JPS6293776A - Information recognizing device - Google Patents

Abstract

PURPOSE:To register a feature of a handwritten information pattern whose recognition is difficult, by inputting said pattern from an input device so that it can be recognized continuously and making it correspond to the recognized result, in case the handwritten information pattern whose recognition is difficult as they are is a cursive character or a continuously written character, a peculiar character, an abbreviation, etc. CONSTITUTION:Although 'nin' (recognition) is written in a character input frame 8, and 'shiki' (identification) is written in a character input frame 9, this 'nin' is not recognized exactly since it is a cursive character, and '?' is displayed on a display part 7. 'nin' of a cursive character of the character input frame 8 is registered so that 'nin' of the cursive character which is written in the character input frame 8 can be recognized as 'nin' of a standard pattern. That is to say, 'nin' is written carefully by a square style character in the character input frame 8, and after confirming that it has been recognized exactly, a learning mode selecting part 6 is inputted with an input pen 2. Next, 'nin' of a square style character which can be recognized is written again in the character input frame 8, and a feature of 'nin' of the cursive character is registered in a dictionary part 16 by executing a pen-touch to the learning mode selecting part 16 with the input pen 2.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an information recognition device that recognizes information such as handwritten characters and symbols.

[Prior art] Various information recognition methods have been proposed for conventional information recognition devices, and there are reports that if the measures are carefully written, a recognition rate of 99% or more can be achieved in real time. When you enter handwritten characters that are difficult to recognize, enter the character code to register the handwritten characters that are difficult to recognize. ＝I was doing it.

[Problems to be solved by the invention] However, this type of conventional device still has severe restrictions on handwriting, and the next step is to recognize characters, cursive, abbreviations, nine characters unique to individuals, etc. in real time, and achieve a high recognition rate. We have not yet reached that stage.

Therefore, characters written by the user are sometimes not recognized correctly because of the following characters, cursive characters, 9 characters, and abbreviations, and may be misrecognized or rejected. However, even in such cases, most characters can often be recognized correctly if the user carefully rewrites the characters.

By the way, in an information recognition device that extracts the characteristics of input handwritten characters, symbols, figures, etc. and compares them with standard patterns registered in the dictionary section and recognizes them, the continuation of an individual's character is a letter or 9 letters. , or abbreviations, etc., it is conceivable that the user registers these character patterns in the dictionary section in correspondence with the corresponding character codes. That is, it is conceivable that the information recognition device is equipped with a learning function.

However, in the conventional technology, it is not easy for a user to register a character pattern. Figure 5 shows a conventional example of this type.
51 is a tablet, 52 is an input bench, 53 is a character input section, and 54 is a character code input section. In this example, the user writes an input cover pattern in the area of the character input section 53 using the input pen 52, and also inputs a corresponding character code from the character code input section 54 using the input pen 52 to register. For example, by using the input ben 52 to write ``recognition'' in cursive in the character input section 53, and inputting the character code 4727 corresponding to the standard character pattern ``recognition'' in the character code input section. , it is possible to register the cursive character "EI" to be recognized as the standard character pattern "Ken". However, such character codes are not familiar to general users, and it is therefore troublesome for users to look up the character codes in a character code table or the like.

FIG. 6 is a conceptual diagram expressing the correspondence between the character code 4727 and the character pattern 21 magazine in the dictionary section.

The purpose of the present invention is to
According to the present invention, the user can register the user's characters, symbols, or graphic patterns through simple operations. An object of the present invention is to provide an information recognition device.

[Means for solving the problem] As a means for solving this problem, for example, the information recognition device of the embodiment shown in FIG. 1 temporarily stores the characteristics of the handwritten information pattern input from the input device. registering the features of the first handwritten information pattern inputted from the input device with the recognition results of the second handwritten information pattern inputted from the human blade device in a dictionary section; registration selection means for selecting whether or not to register.

[Operation] In the above configuration, when the handwritten information pattern that is difficult to recognize as it is is in cursive, continuous letters, abbreviations, abbreviations, etc., it is input from the input device and corresponds to the recognition result so that it can be continuously recognized. Then, the characteristics of handwritten information patterns that are difficult to recognize are registered in the dictionary section.

[Embodiments] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is an external view of the information recognition device of this embodiment. In the figure, 2 is an input pen, 3 is a character input area, 5 is an information recognition device, 6 is a learning mode selection section, and 7 is a display section. be.

When the user writes characters, symbols, figures, etc. in the character input frame of the character input area 3 with the input pen 2, the knowledge result is displayed on the display section 7.
will be displayed. Furthermore, by touching the learning mode selection section 6, it is possible to register it in the dictionary section.

FIG. 2 is a configuration diagram of this embodiment, in which 10 is an input device, 1
1 is a function determination unit, 12 is a feature extraction unit, 13 is a temporary storage buffer unit, 14 is a learning pattern buffer unit,
15 is a recognition section, 16 is a dictionary section, 17 is a code buffer section, 18 is an output section, and 19 is a learning processing section.

FIG. 3 is a flowchart showing the processing procedure of this embodiment. Among the steps, steps S1 and S2 are performed by the function determination section 11, steps 53 to S7 are performed by the feature extraction section 12, steps S8 and S9 are performed by the recognition section 15, and step SIO is performed by the learning processing section 19.

The processing procedure in this embodiment will be explained based on FIG. Information input from the input device 10 is sent to the function determination section 11. The function determination unit 11 determines whether the input information is on the character input area 3 in step S1, and if the input information is on the character input area 3, the function determination unit 11 accumulates information on the person's cover pattern and stores it in the feature extraction unit. 12 and proceeds to step S3. In step S3, the feature extraction unit 12 extracts the characteristics of the human kabataan, and in step S4, determines the position of the character input frame in which the input kabataan has been written, and determines the position of the character input frame in which the input kabataan has been previously written. When the character, symbol, figure, etc., are stored in the temporary storage buffer section 13, it is determined whether the characteristics of the character, symbol, figure, etc. are stored in the temporary memory buffer section 13. In step S5, if the feature is stored in the temporary storage buffer section 13, the process proceeds to step S6, where the feature is copied to the learning pattern buffer 14, and in step S7, the feature extracted in step S3 is stored in the temporary storage buffer section 13. to be memorized. In step S5, if the data is not stored in the temporary storage buffer unit 13, the process directly advances to step S7. After that, the feature extraction section 12 sends the features to the recognition section 15, and the process proceeds to step S8. In the recognition unit 15,
The features of the human kataan are recognized by comparing them with the standard patterns in the dictionary section 16, and the recognition results are stored in the code buffer section 17 and sent to the output section 18, and the process returns to the start.

In step S1, if the input information is not on the character input area 3, the process proceeds to step S2, and if it is on the learning mode selection section 6, the process proceeds to step SIO, where the input information is set to the characteristics of the input pattern stored in the learning pattern buffer section 14. , and the recognition results stored in the code buffer section 17 are registered in the dictionary section 16 in correspondence with each other, and the process returns to the start.

FIGS. 4(a) and 4(b) show an operation example of this embodiment. The user proceeds with the operations in the order shown in FIGS. 4(a) and 4(b). In FIG. 4(a), "certification" is written in the character input box 8,
Although the character ``Ki'' is written in the character input box 9, it is not recognized correctly because it is in cursive, and ``?'' is displayed on the display section 7. In Fig. 4(b), the cursive ``Yetsu'' in the character input box 8 is registered so that the cursive ``Ken'' written in the character input box 8 is recognized as the standard pattern ``Ki''. . That is, after carefully writing ``recognize'' in the character input box 8 using the measure method and confirming that it has been correctly recognized, the user enters the input into the learning mode selection section 6 using the input bezel 2.

As is clear from the flowchart in Figure 3, character input frame 8
This time, write down the recognizable ``recognizable'' font again, and touch the learning mode selection area 6 with the input pen 2 to write ``recognize'' in the cursive font in the character input frame 8 in FIG. 4(a). The characteristic is that the character code 4727, which is the standard pattern of the recognition result of the measure law in Figure 4 (b), is paired with the dictionary section 1.
6 will be registered.

In addition, as for the order of designation of the learning mode, when a handwritten information pattern that is difficult to recognize is input from the input device, the learning mode is selected, and then input correctly using the input device 2 and recognized.
A handwritten information pattern that is difficult to recognize may be registered in the dictionary section 16 by associating its features with the recognized pattern.

In the above example, the method of specifying the learning mode is to touch the learning mode selection section 6 with the input bezel 2. You may specify it in writing or by voice.

Further, when the recognition unit 15 outputs not one candidate but a plurality of candidates, a means for allowing the user to select from among the plurality of candidates may be provided, and the candidate selected by the user may be used as the recognition result.

As is clear from the above description, in this embodiment, the user can register the characteristics of the user's writing patterns such as characters, symbols, figures, etc. in the dictionary section with a simple operation.

[Effects of the Invention] As described above, according to the present invention, writing patterns such as letters, symbols, or figures can be changed by simply re-inputting them using accurate knowledge that the user naturally possesses. can be registered in the dictionary section.

Therefore, registration can be done much more efficiently than searching for character codes in a code table.

[Brief explanation of drawings]

FIG. 1 is an external view of an embodiment according to the present invention, FIG. 2 is a block diagram of the embodiment, FIG. 3 is a flowchart of the control section of the embodiment, and FIG.
a) and (h) are diagrams showing the operating procedure of this embodiment, Figure 5 is an external view of a conventional device, and Figure 6 is a diagram showing the relationship between character patterns and character codes in the dictionary section of the conventional device. be. In the figure, 2... input pen, 3... character input section, 5...
・Information recognition device main body, 6...Learning mode selection section, 7.
... Representation section, 10... Input device, 11... Function determination section, 12... Feature extraction section, 13... Temporary storage no. , 15... recognition section, 16... dictionary section, 17...
・Code buffer section, 18... Output section, 19... Learning processing section.

Claims (3)

[Claims] (1) In an information recognition device that extracts the features of a handwritten information pattern input from an input device, compares it with a standard pattern registered in a dictionary section, and outputs the recognized recognition result to a predetermined output device, the input a storage means for temporarily storing characteristics of the handwritten information pattern input from the device;
Associating the characteristics of the first handwritten information pattern input from the input device with the recognition result of a second handwritten information pattern input at the same position of the input device as the input device where the first handwritten information pattern was input. and registration selection means for selecting whether or not to register the information in the dictionary section. (2) Registration is carried out by inputting a first handwritten information pattern and a second handwritten information pattern using an input device, and then selecting registration using a registration selection means to transfer the characteristics of the first handwritten information pattern to the second handwritten information pattern. 2. The information recognition device according to claim 1, wherein the recognition result of the handwritten information pattern is registered in the dictionary section in association with the recognition result of the handwritten information pattern. (3) Registration is performed by inputting a first handwritten information pattern from the input device, selecting registration using the registration selection means, inputting a second handwritten information pattern from the input device, and then inputting the first handwritten information pattern from the input device. 2. The information recognition device according to claim 1, wherein features of a handwritten information pattern are registered in association with a recognition result of the second handwritten information pattern.