JP2002358521A - Device, method and program for registering and identifying document format - Google Patents

Abstract

PROBLEM TO BE SOLVED: To identify a document feature while covering the existence of not only a ruled line but also a barcode, etc. SOLUTION: A slip format is registered by performing filtering for weakening the feature in the prescribed part of inputted document information (1), extracting the feature values of a ruled line pattern from filtered document information (2) and registering the extracted featured values of the ruled line s format information of the document. The document format is identified by performing filtering for weakening the feature in the prescribed part of the inputted document information (1), extracting the featured values of the input ruled line pattern from filtered input document information (2) and comparing the featured values of the input ruled line pattern with the pre-registered format information featured values so as to decide the format of the input ruled line pattern (3).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to a form format registration / identification apparatus, method and program.

[0002]

2. Description of the Related Art Many forms, such as bills and invoices, used in business transactions and business processes have a fixed format. For example, when a predetermined item is printed on a form of such a fixed form by a printer or the like, if the form type of the form can be identified based on features such as ruled lines described in the form, a place where the predetermined item is printed is automatically determined. This is convenient in that it can be set up efficiently and the efficiency of office work can be increased.

In recent years, in many cases, information on a predetermined item is converted into a barcode, which is pasted on a form or directly printed. In such a case, even if the ruled line itself is exactly the same between a plurality of forms, the subsequent handling is often different between the one with the bar code attached and the one without the bar code attached. It is desirable to treat it differently depending on the presence or absence of the code. For example, after distinguishing a form having a barcode from a form having no barcode, when reading information described in the barcode as a subsequent operation only for a form having a barcode, the barcode It is necessary to treat it differently depending on the presence or absence.

[0004] The identification of the form type is performed according to the following procedure. That is, a plurality of form formats are registered in advance (hereinafter, this procedure is referred to as “form form registration”). And read the form,
The ruled line pattern of the form is extracted, and the extracted ruled line pattern is compared with each of the plurality of registered form formats to calculate the degree of similarity. Anything exceeding that is determined as the form format of the read form (hereinafter, this procedure is referred to as “form format identification”). FIG. 2 is a functional block diagram showing the configuration of a conventional example of such a device for identifying a form type.

[0005] The form format registration in the conventional form automatic identification device 3 is performed as follows. That is, the form information is read by the form information input unit 31, and the read form information is given to the ruled line pattern extracting unit 33. In the ruled line pattern extraction unit 33, a ruled line pattern is extracted from the form information, and the extracted ruled line pattern is registered as a form format by the format registration unit.

On the other hand, the form format identification in the conventional form format registration / identification device 3 is performed as follows. That is, similar to the format registration, the form information is read by the form information input unit 31,
The read form information is provided to the ruled line pattern extraction unit 33. In the ruled line pattern extraction unit 33, an input ruled line pattern is extracted from the form information, and given to the identification unit.

On the other hand, the format reading unit 34
The registered form formats are read out one by one and sent to the identification unit 36. In the identification unit 36, the input ruled line pattern and the plurality of registered formats are subjected to comparison one by one, and the similarity between the input ruled line pattern and each registered format is calculated. The registered form format having the highest similarity to the input ruled line pattern and exceeding the predetermined threshold value is determined as the form format of the input ruled line pattern. Then, the determined form format is displayed on a display unit (not shown) and used for user confirmation and the like.

However, in such a conventional apparatus, additional items other than ruled lines, for example, bar codes, characters, symbols, etc. (hereinafter, also referred to as “bar codes, etc.”) are included in the form.
Is described, such a barcode or the like is recognized as a part of the ruled line. Therefore, although the ruled line itself is completely the same, it is recognized as a completely different ruled line pattern. For example, the forms described in FIGS. 3A to 3C are all recognized as different ruled line patterns in the conventional device, although the ruled lines themselves are exactly the same. This is because a bar code is written at the lower left in FIGS. 3B and 3C, and FIG. 3B and FIG.
(C) is different from the barcode pattern. On the other hand, since the determination of the ruled line pattern is made by observing the entire form including the barcode pattern in the ruled line, FIG.
This is because (C) is all recognized as different ruled line patterns.

In order to solve such a problem, another conventional apparatus employs a technique in which a bar code or the like is subjected to a masking process so as to treat the bar code or the like as not having a bar code or the like. According to this technique, the form on which the barcode shown in FIGS. 3B and 3C is described is handled in the same manner as a form without a barcode or the like by the mask processing.
As a result, FIGS. 3A to 3C are all recognized as ruled line patterns of the same type.

[0010]

However, in an apparatus for performing such a mask process, the presence or absence of a bar code or the like is completely ignored. For this reason, there is a problem in that it is not possible to perform subsequent processing, such as reading information such as a barcode on a form having a barcode or the like after distinguishing the form based on the presence or absence of a barcode or the like.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a form format registration / identification apparatus capable of identifying form characteristics including not only ruled lines but also the presence or absence of a barcode or the like.

[0012]

Means for Solving the Problems (A) The first invention provides
(1) means for performing filtering for weakening a feature on a predetermined portion of the input form information; (2) means for extracting a feature amount of a ruled line pattern from the filtered form information; Means for registering the characteristic amount of the ruled line pattern as format information of the form.

Thus, since a ruled line pattern reflecting the presence or absence of a barcode or the like can be registered, further processing such as reading a barcode can be performed after the ruled line pattern of the form is identified.

(B) The second invention provides (1) means for performing filtering for weakening a feature on a predetermined portion of the input form information, and (2) an input ruled line pattern from the input form information on which the filtering has been performed. And (3) means for determining the format of the input ruled line pattern by comparing the featured value of the input ruled line pattern with the characteristic amount of the format information registered in advance. The present invention provides a form format registration / identification device characterized by the following.

In this way, since a ruled line pattern reflecting the presence or absence of a barcode or the like can be identified, it is possible to perform further processing such as reading a barcode after identifying the ruled line pattern of a form.

(C) The third invention provides (1) means for extracting a feature amount of an input ruled line pattern from input form information, and (2) filtering information for executing filtering for weakening features. A form format registration / identification apparatus characterized by comprising means and (3) means for adding generated filtering information to the extracted feature amount of a ruled line pattern and registering the information as format information.

Thus, at the time of format registration, filtering information is added to the ruled line pattern and registered, so that the number of registered formats can be reduced.

(D) The fourth invention provides (1) means for extracting a feature amount of an input ruled line pattern from input form information, and (2) filtering of one or more registered format information added thereto. Means for sequentially reading the information together with the information; and (3) means for executing the same filtering based on the read filtering information on both the extracted feature of the input ruled line pattern and the read feature of the registered format information. (4) means for determining the format of the input ruled line pattern by comparing the characteristic amount of the input ruled line pattern after filtering is executed with the characteristic amount of the registered format information; An identification device is provided.

Since filtering is performed at the time of identifying a ruled line pattern, a large number of ruled line patterns can be handled with a small number of registered formats.

(E) In a fifth aspect based on the first to fourth aspects, the filtering is any one of a blurring process, a mosaic process, and an emphasis process, or a combination thereof. Provide a registration and identification device.

Thus, since various types of filtering can be employed, the type of filtering most suitable for the filtering target can be selected.

(F) A sixth aspect of the present invention is to (1) perform filtering for weakening a feature on a predetermined portion of the input form information, and (2) extract a characteristic amount of a ruled line pattern from the filtered form information. (3) registering a feature amount of the extracted ruled line pattern as format information of the form;
Provide an identification method.

Thus, since a ruled line pattern reflecting the presence / absence of a barcode or the like can be registered, it is possible to perform further processing such as reading a barcode after identifying the ruled line pattern of a form.

(G) According to a seventh aspect of the present invention, (1) filtering for weakening a feature is performed on a predetermined portion of the input form information, and (2) characteristics of an input ruled line pattern based on the filtered input form information. (3) Form format registration / identification wherein the format of the input ruled line pattern is determined by comparing the feature amount of the input ruled line pattern with the feature amount of the format information registered in advance. Provide a way.

Thus, since a ruled line pattern reflecting the presence / absence of a barcode or the like can be identified, it is possible to perform further processing such as reading a barcode after identifying the ruled line pattern of a form.

(H) According to an eighth aspect of the present invention, (1) extracting a feature amount of an input ruled line pattern from input form information, (2) generating filtering information for executing filtering for weakening a feature, and (3) A) providing a form format registration / identification method characterized in that generated filtering information is added to the extracted feature amount of a ruled line pattern and registered as format information;

Thus, at the time of format registration, filtering information is added to the ruled line pattern and registered, so that the number of registered formats can be reduced.

(I) The ninth invention provides (1) extracting a feature amount of an input ruled line pattern from input form information, and (2)
One or more registered format information is sequentially read out together with the filtering information added thereto, and (3) both the extracted feature amount of the input ruled line pattern and the feature amount of the read-out registered format information are read out. Perform the same filtering based on the filtering information,
(4) A form format registration / identification method characterized by determining the format of an input ruled line pattern by comparing the characteristic amount of an input ruled line pattern after filtering and the characteristic amount of registered format information.

Since filtering is performed at the time of identifying a ruled line pattern, a large number of types of ruled line patterns can be handled with a small number of registered formats.

(J) The tenth invention is based on the sixth to ninth inventions, wherein the filtering is any one of blur processing, mosaic processing and enhancement processing or a combination thereof. Provides registration and identification methods.

Thus, since various types of filtering can be adopted, the type of filtering most suitable for the filtering target can be selected.

(K) In the eleventh invention, (1) filtering for weakening a feature is performed on a predetermined portion of the input form information, and (2) the feature amount of the ruled line pattern is calculated from the filtered form information. (3) Form format registration and registration, characterized in that the feature amount of the extracted ruled line pattern is registered as format information of the form.
Provide an identification program.

Thus, since a ruled line pattern reflecting the presence or absence of a barcode or the like can be registered, further processing such as reading a barcode can be performed after the ruled line pattern of the form is identified.

(L) The twelfth invention provides (1) filtering for weakening a feature in a predetermined portion of the input form information, and (2) characteristics of an input ruled line pattern from the filtered input form information. (3) Form format registration / identification wherein the format of the input ruled line pattern is determined by comparing the feature amount of the input ruled line pattern with the feature amount of the format information registered in advance. Provide a program.

In this way, since a ruled line pattern reflecting the presence or absence of a barcode or the like can be identified, it is possible to perform further processing such as reading a barcode after identifying the ruled line pattern of a form.

(M) According to a thirteenth aspect, (1) a feature amount of an input ruled line pattern is extracted from input form information,
(2) A form format characterized by generating filtering information for executing filtering for weakening the feature, and (3) adding the generated filtering information to the extracted feature amount of the ruled line pattern and registering the information as format information. Provide a registration and identification program.

Thus, at the time of format registration, filtering information is added to the ruled line pattern and registered, so that the number of registered formats can be reduced.

(N) A fourteenth aspect of the present invention is to (1) extract a feature amount of an input ruled line pattern from input form information;
(2) One or more registered format information is sequentially read out together with the filtering information added thereto, and (3)
The same filtering based on the read filtering information is performed on both the extracted feature amount of the input ruled line pattern and the feature amount of the read registered format information, and (4) the input ruled line pattern after the filtering is executed. A form format registration / identification program characterized in that a format of an input ruled line pattern is determined by comparing a feature value with a feature value of registered format information.

Thus, since filtering is performed at the time of identifying a ruled line pattern, it is possible to handle a large number of ruled line patterns with a small number of registered formats.

(O) A fifteenth invention according to the eleventh to fourteenth inventions, wherein the filtering is any one of a blur processing, a mosaic processing and an emphasis processing or a combination thereof. Provide a registration and identification program.

As described above, since various types of filtering can be adopted, it is possible to select the type of filtering most suitable for the filtering target.

[0042]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(A) First Embodiment A form format registration / identification apparatus according to a first embodiment of the present invention will be described with reference to FIG. 1 and FIGS.

(A-1) Configuration of First Embodiment The configuration of a form format registration / identification device according to the first embodiment will be described with reference to FIG. FIG. 1 is a functional block diagram of such a configuration.

The form format registration / identification device 1 of the first embodiment comprises a form information input unit 11, a filtering execution unit 12, a ruled line pattern extraction unit 13, a format registration unit 14, a format reading unit 15, and an identification unit 16. Is done.

Here, the form information input section 11 is used for information described on the form (refers to all information described on the form, specifically, ruled lines, diagrams, characters, symbols, etc.).
This is also what the filtering execution unit 12 gives the read form information.

The filtering execution section 12 executes filtering on a predetermined portion in the given form information, which acts to weaken the features in the portion.

The ruled line pattern extracting unit 13 extracts the characteristic amount of the ruled line pattern from the form information after the execution of the filtering, and provides the extracted characteristic amount of the ruled line pattern to the format registration unit 14 and the identification unit 16.

The format registration unit 14 stores the feature amount of the ruled line pattern as a registration format.

The format reading section 15 sequentially reads out the registered formats stored in the format registration section 14 and gives it to the identification section 16.

The discriminating unit 16 calculates a characteristic amount of a ruled line pattern relating to an input (hereinafter, also referred to as an “input ruled line pattern”)
The format of the input ruled line pattern is determined by comparing with the feature amount of the read registered format information. The identification unit 16 stores a threshold value for determining whether a registered format listed as a candidate for the input ruled line pattern format is recognized as the format of the input ruled line pattern.

(A-2) Processing Operation When Registering Form Format Referring to FIGS. 4A, 5 and 6, the form format registration / identification device 1 according to the first embodiment performs the processing when registering the form format. The processing operation will be described. Here, FIG. 4A is a flowchart showing a flow of processing at the time of registering a form format, FIG. 5 is a diagram showing a filtering target area, and FIG. 6 is a diagram showing filtering of various forms by the present embodiment. It is a figure which shows the change before and after execution at the time of execution.

First, the form information read unit 11 reads form information described in the form (S11), and the read form information is provided to the filtering execution unit 12.

In the filtering execution unit 12, filtering is performed on the given form information. The filtering performed here is a filtering function that weakens the characteristics of a predetermined portion in the form information.

The filtering is performed in the following procedure. First, as shown in FIG.
The filtering target area 101 is specified in advance in a part of 0. Here, the form information display area 100 is an area in a display unit (not shown) that displays the read form information. The filtering target area 10
1 indicates an area to be filtered. It should be noted that although all of the read form information is actually displayed in the form information display area 100, only the ruled lines in the form information are indicated by broken lines in FIG. This is because the ruled line and the filtering target area 101
This is to make it easier to grasp the positional relationship between the objects.

The type of filtering is specified in advance. In the present embodiment, a blurring process is employed as a type of filtering. When the blurring process in the present embodiment is executed, the filtering target area 1
01 includes a line diagram, the inside of the filtering target region 101 changes to a black-out state, whereas if there is no such line diagram, there is no change in the region 101. It does not happen and remains as it is.
That is, when the blurring process is performed, if there is a diagram or the like in the filtering target region 101, the diagram or the like in the region 101 is weakened in such a manner that the inside of the region 101 is painted black. As described above, in the blurring process, it is only a matter of whether or not there is a diagram in the area, and the shape of the diagram itself does not matter.

FIG. 6 shows a specific example of a change when the blurring process is performed on a form having various characteristics. Here, FIG. 6 (A1), FIG. 6 (B1), and FIG.
FIGS. 6A and 6B show states before the execution of the blurring process.
2) and 6 (C2) show the state after the execution of the blurring process. In order to make a comparison with the prior art, the forms shown in FIGS. 6A1, 6B1 and 6C1 are shown in FIGS. 3A and 3B, respectively. ) And the form shown in FIG. 3C.

As shown in FIG. 6 (A2), FIG.
Are not described in the filtering target area 101 even if the blur processing is executed,
No change occurs. On the other hand, as shown in FIG. 6B2 and FIG. 6C2, when there is a description such as a diagram in the filtering target area 101, it becomes a rectangle whose inside is painted black. This change occurs because the form in FIG. 6 (B1) and FIG. 6 (C1) includes a barcode in the rectangular filtering target area 101, so that the entire rectangular filtering target area 101 is subjected to the blurring processing. This is because it is painted black.

Next, in the ruled line pattern extraction unit 23,
A ruled line pattern is extracted from the thus obtained form information after the execution of the blurring process (S14). Here, in extracting the ruled line pattern, not only the characteristics of the ruled line but also the characteristics of the filtered portion (in the present embodiment, the portion painted black as a result of the execution of the blurring process) are considered. The extraction of the ruled line pattern is performed by extracting the feature amount of the ruled line pattern. In the case of the present embodiment, the obtained ruled line pattern is as shown in FIG.
(B2) and 6 (C2) shown in FIG. 6 (B2) and FIG. 6 (C2), which have a filled rectangle at the lower left of the figure.

Thereafter, the obtained feature amount of the ruled line pattern is registered in the format registration unit 14 as a form format (S14).

As described above, according to the present embodiment, by executing the blurring process, a ruled line pattern which differs depending on whether or not a barcode exists in a predetermined area while registering a barcode pattern regardless of the barcode pattern is registered. Will be done.

(A-3) Operation at the time of form format identification processing The operation at the time of form format identification processing will be described with reference to FIG. Here, FIG. 4B is a flowchart showing the operation during the form format identification processing.

First, a ruled line pattern of an input form (hereinafter referred to as an "input ruled line pattern") is extracted by the same procedure as in the format registration process (S21 to S21).
S23). Specifically, the form information is read by the form information reading unit 11 (S21), filtering is performed on a predetermined area in the form information by the filtering execution unit 12 (S22), and the filtering is performed by the ruled line pattern extracting unit 23. The ruled line pattern is extracted from the form information (S23).

Next, the input ruled line pattern is identified (S24 to S34). In the following description, it is assumed that the format registration unit 14 stores n (n is a natural number) registered formats. First, 1
The second registered format is read by the format reading unit 15 (S24, S25), the registered format read by the identification unit 16 is compared with the input ruled line pattern, and the similarity between them is calculated (S2).
6). For the second and subsequent registration formats, comparison and similarity are calculated in the same procedure (S27).
Repeated until the n-th registration format (S2
8). In this way, the similarity to the input ruled line pattern is calculated for all of the n registered formats.

The identification unit 16 selects a registration format indicating the maximum value among the n similarities obtained as a candidate for the format of the input form. The selected format candidate is compared with the threshold value stored in the identification unit 26 (S29). If the maximum similarity exceeds the threshold, the registration format indicating the maximum similarity is determined as the format of the input ruled line pattern (S3).
0), the determined format is displayed on a display unit (not shown) (S31). On the other hand, if the maximum similarity is equal to or smaller than the threshold value, it is determined that there is no corresponding format (S32), and the fact that there is no corresponding format is displayed on a display unit (not shown) (S33). . Here, the threshold value is provided because, when the similarity between the input ruled line pattern and the registered format to be compared is low, even if the maximum similarity is indicated, the registered format is This is because it is not appropriate to recognize the format of the input ruled line pattern.

The similarity can be calculated by a normal pattern matching method (for example, a method using the minimum distance between patterns).

After the format of the input ruled line pattern is determined by the above procedure, the user can perform different processing according to the presence or absence of the barcode. For example, for a form determined to have a barcode, an operation such as reading information recorded on the barcode can be performed.

(A-4) Effects of the First Embodiment As described above, in the first embodiment, when a form is registered, filtering is performed on an area in which a barcode or the like is written, thereby reducing the bar. The configuration in which the presence or absence of a code or the like is reflected in the identification of the ruled line pattern, that is, the configuration in which the pattern such as a barcode is not considered but the presence or absence of a barcode or the like is considered in identifying the ruled line pattern. This allows
Problems of the prior art in recognizing ruled line patterns, namely,
The rule lines themselves are exactly the same.Solving the point that the rule line pattern is recognized as different depending on the bar code pattern for multiple forms, and the rule line pattern is recognized completely ignoring the existence of the bar code Is done.

Further, it is possible to perform further advanced processing such as reading barcode information on a form on which a barcode or the like is written.

(B) Second Embodiment In the first embodiment, filtering is performed on an area in which a barcode or the like is written, and in the case of a barcode, the area is blacked out. This makes it possible to register and identify a ruled line pattern in consideration of the presence or absence of a barcode or the like.

However, in this configuration, the location, the shape, the size, etc., of the location where the barcode or the like is attached (ie, the filtering target area) are slightly different, and are recognized as completely different ruled line patterns. As a result, a number of mutually similar registration form formats having only slight differences are generated, which may cause a problem that the registration procedure and the registration procedure become complicated.

In order to solve such inconveniences, in the second embodiment, only the filtering information is added to the registration format at the time of format registration, and the items (parameters) of the filtering information are used to identify the ruled line pattern of the input form. Is appropriately specified to perform filtering. With such a configuration, it is possible to cope with various positions, shapes, sizes, and the like of the filter target regions.

(B-1) Configuration of Second Embodiment FIG. 7 shows the overall configuration of a form format registration / identification device 2 according to the second embodiment.

The form format registration / identification device 2 of the second embodiment includes a form information input section 21, a filtering information setting section 22a, a filtering execution section 22b, a ruled line pattern extraction section 23, a format registration section 24, a format reading section 25, And an identification unit 26.

The form format registration / identification apparatus 2 of the second embodiment is largely different from the form format registration / identification apparatus 1 of the first embodiment in the configuration of the filtering information setting unit 2 for generating information on filtering.
2a is newly provided.

Here, the form information input section 21 reads the form information and gives the read form information to the ruled line pattern extracting section 23.

The ruled line pattern extracting unit 23 extracts the characteristic amount of the ruled line pattern from the given form information. Then, the feature amount of the extracted ruled line pattern is provided to the format registration unit 24 together with the filtering information generated by the filtering information setting unit 22a described later.

The filtering information setting section 22a sets filtering information for specifying the content of filtering.

The format registering section 24 stores, as a registration format, a ruled line pattern obtained by adding filtering information to the feature amount.

The format reading section 25 sequentially reads out one or more registered formats stored in the format registration section 24 together with the filtering information, and gives it to the filtering execution section 22b.

The filtering execution unit 22b executes the same filtering on both the input ruled line pattern and the read registered format based on the given filtering information.

The identification unit 26 determines the format of the input form by comparing the characteristic amount of the input ruled line pattern with the characteristic amount of the read registered format information. In addition, the identification unit 26 stores a threshold value for determining whether or not a registered format given as a candidate for the format of the input form is recognized as the format of the input form.

(B-2) Processing Operation at the Time of Form Format Registration Referring to FIGS. 8A and 9, the flow of processing operation at the time of format registration in the form format registration / identification apparatus 2 of the second embodiment will be described. Will be described. Here, FIG.
FIG. 9A is a flowchart showing the flow of processing at the time of format registration, and FIG. 9 is a diagram showing items (parameters) of filtering information.

First, the form information read unit 21 reads the form information described in the form (S 21), and the read form information is given to the ruled line pattern extracting unit 23.

The ruled line pattern extracting unit 23 extracts a ruled line pattern of the form from the provided form information (S42).

Next, filtering information is set by the filtering information setting section 22a. Here, the filtering information refers to information for specifying the content of the filtering. As shown in FIG. 9, specific items (parameters) constituting the filtering information are the type of filtering, the shape of the filter, the size of the filter, and the position of the filter.

After that, the format registration unit 24 adds the filtering information set by the filtering information setting unit 22a to the ruled line pattern extracted by the ruled line pattern extraction unit 23, and registers it as a form format.

As described above, in the second embodiment, filtering is not executed before format registration, and instead, filtering information is set and added to the registered format.

(B-3) Operation in Form Format Identification Processing The form format identification processing operation in the second embodiment will be described with reference to FIG. here,
FIG. 8B is a flowchart showing the operation during the form format identification processing.

First, a ruled line pattern of an input form is extracted according to the same procedure as in the form format registration process (S51, S52). Specifically, the form information is read by the form information reading unit 11 (S51), and the ruled line pattern is extracted from the form information by the ruled line pattern extracting unit 23 (S52).

Next, the input ruled line pattern is identified (S54 to S64). In the following description, it is assumed that n (n is a natural number) registered formats are stored in the format registration unit 24.

First, the format reading unit 25
The first registration format and the filtering information added thereto are read (S53, S54).

Next, each item (parameter) of the filtering information is specified by the user (S55). The parameters of the filtering information are specified by selection from a menu (not shown) or key input from a keyboard (not shown). The parameters are specified by, for example, blurring processing (type of filtering), rectangle (filter shape), vertical 10 dots × horizontal 20 dots (filter size), and the dot at the lower left corner of the filter is the base point of the form information display area. (For example, the point at the lower left corner of the form information display area)
This is performed by selecting an item such as a position shifted 5 dots upward by 10 dots to the right (the position of the filter).

Next, the same filtering based on the specified filtering information is performed on both the input form information and the read registration format by the filtering execution unit 22b (S56).

Next, the identification unit 26 compares the input ruled line pattern that has been subjected to the same filtering with the first registered format, and determines the similarity between the two (S5).
7).

For the second and subsequent registration formats, the above-described comparison and similarity are calculated in the same procedure (S58), and the process is repeated until the n-th registration format (S59). With the above procedure, the similarity with the input ruled line pattern is calculated for all of the n registered formats.

In the discriminating section 26, the one showing the maximum value is selected from the obtained n similarities, and this maximum similarity is compared with the threshold value stored in the discriminating section 26. (S60). If the maximum similarity exceeds the threshold, the registration format indicating the maximum similarity is determined as the format of the input ruled line pattern (S61), and the determined format is displayed on a display unit (not shown) ( S62). On the other hand, when the maximum similarity is equal to or smaller than the threshold, it is determined that there is no corresponding format (S63), and the fact that there is no corresponding format is displayed on a display unit (not shown) (S64).

In the second embodiment having the above configuration and operation, the number of formats to be registered is greatly reduced as compared with the first embodiment. For example, in the first embodiment, in the example shown in FIG. 6, the one shown in FIG.
2) and the format shown in FIG. 6 (C2) needs to be registered, whereas in the second embodiment, the ruled line pattern shown in FIG. 6 (A2) with filtering information added is registered. Then, all forms from FIG. 6 (A1) to FIG. 6 (C1) can be handled. For this reason,
The registration and identification process is simplified. This effect becomes more remarkable as the number of similar forms increases.

(B-4) Effects of the Second Embodiment As described above, in the second embodiment, the filtering is not performed before the format registration, and when the input ruled line pattern is identified, the filtering is performed at the time of format registration. Based on the filtering information thus obtained, a configuration is executed for both the input ruled line pattern and the registered format. Therefore, just by changing the parameters of the filtering information,
Various filterings can be performed. as a result,
The number of formats to be registered can be significantly reduced.

(C) Other Embodiments (C-1) In the above embodiment, the type of filtering is described as a blurring process. However, the type of filtering is not limited to this as long as it has a function of weakening features. Any of enhancement processing and other image processing may be used. In addition,
These filtering types may be applied alone, or may be applied in an appropriate combination.

(C-2) In the above embodiment, the format of the input form is determined by selecting one registered format indicating the maximum similarity as a candidate for the format of the input form, and setting the selected candidate as a threshold value. The comparison was performed. However, the registered format indicating the maximum similarity may not always be optimal as the format of the input form, depending on how the item for determining the similarity is selected. Therefore, a configuration may be adopted in which a plurality of registered formats are selected in order from the registered format in descending order of similarity, and an optimal format is selected from the plurality of candidates by visual observation as a whole. Good.

(C-3) In the above embodiment, the case where the object of filtering is a barcode has been described. However, the object of filtering is not limited to this, and can be used to determine the presence or absence of a seal in a form. . For example, if the stamp field is subjected to filtering as a filtering process, the portion subjected to the stamp process is painted black while the portion with the stamp is black, whereas there is no change if there is no stamp. As a result, it is possible to distinguish between forms that are stamped in the seal box and forms that are not stamped, and for those that are not stamped, it is possible to perform further processing such as urging the person who fills the form to stamp it. Becomes

(C-4) Further, as in the relationship between the withdrawal slip and the deposit slip, only the characters of "withdrawal" and "payment" are different, and the other parts are to be handled in the same form as the same format. If filtering is performed on different character portions such as "payment" and "payment", they can be treated as the same format.

(C-5) Similarly, if it is desired to treat a plurality of forms having the same ruled line characteristic but different destinations as the same format, this can be achieved by executing filtering in the destination column. Become. In this way, for forms with the same ruled line, those with the destination name are distinguished from those without the destination name, and for those with the destination name, it is examined whether the destination name is correctly attached. Becomes possible.

(C-6) Further, the filtering according to the present invention can be applied to a character string to be used for character string identification. For example, if only the first part of the name is filtered, the form can be distinguished by the last name. In addition, if filtering is applied to a part of the address other than the prefecture name, the form can be distinguished by the prefecture name.

[0106]

(1) When registering a form, filtering is performed on an area in which a barcode or the like is written, so that the presence or absence of a barcode or the like is reflected in the identification of a ruled line pattern, that is, a ruled line. In the identification of the pattern, a pattern such as a barcode is not a problem, but the presence or absence of a barcode is a problem. As a result, the ruled line pattern is recognized to be different depending on the barcode pattern for a plurality of forms having exactly the same ruled line,
The disadvantages of the prior art, such as completely ignoring the barcode, can be eliminated. Further, it is possible to perform further advanced processing such as reading barcode information on a form on which a barcode is written.

(2) Filtering based on the filtering information registered at the time of format registration is executed for both the input ruled line pattern and the registered format at the time of identifying the input ruled line pattern, not at the time of format registration. Therefore, various filterings can be performed only by changing the parameters of the filtering information. As a result, the number of formats to be registered can be significantly reduced.

[Brief description of the drawings]

FIG. 1 is a functional block diagram illustrating a configuration of a form format registration / identification device according to a first embodiment.

FIG. 2 is a functional block diagram showing a configuration of a conventional form format registration / identification device.

FIG. 3 is a diagram showing various forms applied to a conventional form format registration / identification apparatus.

FIG. 4 is a flowchart showing an operation of the form format registration / identification device according to the first embodiment.

FIG. 5 is a diagram showing a filtering target area according to the first embodiment.

FIG. 6 is a diagram illustrating states before and after filtering is performed on various forms in the first embodiment.

FIG. 7 is a functional block diagram illustrating a configuration of a form format registration / identification device according to a second embodiment.

FIG. 8 is a flowchart showing an operation of a form format registration / identification device according to the second embodiment.

FIG. 9 is a diagram showing items (parameters) of filtering information according to the second embodiment.

[Explanation of symbols]

11, 21 ... form information input unit, 22a ... filtering information setting unit, 12, 22b ... filtering execution unit, 1
3, 23 ... ruled line pattern extraction unit, 14, 24 ... format registration unit, 15, 25 ... format reading unit, 16,
26 ... Identifier.

Claims (15)

[Claims] 1. A means for performing filtering for weakening a feature on a predetermined portion of input form information, a means for extracting a feature amount of a ruled line pattern from the filtered form information, and an extracted ruled line pattern Means for registering the characteristic amount as form format information. 2. A means for performing filtering for weakening a feature on a predetermined portion of the input form information, a means for extracting a feature amount of a ruled line pattern from the filtered form information, and an extracted ruled line pattern Means for determining a format of a form associated with input form information by comparing the characteristic amount of the form with the characteristic amount of the registered form format information. 3. A means for extracting a feature amount of a ruled line pattern from input form information; a means for generating filtering information for specifying the content of filtering for weakening a feature; , Create the form format information by adding the generated filtering information,
Means for registering the created form format information. 4. A means for extracting a feature amount of a ruled line pattern from input form information, a means for sequentially reading one or more registered form format information together with filtering information added thereto, and an extracted ruled line pattern Means for performing the same filtering based on the read filtering information for both the characteristic amount of the input form information and the characteristic amount of the read registration form format information; By comparing with the feature amount of
Means for determining a format of an input ruled line pattern. 5. The method according to claim 1, wherein the filtering for weakening the feature is any one of a blurring process, a mosaic process, and an enhancement process, or a combination thereof.
A form format registration / identification apparatus according to any one of claims 1 to 4. 6. A predetermined portion of the input form information is subjected to filtering for weakening a feature, a feature amount of a ruled line pattern is extracted from the filtered form information, and the feature amount of the extracted ruled line pattern is extracted. A form format registration and identification method characterized in that the form format information is registered. 7. Filtering for weakening a feature in a predetermined portion of the input form information, extracting a feature amount of an input ruled line pattern from the filtered input form information, A form format registration / identification method, wherein a format of an input ruled line pattern is determined by comparing with a feature amount of format information registered in advance. 8. A feature amount of an input ruled line pattern is extracted from input form information, filtering information for executing filtering for weakening the feature is generated, and the generated filtering information is added to the feature amount of the extracted ruled line pattern. A form format registration / identification method characterized by adding and registering as format information. 9. A feature amount of an input ruled line pattern is extracted from the input form information, one or more registered format information is sequentially read out together with filtering information added thereto, and a feature amount of the extracted input ruled line pattern is extracted. The same filtering based on the read filtering information is performed on both the read feature information of the registered format information, and the feature value of the input ruled line pattern after the filtering is performed is compared with the feature value of the registered format information. By doing
A form format registration / identification method characterized by determining the format of an input ruled line pattern. 10. The filtering for weakening the feature,
The form format registration / identification method according to any one of claims 6 to 9, wherein the form format is any one of a blur process, a mosaic process, and an emphasis process or a combination thereof. 11. A filter for weakening a feature is performed on a predetermined portion of the input form information, a feature amount of a ruled line pattern is extracted from the filtered form information, and the extracted feature amount of the ruled line pattern is calculated. A form format registration / identification program characterized by registering as form format information. 12. A predetermined portion of the input form information is subjected to a filtering for weakening a feature, a feature amount of an input ruled line pattern is extracted from the filtered input form information, and a feature amount of the input ruled line pattern is calculated. A form format registration / identification program characterized in that a format of an input ruled line pattern is determined by comparing with a feature amount of format information registered in advance. 13. Extracting a feature amount of an input ruled line pattern from input form information, generating filtering information for executing filtering for weakening the feature, and applying the generated filtering information to the feature amount of the extracted ruled line pattern. A form format registration / identification program characterized by being added and registered as format information. 14. A feature amount of an input ruled line pattern is extracted from input form information, one or more registered format information is sequentially read out together with filtering information added thereto, and a feature amount of the extracted input ruled line pattern is extracted. The same filtering based on the read filtering information is performed on both the read characteristic information of the registered format information, and the characteristic value of the input ruled line pattern after the filtering is performed is compared with the characteristic value of the registered format information. By
A form format registration / identification program, which determines the format of an input ruled line pattern. 15. The filtering for weakening the feature,
15. The form format registration / registration according to any one of claims 11 to 14, wherein the form processing is any one of a blur processing, a mosaic processing, and an emphasis processing or a combination thereof.
Identification program.