1282947 九、發明說明： 【發明所屬之技術領域】 本發明屬於書寫文字或符號之數位表現相關技藝的領 域，尤指產生邊影導向之灰階字型的一種系統、方法、以 及電腦可讀取媒介裝置。 【先前技術】 目前市面上存在著許多產生漢字之字型產生系統，漢 _ 字字塑係由一群代表漢字之象形圖樣所組成，被組成之漢 字數是成千上萬個，例如通用的中文就包含有兩萬個以上 Ο 傳統的電腦技術用於產生漢字的字型係採用週邊形 狀的表現方法’此種技術描述於八(1<118〇11-界6316}^出版（1985 年）（’Postscript Language Tutorial Co〇kbook，，(Adobe1282947 IX. Description of the invention: [Technical field to which the invention pertains] The present invention pertains to the field of digital expression related techniques for writing characters or symbols, and more particularly to a system, method, and computer readable method for generating gray-scale fonts for silhouette-oriented Media device. [Prior Art] At present, there are many font-generating systems for generating Chinese characters. The Chinese-word-shaped plastic system is composed of a group of pictographic patterns representing Chinese characters. The number of Chinese characters is tens of thousands, such as general Chinese. It contains more than 20,000. The traditional computer technology used to generate Chinese characters uses the expression of the surrounding shape. This technique is described in the eight (1 <118〇11-Boundary 6316}^ publication (1985) ( 'Postscript Language Tutorial Co〇kbook,, (Adobe
System inc·) ’這種方法用組合成文字外形的一些直線段及 鲁=線段儲存起來代表產生字型的資料。這種技術有一些缺 2萬貫先，由於不同的字型外形必需一一儲存來產生成千 其▲個不同文字的形狀，所以記憶體的空間需求相當大。 ^二，字型週邊資料係依高解像度之需求儲存，只能產生 =型像品質的字型，這些資料並不適合於產生低解像度之 型次^外一種產生漢字字型之方法係採用筆劃基礎的字 每二料丄其中文字的每個筆劃是分別被定義成可用來組成 予的子型資料’而每字係由多個筆劃互相疊合或交叉形System inc·) ’ This method uses a combination of straight line segments and ru=line segments that are combined into a text shape to represent the data that produces the font. There are some shortcomings in this technology. Since different font shapes must be stored one by one to produce thousands of different characters, the memory space is quite large. ^ Second, the font surrounding data is stored according to the demand of high resolution, can only produce the type of image quality, these data are not suitable for the generation of low resolution type. The other method of generating Chinese characters is based on strokes. Each word in the text is defined as a subtype of material that can be used to form a ', and each word is superimposed or crossed by multiple strokes.
1282947 成。筆劃基礎的字型資料包含關鍵點、寬度值、特徵點和 曲線比例值，這些資料的組合用來定義每個筆劃的週邊形 狀’有關筆劃基礎字型之建造及顯示已經詳述於美國專利 案號5,852,448、6，151，〇32和6，157,390。筆劃基礎字型技 術適用於低s己憶體谷量，以及在高低解像度兩者並用之需 求下的場合。 再另一種產生漢字字型之方法係採用圖樣基礎的字 型^料，其中文字的每個圖樣是分別被定義成可用來組成 每字的字型資料，而每字係由至少一個圖樣組成，且每個 圖樣係由至少一個筆劃組成。圖樣基礎的字型資料亦如筆 劃基礎之定義一樣係用關鍵點、寬度值、特徵點、和曲線 比例值來表現，有關圖樣基礎字型之技術已經詳述於美國 專利案號 6,501，475 和 6,661，417。 $ 如今隨著面板顯像技術的進步發展，文字已經可以用 灰階影像顯示在手機、掌上型電腦以及其他電子儀器上， 以前只能以黑白二位元階影像的顯示變成可以用累（〇) 到白（ 255)之間的256的灰階影像來顯示，由於用二位 元階來顯示文字的技術用於灰階螢幕會造成文字影像口 質變差，這種品質變差的現象對具有複雜幾何形=之$字 的情況尤其顯著，例如圖一所顯示之中文字9，其^ 之方格Π在灰階顯像時就如圖一（E)所示，這個方格' U 包含四個圖素16a_16d，參看其中之16b，它包含一佔^文 字9的部份區域A1以及沒有佔據文字9的區域A2，用傳 統的技術來產生16b的灰階值係用下式計算： 1282947 - 灰階值=(Α1χ黑+Α2χ白）/(Α1 + Α2) ··· (1) 在以下的敍述裏，方程式（丨）稱為”灰階值計算式，，，圖一 (A)示出用灰階值計算式運算後之文字1〇顯示，其文字 10之筆2〇和筆劃22的部份顯示如圖一（b )所示，圖 (B )示出（4x4 )圖素於I8a-18d的四列以表示筆割2〇 和筆劃22之灰階值，如圖示，筆劃2〇佔有圖素列18&的 _ 50%下半部以及圖素列18b的5〇%上半部，同樣地，筆劃 22佔有圖素列18c的50%下半部以及圖素列18d的5〇%上 半。卩，據此，這四列18a_ 18d的灰階值皆有如下式計算所 得相同的灰階值： 灰階值=0·5χ黑 + 〇·5χ白···（2) 計算結果使得這四列18a-18d是用同一灰階值來整個塗 _ 滿，於是形成分不出是由二個筆劃組成的區塊。圖一（C) 示出用傳統技術顯示圖一（A)文字10的結果，可以看出 這種將筆劃混成一塊的灰階影像確實使得文字顯示品質 嚴重變差。 、 綜上所述，開發一種有別於傳統技術之系統、方法以 及電腦可f買取媒介裝置可同時來顯示高品質之灰階及二 階的字型技術是被渴望的一項需求。 — 1282947 【發明内容】 本务明提供m方法以及電腦可讀 置，用以產生邊影導向之灰階字型影像，取得高品質: 型顯像於點矩陣晝面上以及灰階影像晝面上。 根據本發明的一項揭露，本發明提供-種方法用來定 邊影導向表現之基本筆劃集，並據以建造 。這種方法首 拓樸形狀的筆劃組合成不同文字的筆劃，因 =構=多的文字型狀’根據美國專利案號5,852,梢、 =::，157,390之描述，一群相似拓樸形狀的筆割 之，：有文字集裏所有的筆劃是由至少一個=筆 寬度值、特徵點和曲線比例值^案所述’由關鍵點、 ^0 ㈣錢Μ表現相㈣_生成過 域碼來定:基十二7二本:劃時’可以指定邊影導向區 基本筆劃在生成筆劃時，應進行計算灰階值（亦即 式⑴所計算出的圖素）=的部份區域涵蓋了由方程 的上半部區域;計算灰階值，這樣被定義 型設計師選-適當的邊就稱為一個邊影區域。字 裏，使得藉由該基本筆來生 =定義於每一基本筆劃 王成的所有相似形狀的筆劃皆 1282947 會同樣產生如定義所規範的邊影區域，並能將被產生的邊 影區域呈現於灰階晝面上。本發明的一項揭露乃是在基本 筆®彳的定義中指定邊影區域碼，就已經自動指定所有對應 之相似筆劃的邊影區域。此後，字型設計師就可觀測整個 被生成文字集裏的所有筆劃顯示在灰階晝面的效果。如果 第一次嘗試的效果不如想像中的預期，那麼字型設計師可 以再重新調整邊影區域範圍於基本筆劃之定義裏；一旦效 果被確認滿意後，就可將邊影導向表現之基本筆劃的資料 儲存起來供組字使用。、 很踝枣發明的 v ，+乃沄巴言秌用電腦圖形之 機界面工具’並透過適當之影像分析的技術，於步驟中 s引從事邊料向表現之基本策劃建 =:提供:電腦圖形之人機界面I具 | 、，包含提供各項指令完成本方法之各項揭露。、 現以月之另一項觀點，本發明提供-種方法來呈 ΐ 字。每—個文字係由-個或多個筆 比值，且母個筆劃係由關鍵點、寬度值 = =及二 =邊影叫對於每-個形= 定義二 ==形狀及座標位置，而指定於筆劃 射在各個圖素所占的面積係筆劃的邊影區域投 值，最後’本方法乃是根據這此了圖素的灰階 個晝面上的圖素值。 二被决疋之灰階值來呈現各 1282947 本發明更且提供一種反異化方法於減缓鋸齒狀影像 之異化現象的產生，反異化方法涉及分析每一圖素周遭的 影像拓樸結構，並依照分析後的拓樸結構選一預先定義之 加權值矩陣計算圖素平滑化的動作以產生最佳的反異化 效果，這些被預先選定的加權值矩陣為計算圖素平滑化取 得反異化圖素值的規則依據。 【實施方式】 Φ 圖二示出實施本發明方法之典型的計算環境，其中有 -包括顯示儀27及印表儀28的通用數位電腦26，用以實 現本發明的各種方法。顯示儀28，印表儀28，以及其他 輸出儀（如螢幕晝面顯示器和印表機等）可以顯像、印出、 以及輸出本發明方法所產生之字型。其他為此技藝所熟知 之型式的計算系統，例如網路或主機系統亦可以用來實施 本發明方法。 以下所述本發明用於定義或產生文字所含之筆劃，亦 • 可被瞭解為定義或產生文字所含之圖樣，如發明背景所 述，圖樣為構成文字的單元，此單元由一個或多個筆劃組 成。在本發明裏不管是筆劃或圖樣為基礎，兩者皆是由關 鍵點、寬度值、特徵點、以及曲線比例值加予定義，且用 這些定義的基本筆劃或基本圖樣來決定字型資料以及產 生字型，因此若不會有混淆之顧慮，此後所述’’筆劃’’涵蓋 筆劃和圖樣二者。 圖三(A)示出-基本筆劃30,它係由關鍵點（K1和K2)， 7 1282947 寬度值（Wl)，特徵點（fl，f2，f3，和f4)，以及曲線比例 值（未示出）所定義，如美國專利案號5,852,448、6,151,032， 和6，157,390所述，漢字的特性之一是很多字可共同擁有 很多的比劃，換言之，相似筆劃可用於很多字裏，例如圖 三(A)所示之水平筆劃30可以用在很多字裏也可以重複用 在同一字裏，例如圖一（A)文字10裏的筆劃20和22。且 如上述專利案所述，一組相似筆劃群可用-基本筆劃來代表 其拓樸形狀，每一基本筆劃係首先由關鍵點和寬度值加予 ❿ 定義。 關鍵點（K1和K2)通常被選為在基本筆劃之線上或彎 曲處或彎曲處之切線交點，始得移動關鍵點就可以變基本 筆劃的外形，以產生各式各樣和它相似的筆劃，寬度值(W1) 通常被選在基本筆劃裏之二點間的距離，使得改變寬度值 亦可產生各式各樣不同粗細的相似筆劃。每個基本筆劃皆 有一些方程式來表示怎樣從關鍵點及寬度值決定特徵點， 例如圖三（A)，那些決定特徵點（Π，f2，f3，和f4)的方程 • 式為·. fl= K1+W1/2 f2= K1-W1/2 f3= K2+W1/2 f4= K2-W1/2 ......(3) 因此，特徵點係經由預先定義它和關鍵點以及寬度值 之空間關係所決定，如圖所示，特徵點係選在基本筆劃之 8 1282947 外圖形狀上正轉換切線或曲率方向處。更且，每個基本筆 劃亦也一些預先定義之曲線比例值，它們係根據解像度之 層級所預先建造之數值群，用以產生相鄰二個特徵點之間 在不同解像度時的曲線形狀，有關樹狀結構之曲線比例值 的建造係如上述專利案所描述。 每個基本筆劃係用一指定的筆劃數目來確認，每個被 定義之基本筆劃可藉由移動關鍵點以及改變寬度值來定 義那些對應它的相似筆劃群，一旦關鍵點被移動以及寬度 Φ 值被改變後，就會那些預存之方程式計算出新的特徵點位 置，並經由預存之樹狀結構的曲線比例值計算出相鄰二個 新特徵點間新的曲線形狀於各個不同解像度的層級裏。這 種產生相似筆劃之過程已經詳述於上述專利案裏。 如前述發明背景所述，當上述筆劃被定義成顯像於二 位元階之點矩陣影像晝面用來顯像於灰階影像畫面時，就 會產生品質變差的現象，所以，根據本發明的一項揭露， 每個基本筆劃（以及因此基本比劃所生成之相似筆劃）也增 • 加了一項指定邊影區域碼來定義在筆劃内產生灰階邊影 區域的方式。 圖三(B)示出分別指定邊影區域碼”10”和”01”於二筆 劃34和36裏，一個邊影區域碼用來定義在每個筆劃裏的 灰階邊影區域，當此區域投影到灰階影像晝面時會根據其 落在筆劃裏之格子面積決定此格子圖素的灰階值，例如筆 | 劃34指定-邊影區域碼”10”，表示其上半部，34a係由點 (ΚΙ，Π，f3，f2)所圍成之邊影區域，且由其被指定之灰 1282947 階值來計算其所座落格子之圖素值。換言之，邊影區域亦 係由關鍵點、寬度值、特徵點、以及曲線比例值所圍成的 部份筆劃區域，另外屬於下半部之34b不是一個邊影區 域，且被調整為黑點（不會造成其圖素格子成為灰值）。 筆劃36被指定一邊影區域碼”01”，在本例裏意指其上半部 36a不是邊影區域且被調整為黑點，而下半部36b係由點 群（k、k2、f4、f2)所圍成之邊影區域，當被顯像時需計 算其上圖素格子的灰階值。對熟悉本技藝人士而言，上述 φ 邊影區域碼僅是其中一個範例，事實上邊影區域碼係根據 基本筆劃形狀及它在字型產生上之需求而有多樣的定義 方式。僅是為了目前的描述用，在每一筆劃上所指定之邊 影區域碼意指筆劃内部的部份區塊需要去計算落在其上 之圖素的灰值，每個筆劃之邊影區域碼係被儲存於筆劃的 身份碼（筆劃數目或第幾個筆劃）裏。 圖三（C)出示圖一（A)之文字10裏的二個筆劃20 和22，且它們各別皆有一邊影區域碼”10’’。如圖三（B) • 所示，當影像時會有四個圖素列18a-18d座落其上。由於 每個筆劃的邊影區域碼’’10’’在顯像時會變成灰階值而使得 這二個筆劃在顯像時會如標示39所示，它們在視覺上是 兩個被區隔的筆劃。產生這種結果是因為下半部非邊影區 域的影像是採用指定的灰階值，例如指定其灰階值為白 色，所以，採用適當定義邊影區域導向所表現的筆劃來組 合文字，就可以顯示平滑化的灰階影像字型，譬如圖一（A) 所示文字10之二個筆劃20和22混疊在一起的現象就可 1282947 、由疋義邊影區威導向之基本筆劃的表現來避免掉。 圖四是定義一級帶有邊影導向表現之基本筆劃的流 禾固這些被定義後的基本筆劃就用來定義文字的灰階影 像字型。如方塊40所示，字型設計師顯示一組文字集（顯 不在各種不同的點矩陣尺寸）供定義邊影導向之參考用， 基本上，一字型設計師是一圖形設計家，其任務是創造字 型資料庫供商業用。在文字集的每個文字係由一個或多個 筆劃組成，且所有組成文字的筆劃更可切分成群，每群包 _ 含一組相似筆劃，每組相似筆劃群可用一基本筆劃來代表 它們之拓樸形狀的生成，基本筆劃係用關鍵點、寬度值、 特徵點和曲線比例值為參數所定義之筆形生成的程式。 如方塊42所禾，字型設計師將每個基本筆劃定義成 邊影導向的表現。字型設計師經由視覺確認每個基本筆劃 裏指定一邊影區域的最佳定義，例如觀測一個筆劃和另外 相接或相鄰筆劃之間的狀況來指定邊影區域以避免它們 在低解像時會產生重疊的現象，然後再指定邊影區域碼給 φ 該基本筆劃，在基本筆劃之定義裏被指定的邊影區域碣於 組字時就會呈現在對應此基本筆劃之各個相似筆劃裏。 如方塊44所示’字型設計師將所有用邊影導向表現 之文字顯示出來，每一文字之筆劃皆由其邊影區域碼決定 其灰階影像的顯像；換言之，僅有邊影區域所座落之圖素 格子才會計算該些圖素的灰值。 、 如方塊46所示，字型設計師用其視覺觀看所有被顯 ^ 示之灰階顯像的文字，若這些文字已經被適當地減建成灰 11 1282947 • 階子型，已經沒有筆劃被疊合在一起的現象時，就完成了 本方法之灰階字型的建造，否則再回到方塊42重新對某 些基本筆劃定義成新的邊影區域導向之表現 ，再如方塊44 所顯示’由這些重新定義邊影區域之基本筆劃所生成的那 些灰卩6衫像文字；再如方塊46所示，字型設計師再觀看 這些灰階影像文字的品質是否滿意，反復這樣的動作直至 滿意為止。 根據本發明的另一項觀點，本發明提供一種顯示以邊 # 影導向表現之灰階影像文字的方法，圖五示出這種方法之 一種流程圖，例如圖五所示之方法可用於圖四的方塊4， 圖五所示之步驟係由一儲存於電腦裏的影像顯示程式或 引擎執行，以顯示灰階影像文字於灰階晝面之顯示器上。 如方塊52所示，這種方法係根據點矩陣尺寸建造〆 文字的掃瞄影像，例如圖六（A)所示之一 0x6點矩陣， 其中沿箭頭67所示者為其一條掃瞄線，在範例說明中， 一文字影像64將被顯示於點矩陣60上。 _ 如圖五方塊54所示，對於點矩陣60裏之每個圖素’ 會有一個合適的演算式計晝出邊影導向表現之圖素的炙 值。例如方程式（1)用來計算每個圖素之灰值 ，如方塊今9 所示’這些被计算出的灰值就構成文字的灰階影像。供 疋’依上述被計算出來的灰值直接用顯示文字之灰階#像 會產生邊緣呈鋸齒狀或異化的影像字型，因此也可以捧讳 反異化的方法來降低鋸齒狀的突顯，例如圖五方塊56和 方塊58所示的反異化方法。在影像圖形之處理上有很爹 12 1282947 反異化方法被提出，本發明所提供如圖六(B)所示的一種反 異化方法，其中在3x3圖素區70之九個圖素分別各有如 圖五方塊54所述被計算出來的灰值，而座落在圖素區70 之中心的P5係欲被重新計算其新的灰值之圖素，P5之新 灰值的計算係根據其周圍的八個圖素灰值依其加權多寡 所形成之均勻混合結果，其計算式如下： P5之新圖素值=§[出，值_] —¥ , 其中Wi為圖素I的加權值，且熟悉本技藝人士皆知矩陣 W之加權值通常係呈高斯分佈，亦即，若中心點之加權值 最大，則其他離中心點之加權值係依高斯分佈遞減。 傳統之反異化方法係取固定高斯分佈之加權值矩陣 而本發明之反異化方法係依據筆劃的書寫方向取不同的 加權值矩陣分別計算P5。例如考慮之圖素66和圖素68座 落於文字影像64(圖六(A))的反異化計算方式就如圖六（C) 和圖六（D)所示。 ® 圖六（C)考慮包圍圖素66之文字影像的3x3圖素區， 從計算最左及最右圖素76和78之差值會得到如80所示 之三類：零、正值、或負值的書寫值，例如圖六（C)所示之 書寫值為<0，+，+>，根據書寫值就可以選出數個預複製 之加權值矩陣的其中一個矩陣，例如選加權值矩陣W， (82)，它是預製的加權值矩陣用來計算書寫值<0，+，+> 以產生最佳反異化效果。因此，中心位置的圖素66係採 13 1282947 用被選出之加權值矩陣wl(82)代入方程式（4)裏計算出 P5 (圖素66)之反異化的灰值。 同樣地’圖六（D )考慮壞繞圖素6 8之3 χ3圖素區’ 經計算圖素列83 a，83b，和83 c後得到圖六（D)之書 寫值84為 <-，-，->根據此書寫值84於是選了複製之加 權矩陣w2 ( 86)，用經代入方程式（4)裏計算出p 5 (圖 素68)之反異化的灰值，以達到最佳的反異化效果·熟悉 本技藝之專業人士亦可瞭解到，加權值矩陣的元素亦可採 φ 用其他不同於本發明所揭露的數值，用於不同之書寫值的 狀況裏，也可產生同樣的反異化效果·本發明的各種方法 可藉由一圖形人機界面（GUI)的工具來實施，例如可用 視窗為基礎之作業系統的GUI工具，且一些方法裏的一個 或多個步驟亦可，透過影像方析之技巧自動完成。 一旦本發明的優先實施例己經披露且詳述如上，其他 類似或相當之想法及作法的修飾就很容易被略具本行業 技能的人士所理解，本發明的權利請求範圍包含本實施例 • 以及其他類似的實施例。 【圖式簡單說明】 前述觀點及本發明之既存優點更可從以下之圖式的 描述及解說中窺知。 圖一係用圖例示出採用傳統二階（黑白）點矩陣影像 顯示灰階影像字型會遭遇到的困境。 圖二（A)〜(E)示出用以實施本發明方法之系統或圖形 14 1282947 人機界面工具的方塊圖。 _ 圖三（A)〜(C)示出本發明所揭露之一採用關鍵點、寬 度值、特徵點等定義基本筆劃及其邊影碼的範例。 圖四為本發明所揭露之一定義邊影導向表現灰階影 像字型方法的流程圖。 圖五為本發明所揭露之一顯示邊影導向表現灰階影 像字型方法的流程圖。 圖六（A)〜(D)示出本發明所揭露之一用於降低形成鋸 ❿ 齒邊灰階影像字型的反異化方法。 【主要元件符號說明】 9中文字 10文字 12方格 16a圖素 16b圖素 16c圖素 16d圖素 18a圖素列 18b圖素列 18c圖素列 18d圖素列 20筆劃 22筆劃 27顯示儀 28印表儀 26通用數位電腦 30筆劃 34筆劃 36筆劃 34a上半部 34b下半部 60點矩陣 64文字影像 66圖素 15 1282947 68圖素 70圖素區 78圖素 83b圖素列 84書寫值 67箭頭 76圖素 83a圖素列 83c圖素列1282947 成. The stroke-based font data contains key points, width values, feature points, and curve scale values. The combination of these materials is used to define the perimeter shape of each stroke. The construction and display of the stroke-based basic fonts is detailed in the US patent case. Nos. 5,852,448, 6,151, 〇32 and 6,157,390. Stroke basic font technology is suitable for low s-recovery volume, and where both high and low resolution are used together. Yet another method for generating Chinese character fonts is to use a font based on a pattern, wherein each pattern of the text is defined as a font material that can be used to form each word, and each word is composed of at least one pattern. And each pattern is composed of at least one stroke. The font data based on the pattern is also represented by the key points, width values, feature points, and curve scale values as defined by the stroke basis. The technique for the pattern basic font has been detailed in U.S. Patent No. 6,501,475 and 6,661,417. $ Nowadays, with the development of panel imaging technology, text can be displayed on mobile phones, palmtop computers and other electronic instruments with grayscale images. In the past, only black and white binary image display can be used to become tired (〇 ) to 256 grayscale images between white (255) to display, because the technique of displaying text with two-dimensional order is used for grayscale screens, the vocal quality of text images is deteriorated, and this quality deterioration phenomenon has The case of the complex geometry = $word is particularly significant, for example, the text 9 in Figure 1 is displayed, and the square of the ^ is shown in Figure 1 (E) when the grayscale is displayed. This square ' U contains The four pixels 16a_16d, see 16b thereof, contain a partial area A1 occupying the text 9 and an area A2 not occupying the text 9. The conventional technique is used to generate the grayscale value of 16b, which is calculated by the following formula: 1282947 - Gray scale value = (Α1χ黑+Α2χ白)/(Α1 + Α2) ··· (1) In the following description, the equation (丨) is called “the gray scale value calculation formula,,,, Fig. 1 (A) It shows the text 1〇 after the calculation of the gray-scale value, and the pen of the text 10 is 2〇 And the part of the stroke 22 is shown in Figure 1 (b), and Figure (B) shows the (4x4) picture in the four columns of I8a-18d to indicate the gray scale value of the stroke 2 〇 and the stroke 22, as shown in the figure. It is shown that the stroke 2〇 occupies the lower half of the _50% of the pixel sequence 18& and the upper half of the 〇50% of the pixel array 18b. Similarly, the stroke 22 occupies 50% of the lower half of the pixel array 18c and the pixel The upper half of the 5〇% of the column 18d. According to this, the grayscale values of the four columns 18a-18d have the same grayscale value calculated as follows: Grayscale value=0·5χblack+〇·5χ白·· (2) The calculation results are such that the four columns 18a-18d are all coated with the same gray scale value, so that a block composed of two strokes is formed. Figure 1 (C) shows the conventional technique. Displaying the result of Figure 10 (A), it can be seen that the gray-scale image that blends the strokes into one piece does cause the text display quality to deteriorate badly. In summary, a system and method different from the conventional technology is developed. And the computer can buy a medium device to simultaneously display high-quality grayscale and second-order font technology is a demand that is desired. — 1282947 [Invention 】 The present invention provides a m method and a computer readable set for generating a gray-scale font image of a shadow-oriented, achieving high quality: type imaging on a dot matrix surface and a grayscale image surface. In one disclosure, the present invention provides a method for determining the basic stroke set of the silhouette-oriented performance and constructing it according to the method. The method of the top-top shape of the stroke is combined into a stroke of different characters, because the = structure = more The text type 'according to the description of U.S. Patent No. 5,852, tip, =::, 157,390, a group of similarly shaped pens are cut: all the strokes in the text set are composed of at least one pen width value, feature points And the curve scale value ^ case described by 'key point, ^0 (four) money Μ performance phase (four) _ generated over-domain code to determine: base 12 7 two: stroke time 'can specify the shadow-guided area basic stroke in the generation of strokes When calculating the gray scale value (that is, the pixel calculated by equation (1)) = part of the area covers the upper half of the equation; calculate the gray scale value, so that the definition designer chooses - appropriate The edge is called a silhouette area. In the word, all the similarly shaped strokes defined by each basic stroke will be generated by the basic pen. The 1282947 will also produce the silhouette area as defined by the definition, and the generated shadow area can be rendered. On the grayscale surface. One of the disclosed aspects of the present invention is that the silhouette area code is specified in the definition of the basic pen®, and the silhouette areas of all corresponding similar strokes are automatically designated. After that, the font designer can observe the effect of all the strokes in the generated text set displayed in the grayscale. If the effect of the first attempt is not as expected in the imagination, then the font designer can re-adjust the range of the shadow area in the definition of the basic stroke; once the effect is confirmed, the basic stroke of the silhouette can be directed to the performance. The data is stored for use in the group. The v, + is the invented v, + is the use of computer graphics interface tool 'and through the appropriate image analysis technology, in the steps s cited in the basic planning of the side to the performance of the construction =: provide: computer The graphical human-machine interface I, |, contains various instructions to complete the disclosure of the method. In another aspect of the present month, the present invention provides a method for presenting a word. Each text is composed of one or more pen ratios, and the parent strokes are determined by key points, width values ==, and two = edge shadows for each-shape = definition two == shape and coordinate position, and specified The pen strokes the area occupied by each pixel in the shadow area of the stroke, and finally the method is based on the pixel values of the gray scale surface of the pixel. The grayscale value of the two is used to present each of the 1282947. The present invention further provides an anti-aliasing method for slowing the generation of the alienation phenomenon of the zigzag image, and the de-isolation method involves analyzing the image topology surrounding each pixel, and According to the analyzed topology structure, a predefined weighting matrix is selected to calculate the motion of the pixel smoothing to generate an optimal depolarization effect, and the pre-selected weighting matrix is used to calculate the pixel smoothing to obtain the denormalized pixel. The rule basis for the value. [Embodiment] Φ Figure 2 shows a typical computing environment in which the method of the present invention is implemented, including a general purpose digital computer 26 including a display 27 and a printer 28 for implementing the various methods of the present invention. Display 28, printer 28, and other output meters (e.g., screen top displays and printers, etc.) can image, print, and output the fonts produced by the method of the present invention. Other types of computing systems known to the art, such as network or host systems, can also be used to implement the methods of the present invention. The invention described below is used to define or produce a stroke contained in a character, and can also be understood as defining or generating a pattern contained in the text. As described in the background of the invention, the pattern is a unit constituting a character, and the unit is composed of one or more The strokes are composed. In the present invention, whether based on a stroke or a pattern, both are defined by key points, width values, feature points, and curve scale values, and the basic strokes or basic patterns defined by these definitions are used to determine the font data and The font is generated, so if there is no confusion, the ''stroke'' will cover both the stroke and the pattern. Figure 3 (A) shows - basic stroke 30, which is determined by key points (K1 and K2), 7 1282947 width value (Wl), feature points (fl, f2, f3, and f4), and curve scale values (not One of the characteristics of Chinese characters, as defined in U.S. Patent Nos. 5,852,448, 6,151,032, and 6,157,390, is that many words can share a large number of strokes. In other words, similar strokes can be used in many words, such as Figure 3. The horizontal stroke 30 shown in (A) can be used in many words or in the same word, such as strokes 20 and 22 in Figure 10 (A). And as described in the above patent, a group of similar stroke groups can be represented by a basic stroke to represent its topological shape, and each basic stroke is first defined by a key point and a width value. The key points (K1 and K2) are usually selected as the intersection of the tangent on the line of the basic stroke or at the bend or bend. The key point can be changed to change the shape of the basic stroke to produce a variety of strokes similar to it. The width value (W1) is usually chosen to be the distance between two points in the basic stroke, so that changing the width value can also produce a variety of similar strokes of different thicknesses. Each basic stroke has some equations to indicate how to determine feature points from key points and width values, such as Figure 3 (A), those equations that determine feature points (Π, f2, f3, and f4). = K1+W1/2 f2= K1-W1/2 f3= K2+W1/2 f4= K2-W1/2 (3) Therefore, the feature point is defined by its pre-defined key points and widths The spatial relationship of the values is determined. As shown in the figure, the feature points are selected on the shape of the outer stroke of the basic stroke of the shape of the tangential or curvature. Moreover, each basic stroke also has some predefined curve scale values, which are pre-built value groups according to the level of resolution, to generate a curve shape between different adjacent feature points at different resolutions. The construction of the curve scale value of the tree structure is as described in the above patent. Each basic stroke is confirmed by a specified number of strokes. Each defined basic stroke can define a similar stroke group corresponding to it by moving the key point and changing the width value, once the key point is moved and the width Φ value After being changed, the pre-existing equations calculate the new feature point position, and calculate the new curve shape between the two adjacent new feature points in the hierarchy of different resolutions by the pre-stored tree-like curve scale value. . This process of generating similar strokes has been detailed in the above patent case. As described in the foregoing background, when the stroke is defined to be displayed on a gray matrix image of a dot matrix image of a two-dimensional order, a quality deterioration phenomenon occurs, and therefore, according to the present invention, In one aspect of the invention, each basic stroke (and thus a similar stroke generated by the basic stroke) is also incremented by adding a specified silhouette area code to define the manner in which the grayscale silhouette region is created within the stroke. Figure 3 (B) shows that the shadow area code "10" and "01" are respectively specified in the two strokes 34 and 36, and a silhouette area code is used to define the grayscale shadow area in each stroke. When the area is projected onto the grayscale image, the grayscale value of the lattice element is determined according to the area of the lattice falling in the stroke. For example, the pen|34 is specified-the shadow area code "10", indicating the upper half thereof. 34a is a silhouette area surrounded by dots (ΚΙ, Π, f3, f2), and its specified gray value of 1282947 is used to calculate the pixel value of the lattice in which it is located. In other words, the shadow area is also a part of the stroke area surrounded by the key point, the width value, the feature point, and the curve scale value, and the other part belonging to the lower half 34b is not a shadow area, and is adjusted to a black point ( Will not cause its pixel grid to become gray value). The stroke 36 is assigned a shadow area code "01", which means that the upper half 36a is not a shadow area and is adjusted to a black point, and the lower half 36b is a point group (k, k2, f4, F2) The enclosed shadow area, when it is developed, the gray scale value of the upper pixel grid is calculated. For those skilled in the art, the above-mentioned φ silhouette area code is only one example. In fact, the silhouette area code has various definitions according to the basic stroke shape and its demand for font generation. For the purposes of the present description only, the mask area code specified on each stroke means that some of the blocks inside the stroke need to calculate the gray value of the pixel falling on it, and the shadow area of each stroke. The code system is stored in the signature code (the number of strokes or the first stroke) of the stroke. Figure 3 (C) shows the two strokes 20 and 22 in the text 10 of Figure 1 (A), and they each have a shadow area code "10". As shown in Figure 3 (B) •, when the image There will be four pixel columns 18a-18d located above it. Since the shadow area code ''10'' of each stroke will become a grayscale value during development, the two strokes will be displayed during the development. As indicated by the reference 39, they are visually two spaced strokes. This result is due to the fact that the image of the lower half of the non-shadow region is of a specified grayscale value, for example, the grayscale value is specified as white. Therefore, by appropriately defining the strokes represented by the silhouette of the shadow region to combine the characters, the smoothed grayscale image font can be displayed, and the two strokes 20 and 22 of the character 10 shown in Fig. 1(A) are overlapped. The phenomenon together can be avoided by the performance of the basic strokes of the Weiyi Shadow Zone. The fourth figure is to define the basic strokes of the first stroke with the silhouette of the shadow-oriented performance. Grayscale image font used to define text. As shown in block 40, font design A set of text sets (not shown in various dot matrix sizes) is displayed for the purpose of defining a shadow-oriented reference. Basically, a font designer is a graphic designer whose task is to create a font database for commercial use. Each text in the text set consists of one or more strokes, and all the strokes that make up the text can be divided into groups. Each group of packages _ contains a set of similar strokes. Each group of similar strokes can be represented by a basic stroke. The generation of their topological shapes, the basic strokes are the program generated by the pen shape defined by the key points, width values, feature points and curve scale values. As shown in block 42, the font designer defines each basic stroke. The performance of the shadow-oriented orientation. The font designer visually confirms the best definition of the specified shadow area in each basic stroke, such as observing the condition between a stroke and another adjacent or adjacent stroke to specify the shadow area. Avoid overlapping them in low resolution, and then specify the shadow area code to φ the basic stroke, the specified shadow area in the definition of the basic stroke. When the group character is displayed, it will be presented in the corresponding strokes corresponding to the basic stroke. As shown in block 44, the font designer will display all the characters that are directed by the silhouette, and each stroke of the character will be shadowed by the silhouette. The area code determines the image of the grayscale image; in other words, only the pixel grid in which the silhouette area is located will calculate the gray value of the pixels. As shown in block 46, the font designer uses its vision. View all the grayscale images that have been displayed. If the text has been properly reduced to ash 11 1282947 • The sub-type, there is no phenomenon that the strokes are stacked together, and the gray of the method is completed. The construction of the step type, otherwise return to block 42 to re-define some of the basic strokes into a new silhouette-oriented representation, and then as shown in block 44, 'the ones generated by these basic strokes that redefine the silhouette area The ash 6 shirt is like a text; as shown in block 46, the font designer then watches whether the quality of the grayscale image text is satisfactory, and repeats such action until it is satisfied. According to another aspect of the present invention, the present invention provides a method of displaying grayscale image characters represented by edges, and FIG. 5 is a flow chart showing such a method. For example, the method shown in FIG. The block 4 of FIG. 4, the steps shown in FIG. 5 are executed by an image display program or engine stored in the computer to display the grayscale image text on the display of the grayscale surface. As shown in block 52, this method constructs a scanned image of 〆 text according to the size of the dot matrix, such as a 0x6 dot matrix as shown in FIG. 6(A), wherein one of the scanning lines is indicated by arrow 67. In the example description, a text image 64 will be displayed on the dot matrix 60. _ As shown in Fig. 5, block 54, there is a suitable formula for each pixel in the point matrix 60 to calculate the value of the pixel of the edge-directed representation. For example, equation (1) is used to calculate the gray value of each pixel, as shown in Figure 9. The calculated gray values constitute the grayscale image of the text. For the gray value calculated by the above, the gray scale # of the display text directly produces a jagged or dissimilated image font, so that the anti-aliasing method can also be used to reduce the jagged appearance, for example, The inverse alienation method shown in blocks 56 and 58 of FIG. There is a very large 12 1282947 anti-aliasing method in the processing of image graphics. The present invention provides an anti-aliasing method as shown in FIG. 6(B), in which nine pixels in the 3x3 pixel region 70 are respectively The gray value calculated in block 5 of Fig. 5, and the P5 system located at the center of the pixel area 70 is to be recalculated as the pixel of the new gray value, and the calculation of the new gray value of P5 is based on the surrounding The average blending result of the eight pixel gray values according to their weighted weights is calculated as follows: The new pixel value of P5 = § [out, value_] - ¥, where Wi is the weighted value of pixel I, It is well known to those skilled in the art that the weighting value of the matrix W is usually Gaussian, that is, if the weight value of the center point is the largest, the weighting values of other off-center points are decremented according to the Gaussian distribution. The traditional anti-aliasing method takes a weighted value matrix of a fixed Gaussian distribution. The anti-aliasing method of the present invention calculates P5 by taking different weighting matrix according to the writing direction of the stroke. For example, the calculation of the inverse alienation of the pixel 66 and the pixel 68 in the text image 64 (Fig. 6(A)) is as shown in Fig. 6(C) and Fig. 6(D). ® Figure 6 (C) considers the 3x3 pixel region surrounding the text image of pixel 66. From calculating the difference between the leftmost and rightmost pixels 76 and 78, you will get three categories as shown in 80: zero, positive, Or a written value of a negative value, for example, a written value of <0, +, +> as shown in Fig. 6(C), and one of a plurality of pre-replicated weighting value matrices may be selected according to the written value, for example, The weighted value matrix W, (82), which is a pre-made weighted value matrix is used to calculate the written value <0, +, +> to produce the best anti-aliasing effect. Therefore, the central position of the pixel 66 is taken from 1 1 282 947 and the selected weighted value matrix wl (82) is substituted into equation (4) to calculate the inverse anamorphic gray value of P5 (pixel 66). Similarly, Figure 6 (D) considers the bad-around pixel 6 8 3 χ 3 pixel region ' After calculating the pixel columns 83 a, 83b, and 83 c, the written value 84 of Figure 6 (D) is <- , -, ->; according to this written value 84, then the weighted matrix w2 (86) of the copy is selected, and the gray value of the inverse dissimilation of p 5 (pixel 68) is calculated by substituting into equation (4) to achieve the most Good anti-alienation effect. Those skilled in the art can also understand that the elements of the weighting matrix can also be used in other conditions than the values disclosed in the present invention, and can also be generated in different writing values. The same anti-aliasing effect. The various methods of the present invention can be implemented by a graphical human machine interface (GUI) tool, such as a GUI tool for a window-based operating system, and one or more steps in some methods. Yes, it can be done automatically through the techniques of image analysis. </ RTI> </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; </ RTI> <RTIgt; And other similar embodiments. BRIEF DESCRIPTION OF THE DRAWINGS The foregoing and other advantages of the present invention will be apparent from the description and accompanying drawings. Figure 1 illustrates the dilemma encountered in displaying grayscale image fonts using traditional second-order (black and white) dot matrix images. Figures 2(A)-(E) show block diagrams of a system or graphics 14 1282947 human interface tool for implementing the method of the present invention. _ Figures 3(A) to (C) show an example of the definition of a basic stroke and its side shadow code using key points, width values, feature points, etc., as disclosed in the present invention. FIG. 4 is a flow chart of a method for defining a silhouette-directed grayscale image font according to one embodiment of the present invention. FIG. 5 is a flow chart showing a method for displaying a gray-scale image font according to one of the aspects of the present invention. Figures 6(A)-(D) illustrate one embodiment of the present invention for reducing the anti-aliasing method for forming a gray-scale image font of a saw blade. [Main component symbol description] 9 Chinese characters 10 characters 12 squares 16a pixels 16b pixels 16c pixels 16d pixels 18a pixels 18b pixels 18c pixels 18d pixels 20 strokes 22 strokes 27 display instrument 28 Printer 26 universal digital computer 30 strokes 34 strokes 36 strokes 34a upper half 34b lower half 60 points matrix 64 text image 66 pixels 15 1282947 68 pixel 70 pixel area 78 pixel 83b figure column 84 writing value 67 Arrow 76 pixel 83a pixel column 83c pixel column
1616