JPH11242718A - Character recognizing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To flexibly facilitate a countermeasure to the change of a recognizing environment by improving recognizing precision by the weak point complementation of a recognizing means in a character recognizing device. SOLUTION: Each recognizing means 1 is provided with a dictionary 3 for integration for integral use with another recognizing means in addition to a specific dictionary 2 for separation. When a recognizing speed is taken priority, one recognizing means 1 is separately used, and the specific dictionary 2 for separation is used as the dictionary. When recognizing precision is taken priority, two or more recognizing means 1 are integrally used, and the specific dictionary 2 for separation is used as the dictionary of the main recognizing means, and the dictionary 3 for integration for integral use with the main recognizing means is used as the dictionary of the subordinate recognizing means.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character recognition device, and more particularly to a character recognition device having a plurality of recognition means and capable of integrally using two or more recognition means.

[0002]

2. Description of the Related Art There are known a character recognizing apparatus of a type using a single recognizing means and a character recognizing apparatus of a type using a plurality of recognizing means. The latter character recognition device performs character recognition on the same character image by means of different recognition methods and integrates the respective recognition results to obtain a final recognition result. No. 7-271899, H. Yamagata et al., "A.
Handwritten Character Recognition of Multiple
Classifiers ", IEICETRANS, INF. &
SYST., Vol. E79-D, No. 5, MAY 199
6. L. Xu et al., “Methods of Combining Multiple”
Classifiers andTheir Applications to Handwriti
ng Recognition ”, IEEE TRANSACTION
SON SYSTEMS, MAN, ANDCYBERN
ETICS, VOL.22, NO.3, MAY / JUNE
1992 and many others.

[0003]

Heretofore, a unique dictionary has been prepared for each recognition means. Even when a plurality of recognition means are used in an integrated manner, each recognition means uses the unique dictionary. In other words, the dictionary is handled in the same manner as when the recognition means is used alone.

[0004] The integrated use of two or more recognition means is:
It is intended to improve the recognition accuracy as a whole by complementing the weakness of a certain recognition means with another recognition means. An object of the present invention is to provide a character recognition device in which a configuration related to a dictionary is improved so that this weak point complementation can be achieved more effectively. Another object of the present invention is to provide a character recognition apparatus that can flexibly cope with changes in the recognition environment (for example, changes in the importance of recognition speed and recognition accuracy, changes in the types of characters and fonts to be recognized). Is to do.

[0005]

In order to achieve the above object, a character recognition apparatus according to the present invention has a plurality of recognition means, and two or more recognition means can be selected for integrated use. Wherein each recognizer has, in addition to its own dictionary, a dictionary for integrated use with other recognizers,
If two or more recognition means are used together,
The main recognition means of the one or more recognition means uses its own dictionary, and the other sub-recognition means uses a dictionary for integrated use with the main recognition means.

The character recognition device according to the second aspect of the present invention has the configuration of the character recognition device according to the first aspect of the present invention, and further includes a character recognition device that is erroneously recognized or recognized by a main recognition unit among two or more recognition units that are integrated and used. Means for collecting a data sample not to be processed, and utilizing the collected data sample for the secondary recognition means of the two or more recognition means.
Means for creating or updating a dictionary for integrated use with the main recognition means.

According to a third aspect of the present invention, there is provided a character recognition apparatus, comprising: a plurality of recognition means; a plurality of independent dictionaries specific to each of the plurality of recognition means; and another recognition means for each of the plurality of recognition means. A plurality of integration dictionaries for integrated use with the unit, means for designating priority of recognition speed or priority of recognition accuracy, and single use of one recognition means selected when priority of recognition speed is designated Means for performing character recognition by the user and the two selected when priority is given to recognition accuracy.
Means for performing character recognition by the integrated use of one or more recognizing means, and means for selecting a single dictionary specific to the recognizing means as a dictionary used by the recognizing means in the case of character recognition by single use of one recognizing means. And, in the case of character recognition by integrated use of two or more recognition means, selecting a single dictionary unique to the main recognition means as a dictionary used by one main recognition means of the two or more recognition means, As a dictionary used by each of the other sub-recognition units, the sub-recognition unit includes a unit for selecting an integration dictionary for integrated use with the main recognition unit.

A character recognition device according to a fourth aspect of the present invention has means for designating a recognition target in addition to the configuration of the character recognition device according to the third aspect of the present invention. Recognition means used alone or integrated is selected.

According to a fifth aspect of the present invention, in addition to the configuration of the character recognition apparatus according to the third aspect of the present invention, there are provided means for designating a recognition target and means for designating a recognition accuracy level. In the case where the priority of the recognition speed is instructed, when the recognition accuracy level is instructed, the recognition means having the fastest recognition speed among the recognition means satisfying the instructed recognition accuracy level for the instructed recognition target Is selected, and when the recognition accuracy level is not specified, a recognition unit having the highest recognition accuracy for the specified recognition target is selected.

A character recognition apparatus according to a sixth aspect of the present invention has a structure for designating a recognition target and a means for designating a recognition speed level, in addition to the configuration of the character recognition apparatus according to the third aspect of the invention. When priority is given to the recognition accuracy, when the recognition speed level is specified, the recognition unit having the highest recognition accuracy among the recognition units satisfying the specified recognition speed level for the specified recognition target. Is selected, and when the recognition speed level is not set, the combination of the recognition means having the highest recognition accuracy for the specified recognition target is selected.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of a character recognition apparatus according to the present invention, as shown in FIG. 1, includes a plurality of recognition means 1, a plurality of single dictionaries 2, a plurality of integrated dictionaries 3, and Control means 4, system control means 5, integration dictionary creating / updating means 6, collected data storage means 7, instruction input means 9
And configuration control data storage means 8.

The independent dictionary 2 is a dictionary specific to a specific recognition means. The integration dictionary 3 is a dictionary prepared corresponding to a combination of a main recognition unit and a secondary recognition unit in the case of integrated use, and is used as a dictionary of a recognition unit in accordance with the case of integrated use as described later. . When the main recognition unit has weaknesses, that is, character image data that cannot be recognized or is erroneously recognized, by preparing an integration dictionary 3 that enables such character image data to be recognized by the secondary recognition unit, The weakness of the main recognition means can be effectively supplemented, and the recognition accuracy can be improved as a whole.

A single dictionary 2 which is a basic dictionary of each recognition means 1
Is created so as to enhance the overall recognition performance of each recognition means, so the weak points can be compensated by additionally registering the features of the character images that are misrecognized or rejected as described above in the single dictionary. Attempting to do so may adversely affect overall recognition performance, such as an increase in misidentified characters. Such a problem can be avoided by a method of complementing the weak points with the integration dictionary 2.

The integration dictionary creating / updating means 6 is a means for coordinating with the recognition control means 4 to create or update the integration dictionary 3 for compensating for the weaknesses of the main recognition means as described above. Yes, the details of which will be described later. The collected data storage means 7 is a means for temporarily storing data samples collected for creating or updating the integration dictionary.

The instruction input means 9 is used by the user of the character recognition apparatus to instruct a recognition operation or an integrated dictionary creating / updating operation, an instruction to give priority to recognition accuracy or a priority to recognition speed, a minimum required recognition accuracy level or recognition speed. Level indication, indication of recognition target (one character type such as numbers, alphabets, hiragana, katakana, symbols, kanji, or a mixture of two or more character types, font type, etc.) The system control means 5
This is a means for inputting to.

The system control means 5 is a means for controlling the selection of the recognition means and the dictionary, and switching between the independent use and the integrated use of the recognition means in accordance with the instruction input from the instruction input means 9. Details will be described later. The configuration control data storage unit 8 stores configuration control data referred to by the system control unit 5. The contents of the configuration control data will be described later in connection with the operation of the character recognition device.

According to the control of the system control means 5, the recognition control means 4 includes one recognition means and one single dictionary, or two (generally two or more) recognition means and one single dictionary and Using one (generally one or more) integration dictionary, character image data is fetched from outside and character recognition is performed. When one recognition unit is used alone, the recognition result is directly sent to the outside. In the case where two (generally two or more) recognition means are integrated and used, the recognition results of the respective recognition means are integrated, and the integrated result is output to the outside.

This character recognition device is realized by dedicated hardware, or is connected to an external input device such as a CPU 100, a memory 101, an input device 103 such as a keyboard or an image scanner as shown in FIG. It is realized on a computer having a general configuration in which an input interface 104 for connection, an output device 105 such as a printer or a display, and an auxiliary storage device 102 such as a hard disk device are connected by a bus 107.

The correspondence between FIG. 1 and FIG. 2 is as follows.
The plurality of recognition units 1, the recognition control unit 4, the system control unit 5, and the integration dictionary creating / updating unit 6 each include a memory 1
01 is realized by the corresponding programs P1, P4, P5, and P6. The independent dictionary 2 and the integrated dictionary 3 are stored in the memory 101 as data files or table data.
Put on top. Specific areas of the memory 101 (or the auxiliary storage device 7) are allocated as the collected data storage unit 7 and the configuration control data storage unit 8. The input device 103 is used as the instruction input unit 9. The character image data is input from the outside via the input interface 104 or from the auxiliary storage device 102, for example. Although not shown, an operating system and the like necessary for the operation of the computer are naturally loaded in the memory 101.

FIGS. 3 to 5 are flowcharts showing the control flow of the system control means 5. The operation of the character recognition device will be described with reference to this flowchart and FIG. 1 (FIG. 2 as necessary).

The system control means 5 first checks whether the user's instruction is a recognition operation or an integration dictionary creation or update operation (step S1). When the recognition operation is instructed, it is determined whether the user's instruction is prioritized in recognition accuracy or in recognition speed (step S2).

A case where the recognition speed priority is instructed will be described below. In the case of giving priority to the recognition speed, one recognition unit 1 is used independently. The system control unit 5 checks whether or not the recognition accuracy level is specified by the user for selecting the recognition unit to be used (step S3). When the recognition accuracy level is not specified, one recognition unit 1 that can expect the highest recognition accuracy for the recognition target specified by the user based on the configuration control data stored in the configuration control data storage unit 8. A selection is made (step S4). When the recognition accuracy level is specified, the recognition unit 1 with the highest recognition speed capable of expecting a recognition accuracy higher than the specified recognition accuracy level is selected for the recognition target specified by the user based on the configuration control data. (Step S
5). Next, based on the configuration control data, an independent dictionary 2 specific to the selected recognition means 1 is selected (step S6),
It instructs the recognition control means 4 to start character recognition processing by using one recognition means alone (step S7).

The configuration control data includes information indicating the recognition accuracy and the recognition speed that can be expected for each recognition target of each recognition means when the recognition means is used independently, and information indicating the correspondence between the recognition means and the single dictionary 2. . Information on the recognition accuracy and the recognition speed is set in advance based on a recognition experiment using test data of each recognition target.

In the recognition control means 4 which has started the character recognition processing by using one recognition means alone, the recognition means 1 selected by the system control means 5 and the single dictionary 2 selected are used to execute external recognition. The character recognition processing of the fetched character image data is performed, and the recognition result is output to the outside.

Next, a case where the user gives an instruction to give priority to recognition accuracy will be described. In the case of giving priority to the recognition accuracy, two (generally two or more) recognition means are integratedly used.
In the system control means 5, first, it is checked whether or not a recognition speed level has been instructed by the user in order to select a recognition means to be integratedly used (step S8). When the recognition speed level is not indicated, based on the configuration control data,
A combination of the two recognizing means 1 that can expect the highest recognition accuracy for the recognition target specified by the user is selected (step S9). When the recognition speed level is instructed, two recognition units 1 with the highest recognition accuracy that can expect a recognition speed equal to or higher than the instructed recognition speed level for the recognition target instructed by the user based on the configuration control data. Is selected (step S10). Next, referring to the configuration control data, of the two recognition means 1 selected, the recognition means having the higher recognition accuracy when the specified recognition target is used alone (generally, two or more recognition means 1). The one having the highest recognition accuracy among the means) is set as the main recognition means, and its unique single dictionary 2 is selected (step S11). Next, referring to the configuration control data, the other recognition means (generally, each of the remaining recognition means excluding the main recognition means) is used as a subordinate recognition means for integration with the main recognition means for integrated use. The dictionary 3 is selected (step S12). Then, it instructs the recognition control means 4 to start the character recognition processing by the integrated use of the two recognition means (step S13).

In the configuration control data, in the case of using two recognition means in an integrated manner, for each combination of the recognition means, a distinction is made between the main recognition means and the subordinate recognition means, and the recognition accuracy and recognition expected for each recognition object. Information indicating the speed and the correspondence with the integration dictionary to be used is included. This is also set in advance based on a recognition experiment using test data.

In the recognition control means 4 which has started the character recognition processing by integrated use, the character recognition of the character image data is performed by using the selected main recognition means 1 and the selected single dictionary 2, Character recognition of the same character image data is performed by using the barrel recognition means 1 and the selected integration dictionary 3, the respective recognition results are integrated, and the integrated recognition result is output to the outside. In addition, the integration process may be performed by various known methods, and also uses the distance between the input character image and the dictionary as described in the specification of Japanese Patent Application No. 9-195444 of the present applicant. A higher performance integration method may be used.

Next, a case where the user instructs the operation of creating or updating the integration dictionary will be described. In this case, the system control means 5 selects the main recognition means 1 designated by the user and its own single dictionary 2 (step S14), and instructs the recognition control means 4 to use the main recognition means 1 alone. Instructs the start of the test character image data recognizing process, and instructs the integration dictionary creating / updating means 6 to start collecting data samples (step S15). The recognition control unit 4 captures test character image data for the recognition target specified by the user (for example, reads from the auxiliary storage device 102 in FIG. 2), and specifies the designated main recognition unit 1 and the selected single dictionary. 2 to perform character recognition. The integration dictionary creating / updating unit 6 causes the collected data storage unit 7 to store the character image data whose character recognition result is erroneously recognized or rejected (unrecognizable) via the recognition control unit 4. In this way, a sample of character image data that the main recognition means misrecognized or could not recognize is collected.

When this is completed, the system control means 5
Selects the integration dictionary 3 for integrated use of the secondary recognition means specified by the user and the main recognition means (step S16), and performs the operation of creating or updating the integration dictionary by creating or updating the integration dictionary. Instruct update means 6 (step S
17). The integration dictionary creating / updating unit 6 creates or updates the corresponding integration dictionary 3 based on the character image data collected in the collected data storage unit 7.

When instructed to create an integration dictionary, the integration dictionary creating / updating means 6 sequentially inputs the collected character image data to the recognition control means 4, for example. The characteristic amount of the character image data is extracted by the secondary recognizing means 1, and the selected integration dictionary 3 is created using the characteristic amount. When an update of the integration dictionary is instructed, the integration dictionary creation / update means 6
For example, inputting the collected character image data to the recognition control means 4 in order, and causing the recognition control means 4 to perform character recognition using the designated secondary recognition means and the selected integration dictionary 3. The feature amount extracted for the character image data whose recognition result is erroneously recognized or rejected is registered as it is or after statistical processing, and is additionally registered in the integration dictionary 3. Modify the contents of Thus, the user can create or update the integration dictionary for effectively reinforcing the weaknesses of the main recognition means as needed.

Incidentally, means for creating or updating the independent dictionary 2 may be added to the character recognition device.

[0032]

According to the first aspect of the present invention, there is provided a character recognition apparatus comprising:
If you use two or more selected recognition means,
The dictionary for making the weakness of the main recognition means, that is, the character image data which cannot be recognized or misrecognized, recognizable by the main recognition means is a dictionary for integrated use of the main recognition means of the main recognition means ( By preparing as an integrated dictionary, the weakness of the main recognition means can be reduced without modifying the unique dictionary (single dictionary) of the main recognition means, which may deteriorate the overall recognition performance of the main recognition means. It is possible to complement effectively and improve recognition accuracy as a whole. In addition, the dictionary used for the secondary recognition means is not fixed, but is selected according to the combination of the secondary recognition means and the main recognition means, so that the recognition means are integrated in various combinations according to the recognition environment. Even if used, the weaknesses of the main recognition means can be effectively supplemented to achieve high recognition accuracy.

The character recognition apparatus according to the second aspect of the present invention is intended to effectively supplement the weakness of the main recognition means in the integrated use of a necessary combination of recognition means in response to a change in the recognition environment on the user side. The dictionary can be created or updated to improve recognition accuracy or maintain high recognition accuracy.

The character recognition device according to the third aspect of the present invention can perform high-speed character recognition by using one recognition unit independently by giving an instruction to prioritize the recognition speed when the recognition speed is important. When recognition accuracy is important, by giving priority to recognition accuracy, high-precision character recognition can be performed by the integrated use of two or more recognition means.

The character recognition apparatus according to the fourth aspect of the present invention can specify a recognition target to be processed, and can perform character recognition by use of the optimum recognition means for the recognition target alone or in combination. Performance can be maximized for various recognition targets.

The character recognition apparatus according to the fifth aspect of the present invention can perform the fastest possible character recognition for a recognition target to be processed when the recognition speed is the most important, and the recognition speed is also important. However, if recognition accuracy of a certain level or more is required, the recognition target to be processed can be made to perform the fastest character recognition that can maintain the recognition accuracy of that level or more.

In the character recognition device according to the present invention, when the recognition accuracy is the most important, the recognition target to be processed can be subjected to the highest possible character recognition, and the recognition accuracy is also important. However, when a recognition speed higher than a certain level is required, the recognition target to be processed can be made to perform the highest-accuracy character recognition capable of maintaining the recognition speed higher than that level.

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of a character recognition device of the present invention.

FIG. 2 is a block diagram showing an example in which the character recognition device of FIG. 1 is realized by software on a computer.

FIG. 3 is a flowchart showing a part of a control flow of a system control unit.

FIG. 4 is a flowchart showing a part of a control flow of a system control unit.

FIG. 5 is a flowchart showing a part of a control flow of a system control unit.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 recognition means 2 independent dictionary 3 integration dictionary 4 recognition control means 5 system control means 6 integration dictionary creation / update means 7 collected data storage means 8 configuration control data storage means 9 instruction input means 100 CPU 101 memory 102 auxiliary storage device 104 input interface 105 output device P1 recognition program P4 recognition control program P5 system control program P6 dictionary for creating and updating integration

Claims (6)

[Claims] 1. A character recognition device having a plurality of recognition means and capable of selecting two or more recognition means and using them in an integrated manner,
Each recognizing means has, in addition to its own dictionary, a dictionary for integrated use with other recognizing means, and when two or more recognizing means are used integrally, the two or more recognizing means are used. A character recognition apparatus, wherein the main recognition means uses a dictionary specific to the main recognition means, and the other sub-recognition means uses a dictionary for integrated use with the main recognition means. 2. A means for collecting data samples that are misrecognized or not recognized by a main one of the two or more recognition means used in an integrated manner, and the two or more recognition means utilizing the collected data samples. 2. A character recognition apparatus according to claim 1, further comprising means for creating or updating a dictionary for integrated use of the secondary recognition means in the means with the primary recognition means. 3. A plurality of recognition means, a plurality of independent dictionaries specific to each of the plurality of recognition means, and a plurality of integrations for integrated use of each of the plurality of recognition means with each other recognition means. Means for instructing priority of recognition speed or recognition accuracy, means for performing character recognition by solely using one recognition means selected when priority of recognition speed is instructed, A means for performing character recognition by the integrated use of two or more recognition means selected when priority of accuracy is instructed, and a dictionary used by the recognition means for character recognition by single use of one recognition means. In the case of character recognition by means of selecting an independent dictionary specific to the recognition means and the integrated use of two or more recognition means,
A single dictionary unique to the main recognition means is selected as a dictionary used by one main recognition means of the one or more recognition means, and a dictionary used by each of the sub recognition means is selected as a dictionary used by the other sub recognition means. Means for selecting an integration dictionary for integrated use with a main recognition means. 4. It has a means for indicating a recognition target,
4. The character recognition device according to claim 3, wherein a recognition unit used alone or integrated is selected according to the specified recognition target. 5. A device for instructing a recognition target and a means for instructing a recognition accuracy level, wherein when priority is given to the recognition speed, an instruction is given when the recognition accuracy level is instructed. The recognition means with the highest recognition speed is selected from among the recognition means that satisfies the specified recognition accuracy level for the recognized recognition object, and the recognition means with the highest recognition accuracy for the specified recognition target when the recognition accuracy level is not specified. The character recognition device according to claim 3, wherein is selected. 6. A device for instructing a recognition target and a means for instructing a recognition speed level, wherein when priority is given to recognition accuracy, an instruction is given when the recognition speed level is instructed. The combination of the recognition means with the highest recognition accuracy among the recognition means satisfying the specified recognition speed level for the recognized recognition target is selected, and when the recognition speed level is not specified, the recognition accuracy for the specified recognition target is the highest. 4. The character recognition device according to claim 3, wherein a combination of recognition means is selected.