時, 處理為文字 得底色設 内縮(步 步驟2 3 ) 塊可呈現 取的一個 或圖形( 然包含較 準或於該 實施一次 即記錄該 區塊時, 定(步 驟22 ) ,之後 數個個 底色( 步驟27 小範圍 個別區 自動偵 圖形區 即記錄 、實施縱 別區塊 步驟 ),如果 的圖像與 塊内再選 測圖文; 塊(步驟 該文字區The sampling point in the image scattering method described above, the method proposed by the present invention is as shown in the fifth figure, which is to multiply the adjacent pixel by the sampling point mode (reference numeral 5) which is shown in the fifth figure by sixty-eight. One of the formulas is this. The invention proposes that, in order to detect the background color, the background color has been internally cut (step 24) (step 25), 26), and the text cannot be distinguished by the bottom, so the other background color is taken. , sequence, and if it can be 2 9) 'If the self-test graphic is out of the block, the block will be resolved, and then the color of the needle color may be taken back to the step to distinguish the block discriminating motion detection graphic area. The margin of the block in the block is attempted to be cross-cut (the original region is based on whether the individual block is still the background color of the individual block. If the image is again, the image is set to the bottom of the text. Step 2 2) The block can present one or a graph (which is included in the calibration or when the block is recorded once (step 22), and then several background colors (step 27 small range individual area automatically The detection graphics area is the step of recording and implementing the vertical block), if the image and the block are selected, the block is replaced; the block (steps the text area)
12700141270014
當然、在步驟23中,如 ^ 亦進入步驟2 4中實施縱切;二二無法進行橫切時,則 法進行縱切時,則仍進入步=9r ·、:如果在步驟24中,無 步驟23中無法實施橫切且在舟’特別的時候,如果在 步驟27中無法分辨文字或圖二、 無法實施縱切,又在 步驟於第二圖中未示)。夕牯,則記錄為圖形區塊(此 關於第二 圖的示意,而 先以選取的底 域,如此可以 為對底色的影 的内縮後,對 許多的個別獨 設備進行是否 或疋圖像,則 切、縱切及判 個文稿區塊 或圖像的區 的時間,因 成弟一階段 如此將行成 軟體或硬體 分辨是文字 内縮、橫 圖所述的自動偉測圖文方法,更 得-個較佳的理解,該:中,Ϊ 色内縮,以去^ 衣除不包含任何文字 減少後續半色%々θ 1彳』又子 你士说μ 巴调或疋二值法處理 像地理將增加整體處理時間;完 該^縮的^塊實施橫切與縱切, 立區,,每—獨立區塊可以利用 為文字或圖像的判斷,如果不能 可針對個別獨立的區塊重複前述 讀程序。 :上所述僅為本發明之較佳實施例而 =發明之申請專利範®;凡其他為脫離本發明所Π艮 :神下所完成之等效改變或修飾,均應 專利範圍。 你r $之申凊Of course, in step 23, if ^ is also carried out in step 24 to perform slitting; if the second is unable to perform crosscutting, then if the slitting is performed, then step = 9r is still entered. If: in step 24, no In step 23, cross-cutting cannot be performed and when the boat is 'special, if the text or figure 2 cannot be distinguished in step 27, the slitting cannot be performed, and the step is not shown in the second figure). In the evening, it is recorded as a graphic block (this is a schematic diagram of the second figure, and the bottom field is selected first, so that after the indentation of the shadow of the background color, whether or not a plurality of individual devices are performed or not For example, the time for cutting, slitting, and judging the area of a document block or image, because of the stage of the younger brother, will be a soft or hard way to distinguish the text, which is the textual indentation and the horizontal picture. The method, more preferably - a better understanding, the: in, Ϊ color indentation, in order to go to the clothing, except that it does not contain any text to reduce the subsequent half-color % 々 θ 1 彳 』 』 你 你 你 说 说 说 巴 巴 巴 巴 巴Value method processing like geography will increase the overall processing time; after the completion of the shrinking block implementation cross-cutting and slitting, vertical area, each-independent block can be used as a text or image judgment, if not for individual independence The above-mentioned reading procedure is repeated for the block: The above is only the preferred embodiment of the present invention = the patent application of the invention; where the other is the departure from the invention: the equivalent change or modification performed by God , all should be patented. Your application for r $
第11頁 9, 1270014 圖式簡單說明 第一圖顯示本發明方法之整體示意圖; 第二圖顯示本發明之自動偵測圖文方法與細部步驟; 第三圖為本發明之第一圖之一個實施示意圖; 弟四圖為本發明之第-一圖之貫施不意圖,以及Page 11 9, 1270014 Brief Description of the Drawings The first figure shows the overall schematic diagram of the method of the present invention; the second figure shows the automatic detection graphic method and detailed steps of the present invention; the third figure is one of the first figures of the present invention. Schematic diagram of implementation; the fourth figure of the invention is the intention of the first figure of the invention, and
第五圖為本發明之影像半色調之打散法取樣點模式及 其計算公式。 主要部分之代表符號:The fifth figure is the sampling point mode of the image halftone of the present invention and its calculation formula. Representative symbols of the main part:
標 號1 1 原 稿 底 色 標 號12 圖 文 分 離 標 號13 半 色 調 處 理 標 號14 二 值 法 處 理 標 號1 5 圖 文 合併 步 驟2 1 選 取 底 色 步 驟2 2 内 縮 步 驟2 3 橫 切 步 驟24 縱 切 步 驟25 形 成 個 別 區 塊 步 驟26 第 _ 一 雷 射 二 極 步 驟27 選 取 底 色Reference numeral 1 1 Original background color label 12 Graphic separation number 13 Halftone processing label 14 Binary processing label 1 5 Graphical combining step 2 1 Selection of background color 2 2 Retraction step 2 3 Cross-cutting step 24 Longitudinal cutting step 25 Forming individual blocks Step 26 First _ Laser diode step 27 Select background color
第12頁Page 12
LO 1270014 圖式簡單說明 步驟28 :記錄文字區塊 步驟2 9 :記錄圖形區塊 標號5 0 :打散法取樣點模式LO 1270014 Schematic description of the steps Step 28: Record the text block Step 2 9: Record the graphic block Label 5 0: Breaking sampling point mode
11111 第13頁11111第13页
Claims (1)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW92119536A TWI270014B (en) | 2003-07-17 | 2003-07-17 | Method of image dithering process for detecting photo and character automatically |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW92119536A TWI270014B (en) | 2003-07-17 | 2003-07-17 | Method of image dithering process for detecting photo and character automatically |
Publications (2)
Publication Number | Publication Date |
---|---|
TW200504614A TW200504614A (en) | 2005-02-01 |
TWI270014B true TWI270014B (en) | 2007-01-01 |
Family
ID=38318530
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW92119536A TWI270014B (en) | 2003-07-17 | 2003-07-17 | Method of image dithering process for detecting photo and character automatically |
Country Status (1)
Country | Link |
---|---|
TW (1) | TWI270014B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI401669B (en) * | 2008-04-11 | 2013-07-11 | Novatek Microelectronics Corp | Image processing circuit and method thereof for enhancing text displaying |
-
2003
- 2003-07-17 TW TW92119536A patent/TWI270014B/en not_active IP Right Cessation
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI401669B (en) * | 2008-04-11 | 2013-07-11 | Novatek Microelectronics Corp | Image processing circuit and method thereof for enhancing text displaying |
Also Published As
Publication number | Publication date |
---|---|
TW200504614A (en) | 2005-02-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP3305772B2 (en) | Method for detecting handwritten instruction image using morphological technique | |
JP5274495B2 (en) | How to change the document image size | |
US7751648B2 (en) | Image processing apparatus, image processing method, and computer program | |
US8494280B2 (en) | Automated method for extracting highlighted regions in scanned source | |
US7729013B2 (en) | Image processing apparatus, image processing method, and computer program product | |
JP3601658B2 (en) | Character string extraction device and pattern extraction device | |
US9641705B2 (en) | Image forming apparatus for reading indicia on a sheet and inserting images on a subsequent printed sheet at a location corresponding to the location of the read indicia | |
JP5262493B2 (en) | Image processing apparatus, image processing method, program, and computer-readable storage medium | |
JP2009015820A (en) | Image processing apparatus, image forming apparatus, image transmitting apparatus, image reading apparatus, image processing system, image processing method, image processing program, and recording medium therefor | |
TWI270014B (en) | Method of image dithering process for detecting photo and character automatically | |
US20210286946A1 (en) | Apparatus and method for learning text detection model | |
US10887491B2 (en) | Image processing apparatus for processing of highlighted regions | |
JP2003298799A (en) | Image processor | |
CN107016320B (en) | Method for improving image security level identification accuracy rate based on Chinese lexicon | |
JP2003087562A (en) | Image processor and image processing method | |
Ramponi et al. | Digital automated restoration of manuscripts and antique printed books | |
US7933446B2 (en) | Method and apparatus for automatically detecting image and character data | |
US10560606B2 (en) | Image processing device, image processing method, and non-transitory computer readable medium | |
Nimbs | NudiKey: an illustrated, interactive identification key to the families of Australian heterobranch sea-slugs (Mollusca: Gastropoda) | |
JP2008154216A (en) | Image processing method and device, image forming device, document reading device, computer program, and recording medium | |
JP2010183487A (en) | Color signal processor, color signal processing method, program, and recording medium | |
JP2006072839A (en) | Image processing method, image processing apparatus, image processing program and recording medium | |
JP2019153230A (en) | Information processor and information processing program | |
Kamal et al. | Braille to Text Translation for Bengali Language: A Geometric Approach | |
US8340434B2 (en) | Image processing apparatus, image processing system and computer readable medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
MM4A | Annulment or lapse of patent due to non-payment of fees |