JPH0766417B2 - Character recognition device - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improved character recognition system.

2. Description of the Related Art Conventionally, a character recognition device observes a character pattern by photoelectric conversion means and converts it into an electric signal, and the character pattern is binarized and normalized in size, as shown in FIG. The pattern is divided into regions where the pattern exists, and the number of pixels of the character part for each small region, the direction contribution degree, or the direction index of the character outline part is totaled and used as a feature amount to refer to a standard pattern created in advance. Many perform character identification processing.

Problems to be Solved by the Invention However, since the recognition device collects the pixels of the character part, the direction contribution degree, the direction index, and the like for each small region, for a character having large fluctuation such as a handwritten character, Since the area in which the pixels of the character portion and the contour points exist varies, there is a problem that the recognition ability is weak because of the fluctuation of the character line, especially the deviation of the position.

In view of the above-mentioned problems, the present invention provides a character recognition device that is stable even for characters that greatly change such as handwritten characters.

Means for Solving the Problems In order to solve the above problems, the character recognition device of the present invention is
A means for extracting a directional index indicating a tangential direction from each point of the contour portion of the input character pattern, and a region in which the character pattern exists is divided into small areas, and a small area of each point of the character contour portion existing in each small area. Means for determining the degree of influence of the direction index of each contour portion point on the small area and the small areas around the small area for each value of the direction index according to the position within the area, and for each small area to its own area It is equipped with means for collecting the influence degree of the direction index of each direction index to obtain a feature amount.

Effect The present invention has the above-described configuration to determine the influence degree of the direction index of each point of the contour portion of the character pattern with respect to the small area and the surrounding area of the small area according to the position of each point of the contour portion in the small area. Specified, and for each small area, the degree of influence on its own area is aggregated and used as a feature amount. However, it is possible to extract a stable feature amount that is not easily affected, and the character pattern can be recognized accurately.

Example Hereinafter, a character recognition device according to an example of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of a character recognition device according to an embodiment of the present invention. In FIG. 1, reference numeral 11 is a photoelectric conversion unit, which converts the input character pattern 10 into an electric signal. A preprocessing unit 12 binarizes the character pattern converted into an electric signal into a character portion and a background portion, and further normalizes the size. Reference numeral 13 is a direction index extraction unit, which extracts the direction indices in the four directions shown in FIG. 2 from each point of the character contour portion by tracing the contour portion of the preprocessed character pattern. For example, in FIG. 3 (a), the input character "a" is pre-processed and the size is normalized to 40 pixels x 40 pixels. By tracing the character contour portion from the starting point 31, four direction indices are given to each point of the character contour portion as shown in FIG. 3 (b). Reference numeral 14 denotes a direction index influence degree extraction unit that divides the area where the character pattern exists into n × n, and determines the degree of influence of the direction index on the small area and the surrounding area of each small area with respect to the contour points in each small area. To extract. Figure 4 shows the size of the small area
This is an example of how to determine the degree of influence of each position of a character contour point at the time of 10 pixels × 10 pixels on a small area where the contour point exists. FIG. 4 (a) is an example of how to determine the degree of influence with respect to the direction index 1. When the degree of influence is α, if y ≦ 5, the degree of influence on the lower area is y ≧ 6. 1-α. For example, in FIG. 5, the degree of influence of the direction index 50 of the contour point on each small region is x = 9, y = 9, so
From FIG. 5A, the area 51 is 0.7 and the area 52 is 0.3. FIG. 4B is an example of how to determine the degree of influence for the direction index 3, and when the degree of influence is α and x ≦ 5, x ≧ 5 to the left region.
If 6, the degree of influence on the right region is 1-α. FIG. 4 (c) is an example of how to determine the degree of influence for the direction indexes 2 and 4, and the degree of influence on the surrounding area is i) If x ≦ 5, y ≦ 5, 0.1 × ( 5-x), 0.1 × (5-
y) ii) If x ≦ 5, y ≧ 6, 0.1 × (5-x) to the left area and 0.1 × (y− to the lower area.
6) iii) If x ≧ 6 and y ≦ 5, 0.1 × (x−6) to the right area and 0.1 × (5- to the upper area.
y) iv) If x ≧ 6, y ≧ 6, 0.1 × (x−6) to the right area and 0.1 × (y− to the lower area.
6) For example, in FIG. 5, the degree of influence of the direction index 53 of the contour point on each small area is x = 3, y = 2, so that from FIG. 4 (c), the area 54 is 0.9, the area 55 is 0.3, and the area 55 is 0.3. 56 to 0.
It becomes 2. If there is no corresponding surrounding area, the degree of influence is added to the area where the contour point exists. Reference numeral 15 is a direction index influence degree totaling unit, and based on the influence degree of the direction index of each point of the character contour portion extracted by the direction index influence degree extraction unit 14 on each small area, the direction index influence degree is calculated for each small area. The degree of influence on the own area is totaled and used as a feature amount. For example, in the example shown in FIG. 5, there are 64 feature quantities based on the 4 × 4 small area and the directional indices of the four directions. 16
Is an identification processing unit, which calculates the distance between the feature amount obtained by the direction index influence degree aggregation unit 15 and the standard pattern 18 created in advance, and outputs the closest character type as a recognition result.

As described above, according to the present embodiment, the means for extracting the direction index from each point of the contour portion of the character pattern and the area in which the character pattern exists are divided into small areas, and the points of the character contour in each small area are divided. A means for determining the degree of influence of the direction index on the small area and the surrounding area of the small area according to the position in the area, and the degree of influence of each direction index on the own area is calculated for each small area. By providing the means for determining the amount, it is possible to extract a stable feature amount even for a character having large fluctuation such as a handwritten character, and the character pattern can be recognized accurately.

In the above embodiment, the size of the small area is set to 10 pixels × 10 pixels, but this may be any size, and the influence degree and the surrounding area to which the influence degree is applied may be determined in another way.

EFFECTS OF THE INVENTION As described above, according to the present invention, the degree of influence of the direction index of a character contour point is set to a small area where the character contour point exists and its surrounding area, and each direction index to the own area for each small area. Since the influence degree of is aggregated, stable feature quantities can be extracted even when the small area where the character contour points are present changes.
It is possible to realize a character recognizing device with high accuracy even for a character having large fluctuations such as handwritten characters, which is extremely useful in practice.

[Brief description of drawings]

FIG. 1 is a block configuration diagram of a character recognition device according to an embodiment of the present invention, FIG. 2 is an explanatory diagram showing a direction index, and FIG. 3 (a) is up to preprocessing for an input character "A". FIG. 3 (b) is an explanatory view showing the direction index of each point of the contour portion extracted from the character pattern of FIG. 3 (a), and FIG. 4 shows in each small area. Is an explanatory view showing an example of how to determine the influence degree of the direction index of FIG. 5, FIG. 5 is an explanatory view showing an example of the influence degree of the direction index of the character contour point, and FIG. It is explanatory drawing which shows the example divided. 10 …… Character pattern, 11 …… Photoelectric conversion diagram, 12 …… Preprocessing section, 13 …… Direction index extraction section, 14 …… Direction index influence degree extraction section, 15 …… Direction index influence degree aggregation section, 16… ... Identification processing unit,
17 …… recognition result, 18 …… standard pattern.

Claims (1)

[Claims] 1. A means for extracting a directional index indicating a tangential direction from each point of an outline portion of an input character pattern, and an area in which a character pattern exists is divided into small areas, and character synchronization conditions existing in each small area. According to the position of each point of the contour part in the small area, the degree of influence of the direction index of each point of the contour part on the small area and the small areas around the small area is aggregated for each direction index to be a special amount. A character recognition device comprising: a minute amount extraction means; and a means for performing character identification processing by referring to an output of the extra minute amount extraction means and a standard pattern created in advance.