1352298 九、發明說明： 【發明所屬之技術領域】 本發明係關於一種產生單線字型的方法，尤指一種由外框字型 產生單線字型的方法。 【先前技術】 k著資s孔電子業的蓬勃發展’手持式裝置(p0r^abie device)及消 % 費性電子產品(consuming electronics)也越來越熱門。而這些產品例 如行動電話、個人數位助理(personal digital assistant, pDA)、以及 • 媒體播放器等等皆須要設置顯示相關資訊的顯示裝置，以提 -供使用者充足的操作資訊。但也因為這些產品的體積皆不大，其 顯示裝置亦不大，因此只適合顯示較小的字型。以單線(center lined) 筆劃所產生的字型，技術上每筆劃是以單線為骨幹，由各筆劃組 成一個個字之後，經過縮小再由骨幹長肉變粗，來產生不同尺寸 鲁的子型。單線字型係指黑體、圓體、海報體等一直線間只有一寬 度值的字型，並不適祕料體等，-直_具有多種粗細寬度 •的字型。請參考第1 ®，第2圖，以及第3圖。第i圖係為一黑 體之單線字型的示意圖，第2圖係為一圓體之單線字型的示音、圖， 而第二圖係為-海報體之單線字型的示意圖。由單線筆劃所產生 的J、字型’速度快又品質好，在螢幕上的可讀性非常高，很適合 應用於目前大量製造的手持式裝置及消費性電子產品巾。° • 但目前—般的辭字魏商大部分都經由設計外框筆劃來組 5 成漢字字型，以盆古 __ αέΒ, '、便顯示各種大小之字型及使用少數筆劃就可 圖係為子，故。請參考第4圖、第5圖、以及第6圖。第4 為笛 > 框筆劃所構成之外框字型，，凸，，之實例圖，第5圖則 同7圖之外框筆劃塗黑後所組成之外框字型”凸，，之實例圖。第G 二，系為第4 ^、第5圖中”凸，，字之筆劃順序圖。由第6圖中可知 凸/字係由五筆劃所構成，帛丨筆劃係為—短直線，’丨”，第2 為劃係為，橫線第3筆劃係為—短直線，，卜第4筆劃係 :3 ’第5筆劃係為—長橫線，，—，，。每個外框字型的資料格式 系以下列表丨的方式儲存於記憶體中： 字 字 +參數 +參數 -~~-- ¥2筆筆劃序 號+參數 系2筆筆劃序 號+參數 第S1筆 號+參數 % S2 號+參數 第Ν個 字 ^--^ +參數 ------— 第2筆筆劃序 號+參數 第 號+參數 表1 以日本工業規格(Japanese Industrial Standards)中所制定之基本 漢字編碼系統JIS X 〇2〇8為例，JIS X 0208的字集共約7〇〇〇個 字，由共約65000筆筆劃所組成。但是，對外框字型而言，通常 只需由200〜300筆基本筆劃，再加上參數變化，即可變化出該約 1352298 7_個字集中所有65_筆筆劃。故表丨中之，，軸序 200 300筆基本筆劃之筆劃序號。表丨中之”參數，，則包含割^ 關鍵點之位置及寬度值。請參考第7圖，第7圖係為，，凸，，字之里 筆sj云中之關鍵點之位置及寬度值之資訊。” J4 華^ 查|| ,， 令有5個關鍵點及4個寬度值。5個關鍵點包含—起始^ 關鍵點），-結束點(第5關鍵點），以及3個轉折點(第2關鍵點，1 第3關鍵點’以及第4關鍵點）；4個寬度值則包含介於關’ 點以及第2 _點之間、表示該段線段之寬度之—第丨、寬度值鍵 _點以及第3 _點之間、表示該段線段之=之丄 第2寬度值，介於第3關鍵點以第 : 段之實4 纖之間、表不該段線 見^之1 3讀值，以及介於第4 __及第5關鍵點 之間、表不該段線段之寬度之一第4寬度值。 則如下列 而儲存於外框軸表中的每一外框筆_:#料 表2 : 筆劃序號=1 外框描述权式1(麥數 聿蚤彳序號=2 外框描述程式2(參數— ------- 呈式 Μ(參數— --——---------- 表 2 '-- 例如黑體、圓 表2㈣外框描述裎式侧來產生獨的字型， 7 楷書體等等，馨 寻，亦即 横筆直線 — 的不门筆3彳，例如直線、斜線、或橫線等 指述遠筆劃外框的程式。下面的例子即是一個產生— 的簡單的外框描述程式的例子： 參數，鍵點··_，_，寬度值：w。 步^ 1.從關鍵點pntl向上移動w/2，產生 newpntl 〇 步驟2··從newpntl向下移動w，產生。 步驟3:從關鍵‘點pnt2向上移動w/2，產生丽邮。 步驟4:從newpnt3向下飾w，產生。 步驟5:從newpntl晝直線到newpnt3。 步驟6··從newpnt3晝直線到newpnt4。 步驟7:從newpnt4畫直線到newpnt2。 步驟8:從newpnt2畫直線到newpnt卜 步驟9:結束。 若起始點和結束點之間不為一直線，而為一曲線，則該曲線依 貝氏曲線(Bezier Curve)原理，由分別相切於該起始點和該結束點 之兩條直線所相交之控制點，來描述該曲線。請參考第8圖。第8 圖即為一曲線段(起始點A1，結束點A3)上，根據貝氏曲線指定控 制點A2所繪之曲線段之示意圖。請再參考第9圖。第9圖係為,， 介”字中根據貝氏曲線所定義之控制點，描述左下撇，，/，，的示意 圖。在第9圖中，，，介”字左下撇的上邊緣係為直線段a1〇_au_a12 再加上曲線段A12-A13-A14。曲線段A12-A13之起始點A12和結 1352298 束點A13都在該筆劃上，而由貝氏曲線所得之控制點A15，則不 位於該筆劃上’但可透過控制點A15來控制曲線段A12 Al3的弧 度。同理，曲線段八13-八14之起始點A13和結束點A14都在該筆 劃上，而由貝氏曲線所得之控制點A16，則不位於該筆劃上，但 可透過控制點A16來控制曲線段A13_A14的弧度。 相同的-筆劃序號，例如”一”之筆劃序號為！，加上不同的關 • 鍵點位置及寬度值’可以形成不同的，，-”。請參考第1〇圖，第1〇 圖係為不同的關鍵點位置及寬度值所形成的不同的，，—，，之示意 圖由上述可知’對外框字型而言，由2〇〇〜·筆基本筆劃，再 加上參數憂化(關鍵點及寬度值），即可變化出JIS X 〇2〇8約7〇〇〇 個字集中所有約65000筆筆劃。 〜若要#棄目前通用的外框字型’專為適合小螢幕顯示，重新設 十，·且單線字型’曠日費時且工程浩大。有鑑於此，本發明提出 -種可以由外框字型產生單線字型的方法，並建立—適當 筆劃樣本’以減少其所佔據之記髓_存空間，再由單線筆叫 2中擷取單線_，組成—單線字型，再根據較大字型時的單 ^型所對應的外框字型’修正單線字型，歸根據所欲顯示字 寬之單^ ^，再加上微調資訊，修正被加 早線子型’使所產生之單線字型更加完美。 1352298 【發明内容】 、本發明係揭露-種由外框字寵生單線字型的方法，包 奴儲存單線軸樣本的數目，㈣框字群姆对產x 劃樣本；由該單線筆舰本帽取單線筆成 約 =型所對應之外框字型，修正該單線字型;根據欲顯： 予㈣粗細資料’加寬該修正後的單線字型，以產纽 車線字型；以及加入微調資訊(hinting)，修正該具有寬度之單^字 【實施方式】 本發明所提供之方法，即根據約6獅筆軸之外框字型筆 釗’產生較少之單線筆劃樣本(因為本發明所提供之字體，係為了 用於較小_科幕，料需產生財約6聊筆_之單線 子堂）再由單線筆劃樣本，組成單線字型。在下面所舉之實施例 鲁中’所產生之單線筆劃樣本約為一萬個左右，接著由單線筆劃樣 •本巾擷取適自之筆触字’再根據較大字斜單線字型所對應的 •外框字型’修正單線字型，接著根據所欲顯示字_粗細資料， 長出-適當的寬度，再經由微調技術，修正所加寬之單線字型。 請參考第11圖。第u圖係為根據本發明之方法所緣之一流程 圖。第11圖中’本發明所提供之由外框字型產生單線字型的方法， 係包含下列步驟： 1352298 步12 .根據所欲儲存的單線筆劃樣本數目，選取一適當之字型 大小。 步驟14 ·根據所選取之適畲的字型大小，從基本外框筆劃的資訊 中’產生具有適當數目之單線筆劃樣本。其中該基本外 框筆劃的育訊包含200〜3〇〇筆基本外框筆劃之相對應之 筆劃序號、複數個關鍵點位置、及寬度值等資訊。 • 步驟16 :根據所產生之單線筆劃樣本組成單線字型。 • 步驟18 :在較大字型256像素*256像素時，將所產生之單線字型 與外框字型做比較，修正該單線字型。 步驟20 :根據欲顯示判的粗細#料，產生具有適當寬度之單線 字型。 步驟22 :加人賴㈣，修正該具有適#寬度之單線字型。 在本發明所提供之由外框字型產生單線字型的方法中， 200〜30?筆左右之基本外框筆射每一基本外框筆劃均包含一相 二U序t冷复數個關鍵點位置、及複數個寬度值等資訊。 月J所述4 200〜3〇〇筆左右之基本外框筆劃可以變化出所有外 2M0 ^ 奸之而求而定’本實施例所提供之數字，只是一參考值， 凡付合本發，所赴之適雜量的 =蓋的範圍姆第_的流程圖，= 順序’凡符合本發明精神，各步驟間不同的實施順序=為斤= 明所涵蓋之範疇。 本數量，馳適當大奴押。在衫射線葦劃樣 字集為例’每-個字原本係以256像素 2 〇 係選取字型大小為32像素*32料，4筆；。本貫施例 叙田/、萬多筆之筆割中，姦从 里、力為近-萬個的單線筆劃樣本。若選擇 一 像素，則產生的單線筆劃樣本數量 J、為64像素吩 數目之單線购本，軸計者本身^二適當 及系統效料等各方面财量；本發騎舉㈣財， 列子’亚不限定於該數字，凡符合本發明之精 =山 數值，都屬於本發明的範圍。告缺 Μ 戈入的適當 本數越少，雖然資料」 翁的字型越小，筆劃樣 劃誤差也魏、、里7所佔5己憶體的空間少，但所產生的筆 、，因此需在筆触本數和所產生的_誤差之間，1352298 IX. Description of the invention: [Technical field to which the invention pertains] The present invention relates to a method for generating a single-line font, and more particularly to a method for generating a single-line font from a frame font. [Prior Art] k is booming in the development of the s hole electronics industry. The handheld devices (p0r^abie device) and consuming electronics (consuming electronics) are also becoming more and more popular. These products, such as mobile phones, personal digital assistants (PDAs), and media players, need to be provided with display devices for displaying relevant information to provide sufficient information for the user to operate. However, because these products are not large in size, their display devices are not large, so they are only suitable for displaying smaller fonts. The font produced by the center lined stroke is technically characterized by a single line as the backbone. After each stroke is composed of one word, it is reduced and then thickened by the long bones to produce different sizes of Lu. . The single-line type refers to a font with a width value of a black body, a round body, a poster body, and the like, and is not suitable for a body, etc., and has a font size of a plurality of thicknesses. Please refer to 1 ® , 2 , and 3 . The i-th image is a schematic diagram of a single-line font of a black body, the second figure is a single-line type of sound and diagram of a circle, and the second figure is a schematic diagram of a single-line font of a poster body. The J and fonts produced by single-line strokes are fast and of good quality, and are very readable on the screen. They are suitable for hand-held devices and consumer electronics towels that are currently mass-produced. ° • But the current general-style Wei merchants mostly use the outer frame strokes to form a group of 5 Chinese characters, which can be displayed in various sizes and with a few strokes. It is a child, so. Please refer to Figure 4, Figure 5, and Figure 6. The fourth is the flute > box strokes constitute the outer frame font, convex, and the example map, the fifth map is the same as the 7 outer frame strokes, and the outer frame type is convex, Example figure. The second part is the stroke sequence of the word "convex" in the 4th and 5th pictures. It can be seen from Fig. 6 that the convex/word system is composed of five strokes, the stroke is a short straight line, '丨', the second is a stroke, and the third stroke of the horizontal line is a short straight line, 4 strokes: 3 'The fifth stroke is - long horizontal line,, -,,. The data format of each outer frame font is stored in the memory in the following list: Word + parameter + parameter -~ ~-- ¥2 stroke number + parameter system 2 stroke number + parameter S1 pen number + parameter % S2 number + parameter number one word ^--^ + parameter ------ - 2nd stroke number +Parameter No. + Parameter Table 1 Take the basic Chinese character encoding system JIS X 〇2〇8 as defined in Japanese Industrial Standards. The JIS X 0208 word set is about 7 characters. A total of about 65,000 strokes are composed. However, for the outer frame font, usually only 200 to 300 basic strokes, plus parameter changes, can be changed to about 1352298 7_words all 65_ pens Strokes. Therefore, in the table, the sequence number of 200 300 basic stroke strokes. In the table, the "parameters," include cutting ^ key points Position and width values. Please refer to Figure 7, which is the information on the position and width of the key points in the pen sj cloud. J4 Hua ^ check || ,, there are 5 key points and 4 width values. 5 key points include - start ^ key point), - end point (5th key point), and 3 turning points (first 2 key points, 1 3rd key point 'and 4th key point'; 4 width values include between the 'point' and the 2nd point, indicating the width of the line segment - the 丨, width value key Between the _ point and the 3rd _ point, the second width value of the segment of the segment is indicated by the third key point between the real segment of the third segment and the segment of the segment. The read value, and the fourth width value between the 4th __ and the 5th key point, which is the width of the segment line segment. Then each outer frame pen stored in the outer frame axis table is as follows_ :#Material Table 2: Stroke Number=1 Frame Description Weight 1 (Mc Number ==2 Frame Description Program 2 (parameters - ------- Presentation Μ (parameters -- --- ---------- Table 2 '-- For example, the black body, the round table 2 (four) the outer frame describes the side of the 来 type to produce a unique font, 7 楷 book body, etc., Xin Xun, that is, the straight line straight line - 3 不, such as straight lines, diagonal lines, or horizontal lines, etc. The outline of the program. The following example is an example of a simple outline description program that generates -- parameters, key points ··_, _, width value: w. Step ^ 1. Move up from the key point pntl w /2, generate newpntl 〇 Step 2· Move down w from newpntl, generate. Step 3: Move w/2 up from the key 'point pnt2 to generate the privilege. Step 4: Decorate w from newpnt3, generate. 5: From newpntl昼 straight line to newpnt3 Step 6··From newpnt3昼 straight line to newpnt4 Step 7: Draw a line from newpnt4 to newpnt2 Step 8: Draw a line from newpnt2 to newpnt Bu Step 9: End. The end point is not a straight line, but a curve, the curve is based on the Bezier curve principle, and the control points respectively intersecting the two straight lines of the starting point and the ending point Describe the curve. Please refer to Figure 8. Figure 8 is a schematic diagram of the curve segment drawn by the control point A2 according to the Bayesian curve on a curve segment (starting point A1, ending point A3). Please refer to the Figure 9. Figure 9 is the control point defined by the Bayesian curve in the word A description of the lower left 撇, , /, ,. In Fig. 9, the upper edge of the lower left 撇 is the straight line segment a1〇_au_a12 plus the curve segment A12-A13-A14. Curve segment A12 -A13 starting point A12 and knot 1352298 beam point A13 are on the stroke, and the control point A15 obtained from the Bayesian curve is not located on the stroke 'but can control the curve segment A12 Al3 through the control point A15 radian. Similarly, the starting point A13 and the ending point A14 of the curve segment VIII 13-eight 14 are all on the stroke, and the control point A16 obtained by the Bayesian curve is not located on the stroke, but can be transmitted through the control point A16. Controls the radians of the curve segment A13_A14. The same - stroke number, for example, the stroke number of "one" is! , plus different key points and width values 'can be formed differently, -". Please refer to the first figure, the first picture is the difference between the different key position and width value, The schematic diagram of -,,, is known from the above. For the external frame font, the basic strokes of 2〇〇~· pens, plus the parameter worry (keypoint and width value), can change JIS X 〇2〇 8 about 75,000 strokes in about 7 words. ~ If you want to # abandon the current common frame fonts, it is designed for small screen display, re-set ten, and single-line fonts. In view of this, the present invention proposes a method for generating a single line font from the outer frame font, and establishing an appropriate stroke sample to reduce the space occupied by the memory, and then by the single line pen 2 Take a single line _, composition - single line font type, and then according to the outer frame font type corresponding to the single type of the larger font type, modify the single line font type, according to the single word width of the desired width ^ ^, plus fine-tuning Information, corrections are added to the early line type 'to make the resulting single line font more perfect. 1352298 [Description of the Invention] The present invention discloses a method for selecting a single-line font from a frame, a number of samples for storing single-axis samples, and (4) a sample for a group of x-slices; The cap takes a single-line pen to the outer frame type corresponding to the type=correction, and corrects the single-line font type; according to the desire to display: (4) the thickness information 'widens the corrected single-line font type to produce the car line font type; and adds fine adjustment Hinting, correcting the single word with the width [Embodiment] The method provided by the present invention, that is, generating fewer single-line stroke samples according to about 6 lion pen-axis outer frame type pens (because the present invention The fonts provided are for the smaller _cosmetic screens, and it is required to produce a single line of texts for each of the six lines of strokes, and then consist of a single line of strokes to form a single line type. In the example below, the example is used. The sample of the single-line strokes produced is about 10,000, and then the single-line strokes are used. • This towel captures the appropriate strokes and then corrects the single-line font according to the larger-line oblique single-line font. , then according to the desired words _ thick and small capital , grow out - the appropriate width, and then modify the widened single line font by fine-tuning technique. Please refer to Figure 11. Figure u is a flow chart of the method according to the present invention. In Figure 11 The method for generating a single line font from the outer frame font according to the present invention comprises the following steps: 1352298 Step 12. According to the number of single line stroke samples to be stored, select an appropriate font size. Step 14 · According to the selected The appropriate font size, from the information of the basic outer frame strokes, produces a sample with a suitable number of single-line strokes, wherein the basic outer-frame stroke education contains 200~3 〇〇 pen basic outer-frame strokes corresponding to Information such as stroke number, multiple key position, and width value. • Step 16: Form a single line font based on the generated single line stroke samples. • Step 18: When the larger font is 256 pixels*256 pixels, compare the generated single line font with the outer frame font and correct the single line font. Step 20: According to the thickness of the material to be displayed, a single line font having an appropriate width is produced. Step 22: Add the person (4) and correct the single line font with the width of #. In the method for generating a single line font by the outer frame font type provided by the present invention, the basic outer frame strokes of about 200 to 30 strokes each of the basic outer frame strokes include one phase and two U sequences, and a plurality of cold key points. Information such as location, and multiple width values. The basic frame strokes of the 4 200~3 〇〇 pens described in the month J can be changed by all the other 2M0 ^ 之 之 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' 本 本 本 本 本 本 本 本 本The flow of the appropriate amount = the scope of the cover _ the flow chart, = the order 'in the spirit of the present invention, the different implementation order between the steps = the scope covered by the jin = Ming. The number, the appropriate large slave. In the case of the ray 苇 苇 ’ ’ ’ ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' The basic example of the Narita /, more than a pen of the pen, from the inside, the force of nearly 10,000 single-line stroke samples. If a pixel is selected, the number of single-line stroke samples produced is J, the single-line purchase of the number of 64-pixel orders, the axis meter itself, the appropriateness of the system, and the system effect; the hair riding (four), the column' The present invention is not limited to this number, and any numerical value corresponding to the present invention is within the scope of the present invention. The lack of the appropriate number of Ge's entry, although the data "Weng's font size is smaller, the strokes of the strokes are also less, Wei, Li, 7 accounted for less space, but the resulting pen, therefore Between the number of strokes and the resulting error,
=：可Γ接受的平衡點。在較大的字型物像物6像素 月^下’右不同的橫筆直線的長度分別為，舉例來說H 像素:像素’則縮小至字型為32像素％像素時(256像伽 的長度分別為_、 四捨五入社彳 錄，但像素縣整數個，若採取 筆亩力'〜四種在256像素*256像素時，不同長度的橫 言綠。m 32像素32像树，全部魏烟長度6像素的橫筆 、’。5月多考第12圖及第13圖。第12圖係為在256像素*256 12 ^^8 像素時，所顯示的，，一，，字和”三，’字。由第12圖中可看出，，，一，，字 的毛、筆筆劃和”二”字的3條橫筆筆劃，在顯示字型為像素 象素4•，長度都不同。第丨3圖則為在顯示字型為32像素*32像素 τ所顯示的一”字和”三”字。此時”一”字的横筆筆劃和，，三，，字最 下的&葦葦劃長度變成相同了。因此，在顯示字型為32像素*32 . 像素時，如果將關鍵點的距離與寬度值均相同的筆劃，視為相同 的筆劃樣本，透過此做法，可將原本約65000筆的筆劃，減少為 鲁數量约為近一萬個的單線筆劃樣本。 而只際上的做法，係在32像素*32像素的字型大小下，利用外 框筆劃貝讯(包含筆劃序號、參數值)去形成π8χ〇2〇8字集中的所 有再透過由外框筆劃轉為單線筆劃的方法，儲存每個字的單 線筆劃。若如前述，該字的某一筆單線筆劃，與前面所產生的、 諸存的單線子型的其巾—筆船目同⑽鍵點的距離與寬度值均相 • 同）财再齡轉線雜。如此制將整個字集的字形成之後， 產，對應於字集中所有字的適當數量(根據本實施例約為一萬字） 的早線筆劃樣本。此時字集中每個字的單線賴序號(約一萬個）， 已經與原外框筆劃的序號不同(約200〜300個）。由外框筆劃轉為單 線葦劃的方法，這裡介紹兩種： ^取該外框筆劃寬度的中點連成單線筆劃。請參考第14圖。第 14圖係為―由—曲線之外框筆劃寬度的中點轉換為單線筆 . 晝J之示意圖。第Μ圖中包含由貝氏曲線所得之控制點m、 13 1352298=: The balance point that can be accepted. In the larger font object, 6 pixels per month, the lengths of the right different horizontal straight lines are, for example, H pixels: pixels are reduced to 32 pixels per pixel (256 image gamma) The lengths are respectively _, four rounds of the social record, but the pixel counts an integer, if the pen is used to force ~ ~ four in 256 pixels * 256 pixels, different lengths of the horizontal green. m 32 pixels 32 like trees, all Wei smoke 6-pixel horizontal pen, '. May multi-test 12th and 13th. The 12th picture is displayed at 256 pixels * 256 12 ^ ^ 8 pixels, one, one, and three , 'word. As can be seen from Fig. 12,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, The third figure is a "word" and "three" word displayed in the display font of 32 pixels * 32 pixels τ. At this time, the "one" word of the horizontal stroke strokes, and, three, the bottom of the word & The length of the stroke becomes the same. Therefore, when the display font is 32 pixels * 32 pixels, if the stroke of the key point and the width value are the same, the phase is regarded as With the same stroke sample, this method can reduce the original strokes of about 65,000 strokes to a single-line stroke sample with a quantity of nearly 10,000. The only practice is to use 32-pixel * 32-pixel fonts. Under the size, use the outer frame strokes (including the stroke number and parameter value) to form all the π8χ〇2〇8 word sets and then transfer the single-line strokes from the outer frame strokes to the single-line strokes, and store the single-line strokes of each word. As mentioned above, a single line stroke of the word is the same as the distance and width value of the same (10) key point of the single-line sub-type of the existing one-line type. . After forming the words of the entire word set in this way, an early line stroke sample corresponding to the appropriate number (about 10,000 words according to the present embodiment) of all the words in the word set is produced. At this time, the single-line serial number (about 10,000) of each word in the word set is different from the original outer frame stroke number (about 200 to 300). The method of changing from the outer frame stroke to the single line stroke is introduced here: ^ The midpoint of the stroke width of the outer frame is connected into a single line stroke. Please refer to Figure 14. Figure 14 is a schematic diagram of the transition from the midpoint of the stroke width of the outer curve to the single-line pen. The figure contains the control points m, 13 1352298 obtained from the Bayesian curve.
2. B2、B3、B4、B5、B6、B7、B8、B9、ΒΙΟ 等 i〇 點，其中 m、B2、B3為一外框筆劃的上邊緣，B4、B5、B6為同一外 框筆劃的下邊緣。m、B2形成—曲線段m_B2(如第8圖中 之A卜A3)，而由貝氏曲線所得之另一控制點B7不在該曲 _上(如第8圖中之A2)，但可透過控制點B7來控制曲線 段BUB2的弧度。B2、B3形成一曲線段防則，同理，另 -控制點B8不在該曲線段上，但可透過控制點β8來控制曲 線段B2_B3的弧度。同理，B4、B5形成另一曲線段b4_b5 , B9為不在該曲線段上之另一控制點，β5、B6形成另一曲線 段B5-B6，而刪為不在該曲線段上之另—控制點。此做法 為將m與B6相連取中點C1，B2與β5相連取中點c2，扮 與B4相連取甲點C3，β7與_相連取令點控制點a，別 ”則相連取中點控制點cs ’再根據控制點a決定曲線段 ⑽’根據控制點C5決定曲線段⑽，即得所求之單線 取該外框筆劃-邊的控制點連成單線筆劃。請參考第 =圖係為-取-曲線之外框筆劃-邊的控制點‘ B = 意圖。第15圖中，包含由貝氏曲線所得之控制點 -第圖 Β5、Β6、Β7、Βδ、Β9、Β1。、， 二筆:了綱_再_。權例中所選取 Α射 剔6，御卩㈣線Β4·Β5 ' Β恤 ‘、’…之轉_。此法是指定—邊的起始與結束的控制 14 1352298 =齡於中間的關鍵點相連’即為單線筆劃，但是因為此 該外框筆劃的一邊，來連成單線筆劃，而實際上外 :I亥筆劃具有寬度，依此法所得並非真正該外框筆割 々心線’為了降低誤差，應用此法時建議將 寬度先設為零。 萆^之2. B2, B3, B4, B5, B6, B7, B8, B9, ΒΙΟ, etc. i, where m, B2, and B3 are the upper edges of a frame stroke, and B4, B5, and B6 are the same frame strokes. Lower edge. m, B2 form - curve segment m_B2 (such as A A3 in Fig. 8), and another control point B7 obtained from the Bayesian curve is not on the curve (as in A2 in Fig. 8), but is transparent Control point B7 controls the curvature of curve segment BUB2. B2 and B3 form a curve segment defense. Similarly, the control point B8 is not on the curve segment, but the curvature of the curve segment B2_B3 can be controlled by the control point β8. Similarly, B4 and B5 form another curve segment b4_b5, B9 is another control point not on the curve segment, β5, B6 form another curve segment B5-B6, and is deleted as another control not on the curve segment. point. This method is to connect m and B6 to take the midpoint C1, B2 and β5 are connected to take the midpoint c2, and B7 is connected to take the C3, β7 and _ are connected to the control point a, and the other is connected to the midpoint control. Point cs 'and then determine the curve segment (10) according to the control point a'. According to the control point C5, the curve segment (10) is determined, that is, the single line taken from the outer frame stroke-side control point is connected into a single line stroke. Please refer to the figure = - take - curve outside the box stroke - edge control point ' B = intention. Figure 15 contains the control points obtained from the Bayesian curve - Figure Β5, Β6, Β7, Βδ, Β9, Β1.,, Pen: The outline _ re- _. In the case of the right selected Α 剔 6 6, 卩 卩 四 四 四 四 Β ' ' ' ' ' ' ' ' ' ' ' ' ' 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 1352298 = The key points in the middle of the connection are 'single line strokes, but because of the side of the outer frame strokes, they are connected into a single line stroke, but in fact: I Hai strokes have a width, according to this method, the income is not really Box pen cutting heart line 'In order to reduce the error, it is recommended to set the width to zero first when applying this method. 萆^之之
據according to
請注意：此時產生的單線字型的資料格式，與外框字型的資料 不同的是記錄單線的筆劃序號(約—萬個左右)與原外框 偏的序齡實關為勘〜到目)得酵杨_下第一關 鍵點(起始點），其侧鍵點與寬度值均不存在了。由於參數只有起 咖單線筆劃不具有參數變化能力，僅能依原比例縮放。 著由這些單線筆劃樣本組字，每個字根據它的單線筆劃序號， =取_•對應的單線筆麵本，並依照參數(第—_點)的位置， :、'、此筆劃的起始位置。請參看第16圖，帛16圖的上方即為依 此法所產生的單線字型之示意圖。 =呈此法所得之單線輔樣本，會有餅誤差，因此在組成字時 =因誤差而產生字型品f變差關題。字型品質變差的原因有 a•位置誤差：因為所取得之單線筆劃樣本的擺放位置產生誤差，導 致接筆上產生不相連或過於凸出的問題(凸筆）。 b.長度誤差：因為所取得之單線筆劃樣本的長度有誤差，導致該筆 15 丄叫298 劃太短或太長。 «月 '考第17 ϋ帛π圖料—因為所取得之單線筆劃樣本太 短’而導致接筆不相連的，，娃，，字的示意圖。圖中圓圈處即接筆不 相連之處。但有時此問題可在增加單線字型寬度之後，獲得一此 改善。 二 • 解决因為單線筆劃樣本誤差而導致的字型品質變差的問題，係 透過自動分析縣健自動做最祕正程式，進行修正。請參考 第is圖第18圖係為自動分析誤差值並自動做最佳修正的詳細 步驟流程圖： 步驟励：分析單線字酬對應之外框字型在字型大小為256像素 *256像素時，各筆劃之間的相互關係。 鲁步驟搬：分析該單線字型在字型大小為256像素挪像素時， 各筆劃之間的相互關係。 •步驟刚：比較兩者_之_相互_，若有不同時，根據該單 線字型所對應之外框字型自動修正該單線字型。 自動分析誤差值並自動做最佳修正係透過比較Μ6像素*256 像素時(顯示字魏大時)，單線字_對叙外框相和單線字型 之:同，來進行自動修正。因為在較大字型時，若外框字型和單 .線字型有不同處，會最_，因此校正較大字型時的單線字型丨 16 若在較大字贿，外框字型和單線相都沒有不同處，則縮小為 幸又小予型時’例如64像素*64像素、32像素*32像素等就不會不 同。請參考第19圖，帛19圖即為經過此自動分析誤差值並自動 做最佳修正步驟後的原第Π_，娃，，字的示意圖。圖中圓圈處即 經過修正後’接筆重新相連之處。Please note that the data format of the single-line font generated at this time is different from the data of the outer frame font type. The number of strokes recorded on the single line (about 10,000 or so) and the original order of the original outer frame are recorded as The first key point (starting point) of the yeast leaves, the side key points and width values do not exist. Since the parameters only have the ability to change the parameters of the single-line strokes, they can only be scaled according to the original scale. According to the sample number of these single-line strokes, each word is based on its single-line stroke number, = _ corresponding to the single-line pen-based book, and according to the position of the parameter (the -_ point), :, ', the start of the stroke Starting position. Please refer to Figure 16. The top of Figure 16 is a schematic diagram of the single-line font generated by this method. = The single-line auxiliary sample obtained by this method will have a pie error, so when the word is formed, the font f is deteriorated due to the error. The reason for the deterioration of the font quality is a•Position error: because the position of the single-line stroke sample obtained is in error, causing a problem of disconnection or over-bulging on the pen (bull). b. Length error: The length of the single-line stroke sample obtained is inaccurate, causing the pen 298 298 to be too short or too long. The «month' test is the 17th ϋ帛 π picture material—the picture of the word is not connected because the sample of the single line stroke obtained is too short'. The circle in the figure is where the pens are not connected. But sometimes this problem can be improved after adding a single line font width. 2. Solve the problem of poor font quality due to sample error of single-line strokes. It is automatically corrected by automatically analyzing the county health automatically. Please refer to the 18th picture of the is picture for the detailed steps of automatically analyzing the error value and automatically making the best correction. Step excitation: analyze the single line word corresponding to the outer frame type when the font size is 256 pixels * 256 pixels , the relationship between the strokes. Steps to move: Analyze the relationship between the strokes of the single-line font when the font size is 256 pixels. • Step just: Compare the two _ _ mutual _, if there is a difference, automatically correct the single line font according to the outer frame font corresponding to the single line font. Automatically analyze the error value and automatically make the best correction. By comparing Μ6 pixels*256 pixels (display word Wei Da), the single line word _ pairs the outer frame phase and the single line font type: the same, to automatically correct. Because in the case of larger fonts, if there is a difference between the outer frame font and the single-line font, it will be the most _, so the single-line font 校正16 when correcting the larger font is used in larger letters, outer frame fonts and single lines. If there is no difference in the phase, then it will be different if it is reduced to a small size. For example, 64 pixels * 64 pixels, 32 pixels * 32 pixels, etc. will not be different. Please refer to Figure 19, Figure 19 is a schematic diagram of the original Π_, 娃, 字, after this automatic analysis of the error value and automatically doing the best correction step. The circle in the figure is corrected and the pen is reconnected.
前述所產生單線字型係不具有任何寬度之單線字型，接下來即 ，依照筆劃的粗細資料，例如粗黑體、中黑體、或細黑體，以及 子里本身的大小’適當地增加單線字型的寬度，例如16像素％ =素以下的字型，標準寬度設為—點寬，16像素％像素〜^像素 %像素的字型’鮮寬度設為峨ft，31像素*31像素〜63像素 63的予型’標準寬度設為三點寬等等。料色之單線字型來說， =單寬度的方法為往某方向固絲—寬度4為灰階之 1子筆畫:字型(卿獅。她咖心_，則依照與該不具有任何 線子型的早線的距縣決定所長寬度的灰階顏色。請繼 、貝> —· 16®。第16_中間係為—長了寬度的單色單線字型， 以及第I6圖的下方料—長了寬度的灰階單線字型之示意圖 =線字型”―”，往右、下、以及右下方各長出—點The single line font produced in the foregoing does not have a single line font of any width, and then, according to the thickness information of the stroke, such as the bold black body, the middle black body, or the fine black body, and the size of the sub-self itself, the single line font is appropriately added. The width, for example, 16 pixels% = the font below the prime, the standard width is set to - dot width, 16 pixel % pixel ~ ^ pixel % pixel font 'fresh width is set to 峨 ft, 31 pixels * 31 pixels ~ 63 pixels The pre-type 'standard width of 63' is set to three points wide and so on. For the single-line type of material color, the method of single width is to fix the wire in a certain direction - the width 4 is the gray stroke of 1 sub-stroke: font type (Qing lion. Her coffee heart _, then there is no line according to the line The early line of the subtype determines the grayscale color of the width of the length. Please follow, Bay > - 16®. The 16th intermediate is a monochrome single line font with a width, and below the I6 figure. Material - a schematic diagram of a gray-scale single-line font with a wide width = line type "-", grows to the right, bottom, and bottom right - point
圖。第20圖中’左上方，’一，，字的每—點，依照第扣 圖中右上方所顯示的方式，朝右、下、以及右下方各長出一I ==:左下方之具有寬度的，，-，，字。此即是單叫 ==:繼續參考第21圖及第22圖。第21圖係為 、的早色曲線。第22圖為根據第2i圖之原始灰階單色 17 1352298 曲線’長出一寬度之後的灰階單色曲、線之示意圖。灰階單色字型 長出寬度的做法即如第2卜22圖中所示，依據第21圖中，每個 與曲線相鄰點財心點和該曲線的距離，來決定該格的灰階 顏色。距離越近的使用較深的灰階顏色，距離越遠的使用較淺的 灰階顏色，而形成如第22圖中具有一寬度的灰階單線字型的曲線。 比較完256像素*256像素時，單線字型所對應之外框字型和該 單線字型的㈣，進而修正單線字型之後，接下來需考慮當字型 縮小之後’所纽的其他字批f不佳關題。—般字在縮放之 後’都會產生字型品質不佳_題’解決之道即是相對於字型的 母種大小，都加人微_lnting)_訊，來微調在辭型大小時， 微_轉由㈣之自動分析縣餘自動做最佳 =式所產生。請參考第23圖，第23圖即為_，賞，字，在Μ 像素*14像素時，存在的2整; 解決之道即是經過該自動八/而導致糊筆的示意圖。 筆劃不#捕 樣式進行外框箱分析，判斷此2 筆查⑷4像糾纖Μ動分析是否要加人該2筆射的某一 』在4像素丨4像素時，應 直接將該筆__資訊。帛2 ，或者加入 蹲第Μ =才不會導致糊筆的產生。請參考第 的修正，將相^ _ 斤疾差值並自動做最佳修正程式所做 示意圖。 4的1"方橫筆筆劃下移後，所得的，，廟”字的 1352298 加入微調資訊後，最終產生的單線字型的資料格式為： 第1個字 第1筆筆劃序 號+起始關鍵 點+微調資訊 第2筆筆劃序 號+起始關鍵 點+微調資料 訊 第2個字 第1筆筆劃序 號+起始關鍵 點+微調資訊 第2筆筆劃序 號+起始關鍵 點+微調資訊 第S1筆筆劃 序號+起始關 鍵點+微調資 訊 筆劃— 序號+起始關 鍵點+微調資 訊 第N個 字 第1筆筆劃序 號+起始關鍵 點+微調資訊 第2筆筆劃序 號+起始關鍵 點+微調貧訊 第S3筆筆劃 序號+起始關 鍵點+微調資 訊 八中微為貝5fl包含絲劃在某—字型大小時，例如32像素*32像 素^•，轉劃會被轉、刪除、或以另—賴軸的資訊。微調 資訊的資料格式如下： 微調處 針對字 處理方 1¾理值 針對字 處理方 處理佶...... 理數 型大小 —---- 式 ---- ----— 型大小 — 式 _ 19 廇(序號） 筆劃序號=4 微調處理數 針對字型大小 處理方式 處理值 1 14德去* 1 λ你 -- 素 位移 往下移ΥΊ^ -----— ------ 經過如是微娜正後，所顯示之單線字赌更具可讀性及美觀。Figure. In the 20th figure, 'the top left, 'one, every point of the word, according to the way shown in the upper right of the figure, one for the right, the lower, and the lower right, I ==: the lower left has Width,, -,, word. This is the single call ==: continue to refer to Figure 21 and Figure 22. Figure 21 is an early color curve. Fig. 22 is a schematic diagram showing the gray-scale monochrome curved lines and lines of the original gray-scale single color 17 1352298 according to Fig. 2i. The method of growing the width of the gray-scale monochrome font is as shown in Fig. 2, and the gray of the grid is determined according to the distance between the fiscal point and the curve adjacent to the curve in Fig. 21. Order color. The closer the distance is to use the darker grayscale color, the farther the distance is, the lighter grayscale color is used, and the curve of the grayscale single line font having a width as shown in Fig. 22 is formed. When comparing 256 pixels * 256 pixels, the single-line font corresponds to the outer frame font and the single-line font (4), and then the single-line font is modified. Next, it is necessary to consider the other word batches after the font is reduced. f is not good. - After the word is zoomed, it will produce poor font quality. The problem is that the solution is relative to the size of the font, and it adds micro_lnting__ to fine-tune the size of the dictionary. _ Turned to (4) Automatic analysis of the county automatically generated by the best = formula. Please refer to Figure 23, which is _, reward, word, when Μ pixel * 14 pixels, there are 2 whole; the solution is the schematic diagram of the paste pen through the automatic eight /. The stroke does not take the style of the trap and analyzes the outer box. It is judged whether the 2 strokes (4) 4 image analysis and the analysis of the fiber optics must add one of the 2 shots. When 4 pixels 丨 4 pixels, the pen should be directly __ News.帛 2, or join 蹲 Μ = will not lead to the production of paste. Please refer to the corrections in the section, and the schematic diagram of the best correction program will be automatically performed. 4's 1" square pen strokes moved down, the resulting, temple" word 1352298 added fine-tuning information, the resulting single-line font data format is: 1st word 1st stroke number + start key Point + fine tune information 2nd stroke number + start key point + fine tune information 2nd word 1st stroke number + start key point + fine tune information 2nd stroke number + start key point + fine tune information S1 Pen stroke number + starting key point + fine-tuning information stroke - serial number + starting key point + fine-tuning information N-th word 1st stroke number + starting key point + fine-tuning information 2nd stroke number + starting key point + Fine-tuning the poor news S3 strokes serial number + starting key points + fine-tuning information eight in the micro-be 5f containing silk in a certain font size, such as 32 pixels * 32 pixels ^ •, the transfer will be transferred, deleted, Or use the information of the other axis. The data format of the fine-tuning information is as follows: The fine-tuning is for the word processing side. The processing value is processed for the word processing side... The number of the type is ----------- - ----- Type size - Type _ 19 廇 (serial number) Stroke number = 4 The number of fine-tuning processing is processed according to the font size processing method. 1 14 de to * 1 λ you -- the displacement of the prime moves down ---- ^ ----- ------ After the appearance of the micro-Nan, the display Single-line word bets are more readable and beautiful.
請注意：上述實施例係以nsx〇2〇8字集為例說明，但本發明 並不限只能麵_字集，只要符合本㈣之精神，其他字制 如中文繁體(Big 5)料，均為本發明所涵蓋的範圍。 练上所述’本發明係提出一由外框字型產生單線字型的方法， 將現有的外框字麵4，娜―適當料型大小，產生適當的單 線„，再由被產⑽單線筆舰本中，齡適當的筆劃組 j早線子^。再根雜大字斜的外框钟，修正触成的單線 子型’接者根據字型的粗細資料，加寬單線字型，最後加入微調 的資訊，修正加寬後的單線村，使所產生之單線字型更加完美。 ^㈣的妓可以不时新建立單線師㈣訊，直接由原本具 的框筆劃貪訊，獲得單線筆劃資訊，以彌補傳統方法的不足。 20 1352298 以上所述僅為本發明之較佳實施例，凡依本發明申請專利苑 圍所做之均等變化與修部，皆應屬本發明之涵蓋範園。 【圖式簡單說明】 第1圖係為-黑體之單線字型的示意圖。 第2圖係為―圓體之單線字型的示意圖。 第3圖係為—海報體之單線字型的示意圖。 第4圖係為-由外框筆劃所構成之外框字型，，凸，，之實例圖。 =圖則為第1圖之外框筆劃塗黑後所組成之外框字型，，凸，，之實 第6圖係為第4圖、笛《闽心， 々乐，弟5圖中”凸，，字之筆劃順序圖。 第7圖係為第4圖、裳$ (^心， , "凸’字之第4筆劃，巧，冲之關鍵點之 位置及見度值之資訊。 第8圖即為—曲線段(起 點A1 ’結束點A3)上，根據貝氏曲線指 卜々控繼A2Wf之曲線段之示意圖。 第9圖係為，’介，，字中# / X 、氏曲線所定義之控制點描述左下撇” /的示意圖。 第1〇圖係為根據不同的關 -”之示意圖。趣⑽度值，_成的不同的” ^圖係、為根據本翻之方法崎之 第在顯示字型―像素Z；顯示的”-，，字和” 第13圖係為像㈣物，_爾，-，，字和” 弟14圖你i * 、為—由-鱗之外框箱寬度的中轉換為單線筆割之 示意圖。 # θ係為取線之外框賴—邊的控制，轉換為單線賴 . 之示意圖。 弟16圖為？—單線字型’-長了寬度的單色單線字型，以及一 長了寬度的灰階單線字型之示意圖。 17圖係為-因為所取得之單線_樣本太短，㈣致接筆不相 ^ 連的’’娃”字的示意圖。 圖係為自動分析誤差值並自動做最錄正的流程圖。 9圖即為經過自動分析誤差值並自動做最佳修正步驟後的原第 17圖的”娃，，字之示意圖。 第2〇圖·;根據單線字型，，—，，，往右、下、以及右下方各長出-點 的寬度的示意圖。 # 第21圖係為一原始灰階的單色曲線。 '第22圖秘據第21圖之原始灰階單色曲線，長出一寬度之後的 灰階單色曲線之示意圖。 第23圖即為一”瘤，，字，在14像素*14像素時，存在的2筆橫筆劃 相連，而導致糊筆的示意圖。 ★圖U動分析誤差值並自動做最絲正程柄做的修正 後，所得的，，廇，，字的示意圖。 〆 22 1352298 【主要元件符號說明】 Α1, Α2，Α3 ,Β 1，B2,B3 ,B4,B5，B6,B7，B8,B9,B 10,C 1,C2,C3，C4，C5，A10，A11，A12，A13，A14，A15，A16 12,14,16,18,20,22,100,102,104 控制點 步驟 23Please note that the above embodiment is exemplified by the nsx〇2〇8 word set, but the present invention is not limited to the _word set, as long as it conforms to the spirit of (4), other characters such as Chinese Traditional (Big 5) material All are within the scope of the invention. In the present invention, the present invention proposes a method for generating a single-line font from the outer frame font, and the existing outer frame font 4, Na-suitable material size, to generate an appropriate single line „, and then produced (10) single-line pen In the shipbook, the appropriate stroke group j is early line ^. The outer frame clock with the big word is slanted, and the single-line subtype of the modified one is adjusted according to the thickness of the font, the single line font is widened, and finally added. Fine-tuning the information, correcting the widened single-line village, so that the single-line fonts produced are more perfect. ^ (4) The 妓 can newly establish a single-line division (4) news from time to time, directly from the original box strokes, get single-line stroke information, In order to make up for the deficiencies of the conventional method. 20 1352298 The above description is only a preferred embodiment of the present invention, and all the equivalent changes and repairs made by the patent application court according to the present invention should belong to the scope of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic diagram of a single-line font of a black body. Fig. 2 is a schematic diagram of a single-line font of a circle. Fig. 3 is a schematic diagram of a single-line font of a poster body. 4 graphics is - composed of outer frame strokes The outer frame font, convex, and the example map. = The plan is the outer frame of the outer frame of the first picture, which is composed of the outer frame, and the convex, and the real picture is the fourth picture. Flute "闽心, 々乐, brother 5 in the picture" convex, the stroke sequence of the word. Figure 7 is the fourth figure, the fourth stroke of the $ (^心, , "convex' word, the information of the position and the value of the key point of the rushing. The 8th figure is the curve section ( At the starting point A1 'end point A3), according to the Bayesian curve, the schematic diagram of the curve segment following the A2Wf is shown. The figure 9 is the control point defined by the '#, X, and the curve in the word.撇" / The schematic diagram. The first diagram is based on the different -" diagram. Interest (10) degree value, _ into the different" ^ map, according to the method of this method of the display of the first word - Pixel Z; displayed "-,, and word" Figure 13 is for the like (four) object, _ er, -,, and "different 14" you i *, for - by - scale outside the box width Schematic diagram of the single-line pen cutting. # θ is the control of the outer frame of the line, and is converted into a single line. The picture of the brother 16 is - single line type '- monochrome single line font with long width And a long-width gray-scale single-line font. The 17-graph is - because the single line _ sample obtained is too short, (4) the word "'wa" that is not connected to the pen Intentions The graph is a flowchart for automatically analyzing the error value and automatically making the most recorded. The 9 graph is the schematic diagram of the original figure 17 after the automatic analysis of the error value and the automatic correction step. Figure 2: According to the single-line font, -,,, to the right, the bottom, and the bottom right, the width of each point - the width of the point. #第21图 is a monochromatic curve of the original gray level. Figure 22 is a schematic diagram of the original gray-scale monochromatic curve of Figure 21, which shows a gray-scale monochromatic curve after a width. Figure 23 is a "tumor," word, at 14 pixels * 14 pixels, The existing two strokes are connected, which leads to the schematic diagram of the paste. ★ Figure U moves the error value and automatically makes the correction of the most straight handle, resulting in the schematic diagram of the resulting, 廇, and 。22 1352298 [ Explanation of main component symbols Α1, Α2, Α3, Β 1, B2, B3, B4, B5, B6, B7, B8, B9, B 10, C 1, C2, C3, C4, C5, A10, A11, A12, A13, A14, A15, A16 12, 14, 16, 18, 20, 22, 100, 102, 104 Control point step 23