JPS5880781A - Font discrimination method - Google Patents

Font discrimination method

Info

Publication number
JPS5880781A
JPS5880781A JP56179225A JP17922581A JPS5880781A JP S5880781 A JPS5880781 A JP S5880781A JP 56179225 A JP56179225 A JP 56179225A JP 17922581 A JP17922581 A JP 17922581A JP S5880781 A JPS5880781 A JP S5880781A
Authority
JP
Japan
Prior art keywords
pattern
italic
character pattern
font
target character
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP56179225A
Other languages
Japanese (ja)
Inventor
Koichi Ejiri
公一 江尻
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ricoh Co Ltd
Original Assignee
Ricoh Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ricoh Co Ltd filed Critical Ricoh Co Ltd
Priority to JP56179225A priority Critical patent/JPS5880781A/en
Publication of JPS5880781A publication Critical patent/JPS5880781A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10Character recognition
    • G06V30/24Character recognition characterised by the processing or recognition method
    • G06V30/242Division of the character sequences into groups prior to recognition; Selection of dictionaries

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Character Discrimination (AREA)

Abstract

PURPOSE:To ensure the accurate and quick discrimination of fonts for an optical character reader, by obtaining the value corresponding to the number of picture elements for each vertical column of a subject pattern and then discriminating whether the font is slant or not for the subject pattern in reference to the emerging frequency of the above-mentioned value. CONSTITUTION:The contour picture elements which are continuous toward the lower left side are noticed among the black picture elements. Then a picture element Cmn is defined by the number (m) of the vertical column and the number (n) of the row within the same vertical column. Thus the picture elements at the left contour of an italic coincide with the parts C11-C101 to which the hatching is applied; while the picture elements of the right side are shown by the parts C11-C81 respectively. On the other hand, the picture elements of the left and right contours of a non-italic type are shown by the parts C11-C71 and C11-C61 respectively. Here the product of the picture element number Nn of each numerical value (n) and the value (n) is set at Na at the left of the contour and at Nb at the right of the contour respectively in terms of the picture element of n>=2. Thus N=Na+Nb is satisfied. As a result, the italic N shows a value much larger than that of the non-italic N.

Description

【発明の詳細な説明】 (1)発明の分野 本発明は光学文字読取(OCR)に関し、特に対抜文字
パターンのフォントが斜体であるか否かを認識するため
の特徴抽出をおこなうフオレト識別方法に関する。
DETAILED DESCRIPTION OF THE INVENTION (1) Field of the Invention The present invention relates to optical character reading (OCR), and in particular to a font identification method for extracting features for recognizing whether or not the font of a diagonal character pattern is italic. Regarding.

(2)発明の背景 従来より光学文字読取においては文字の認識処理に長い
時間を必要とし、捷だ対象となる文字が傾いているよう
な場合には、隣接する文字との区別(切出し)をおこな
うのが細しい。後者の識別をおこなうための方法として
は特開昭55−41512号に開示された「文字切出し
方式」が知られている。その方法は、簡単に説明すれは
、ある文字パターンの切出しをするためのパターン投影
方向を、文字パターン列の方向に垂直な方向から回転さ
せて、対象文字パターンと隣接する文字パターンとの投
影パターンの切れ目を捜して対象文字パターンを切出す
ようにしたものである。この方法と同様にして、対象文
字パターンの投影方向を回転させて投影パターン長が短
くなる方向を捜すことにより、文字パターンの傾きを判
定し文字パターンのフォントがイタリック(斜体)であ
るか否かを判別できる。す々わち第1図に示すように斜
体で々いフォントの投影長lは投影方向がパターン列方
向に垂直の場合に最短ωm1n)となり、フォントが斜
体の場合には投影方向をある角度に傾けたときに最短と
なるので、投影方向を少しずつ変化させてその毎に投影
長沼を求め、−6−gmlnとなる投影方向からフォン
トを判別できる。しかしながらこのような方法は、投影
方向を変化させて投影長沼を求める、という処理を繰り
返し行なう必要があり、処理時間が長くかかつてしまう
(2) Background of the Invention Traditionally, optical character reading requires a long time for character recognition processing, and when the characters to be shuffled are tilted, it is difficult to distinguish them from adjacent characters (cut out). It's difficult to do. As a method for performing the latter identification, a ``character extraction method'' disclosed in Japanese Patent Laid-open No. 55-41512 is known. The method is simply to rotate the pattern projection direction for cutting out a certain character pattern from the direction perpendicular to the direction of the character pattern row, and to create a projection pattern between the target character pattern and the adjacent character pattern. The target character pattern is extracted by searching for breaks in the character pattern. Similar to this method, by rotating the projection direction of the target character pattern and searching for the direction in which the projected pattern length becomes shorter, the inclination of the character pattern is determined and whether the font of the character pattern is italic or not. can be determined. That is, as shown in Figure 1, the projection length l of a large italic font is the shortest when the projection direction is perpendicular to the pattern row direction, and when the font is italic, the projection length l is at a certain angle. Since it is the shortest when tilted, the projection direction is changed little by little and the projection Naganuma is obtained each time, and the font can be determined from the projection direction that is -6-gmln. However, in this method, it is necessary to repeatedly perform the process of changing the projection direction and obtaining the projected Naganuma, resulting in a long processing time.

(3)発明の目的 文字のフォントが斜体か否かを短時間で判別し、斜体文
字の認識処理を高速化すること。
(3) Purpose of the invention To quickly determine whether the font of a character is italic or not, and to speed up recognition processing of italic characters.

(4)発明の要約 上記目的を達成するため本発明においては、対象文字パ
ターンの輪郭画素のうち左下!ll(又は右上9)に連
続する画素の情報について、対象パターンの縦列の各列
毎の画素数に対応する値を求め、それらの値の出現頻度
を参照して対象パターンのフォントが斜体か否かを判別
する。これによれば繰り返し処理を行々わずに済むので
短時間でフォントを判別しうる。
(4) Summary of the Invention In order to achieve the above object, the present invention uses the lower left of the outline pixels of the target character pattern! Regarding the information of pixels consecutive to ll (or upper right 9), find the value corresponding to the number of pixels in each vertical column of the target pattern, and check whether the font of the target pattern is italic by referring to the frequency of occurrence of these values. Determine whether According to this, it is not necessary to perform repeated processing, so that the font can be determined in a short time.

第2a図および第2b図はイタリック体の文字(アルフ
ァベット)Cを、第2c図および第2d図はイタリック
体でない文字Cをそねそれ画素単位(図面中の破線で分
割された大きさ)で黒・白の情報により示した図であり
、そねそれ実線(輪郭の線)の内側が黒を示す。第2a
図〜第2d図に示される黒画素のうち輪郭画素でありし
かも左下りに連続するものに着目し、縦列の番号mと同
一縦列内の行の番号nで画素Cmnを定義すると、イタ
リック体の輪郭左側の画素は第2a図のハツチングを施
した部分のC1l〜C61で表わされる。
Figures 2a and 2b show the italic letter (alphabet) C, and Figures 2c and 2d show the non-italic letter C in pixel units (sizes divided by dashed lines in the drawings). This is a diagram showing black and white information, and the inside of the solid line (outline line) indicates black. 2nd a
Among the black pixels shown in Figures to Figures 2d, we focus on those that are contour pixels and are continuous in the downward left direction, and define the pixel Cmn by the column number m and the row number n in the same column. Pixels on the left side of the outline are represented by C11 to C61 in the hatched area in FIG. 2a.

同様にイタリック体の輪郭右側の画素は第2b図のC1
1−Cstで表わされ、イタリック体でない文字の輪郭
左側および輪郭右側の画素はそれぞれ第20図のC1z
 −C7+および第2d図のC1l〜C61で表わされ
る。ここで第2a図と第2b図のハラチングを施した画
素Cmnのうちnが°2以上の画素について数値n毎の
画素数Nnを数えてNnと数値nの積Na (第2a図
)およびNb (第2b図)とし、N = Na 十N
bとすると第1表が得られ、同様に第2c図と第2d図
のハツチングを施した画素について第2表が得られる。
Similarly, the pixel on the right side of the outline in italics is C1 in Figure 2b.
1-Cst, and the pixels on the left side and right side of the outline of a non-italic character are respectively C1z in Figure 20.
−C7+ and C11 to C61 in FIG. 2d. Here, among the pixels Cmn subjected to halaching in Figures 2a and 2b, count the number of pixels Nn for each numerical value n for pixels where n is 2 or more, and calculate the product Na of Nn and numerical value n (Figure 2a) and Nb (Figure 2b), N = Na 1 N
b, Table 1 is obtained, and similarly Table 2 is obtained for the hatched pixels in FIGS. 2c and 2d.

第  1  表 □※n=3.4のみのNトータノ岬N−13第  2 
 表 ※n=3.4のみのNトータルTN=6第1表および第
2表から、このようにして得らわるNの値は2≦1]≦
4においてはイタリック体とそうでないものとで大きく
異なることが明らかであり、この数値Nをフォント識別
のための特徴データとしうる。
1st Table □*n=3.4 only N Totano Misaki N-13 2nd
Table *N for n=3.4 only Total TN=6 From Tables 1 and 2, the value of N obtained in this way is 2≦1]≦
4, it is clear that there is a large difference between italic fonts and non-italic fonts, and this numerical value N can be used as feature data for font identification.

(5)  発明の実施例 以下、図面を参照して本発明の詳細な説明する。(5) Examples of the invention Hereinafter, the present invention will be described in detail with reference to the drawings.

第3a図に示すパターンを認識する場合について説明す
ると、捷ず第3b図に示すように画素4つに対応する大
きさの正方形の窓Wを想定し、この窓Wをパターンの右
−I一方から走査して窓W内の各部Wl、 W2. W
3およびW4に現わ才する画素の黒・白の情報を読取っ
て処理をおこなう。パターンの輪郭左側の特徴を抽出す
る場合、第3C図に示すようにWl、 W2. W3お
よびW4がそJlそね白、白、白および黒となる寸で窓
Wを走査する。すると寸ず第3d図の位置P】が検知さ
れる。ここでカウントする(nとする)画素は窓WのW
4部分に現われる画素(黒)であり、Plの位置では第
3a図に示すC1lである。Plの位置からCI+の属
する列について窓Wを下に移動させ々から所定画素のカ
ウントをすることに々るが、輪郭左側の画素を切出すた
め次の2つの条件を与える。その1つは窓WのW3が黒
でないこと、もう1つはW3とW4の両者が白でないこ
とである。窓Wを下げて2つの条件が満たされていれば
nをインクリメント(+1)してn≧2の場合に第1表
に示したようなテーブルのそのnの欄のNnをインクリ
メントシ、いずれかまたは両方の条件が渦ださねでいな
い場合にはその列のカウントを終了しnを1にリセット
し、画像の境界に沿って再び窓を走査させる。第3d図
の位置PIから窓Wを一画素分だけ下げると、黒の画素
Celが窓WのW3に現われ1つの条件が満たされなく
なるので窓Wを走査させて次の位置P2を検知する。位
置P2から窓を一画素分だけ下げた場合、W3に現われ
る画素Ce2は白々ので条件が満たされ、nを1から2
にカウントアツプしてテーブルのn二2のNnをインク
リメントする。同様にして窓Wを走査すると第3a図に
示す画素C+ 1% C71がカウントさね、結果的に
前記第1表の輪郭左側の欄と同一の特徴データが得ら牙
する。パターン輪郭右側の特徴を抽出する場合、第4a
図に示すようにWl。
To explain the case of recognizing the pattern shown in Fig. 3a, assume a square window W of a size corresponding to four pixels as shown in Fig. 3b, and place this window W on the right side of the pattern. Each part Wl, W2 within the window W is scanned from . W
The black/white information of pixels appearing in pixels 3 and W4 is read and processed. When extracting the features on the left side of the pattern outline, as shown in FIG. 3C, Wl, W2. The window W is scanned so that W3 and W4 become white, white, white, and black. Then, the position P shown in Fig. 3d is immediately detected. The pixels to be counted (referred to as n) here are W of window W.
It is a pixel (black) that appears in the fourth part, and at the position of Pl, it is C1l shown in FIG. 3a. The window W is moved downward for the column to which CI+ belongs from the position of Pl and a predetermined number of pixels are counted.The following two conditions are given in order to cut out the pixels on the left side of the outline. One is that W3 of window W is not black, and the other is that both W3 and W4 are not white. If the window W is lowered and two conditions are met, n is incremented (+1), and if n≧2, Nn in the n column of the table shown in Table 1 is incremented. Or, if both conditions are not swirly, stop counting for that column, reset n to 1, and scan the window again along the image border. When the window W is lowered by one pixel from the position PI in FIG. 3d, a black pixel Cel appears in the window W3 and one condition is no longer satisfied, so the window W is scanned to detect the next position P2. When the window is lowered by one pixel from position P2, the pixel Ce2 that appears in W3 is white, so the condition is met, and n is changed from 1 to 2.
Count up and increment Nn of n22 in the table. When the window W is scanned in the same manner, the pixel C+1% C71 shown in FIG. 3a is counted, and as a result, the same characteristic data as in the column to the left of the outline in Table 1 is obtained. When extracting the features on the right side of the pattern contour, the fourth a
Wl as shown in the figure.

W2. W3およびw4がそれぞれ黒、黒、黒および白
となる位置に窓Wを移動させ、窓WのW3に対応する画
素をカウントする。この場合に輪郭右側の画素を切出す
ため、W4が黒でないことおよびw3とw4の両者が白
でないことの2つの条件を与える。このようにして前記
輪郭左側の特徴を抽出する場合と同様にして処理を行な
うと第4b図に示す画素011〜C61がカウントされ
、結果的に第1表の輪郭右側の欄と同一の特徴データが
得らJする。第5a図および第5b図はそれぞれ上記の
輪郭左側および輪郭右側の特徴抽出処理フローを示すフ
ローチャートである。
W2. The window W is moved to a position where W3 and w4 are black, black, black, and white, respectively, and the pixels of the window W corresponding to W3 are counted. In this case, in order to cut out the pixels on the right side of the outline, two conditions are given: W4 is not black and both w3 and w4 are not white. When processing is performed in the same manner as when extracting the features on the left side of the contour, pixels 011 to C61 shown in FIG. 4b are counted, and as a result, the same feature data as in the column on the right side of the contour in Table 1 is obtained. FIGS. 5a and 5b are flowcharts showing the feature extraction processing flow for the left side of the contour and the right side of the contour, respectively.

このようにして得らねた輪郭左側および輪郭右側の欄の
Nnのそれぞわと数値nの積を求めてNaおよびNl)
とし、そわらの和をNとする(第1表参照)。
Find the product of Nn in the column on the left side of the contour and on the right side of the contour obtained in this way and the numerical value n (Na and Nl)
Let the sum of the numbers be N (see Table 1).

次に、特徴データNを使用してフォントの識別を行なう
が、ここではn = 3と4のみのNを加算したトータ
ルの値TNを参照値と比較する。第1表および第2表に
よねばイタリック体のTNは13、そうで々いもののT
Nは6なので参照値TR,をたとえば10に定めて、 
TN≧Tnの場合に対象パターンを斜体文字の候補とし
て、レジスタR1をインクリメントする。ここで、斜体
文字の候補となったパターンをそのt−を斜体文字とし
て判定してもよいが、認識率を高めるため更に次の処理
を行なう。
Next, the characteristic data N is used to identify the font, but here, the total value TN obtained by adding only N where n = 3 and 4 is compared with a reference value. According to Tables 1 and 2, the italicized TN is 13, and the big T
Since N is 6, set the reference value TR to 10, for example,
If TN≧Tn, the target pattern is set as an italic character candidate, and register R1 is incremented. Here, the t- of the pattern that is a candidate for an italic character may be determined as an italic character, but in order to increase the recognition rate, the following processing is further performed.

1ず、対象パターンの上部および下部の重心となるX座
標(横方向) XIおよびX2を求める。これらの処理
はパターン上部3行(3画素)の黒欄およびパターン下
部3行の黒欄を2分する画素単位のX座標を算出するも
のである。第6a図および第6b図はそねそれ斜体文字
のCおよび非斜体文字のCについて座標Xi、 X2を
示したものであり、これらの図を参照すると座標間の差
Xi −X2は斜体文字の場合に2(画素)、非斜体文
字の場合に0である。そこで、次にXl−X2を参照値
2と比較してXl−X2≧2であればレジスタR1をイ
ンクリメントする。更に、対象文字パターンの属する単
語内において対象文字パターンと隣接する文字パタ−ン
のフォントが斜体候補か否かを調べ、斜体であればレジ
スタR1をインクリメントする。以上の処理を行なった
後レジスタR・1の値を調べ、R+1≧2すなわちTN
≧TR,Xl−X2≧2および隣接文字が斜体候補、の
うち少なくとも2つの条件が満たされていれば対象文字
パターンを斜体文字と判定しそうでなけわば非斜体文字
と判定する。第7図はこれら一連の処理フローの概略を
示すフローチャートである。
1. First, determine the X coordinates (horizontal direction) XI and X2 that are the centers of gravity of the upper and lower parts of the target pattern. These processes calculate the X coordinate in pixel units that divides the black columns of the upper three rows (three pixels) of the pattern and the black columns of the lower three rows of the pattern into two. Figures 6a and 6b show the coordinates Xi and X2 for the italic character C and the non-italic character C, respectively, and with reference to these figures, the difference between the coordinates Xi -X2 is calculated by the italic character C. 2 (pixels) for non-italic characters, and 0 for non-italic characters. Therefore, next, Xl-X2 is compared with reference value 2, and if Xl-X2≧2, register R1 is incremented. Furthermore, it is checked whether or not the font of the character pattern adjacent to the target character pattern in the word to which the target character pattern belongs is an italic candidate, and if it is italic, register R1 is incremented. After performing the above processing, check the value of register R・1 and check that R+1≧2, that is, TN
If at least two of the conditions ≧TR, Xl-X2≧2 and the adjacent character is an italic candidate are satisfied, the target character pattern is determined to be an italic character, otherwise it is determined to be a non-italic character. FIG. 7 is a flowchart showing an outline of a series of these processing flows.

なお、以上の実施例においてはn、TN等を特定の値に
定めて説明したが、とわらの値は対象文字パターン読取
画素構成、パターンの種類等に応じて任意に変更される
。捷だ、パターン走査は右上シとしてもよい。
In the above embodiments, n, TN, etc. have been set to specific values, but the values can be changed arbitrarily depending on the target character pattern reading pixel configuration, the type of pattern, etc. The pattern scan can be done at the top right corner.

(6)発明の効果 以上のとおり本発明によれば、対象パターンの輪郭の情
報から短時間で正確にフォントの斜体・非斜体を識別で
きる。
(6) Effects of the Invention As described above, according to the present invention, it is possible to accurately identify whether a font is italic or non-italic based on information on the outline of the target pattern in a short time.

影方向と投影長の関係を示す平面図、第2a図。FIG. 2A is a plan view showing the relationship between the shadow direction and the projection length.

第2b図、第3a図、第3d図、第4b図および第6a
図は斜体文字のアルファベットCをある画素区分で読取
った情報のパターンを二次元平面上に表わした説明図、
第2C図、第2d図および第6b図は非斜体文字のアル
ファベットCを斜体文字の場合と同等の画素区分で読取
った情報のパターンを二次元平面上に表わした説明図、
第3b図。
Figures 2b, 3a, 3d, 4b and 6a
The figure is an explanatory diagram showing on a two-dimensional plane the pattern of information obtained by reading the italicized alphabet C in a certain pixel division,
Figures 2C, 2d, and 6b are explanatory diagrams showing, on a two-dimensional plane, patterns of information obtained by reading the non-italicized alphabet C in pixel divisions equivalent to those for italicized characters;
Figure 3b.

第3C図および第4a図は実施例において想定する窓W
の概念を示す説明図、第5a図および第5b図はそれぞ
れパターン輪郭の左側および右側から特徴を抽出する処
理フローを示すフローチャート、第7図は1つの実施例
におけるフォント判定の概略の処理フローを示すフロー
チャートである。
3C and 4a are windows W assumed in the embodiment.
FIGS. 5a and 5b are flowcharts showing the processing flow for extracting features from the left and right sides of the pattern contour, respectively. FIG. 7 is a schematic processing flow for font determination in one embodiment. FIG.

W:窓        cel:画素(黒)ce2:画
素(白)Xl:パターン上部重心位置のX座標X2:パ
ターン下部重心位置のX座標 味1図 第2C図     兜2d図 第5a図 崩5b図
W: window cel: pixel (black) ce2: pixel (white) Xl: X coordinate of upper center of gravity position of pattern X2: X coordinate of lower center of gravity position of pattern

Claims (5)

【特許請求の範囲】[Claims] (1)  対象文字パターンの情報を光学的に読取り、
その情報から文字を判別する光学文字認識において、対
象文字パターンの輪郭画素であって左下り又は右上りに
連続する画素に、対象文字パターンの縦列方向で連続す
る輪郭画素の各縦列において対応位置にあるものの数を
求め、そわらの数を対象パターンのフォントが斜体であ
るか否かを判別するための特徴情報とする、フォント識
別方法。
(1) Optically read information on the target character pattern,
In optical character recognition, which distinguishes characters from that information, the contour pixels of the target character pattern that are continuous in the lower left or upper right direction are matched to the corresponding positions in each vertical column of continuous contour pixels in the vertical direction of the target character pattern. A font identification method that calculates the number of certain objects and uses the number of curls as characteristic information for determining whether the font of the target pattern is italic.
(2)対象文字パターンの上部の任意数の画素情報より
パターン上部の横方向重心座標X1を求め、パターン下
部の任意数の画素情報よりパターン下部の横方向重心座
標x2を求め、座標X1と座標X2の差の情報を対象パ
ターンのフォントが斜体であるか否かを判別するための
もう1つの特徴情報とする、前記特許請求の範囲第(1
)項記載のフォント識別方法。
(2) Find the horizontal center of gravity coordinate X1 of the upper part of the pattern from an arbitrary number of pixel information at the top of the target character pattern, find the lateral center of gravity coordinate X2 of the lower part of the pattern from the arbitrary number of pixel information at the bottom of the pattern, and calculate the coordinate X1 and the coordinate Claim 1, wherein the information on the difference in X2 is another feature information for determining whether the font of the target pattern is italic.
font identification method described in ).
(3)  対象文字パターンが属する単語内で対象文字
パターンに隣接する文字パターンの特徴情報を対象文字
パターンの7オント識別のために参照する前記特許請求
の範囲第(1)項又は第(2)項記載のフォント識別方
法。
(3) Claims (1) or (2) refer to feature information of a character pattern adjacent to the target character pattern within a word to which the target character pattern belongs for 7-ont identification of the target character pattern. Font identification method described in section.
(4)対象文字パターンの輪郭画素であって左下り又は
右上りに連続する画素に、対象文字パターンの縦列方向
で連続する輪郭画素の各縦列において、輪郭画素に所定
順に番号nを定め、nのうち特定の番号のものにつき、
各縦列で同じ番号である画素の数Nnに該画素の番号を
乗じた値Nを特徴情報とする前記特許請求の範囲第(1
)項記載のフォント識別方法。
(4) For the contour pixels of the target character pattern that are continuous in the lower left or upper right direction, in each column of contour pixels that are continuous in the vertical direction of the target character pattern, a number n is assigned to the contour pixels in a predetermined order, and n For those with specific numbers,
The feature information is a value N obtained by multiplying the number Nn of pixels having the same number in each column by the number of the pixel.
font identification method described in ).
(5)特定の番号を複数個とし、それぞれの番号の画素
について求めたNの値を加算し、和を特徴情報とする前
記特許請求の範囲第(4)項記載のフォント識別方法。
(5) The font identification method according to claim (4), wherein there are a plurality of specific numbers, the values of N obtained for the pixels of each number are added, and the sum is used as feature information.
JP56179225A 1981-11-09 1981-11-09 Font discrimination method Pending JPS5880781A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP56179225A JPS5880781A (en) 1981-11-09 1981-11-09 Font discrimination method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP56179225A JPS5880781A (en) 1981-11-09 1981-11-09 Font discrimination method

Publications (1)

Publication Number Publication Date
JPS5880781A true JPS5880781A (en) 1983-05-14

Family

ID=16062118

Family Applications (1)

Application Number Title Priority Date Filing Date
JP56179225A Pending JPS5880781A (en) 1981-11-09 1981-11-09 Font discrimination method

Country Status (1)

Country Link
JP (1) JPS5880781A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0134997A2 (en) * 1983-07-15 1985-03-27 Siemens Aktiengesellschaft Method for automatic type-font recognition without prior knowledge of the text contents, utilizing a representation of the text document by desriptors
US10706337B2 (en) 2017-02-27 2020-07-07 Kyocera Document Solutions Inc. Character recognition device, character recognition method, and recording medium

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0134997A2 (en) * 1983-07-15 1985-03-27 Siemens Aktiengesellschaft Method for automatic type-font recognition without prior knowledge of the text contents, utilizing a representation of the text document by desriptors
US10706337B2 (en) 2017-02-27 2020-07-07 Kyocera Document Solutions Inc. Character recognition device, character recognition method, and recording medium

Similar Documents

Publication Publication Date Title
US4903312A (en) Character recognition with variable subdivisions of a character region
KR910007751B1 (en) Character recognition method
KR900007009B1 (en) Character recognition device
US20120219220A1 (en) Method and system for preprocessing an image for optical character recognition
JPH0420226B2 (en)
CN106875546A (en) A kind of recognition methods of VAT invoice
JPH02165392A (en) Universal character section for multifont
JPH05242292A (en) Separating method
DE69813416T2 (en) Devices for determining the line direction, image angle detection and image angle correction.
JPH0256707B2 (en)
US4596038A (en) Method and apparatus for character recognition
JP2926066B2 (en) Table recognition device
JPS5880781A (en) Font discrimination method
CN116824608A (en) Answer sheet layout analysis method based on target detection technology
CN113408532A (en) Medicine label number identification method based on multi-feature extraction
JP2832928B2 (en) Character recognition method
JPH06187489A (en) Character recognizing device
JP2917427B2 (en) Drawing reader
JP3009237B2 (en) Feature extraction method
JPH0252313B2 (en)
JPH03126188A (en) Character recognizing device
JPS63131287A (en) Character recognition system
JPH06139407A (en) Character segmenting method
KR20190087092A (en) Number recognition method and apparatus based on hierarchical analysis of edge property
JPH0262682A (en) Method for recognizing character