JPS6350746B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a figure recognition device, particularly a figure recognition device that recognizes a two-dimensional figure by extracting the movement of a brush tip as a sequence of two-dimensional coordinate points.
The present invention relates to a figure recognition device that performs recognition by extracting the number of vector line segments projected on the axis and the y-axis and the cumulative number for each vector direction.

2. Description of the Related Art Graphic recognition devices have been known that perform graphic recognition while tracking the movement of a brush tip. However, in conventional devices of this kind, it is difficult to check the inclination of a line segment of a figure that is being described, for example, with respect to the x-axis.
The trend is toward approximation using a plurality of polygonal lines or circular arcs, or expressing the line segments using functions, which tend to result in complicated and overly detailed processing.

SUMMARY OF THE INVENTION An object of the present invention is to improve the above-mentioned points and provide a figure recognition device that roughly categorizes figures using a simple method. Therefore, the figure recognition device of the present invention is a figure recognition device that recognizes a two-dimensional figure represented by a two-dimensional coordinate point sequence, based on a two-dimensional coordinate point sequence inputted in chronological order. an x-axis component feature extraction unit that extracts the number of projected vector line segments projected onto the x-axis and the cumulative number for each vector direction based on the x-coordinate point sequence in the two-dimensional coordinate point sequence; a y-axis component feature extraction unit that extracts the number of projected vector line segments projected onto the y-axis and the cumulative number for each vector direction based on the y-coordinate point sequence in the dimensional coordinate point sequence; The present invention is characterized in that the feature quantities extracted by both of the extracting sections are compared with the standard feature quantities read out from the standard category dictionary. This will be explained below with reference to the drawings.

FIG. 1 is an explanatory diagram illustrating feature amounts extracted by the present invention, FIG. 2 is an example of feature amounts extracted corresponding to multiple characters, and FIG. 3 is an example of the present invention. .

In the present invention, as shown in FIG.
is being drawn in the direction of the arrow, the indicated points 2-0, 2-1, 2 in Fig. 1B correspond to the moving speed of the brush tip.
The coordinates of −2, . . . are sequentially extracted. The vector line segment that projects this movement of the brush tip onto the x-axis is shown in Figure 1C.
The illustrated vector line segments are 3-1, 3-2, and 3-3. On the other hand, vector line segments projected onto the y-axis become vector line segments 4-1, 4-2, and 4-3 shown in FIG. 1C.

In the case of the present invention, (i) the number PX of vector line segments 3-1, 3-2, 3-3, etc. projected onto the x-axis, and (ii)
The cumulative number PR of rightward components of the same vector line segment,
(iii) Similarly, the cumulative number PL of leftward components, (iv) the number PY of vector line segments 4-1, 4-2, 4-3, etc. projected onto the y-axis, and (v) the same vector line segments. The cumulative number of upward components PU and (vi) the cumulative number of downward components.
PB is extracted as a feature. 1st
The character 1 shown in the figure is represented by PX=3, PY=3 PR=1, PL=2 PU=2, PB=1.

Figure 2 A to F are letters L, U, C, M, 2,
6, the above vector line segments are shown respectively. FIG. 2 G summarizes the above-mentioned characteristics PX, PY, PR, PL, PU, and PB for each of the above characters and explains the differences between each character.

As can be seen from FIG. 2G, this is sufficient for distinguishing between characters, and the extraction process for obtaining feature quantities is simple and sufficient. Of course, as the number of categories increases, some categories will have the same feature amount, but this method is sufficiently effective for broadly classifying characters and the like.

FIG. 3 shows an embodiment of the invention. In the figure, 5 is an input coordinate storage register, in which the input coordinates (x _i ,
y _i ). Reference numeral 6 denotes an x-axis projection component generation circuit section which reads the x-axis coordinate value point sequences x _i and x _i+1 and extracts the moving direction (+ or -) of the point sequence and the distance between the points. Reference numeral 7 denotes a y-axis projection component generation circuit section, which has substantially the same function as the circuit section 6 described above. 8 is x
The axial projection component information storage unit includes the generation circuit unit 6.
Temporarily retains the information extracted by 9 is y
The information storage unit 8 is an axial projection component information storage unit.
It has essentially the same function as . Reference numeral 10 denotes an x-axis vector line segment generation unit, which generates vector line segments based on the information in the information storage unit 8. Reference numeral 11 denotes a y-axis vector line segment generating section, which has substantially the same function as the x-axis vector line segment generating section 10 described above. Also 12
1 to 17 are counters, respectively, and count the above-mentioned feature quantities PR, PX, PL, PU, PY, and PB, respectively. Reference numeral 18 represents a collation processing unit, 19 represents a standard category dictionary memory, and 20 represents a recognition result output unit. Further, 21 corresponds to an x-axis component feature extracting section according to the present invention, and 22 similarly corresponds to a y-axis component feature extracting section.

The x-axis component feature extraction unit 21 will be briefly explained. The x-axis projection component generation circuit unit 6 examines L _i =x _i+1 −x _i based on the coordinates of each point sequence, and extracts direction information (+) or (-) and distance information L _i do. The x-axis projection component information storage section 8 stores the above direction information and distance information.

The x-axis vector line segment generation unit 10 calculates ΣL _i for components having the same direction information, and if the value exceeds a threshold value, extracts this as one vector line segment. Then, corresponding to the vector line segment, the counter 13 is incremented by 1, and the counter 12 or 14 is incremented by 1.
do.

Needless to say, the dictionary memory 19 stores feature amounts corresponding to each standard category, and sequentially outputs them to the matching processing section 18 for each category. The verification processing unit 18 includes counters 12 to 17.
and the feature amount read from the dictionary memory 19.

As described above, according to the present invention, it is possible to roughly classify figures through simple processing based on a two-dimensional coordinate point sequence input in time series.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram illustrating feature amounts extracted by the present invention, FIG. 2 is an example of feature amounts extracted corresponding to multiple characters, and FIG. 3 is an example of the present invention. . In the figure, 1 is a character, 2 is a point, 3 is an x-axis projection vector line segment, 4 is a y-axis projection vector line segment, 21 is an x-axis component feature extraction section, 22 is a y-axis component feature extraction section, 1
Reference numeral 8 represents a collation processing unit, and reference numeral 19 represents a standard category dictionary memory.

Claims (1)

[Claims] 1. In a figure recognition device that recognizes a two-dimensional figure represented by a two-dimensional coordinate point sequence input in chronological order,
an x-axis component feature extraction unit that extracts the number of projected vector line segments projected onto the x-axis and the cumulative number for each vector direction based on the x-coordinate point sequence in the two-dimensional coordinate point sequence; a y-axis component feature extraction unit that extracts the number of projected vector line segments projected onto the y-axis and the cumulative number for each vector direction based on the y-coordinate point sequence in the dimensional coordinate point sequence; A figure recognition device characterized in that the feature quantities extracted by the above-mentioned extraction units are compared with the standard feature quantities read from a standard category dictionary.