201005553 九、發明說明： 【發明所屬之技術領域】 本發明係關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】 ❹ ❹ 隨著資訊奸㈣蓬紐展，顿輕置(p_bledevice)及消 費性電子產品(consuming dectr〇nics)也越來越熱門。而這些產品例 如行動電話、個人触賴(p⑽nal digital㈣越，pDA)、以及 mp3媒體播放n料皆财設置齡糊資訊醜示裝置，以提 供使用者充足的操作資訊。但也gj為這些產品的體積皆不大，其 顯示裝置亦不大，因此只適合齡較小的字型。以單脉加㈣⑽) 筆劃所產生的字型’技術上每_是以單線騎幹，由各筆書㈣ 成二個個字之後’經過縮小再㈣幹長肉變粗，來產生不同尺寸 的子型。早線字型·旨黑體、龍、触鮮—直制只有一寬 $的字型’並不翻於楷書體等，—直線間具有多種粗細寬度 :子型。請參考第1圖，第2圖，以及第3圖。第1圖係為一里 =單線相的示意圖，第2圖係為—圓體之單線字型的示意圖，、 圖係為-海報體之單線字型的示意圖。由單線筆劃所產生 的小子型，速度快又品質好，在$幕上的可讀 應用於目前大量製造的手持式裝置及消費性電 问。从口 但目 般的漢字字型廠商大部分_由設計外框 筆劃來組 5 201005553 成漢字字型’叫方 以組出所有漢字之:“種大小之字型及使用少數筆劃就可 圖係為—由外框二參考第4圖、第5圖、以及第6圖。第4 為第4圖之外框筆之外框字型，，凸，，之實侧，第5圖則 圖係為第4 、後所組成之外框字型咖 該，，凸，，字传由五筆查中&子之筆劃順序圖。由第6圖中可知 子係由五糊所構成 ❹ 筆劃係為-雖線，μ 猶為⑯直線1，第2 為-” V，，第5、输働為一短直線，，卜第4筆劃係 伤、； 係為一長橫線，，—，，。每術卜框字型的資料格1 係以下列表1的方式儲存於記憶體中：式 第1個 字 +參數 第2筆筆劃序 號十參數 第S1筆筆劃序 號+參數 弟2個 字 第1筆筆劃序號 +參數 第2筆筆劃序 號+參數 第S2筆筆胃 號+參數 弟N個 字 ------- 第1筆筆劃序號 +參數 _______ 第2筆筆劃序 號+參數 第SN筆筆劃序 號+參數 表1 以日本工業規格(japanese Industrial Standards)中所制定之其本 漢字編碼系統JIS X 0208為例，JIS X 0208的字集共約7〇〇〇個 字，由共约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 201005553 ❹ 7000個字集中所有65000筆筆劃。故表1中之”筆劃序號，，即為該 200〜300筆基本筆劃之筆劃序號。表1中之”參數”貝彳包含一筆劃中 關鍵點之位置及寬度值。請參考第7圖，第7圖係為，，凸，，字之第4 筆劃’巧’’中之關鍵點之位置及寬度值之資訊。”凸，，字之第4筆劃，， 中有5個關鍵點及4個寬度值。5個關鍵點包含一起始點(第i 關鍵點）’ 一結束點(第5關鍵點），以及3個轉折點(第2關鍵點， 第3關鍵點，以及第4關鍵點）；4個寬度值則包含介於第丨關鍵 點以及第2關鍵點之間、表示該段線段之寬度之一第丨寬度值， 介於第2關鍵點以及第3關鍵點之間、表示該段線段之寬度之一 1 2寬度值，介於第3關鍵點以及第4關鍵點之間、表示該段線 之見度之一第3寬度值，以及介於第4關鍵點以及第5關鍵點 之間、表示該段線段之寬度之一第4寬度值。 儲存於外框筆劃表中的每一外框筆劃的資料格式，則如下列 表2 : 筆劃序號=2 外框描述程式1(參數=關鍵點，寬度值) 外框描述程201005553 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] ❹ ❹ With the information (4) Pengu exhibition, p_bledevice and consumer electronics (consuming dectr〇nics) are also becoming more and more popular. These products, such as mobile phones, personal touch (p(10)nal digital (p), pDA), and mp3 media playback, are all available to provide users with sufficient operational information. However, gj is also a small size for these products, and its display device is not large, so it is only suitable for smaller fonts. The font generated by a single pulse plus (four) (10) strokes is technically used to make a single line. After each book (four) is made into two words, it is reduced by (four) dry long meat to produce different sizes. Subtype. Early line type, black body, dragon, touch fresh - straight only a wide $ font 'do not turn over the book body, etc.—the line has a variety of thickness width: subtype. Please refer to Figure 1, Figure 2, and Figure 3. The first picture is a schematic diagram of a single line = single line phase, the second picture is a schematic diagram of a single line font of a circle, and the figure is a schematic diagram of a single line type of a poster body. The small sub-type produced by single-line strokes is fast and of good quality, and can be read on the screen for use in mass-produced handheld devices and consumer electronics. From the mouth, but the majority of Chinese character font manufacturers _ by the design of the outer frame strokes to group 5 201005553 Cheng Han font type 'calling party to group all the Chinese characters: "the size of the font and the use of a few strokes can be mapped For the first frame, refer to Figure 4, Figure 5, and Figure 6. The fourth frame is the outer frame of the outer frame of Figure 4, convex, and the real side, and the fifth image is For the 4th and the later, the outer frame type of the word type, the convex, the word is passed by the five strokes and the sub-stroke sequence diagram. From the sixth picture, the sub-system is composed of five pastes. - Although the line, μ is still 16 straight line 1, the second is -" V, the fifth, the input is a short straight line, the fourth stroke is a wound, and the line is a long horizontal line, -,,,. The data box of each frame type is stored in the memory in the following list: the first word + parameter, the second stroke number, the ten parameter, the S1 stroke number, the parameter, the second word, the first sentence Stroke number + parameter 2nd stroke number + parameter S2 pen stomach number + parameter brother N word ------- 1st stroke number + parameter _______ 2nd stroke number + parameter SN pen stroke No. + parameter table 1 Take the Chinese character encoding system JIS X 0208, which is defined in the Japanese Industrial Standards, as an example. The JIS X 0208 word set has a total of about 7 characters, and a total of about 65,000 strokes. Composed of. However, for the outer frame font, usually only 200 to 300 basic strokes, plus parameter changes, can be changed. The 201005553 ❹ 7000 7000 word sets all 65,000 strokes. Therefore, the "stroke number" in Table 1 is the stroke number of the 200-300 basic strokes. The "parameter" in Table 1 contains the position and width of a key point in the stroke. Please refer to Figure 7, Figure 7 is the information on the position and width of the key points in the 4th stroke 'Qiao' of the word. "Convex, the 4th stroke of the word, there are 5 key points and 4 Width values. The five key points include a starting point (i key point) 'one end point (5th key point), and 3 turning points (2nd key point, 3rd key point, and 4th key point); 4 widths The value includes a width value between the second key point and the second key point indicating a width of the segment line segment, between the second key point and the third key point, indicating the width of the segment line segment. One of the 1 2 width values, between the 3rd key point and the 4th key point, indicating the 3rd width value of the visibility of the section line, and between the 4th key point and the 5th key point, Indicates the fourth width value of one of the widths of the segment. The data format of each outer frame stroke stored in the outer stroke list is as follows: Table 2: stroke number = 2 outer frame description program 1 (parameter = key point, width value)
筆劃序號 外框描述程 表2 表2中的外框描述程式係用來產生 不同的字型，例如黑體、圓體、 7 201005553 ‘措書體等等，的不同筆劃，例如直線、斜線、或橫線等等，亦即 描述4筆劃外框的程式。下面的例子即是一個產生一橫筆直線 的簡單的外框描述程式的例子： 參數-關鍵點：pnti，pnt2，寬度值：w。 步驟1:從關鍵點pntl向上移動w/2，產生newpnt卜 步驟2:從newpntl向下移動w，產生加…坪口。 ❹ 步驟3:從關鍵點pnt2向上移動w/2,產生newpnt3。 步驟4:從newpnt3向下移動w，產生newpnt4。 步驟5:從newpntl畫直線到newpnt3。 步驟6:從newpnt3畫直線到newpnt4。 步驟7:從newpnt4晝直線到newpnt2。 步驟8:從newpnt2晝直線到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_A11_A12 w 再加上曲線段A12-A13-A14。曲線段A12-A13之起始點A12和結 8 201005553 束點A13都在該筆劃上，而由貝氏曲線所得之控制點，則不 位於該筆劃上’但可透過控繼A15來控制曲線段Ai2_Ai3的孤 度。同理，曲線段A13_A14之起始點Αη和結束點Ai4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。 之筆劃序號為1，加上不同的關 相同的一筆劃序號，例如: ❺鍵點位置及寬度值，可以形成不同的”―”。請參考第⑽圖，第ι( 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200〜300筆基本筆書j^Stroke number frame description path table 2 The frame description program in Table 2 is used to generate different fonts, such as black body, circle body, 7 201005553 'booking body, etc., different strokes, such as straight lines, diagonal lines, or Horizontal lines and so on, that is, a program that describes 4 strokes. The following example is an example of a simple outline description program that produces a straight line: Parameters - Key Points: pnti, pnt2, width value: w. Step 1: Move w/2 from key point pntl to generate newpnt. Step 2: Move w from newpntl down to generate plus... ❹ Step 3: Move w/2 up from the key point pnt2 to generate newpnt3. Step 4: Move w down from newpnt3 to generate newpnt4. Step 5: Draw a line from newpntl to newpnt3. Step 6: Draw a line from newpnt3 to newpnt4. Step 7: From newpnt4昼 straight line to newpnt2. Step 8: From newpnt2昼 straight line to newpnt Bu Step 9: End. If the starting point and the ending point are not a straight line but a curve, the curve is intersected by two straight lines respectively tangent to the starting point and the ending point according to the Bezier Curve principle. The control point is used to 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 'end point A3). Please refer to Figure 9 again. Figure 9 is a schematic diagram of the lower left 撇', /, and the control points defined by the Bayesian curve in the word "介". In Fig. 9, the upper edge of the left lower jaw of the word is the straight line segment A1〇_A11_A12 w plus the curve segment A12-A13-A14. The starting point A12 and the knot 8 201005553 of the curve segment A12-A13 are on the stroke, and the control point obtained from the Bayesian curve is not located on the stroke 'but the control segment can be controlled by controlling A15 The degree of asymmetry of Ai2_Ai3. Similarly, the starting point Αη and the ending point Ai4 of the curved segment A13_A14 are both on the stroke and the control point A16 obtained by the Bayesian curve is not located on the stroke, but the curve segment A13_A14 can be controlled through the control point A16. The curvature. The stroke number is 1, plus a different number of the same stroke number, for example: ❺ key point position and width value, can form a different "-". Please refer to the figure (10), the figure ι (the picture shows the different key position and width values formed by the different, -,, schematic diagram. From the above, we can see that the external frame type is 200~300 basic Pen book j^
Hi變__及寬度值）’即可變化請X 0氣約700C 個子木申所有約65000筆筆劃。 若要捨棄目前通用的外框字型，專為 · 她綱㈣綠，地-適當之ΐ φ 佔叙記細賴射間，料單線筆* 樣本中擷取單線筆查，丨，έ日士 行田早線筆畫 線字__^ ，再鋪鼓字型時_ 所對應的外框字型，修正單線字型，接糾· 型的粗鈿資料，長出一適當 、斤欲顯示字 寬之單線字型，使所產生之單線字型更加完f，修正被加 201005553 【發明内容】 線字型的方法，包含根據 本發明係揭露一種由外框字型產生單 ^諸存單線筆舰本的數目，由外框字型產生 观本；_單線筆舰本巾單線軸型祀 =型:對應之外框字型，修正該單線字型;根二= 細Μ，加細__物，喻具有寬度之 轉衫，以及加入微言周資訊(hinting)，修正該具有寬度之軍線字 【實施方式】 本發明所提供之方法，即根據約筆筆劃之外框字型筆 W ’產生較少之早線軸縣(因為本發贿提供之字體，係為了 ，用於較小_讀幕，料t產生所有約6_轉劃之單線 字幻再由單線筆劃樣本，組成單線字型。在下面所舉之實施例 :’所產生之單線筆舰本約為—油左右，接著由單線筆劃樣 本中擷取適當之_組字，再㈣較大字麟單線字型所對應的 外框字^修正單線字型，接著根據所欲顯示字型的粗細資料， 出適田的寬度，再經由微調技術，修正所加寬之單線字型。 請參考第11圖。第11圖係為根據本發明之方法所緣之-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： 201005553 步驟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 :根據所選取之適當的字型大小，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。其中該基本夕卜 框筆劃的資訊包含2〇〇〜3〇〇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。 ❹ 步驟！8 :在較大字型256像素*256像素時，將所產生之單線字型 與外框子型做比較，修正該單線字型。 步驟:姆賴种型_細龍，產生具有適當寬度之 字型。 又 ''' 步驟22 ··加人微調資料，修正該具#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 〜_筆左右之基本外框筆劃中每一基本外框筆劃均包含= 镜之賴序號、複數個關鍵點位置、及複數 如前所述，該細〜300筆左右之基本外框 。 框字形之昝剌^m 苓剷化出所有外 mj。該200〜300筆左右之基本外框 計=身之需求而定，本實施例所提供之數字，只=考：： 凡付合本發明之精神，所產生之適當數量的基本外 ^㈣細。紙意：㈣财⑽麵，^本^ 的只知*例，各步，驟66香·κ — 疋本發月 順序，凡符合本發明斧；’並不一疋要依照如圖♦所列的 本發明猜神，各步驟間不同的實施順序，亦為本發 11 201005553 明所涵蓋之範疇。 太^先要f生早線筆劃樣本。首細根據所欲儲存之單線筆佥》 適當大小之字型。在本系統中，例如以^ ϊγΓ，母一個字原本係以256像素*256像素造字。® X _ 予木中所有予之外框筆劃之總數量約為65〇〇 型大小為―像素，由六萬二^ ❹ 里、·，勺為近-萬個的單線筆劃樣本。若選擇字型大小為64 像素，則產生的單線筆劃樣本數量&素64 齡曰4双里、禹夕個。如何選取適當 目之早線筆應本，端視設計者本身的需求 T效能等等各方面的考量驰所舉出的數二一 定於魏字，凡符合本㈣之·，所代入的適當 數值，都屬於本發_朗。當然，所·的字型越小，筆 越Γ資料量少，記憶體的空間少，但所產生的筆 ，差也越大，因猶縣舰本數和職生的箱誤差之間， 2 一個可以接受的平衡點。在較大的字型為256像素*256像素 的情形下’若不同的橫筆直線的長度分別為，舉例來說，48、47、 =去44個像素，卿小至字型為32像素^像鱗物像素⑽ 像素的八分之十不同的橫筆直線的長度分別為48/8=6、 47/㈣、·5.6、44/8=5.5像素，但像素係為整數個，若採取 四捨五入的方式’該四種在祝像素郎像素時，不同長度的橫 ^線，在32像㈣像素時，全部變成相同長度6像素的橫筆 直線。請參考第12圖及第13圖。第12圖係為在256像素， 201005553 ❹ ❹ 像素時，所顯示的，，一”字和，，三，，字。由第12圖中可看出，，，一’’字 和”三’’字的3條橫筆_，在顯示字型_像素⑽ =素時’長度都不同。第13圖則為在顯示字型為％像素㈤像素 广所’’’、頁不# -字和’’二”字。此時，，一，，字的橫筆筆劃和，，三”字最 2㈣筆筆劃長度變成相同了。因此，在顯示字型為％像素％ 的整果將關鍵點的距離與寬度值均相同的筆劃，視為相同 ，透過此做法，可將原本約65G(K) __，減少為 數里約為近一萬個的單線筆劃樣本。 而實際上的做法，係在32像素*32像素的㈣大小下 =劃=包含筆劃序號、參數值)去形細糊8字集中的所 二。框:劃轉為單線筆劃的方法’儲存每個字的單 已儲的某一筆單線筆劃，與前面所產生的、 、、本子型的其中一筆劃相同（關鍵點的距離 同）’則不再鱗魏賴。蝴綱峨^ 於讀巾所有字的適當數量(根據本實麵約為一萬字) ㈣地士 _(約〜個)七卜框筆劃轉為單 線聿^的方法，這齡紹兩種： 取新框筆劃寬度的中點連成單線筆劃。請參考第 14圖传盔__u 圓乐 ^ —由—曲線之外框筆劃寬度的中點轉換為單線筆 里1之不思圖。第14圖中包含由貝氏曲線所得之控制點B1、 13 201005553 B2、B3、B4、B5、B6、B7、B8、B9、B1〇 等 ι〇 點Hi change __ and width value) can change. Please X 0 gas about 700C sub-wood applications all about 65,000 strokes. If you want to abandon the current common frame font, it is designed for her (4) green, land-appropriate ΐ φ 占 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录 录The line of the line of the early line of the line is __^, and when the word is typed, the corresponding frame type is corrected, the single line type is corrected, and the rough type of the type is selected, and a proper amount is displayed. The single-line font type makes the generated single-line font type more complete, and the correction is added 201005553. [Inventive content] The method of line type includes, according to the present invention, a single-line storage single-line pen shipbook generated by the outer frame font type. The number of the frame is generated by the outer frame font; _ single-line pen ship towel single-line type 祀 = type: corresponding to the outer frame font type, correct the single line font; root two = fine Μ, add fine __ object, Having a width of the shirt, and adding hinting information (hinting), correcting the width of the military line word [embodiment] The method provided by the present invention, that is, according to the pen stroke outside the box type pen W 'produced Less early in the spool county (because the font provided by this bribe is for the sake of the smaller _ reading curtain, material t production All single-line characters of about 6_transition are composed of single-line stroke samples, which constitute a single-line font. In the following example: 'The single-line pen ship produced is about - oil, and then by the single-line stroke sample. Take the appropriate _ group word, and then (4) the outer frame word corresponding to the larger character lining single line font ^ correct the single line font type, and then according to the thickness of the desired font type, the width of the field, and then through the fine-tuning technology, the correction The widened single-line font. Please refer to Fig. 11. Figure 11 is the method according to the present invention - the flow chart in Figure 11 provides the single-line font generated by the outer frame font provided by the present invention. The method comprises the following steps: 201005553 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 font size, from the information of the basic outline strokes 'Generate a sample with a suitable number of single-line strokes. The information of the basic-note strokes includes the corresponding stroke number, multiple key positions, and width values of the basic frame strokes of 2〇〇~3〇〇 Step 16: According to the generated single-line stroke sample, form a single-line font. ❹ Step! 8: When the larger font is 256 pixels*256 pixels, compare the generated single-line font with the outer frame sub-type, and correct the single-line. Step: Mray type _ fine dragon, produce a font with the appropriate width. Also ''' Step 22 · Add people to fine-tune the data, modify the single line font with the width of #适#. In the present invention The basic frame type is generated by the outer frame font type. Each basic outer frame stroke in the left and right basic frame strokes includes = mirror number, multiple key point positions, and plural numbers as described above. The basic frame of ~300 pens or so. The frame font is 昝剌^m 苓 化 化 all the external mj. The basic frame of about 200~300 pens = the demand of the body, the figures provided in this embodiment, Only = test:: Where the spirit of the invention is paid, the appropriate amount of basic external ^ (four) is produced. Paper meaning: (4) Cai (10) face, ^ this ^ only know * example, each step, step 66 Xiang · κ - 疋 发 发 , , , , , , 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合 符合The present invention contemplates that the different implementation sequences between the various steps are also within the scope of the present disclosure. Too ^ first want to f live early stroke sample. The first size is according to the single-line pen that you want to store. In the present system, for example, ^ ϊ γ Γ, the mother word is originally 256 pixels * 256 pixels. ® X _ The total number of strokes in the outer frame of the wood is about 65 〇〇. The size is ―pixels, from 62,000 ❹ 、, ·, spoon is nearly 10,000 single-line stroke samples. If the font size is 64 pixels, the number of single-line stroke samples will be & 64-year-old 曰 4 double-living, 禹 个. How to choose the appropriate early line pens, look at the designer's own needs, T-efficiency, etc. The number of the two is determined by the Wei word, which is consistent with this (4) The values are all in this hair_lang. Of course, the smaller the font size is, the less the amount of data is, the less space the memory has, but the pen that is produced is worse, because of the difference between the number of ships in the county and the box error of the students. 2 An acceptable balance point. In the case where the larger font is 256 pixels * 256 pixels, 'if the lengths of the different horizontal lines are respectively, for example, 48, 47, = go to 44 pixels, and the small to the font is 32 pixels ^ Like the scale pixel (10) pixels, the length of the eight-eighth different horizontal line is 48/8=6, 47/(four), ·5.6, 44/8=5.5 pixels, but the pixels are integers, if rounded off The way of the four kinds of horizontal lines of different lengths, when 32 pixels (four) pixels, all become a horizontal line of the same length of 6 pixels. Please refer to Figure 12 and Figure 13. Figure 12 is the display of the 256 pixels, 201005553 ❹ 像素 pixels, a "word and,", three, and words. As can be seen from Figure 12, a ''word and 'three' 'The three horizontal pens of the word' are different in length when displaying the font_pixel (10) = prime. The 13th picture shows the font type as % pixel (five) pixel wide ''', page not #-word and ''two' word. At this time, one, the word of the horizontal stroke stroke and, three words The length of the strokes of the second (four) pen becomes the same. Therefore, in the case where the display font is % pixel %, the strokes with the same distance and width value of the key points are regarded as the same. By doing so, the original approximately 65G(K) __ can be reduced to a few miles. Ten thousand single-line stroke samples. In fact, the practice is based on the size of 32 pixels * 32 pixels (four) = stroke = contains stroke number, parameter value) de-shaped in the 8 word set. Box: The method of transferring to a single-line stroke 'storing a single-line stroke of a single word stored in each word, which is the same as the one of the previous generations of the sub-type (the distance of the key points is the same)' Scale Wei Lai. Butterfly 峨^ The appropriate number of words in the reading towel (about 10,000 words according to the actual face) (4) The method of converting the pen _ (about ~ a) seven-segment strokes into a single line 聿 ^, this age is two: Take the midpoint of the stroke width of the new frame and connect it to a single stroke. Please refer to Figure 14 for the helmet __u round music ^ - by the - the midpoint of the stroke width of the curve outside the box is converted to a single line pen. Figure 14 contains the control points B1, 13 201005553 B2, B3, B4, B5, B6, B7, B8, B9, B1, etc. obtained from the Bayesian curve.
Bl、R9 * 穴》 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。B1、B2形成一曲線段Bl-B2(如第8圖中 X— A 1 、A3) ’而由貝氏曲線所得之另一控制點Β7*在該曲 線丰又上(如第8圖中之Α2)，但可透過控制點Β7來控制曲線 的弧度。Β2、Β3形成一曲線段Β2_Β3，同理，另 控制點Β8不在該曲線段上，但可透過控制點Β8來控制曲 線段Β2七3的弧度。同理，Β4、Β5形成另一曲線段β4_β5， 則為不在該曲線段上之另一控制點，Β5、Β6形成另一曲線 段Β5-Β6，而Β10為不在該曲線段上之另一控制點。此做法 為將Β1與Β6相連取中點C1 ’ Β2與Β5相連取中點q，Β3 與Β4相連取中點C3，Β7與Β1〇相連取中點控制點c4，別 與B9相連取中點控制點C5，再根據控制點C4決定曲線段 1 C2 ’根據控制點C5決定曲線段C2-C3，即得所求之單線 筆劃。 、、 2.取该外框筆劃一邊的控制點連成單線筆劃。請參考第15圖。 第^圖係為-取一曲線之外框筆劃—邊的控制點，轉換為單 線筆劃之示意圖。第15圖中，包含由貝氏曲線所得之控制點 B1 B2、B3、B4、B5、B6、B7、邮、B9、B10 等 1〇 點， 與第14圖相同，其餘結構關係不再贅述。本實施例中所選取 之單線筆劃為下邊緣B4_B5_B6，亦即以曲線则5、出都 為所求之單線筆劃。此法是指定-邊的起始與結束的控制 14 201005553 點，與介於巾間的繼點減，即為單 =指定該外框筆劃的_邊，來連成單線筆劃 框字里的雜财有寬度，依此法所得並非真正該外框筆劃 的中〜線’為了降低縣，朗此法時建議將料 寬度先設為零。 ❹ 參 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 筆5ΓΓ不同的是記錄單線的筆劃序號(約—萬個左右)與原外框 」的顿本貫施例為·〜個)不同，而參數也僅剩下第一關 私點(起始點），其他_點與寬度值均不存在了。由於參數只有起 去取^難4筆劃樣本組字，每個字根據它的單線筆割序號， 為此二=筆=第=第一 _)的位置， 此法所產生的單線字型:::二16圖，第一^ a. 位置誤差： b. 長度誤差： 因為所取得之單線_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）。 因為所取得之單縣舰本的長度有誤差，導致該筆 201005553 劃太短或太長。 r請而參㈣17嶋—目為馨㈣線筆劃樣本太 短=致接筆不相連的，，娃，，字的示意圖。圖中圓圈處即接筆不 相盖之&。但有時此問題可在增加單線字型寬度之後，獲得〜些 改〇 ❹ 解决口為單線筆劃樣本誤差而導致的字型品 f自動讀縣健自_最雜正程式，進⑽正。 =18 _自動分析嶋並自動做最佳修正的詳細 步驟励：分析單線字型所對應之外框字型在字型大小㈣像素 *256像素時，各軸之_相簡係。 、 步驟服：娜在字鼓小為况像素城像素時， 各筆劃之間的相互關係。 步驟刚：比較兩者筆劃之間的相互關係，若有不同時，根據 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祝像素咖 像素時(齡字型最大時)，單線字型觸應之外財型線 之不同’來進行自動修正。因為在鼓字型時，料型= 線字型有不_，會最麵，因崎正鼓扣 矛早 16 201005553 :在，大子型時’外框字型和單線字型都沒有不同處，則縮 較小子型時，例如64像素^ '''… 同。往采m《 32像素*32像素等就不會不 月乂 …第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丨7fe|AA,， 左俚並自動 乂職_第17 _，，娃，，字的示細 經過修正後，接筆重新相連之處。 _處即 ❹ ❹ 别述所產生單線字型係不具有任何寬度之單線字型，接下來 要依照筆劃的粗細資料，例如粗黑體、中黑體、或細里體，以及 字型本身的;M、，適當地增加單線字型的寬度，例如16像素％ 像素以下的子型，標準寬度設為-點寬，16像素％像素〜31像素 *31像素的字型’標準寬度設為兩點寬，31像素*31像素〜音 63的子型’標準寬度設為三點寬等等。對單色之單線字型來 =轉2敢度的方法為往某方向固定長—寬度；若為灰階之 $子筆剎字型(grayscale center-lined font)，則依照與該不具有任何 ^度之單線字型的單線的麟來決定所長寬度的灰階顏色。請繼 續參考第16圖。帛16圖的中間係為—長了寬度的單色單線字=， 以及第16圖的下方係為一長了寬度的灰階單線字型之示意圖。第 20圖為-根據單線字型，，一”，往右、下、以及右下方各長出一點 的寬度的示意圖。第20圖中，左上方，，一”字的每一點，依照第加 圖中右上方所顯示的方式，朝右、下、以及右下方各長出一點， 而形成如第2G圖左下方之具有寬度的，，—”字。此即是單色之單線 字型長寬度的方法。請繼續參考第21圖及第22圖。第21圖係為 一原始灰階的單色曲線。第22圖為根據第21圖之原始灰階單色 17 201005553 曲線，長出一寬度之後的灰階單色曲線之示意圖。灰階單色字型 長出寬度的做法即如第21、22圖中所示，依據第21圖中，每個 與曲線相鄰的格點的中心點和該曲線的距離，來決定該格的灰階 顏色。距離越近的使雜深的細顏色，麟越遠的使用較淺的 灰階顏色，而形成如第22圖中具有一寬度的灰階單線字型的曲線。 比較完256像素*256像素時，單線字型所對應之外框字型和該 ❹ ❹ 單線字型的不同’進*紅單線字叙後，接下輕考慮當字型 =小之後，所纽的其他字型品般字在縮放之 二’都會產生字型品質不佳的問題，解決之道即是相對於字型的 :種大小’都加A_(hinting)㈣訊，來微調在該字型大小時， I:微調資訊亦由前述之自動分析誤差值並自動做最佳 U正私式難生。請參考第23圖，第23 _為n 像素*14像素時，存在的2筆橫筆 2 解決之道即是經射自動八而導致糊筆的示意圖。 軸^ 析程式進行外框箱分析，判斷此 箬:連，接著繼續自動分析是否要加人該2 _中的某一 4像素％像素時，應該被上移或下移的資訊，或者力、 直接將該筆_除的資訊。第23 _ _ 口 的橫筆筆劃，應該被下移 相連筆屬下方 叫第料即是經自的產生。請參考第 的修正，將猶2 ^做最錄正程式所做 示意圖。 下移後，所得的，，肩，，字的 η 201005553 加入微調資訊後，最終產生的單線字賴資料格式為： 第1個字 第1筆筆劃序 號+起始關鍵 點+微調資訊 第 號+起始關鍵 點+微調資料 訊 ❹ ❹ 第2個字 第N個 字 第1筆筆劃序 號+起始關鍵 點+微調資訊 第2筆筆劃序 號+起始關鍵 點+微調資訊 第S1筆筆劃 序號+起始關 鍵點+微調資 訊 第呂^^筆劃 序號+起始關 鍵點+微調資 訊 第1筆筆劃序 號+起始關鍵 點+微調資訊 第2筆筆劃序 號+起始關鍵 點+微調資訊 第S3筆筆劃 序號+起始關 鍵點+微調資 訊 其中微調資訊包含該筆書彳在草一丰 去性系子型大小時，例如32像素*32像 素時，該筆劃會被位移、刪除、$ 資訊的資料格式如下： 另—筆割替換的資訊。微調 微調處 理數 型大小 式 型大小 式 19 201005553 •以上述的”雇”字在14像素*14像素時，2相連筆劃的下方的橫筆 筆劃(第4筆筆劃），應該被下移一格的例子來說，其. 扃(序號) 微調處理數 f 筆劃序號=4 針對字邊大小 處理方式 處理值 1 14像素*14像 素 位移 往下移1格 經過如S微娜正後’所麻之料字赌更具可·及美觀。 、請注意：上述實施例係以JISX〇2〇8字集為例說明，但本發明 並不限只㈣用於該字集，只要符合本發明之精神，其他字集例 如中文繁體(Blg5)等等，均為本發明所涵蓋的範圍。 ^上所述’本發明係提由外框字龍生單線字型的方法， 〇將現^的外框字型資料，選擇-適當的字型大小，產生適當的單 線筆樣本，再由被產生的單線筆舰本巾，選取適當的筆劃組 =線字，。再根據較大字型時的外框字型，修正所組成的單線 子接者根據字型的粗細資料，加寬單線字型，最後加入微調 、二、δ正加見後的單線字型，使所產生之單線字型更加完美。 Π!!的Γ法可以不用重新建立單線字型的資訊，直接由原本具 、匡筆韻訊’獲得單線筆劃資訊，以彌補傳統方法的不足。 20 201005553 . 以上所述僅為本發明之較佳實關，凡依本發明申嗜裒舰 圍所做之均等變化與修飾，皆應屬本發明之涵蓋範圍/ 【圖式簡單說明】 第1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意圖。 第3圖係為一海報體之單線字型的示意圖。 ❹ 帛4圖係為—由外框筆劃所構成之外框字型，，凸，，之實例圖。 第5圖則為第i圖之外框筆劃塗黑後所組成之外框字型，，凸”之實 例圖。 第6圖係為第4圖、第5圖中，，凸，，字之筆劃順序圖。 第7圖係為第4圖、第5圖中，，凸，，字之第4筆劃，，v冲之關鍵點之 位置及寬度值之資訊。 第8圖^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。 弟9圖係為，，介，，字中根據貝氏曲線所定義之控制點描述左下撤” 的示意圖。 第10圖係為根據不同的關鍵點位置及寬度值，所形成的不同的，， —之示意圖。 第11圖係為根據本發明之方法所繪之一流程圖。 第U圖係為在顯示字型256像素巧56像素時，所顯示的，，一，，字和，， 三，，字。 ' 第I3圖係為在顯示字型為32像素*32像素時，所顯示的，，一，，字和，， 21 201005553 一，， 二子0 第Μ圖係為一由一曲線之外框筆劃寬度的令點轉換為單線_之 示意圖。 一 第15圖係為-取一曲線之外框筆劃一邊的控制點轉換為單線筆割 之示意圖。 第16圖為係為—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。 第17圖係為-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。 第18圖係為自動分析誤差值並自誠最絲正的流程圖。 第19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 Π圖的，，娃，，字之示意圖。 第20圖一根據單線字型，，一 的寬度的示意圖。 ❹ 往右、下、以及右下方各長出一點 第21圖係為一原始灰階的單色曲線。 第22圖為根據第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”廇”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。 第圖Pd自動分析誤差值並自動做最佳修正程式所做的修正 後，所得的”廇”字的示意圖。 22 201005553 • 【主要元件符號說明】 A1,A2,A3,B1,B2，B3,B4,B5,B6,B7，B8，B9，B10,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Bl, R9 * Holes B3 is the upper edge of a frame stroke, and B4, B5, and B6 are the lower edges of the same frame stroke. B1, B2 form a curve segment Bl-B2 (such as X-A1, A3 in Fig. 8), and another control point Β7* obtained from the Bayes curve is abundant in the curve (as shown in Fig. 8). Α 2), but the curvature of the curve can be controlled by the control point Β7. Β2, Β3 form a curve segment Β2_Β3. Similarly, the other control point Β8 is not on the curve segment, but the curvature of the curve segment Β2-7 can be controlled by the control point Β8. Similarly, Β4 and Β5 form another curve segment β4_β5, which is another control point not on the curve segment, Β5, Β6 form another curve segment Β5-Β6, and Β10 is another control not on the curve segment. point. This method is to connect Β1 and Β6 to take the midpoint C1 ' Β2 and Β5 to take the midpoint q, Β3 and Β4 to take the midpoint C3, Β7 and Β1〇 to take the midpoint control point c4, do not connect with B9 to take the midpoint Control point C5, and then according to the control point C4, the curve segment 1 C2 'determines the curve segment C2-C3 according to the control point C5, that is, the desired single-line stroke. , 2. Take the control point on the side of the outer frame stroke into a single line stroke. Please refer to Figure 15. The first picture is a schematic diagram of the single-line strokes taken from the control points of the outer stroke of the curve. In Fig. 15, the control points B1 B2, B3, B4, B5, B6, B7, postal, B9, B10, etc. obtained from the Bayesian curve are included, which are the same as those in Fig. 14, and the rest of the structural relationship will not be described again. The single-line stroke selected in this embodiment is the lower edge B4_B5_B6, that is, the single-line stroke is obtained by the curve 5 and the output. This method is the control of the start and end of the specified - edge 14 201005553 points, and the subsequent point subtraction between the towels, that is, single = specify the _ edge of the outer frame stroke, to connect into the single line strokes The width of money is not really the middle-line of the outer frame stroke. In order to lower the county, it is recommended to set the material width to zero first.参 Note: The data format of the single-line font generated at this time is different from the data type pen of the outer frame type. It is the record number of the single line (about 10,000 or so) and the original frame. The example is ~~) different, and the parameter only has the first private point (starting point), and the other _point and width values do not exist. Since the parameter only has to take the 4 strokes of the sample group word, each word is cut according to its single line stroke number, for the position of the second = pen = the first = first _), the single line font generated by this method: : Figure 2, first ^ a. Position error: b. Length error: Because of the error in the position of the single line_sample obtained, the problem of disconnection or over-bulging on the pen (bull) is caused. Because of the error in the length of the single county ship obtained, the 201005553 is too short or too long. r please refer to (four) 17 嶋 - the eyes are fragrant (four) line strokes sample too short = the connection pen is not connected, the baby, the word diagram. In the circle at the figure, the pen is not covered by & However, sometimes this problem can be obtained after adding a single line font width. ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ ❹ =18 _Automatic analysis 嶋 and automatically make the best correction details Step excitation: analysis of the single-line font corresponding to the outer frame font in the font size (four) pixels * 256 pixels, the _ phase of each axis is simplified. Steps: When Na is a small pixel in the pixel city, the relationship between the strokes. Step just: Compare the relationship between the two strokes. If there is a difference, the single-line font is automatically corrected according to the outer font type corresponding to the line font. Automatically analyzing the error value and automatically making the best correction is automatically corrected by comparing the pixel pixels (when the age font is maximum) and the single line font touches the difference of the fiscal line. Because in the drum type, the material type = line type has no _, will be the most, because the original drum button spear early 16 201005553: In the big child type, there is no difference between the outer frame font and the single line font. , when the smaller subtype is reduced, for example, 64 pixels ^ '''... Going to the m "32 pixels * 32 pixels, etc. will not be a month... Figure 19 is the original 丨7fe|AA after the automatic analysis of the error value and automatically do the best correction step, left 俚 and automatically 乂Job _ 17th _,, baby,, the word's details have been corrected, and the pen is reconnected. _处即❹ ❹ 单 所 所 所 所 所 别 别 别 别 别 别 别 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单 单Appropriately increase the width of the single-line font, for example, a sub-type of 16 pixels% or less, the standard width is set to - point width, 16 pixels% pixels to 31 pixels * 31 pixels, the standard width is set to two points wide , 31 pixels * 31 pixels ~ sound 63 sub-type 'standard width is set to three points wide and so on. For a single-line font of monochrome, the method of turning 2 is to fix the length-width in a certain direction; if it is a grayscale center-lined font, it does not have any The single-line lining of the single-line font determines the gray-scale color of the long width. Please continue to refer to Figure 16. The middle of the 帛16 diagram is a monochrome single-line word with a width of Width =, and the lower part of Figure 16 is a schematic diagram of a gray-scale single-line font with a long width. Figure 20 is a schematic diagram of the width of a point that grows to the right, the bottom, and the bottom right according to the single line type. In Fig. 20, the upper left, and each point of the word, according to the In the upper right, lower, and lower right sides of the figure, a little bit grows to form a width, such as the lower left side of the 2G image. This is a single-line character of a single color. The method of width. Please continue to refer to Fig. 21 and Fig. 22. Fig. 21 is a monochromatic curve of the original gray scale. Fig. 22 is a plot of the original gray scale monochrome 17 201005553 according to Fig. 21. Schematic diagram of the gray-scale monochromatic curve after the width. The gray-scale monochrome font grows the width as shown in Figs. 21 and 22, according to the center of each grid point adjacent to the curve in Fig. 21. The distance between the point and the curve determines the grayscale color of the grid. The closer the distance is to the fine color of the depth, the farther the lin is to use the lighter grayscale color, and the shape has a width as shown in Fig. 22. Gray-scale single-line font curve. When comparing 256 pixels * 256 pixels, the single-line font corresponds to The difference between the box type and the ❹ ❹ single line font is 'into the red single line word, then the light is considered to be the font type = small, the other fonts of the new type are in the zooming two' will produce the font The problem of poor quality, the solution is relative to the font: the size of the 'plus A_ (hinting) (four) message, to fine-tune the size of the font, I: fine-tuning information is also automatically analyzed error value by the aforementioned And automatically do the best U is private and difficult to live. Please refer to Figure 23, the 23rd _ is n pixels * 14 pixels, there are 2 pens 2 solution is the automatic shooting caused by the paste The axis analysis program performs the analysis of the outer box, and judges the 箬: 连, and then continues to automatically analyze whether or not to add a 4 pixel % pixel in the 2 _, the information that should be moved up or down, or Force, directly to the pen _ divided information. The 23rd _ _ mouth of the pen strokes, should be moved down to the pen under the genus called the material is the self-generated. Please refer to the first amendment, will be 2 ^ do Schematic diagram of the most recorded program. After moving down, the resulting, shoulder, and word η 201005553 added to the tweak After the news, the resulting single-line word data format is: 1st word 1st stroke number + starting key point + fine-tuning information number + starting key point + fine-tuning information ❹ ❹ 2nd word Nth Word 1st stroke number + starting key point + fine tuning information 2nd stroke number + starting key point + fine tuning information S1 stroke number + starting key point + fine tuning information Lv ^^ stroke number + starting key Point + fine-tuning information 1st stroke number + starting key point + fine-tuning information 2nd stroke number + starting key point + fine-tuning information S3 pen stroke number + starting key point + fine-tuning information, fine-tuning information including the book When the grass is a subtle size, for example, 32 pixels * 32 pixels, the stroke will be displaced, deleted, and the information format of the information is as follows: Another - the information of the replacement. Fine-tuning fine-tuning processing type size and size 19 201005553 • When the above-mentioned "employed" word is at 14 pixels * 14 pixels, the pen strokes (the fourth stroke) below the 2 connected strokes should be moved down one space For example, its 扃 (serial number) fine-tuning processing number f stroke number = 4 for word edge size processing method processing value 1 14 pixels * 14 pixel displacement down 1 grid after S micro-negative The word bet is more beautiful and beautiful. Please note that the above embodiment is described by taking the JISX〇2〇8 word set as an example, but the present invention is not limited to (4) for the word set, as long as it conforms to the spirit of the present invention, other word sets such as Chinese Traditional (Blg5) Etc., are all within the scope of the present invention. ^ The above description of the invention is based on the method of the outer frame word dragon single line font, 〇 will be the outer frame font data, select - the appropriate font size, generate the appropriate single line pen sample, and then be generated For the single-line pen, select the appropriate stroke group = line word. According to the outer frame font type of the larger font, the single-line sub-connector is modified according to the thickness data of the font, and the single-line font is widened. Finally, the single-line font after the fine adjustment, the second and the δ are added is added. The resulting single line font is more perfect. Π!!'s method can re-establish the information of single-line fonts, and directly obtain the information of single-line strokes from the original and 韵 韵 , to make up for the shortcomings of traditional methods. 20 201005553 . The above is only the preferred embodiment of the present invention, and all the equivalent changes and modifications made by the invention according to the present invention should be within the scope of the present invention / [Simple Description] The figure is a schematic diagram of a single line font of a black body. Figure 2 is a schematic diagram of a single-line font of a circle. Figure 3 is a schematic diagram of a single line font of a poster body. ❹ 帛 4 is a diagram of the outer frame font, convex, and example. The fifth picture is an example of the outer frame type, and the convex form formed by the blackout of the outer frame of the i-th picture. The sixth picture is the fourth picture, the fifth picture, the convex, the word The stroke sequence diagram is shown in Figure 4, Figure 5, the convex, the fourth stroke of the word, the position of the key point of the v-rush and the information of the width value. Figure 8 is a curve segment (Starting point A1, ending point A3), according to the Bayesian curve, the schematic diagram of the curve segment drawn by the control point A2. The brother 9 is the control of the word, according to the Bayesian curve. The point describes the diagram of the left lower exit. Figure 10 is a schematic diagram of the different, and - formed according to different key point positions and width values. Figure 11 is a flow chart of a method in accordance with the present invention. The U picture is displayed when the display font is 256 pixels and 56 pixels, and the words one, the word, and the three words are displayed. 'The I3 picture is displayed when the display font is 32 pixels * 32 pixels, and the one, the word and the, 21 201005553 one, the second sub-zero picture is a stroke from the outside of a curve The width of the order point is converted to a single line _ schematic. A picture of Fig. 15 is a schematic diagram of converting a control point on one side of a stroke outside the curve to a single line stroke. Figure 16 is a diagram of a single-line font, a monochrome single-line font with a long width, and a gray-scale single-line font with a long width. Figure 17 is a schematic diagram of the "wa" word that is not connected because the sample of the single-line stroke obtained is too short. Figure 18 is a flow chart that automatically analyzes the error value and is sincere. Figure 19 is a schematic diagram of the original figure after the automatic analysis of the error value and automatically performing the best correction step. Figure 20 is a schematic diagram of the width of a single line font, according to one.长 A little bit grows to the right, down, and bottom right. Figure 21 is a monochromatic curve of the original grayscale. Figure 22 is a schematic diagram of a gray-scale monochromatic curve after growing a width according to the original gray-scale monochrome curve of Figure 21. Figure 23 is a "廇" character. When 14 pixels * 14 pixels, there are 2 horizontal strokes connected, which leads to a schematic diagram of the paste pen. Figure Pd automatically analyzes the error value and automatically corrects the corrections made by the best correction program. 22 201005553 • [Main component symbol description] A1, A2, A3, B1, B2, B3, B4, B5, B6, B7, B8, B9, B10, C1, C2, C3, C4, C5, A10, A11, A12, A13, A14, A15, A16 12, 14, 16, 18, 20, 22, 100, 102, 104 Control point steps
❹ 23❹ 23