JPS6073795A - Character recognition device - Google Patents

Abstract

PURPOSE:To extract picture features correctly and improve the recognition level by coding the position of the connecting points of the picture segment as the detection division and by extracting features from the said coded position information and the information of the direction and length of the division coded segment. CONSTITUTION:Both X Y co-ordinates of point A1 are detected as segment positions. Next, X position is plurally divided by comparing X and Y maximum and minimum values sent from X Y maximum and minimum detection circuit with the detected value by the X position division circuit. For example, three divisions, upper, middle and lower are installed, and upper, middle and lower divisions are coded and outputted as ''0'', ''1'', and ''2'' respectively. In the same manner, Y position is divided by the Y position division circuit. For example, X position division of point A1 becomes X=2 and Y=0. In the figure 6, point B1 becomes X=2, Y=0, point B2 becomes X=2, Y=0, point B3 becomes X=0, Y=1 and point B4 becomes X=0, Y=1. Two figures ''0'' and ''6'' can be more correctly identified.

Description

[Detailed Description of the Invention] [1] Technical Field The present invention provides an online character reading and recognition device that writes characters, symbols, etc. (hereinafter referred to as characters, etc.) on a predetermined tablet using a pen, and reads the written characters. It is related to.

[2] Prior Art When reading characters written with a pen on a predetermined tablet online, information obtained from a pattern input device for handwritten figures is input using a general-purpose computer connected to the human-powered device. , needed to be recognized by its software.

For this reason, it is necessary to rely on complex character recognition steps, and a long time elapses from input to obtaining recognition results, which is often not suitable for the actual situation of online reading.

However, according to the specification of Japanese Patent Application No. 57-185839, which discloses a device for reading characters written with a pen on a tablet in real time, it is said that 1. A so-called stroke order recognition method, which extracts stroke features for each character based on its order of occurrence and recognizes the characters, has been used, and a major step has been taken to solve the above-mentioned problems.

This method is extremely useful and will be briefly described below as an indication of the current state of the art.

FIG. 1 is a diagram showing classification for identifying the direction of line segments of characters, etc., and the pen movement angle θ is divided into eight types. Second
The figure shows the number "4" which is the normalized character size. The extraction of both the features of the first stroke and the second stroke (stroke order) is shown in Figure 3 (
, ) and Φ). In this example, the direction of the line segment is code-encoded, and its length is also determined by the standard dimension C determined from the normalized character size (for example, -side IQII
Divide Li) into 4 equal parts, and if it is 2.5 questions or less, 0'', 2.5 ~
0.5 starvation is °゛1'', 0.5~0.75M is “2”
, Q, 75am or more is given a code code of 8''.

In this way, the first drawing segment is the direction end, so θ-5
And since the length = 3 and the next I/te is the direction -, θ = 0,
It was No. 8, so I decided that the image characteristics at that time would be C04. The second image is θ-6, No. 2, and the image feature is COI. It may be stipulated in advance that when the image features determined in this manner occur in the order of CO4.-COI, the character will be recognized as "4.". FIG. 4 illustrates a character recognition procedure using this method, and FIG. 5 is a diagram showing the configuration of the apparatus.

[4] It is an object of the invention to provide an improved character recognition device that recognizes characters, symbols, etc. written with a pen on a tablet in real time from the characteristics of line segments based on the stroke order without going through the object order of the invention. The purpose is

(5) Structure of the invention The character recognition device of the invention of the present application is disclosed in Japanese Patent Application No. 57-185839.
In addition to the structure of the character reading device based on the conventional technology disclosed in the above-mentioned issue, by adding a means for dividing the position of a line segment into a plurality of parts, image characteristics of characters etc. are extracted more accurately and the recognition level is improved. The configuration is explained below.

As mentioned above, FIG. 5 shows the configuration of a real-time character reading device that precedes the invention of the present application, and in the present invention, this configuration includes an X position detection circuit 33, an X position division circuit 34, a Y position detection circuit 85.
and Y position classification circuit 86 are added, and 0'' and °°6'' belonging to the same θ and 0 code classification as illustrated in FIG.
Even in this case, the differences between the two features can be clearly distinguished, extracted, and identified.

[6] Example An example of the present invention is shown in FIG. 6(a) and Φ).
A case will be explained taking as an example a case where 01+ and 0.01+ and 0.01+ are to be distinguished.

Figure 6(a) shows the number 110 IT, and ) in the same figure shows the number ``6''.The direction θ and length of the line segment are each encoded as explained in the prior art section. In addition to detecting the X and Y coordinates of the line segment intersection point A and -A4, - □In order to simplify the above processing, position divisions are provided for the X and Y coordinates.X position detection in the configuration shown in Fig. 5 Circuit 83, Y
The position detection circuit 35 corresponds to the X-coordinate position and Y-coordinate position of the line segment, respectively.

Both the X and Y coordinates of the point of Al are detected as the position of the line segment (g) in FIG. 7(a). Next, the X position division circuit 34 detects the X, Y
The maximum and minimum values of are compared with the above-mentioned detected value to divide the X position into a plurality of positions. For example, upper, middle, and lower 03 divisions are provided, and the top is encoded as 0, the middle is encoded as 1, and the bottom is encoded as 2 and output.

5- Similarly, the Y position is classified by the Y position classification circuit 36.

For example, the X position division of point AI in FIG. 7(a) is mark 2 (1
), y=o(c). Number n 6 in Figure 6(b)
In IT, point B1 is x=2.7=o, point B2 is mar2.7=0, point B8 is X-ON V = 'N, and point B4 is x=0, 'i=1. become two numbers 110
IT and "6" can be more accurately identified.

Next, an example of the configuration of the apparatus of the present invention is shown in FIG. Referring to the figure, a tablet 1, paper 2, stylus pen 8, coordinate reading circuit 4, character delimiter detection circuit 5, character coordinate memory 6,
XY maximum/minimum detection circuit 7, character size calculation circuit 8, normalization circuit 9, Z coordinate register (t) lo, Z coordinate register (t-t) 11, Z coordinate differentiation circuit 12, X coordinate register (t) 13, X coordinate register (t-1) 14,
X coordinate difference circuit 15, Y coordinate register (t)le, Y coordinate register (t-'t) 17, Y coordinate difference circuit 18,
Line segment direction detection circuit 19, line segment direction register 20, line segment length detection circuit 21, comparison circuit 22, distance addition circuit 23, direction differentiation circuit 24, relative position detection circuit 25, line segment length division circuit 2
6 is a conventionally used circuit 6-. The X position detection circuit 33, the X position classification circuit 34, the Y position detection circuit 35, and the Y position classification circuit 36 are circuits of the present invention, and are added to more accurately extract image characteristics such as characters written on the tablet. It is what was done.

[8] Effects When the character recognition device of the present invention is used, image features are accurately classified according to stroke order without unnecessarily increasing the number of divisions in terms of direction and length of line segments, and highly accurate character recognition can be performed. Furthermore, since minute changes in hand-written characters due to tentacles are removed as noise and only essential feature information is transmitted to the extraction device, the extraction circuit can be simplified and recognition results can be obtained quickly. .

[Brief explanation of the drawing]

FIG. 1 is a diagram illustrating directional divisions of line segments. FIG. 2 is a diagram showing an example of normalized written characters. FIG. 3 is a diagram showing the procedure for extracting both features of the previously normalized character from the stroke order. FIG. 5 is a diagram showing a state transition of image feature extraction in FIG. 4; FIG. 5 is a diagram showing the configuration of a conventional device. FIG. 6 is a diagram showing an example of characters identified by the device t= of the present invention. FIG. 7 is a diagram showing how the positions of line segment intersection points are classified. FIG. 8 is a diagram showing the configuration of the apparatus of the present invention. ■Tablet 2, paper 3, stylus pen coordinate reading circuit 5, character delimiter detection circuit 6, character coordinate memory? , X, Y maximum/minimum detection circuit 8, character size calculation circuit 9, normalization circuit IOl Z coordinate register (1) 11, 2 coordinate register (1-1) 12, 2 coordinate differentiation circuit 13, X coordinate register (1) 14, X coordinate register (1-1) 15, X coordinate difference circuit 16, y coordinate register (1) 17, Y coordinate register (1-1) 18, Y coordinate difference circuit 19, line segment direction detection circuit 20, line segment direction register 21, line segment length detection circuit 22, comparison circuit 23, distance addition circuit 24, direction differentiation circuit 26, line segment length division circuit 27, image feature storage circuit 28, register 29, AND gate 30 , read-only storage device 31, register 32, AND gate 33, x position detection circuit 34, X position division circuit 35, Y position detection circuit. 1) (-6) Figure θ=4 ノ=2 0=O J=2

Claims (1)

[Claims] (1) Normalize the characters, symbols, etc. written on the tablet, detect the direction and length of each stroke segment according to the stroke order, perform segmental coding to extract the image features, and identify the characters from the order of occurrence. The recognition device is characterized in that the positions of intersection points of image line segments are detected and piecewise encoded, and image features are extracted from the encoded position information and information regarding the direction and length of the piecewise encoded line segments. character recognition device.