JP4385536B2 - Character recognition device, character recognition method, character recognition program, and computer-readable recording medium recording the same - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a character recognition device. Specifically, it is a character recognition device that can recognize a font simultaneously with a character code, and can perform character recognition with high accuracy by simply adapting to the use environment.
[0002]
[Prior art]
In the conventional character recognition device, the identification dictionary used for character pattern identification uses a plurality of typical fonts to perform stable character recognition in response to variations in character patterns by fonts, fluctuations such as blurring and collapse. It was created by averaging feature values (usually one each of serif, sans serif, and monospace).
[0003]
On the other hand, a character recognition device using an identification dictionary for each font created with a single font feature amount has been proposed (Japanese Patent Laid-Open No. 11-85908). Such an apparatus can recognize fonts at the same time as character codes, and has a higher recognition rate for fonts associated with an owned identification dictionary than using an identification dictionary in which the features of a plurality of fonts are averaged. There is an advantage that it can be obtained.
[0004]
On the other hand, however, there is a drawback in that the identification rate decreases for fonts other than those for the identification dictionary owned. Therefore, in order to correspond to various usage environments, it is necessary to prepare a large number of font-specific dictionaries, which is costly. In addition, if the user designates a font to be recognized in advance, the operation becomes complicated, If the configuration is such that the font is automatically recognized, the processing becomes extremely complicated. Furthermore, since there are a wide variety of font types, it is difficult to prepare all the identification dictionaries in reality, and there is a problem that it is not possible to cope with an environment using a special font.
[0005]
[Problems to be solved by the invention]
The present invention has been made in view of the above-described problems of the prior art, and the object of the present invention is to perform font recognition simultaneously with a character code, and to easily perform high-precision character recognition adapted to the use environment. An object of the present invention is to provide a character recognition device that can be used.
[0006]
[Means for Solving the Problems]
The above object of the present invention is achieved by the following means.
[0007]
(1) Font detection means for detecting an outline font incorporated in a connected device, and outline font detected by the font detection means Is a font for which no identification dictionary has been created, Character image data acquisition means for acquiring, from the connected device, character image data obtained by rasterizing in the connected device to a certain size for each character based on outline font data for all characters, and the character image Using the character image data acquired by the data acquisition means, a standard pattern creation means for creating a standard pattern for each character of the detected font and a standard pattern created by the standard pattern creation means are registered to create an identification dictionary A character recognition apparatus comprising: an identification dictionary creating means for determining a character code and a font of a character pattern using the identification dictionary created by the identification dictionary creating means.
[0008]
(2) Font detection means for detecting an outline font incorporated in the connected device, and outline font detected by the font detection means Is a font for which no identification dictionary has been created, Character image data acquisition means for acquiring, from the connected device, character image data obtained by rasterizing in the connected device to a certain size for each character based on outline font data for all characters, and the character image Using the character image data obtained by the data obtaining means, the standard pattern creating means for creating a standard pattern for each character of the detected font, and the standard pattern created by the standard pattern creating means is registered and detected. An identification dictionary creating means for creating an identification dictionary for each font; and a character identifying means for determining a character code and a font of a character pattern using the identification dictionary for each font created by the identification dictionary creating means. Character recognition device.
[0009]
(3) The character recognition device according to (2), wherein the character identification unit uses a character code of a standard pattern having a similarity equal to or higher than a predetermined threshold as the character code of the character pattern.
[0010]
(4) The character recognition device according to (2) or (3), wherein the character identification means determines the priority order of the identification dictionary in order of identification frequency.
[0011]
(5) The character identification means includes: As a font for each character block that is divided into columns, blank lines, or separated by line breaks, The font of the identification dictionary with the highest identification frequency is assigned to each character block. In (2) to (4) characterized by being determined Any one of The character recognition device described.
[0012]
(6) a step of detecting an outline font incorporated in the connected device, and the detected outline font; Is a font for which no identification dictionary has been created, For all characters, obtaining character image data obtained by rasterizing in the connected device to a certain size for each character based on outline font data from the connected device, and using the character image data A step of creating a standard pattern for each character of the detected font, a step of registering the standard pattern to create an identification dictionary, and a step of determining a character code and font of the character pattern using the identification dictionary; And a character recognition method.
[0013]
(7) a step of detecting an outline font incorporated in the connected device, and an outline font detected from the connected device Is a font for which no identification dictionary has been created, For all characters, obtaining character image data obtained by rasterizing in the connected device to a certain size for each character based on outline font data from the connected device, and using the character image data A step of creating a standard pattern for each character of the detected font; a step of creating an identification dictionary for each detected font by registering the standard pattern; and a character pattern using the identification dictionary for each font Determining a character code and a font of the character recognition method.
[0014]
(8) Procedure for detecting an outline font incorporated in a connected device, and the detected outline font Is a font for which no identification dictionary has been created, For all characters, using the character image data, a procedure for acquiring character image data obtained by rasterizing in the connected device to a certain size for each character based on the outline font data, and the character image data A procedure for creating a standard pattern for each character of the detected font, a procedure for registering the standard pattern and creating an identification dictionary, and a procedure for determining a character code and font of the character pattern using the identification dictionary; A program for causing a computer to execute.
[0015]
(9) A procedure for detecting an outline font incorporated in a connected device, and an outline font detected from the connected device Is a font for which no identification dictionary has been created, For all characters, using the character image data, a procedure for acquiring character image data obtained by rasterizing in the connected device to a certain size for each character based on the outline font data, and the character image data A procedure for creating a standard pattern for each character of the detected font; a procedure for creating an identification dictionary for each detected font by registering the standard pattern; and a character pattern using the identification dictionary for each font A program for causing a computer to execute a procedure for determining a character code and a font.
[0016]
(10) A procedure for detecting an outline font incorporated in a connected device, and the detected outline font Is a font for which no identification dictionary has been created, For all characters, using the character image data, a procedure for acquiring character image data obtained by rasterizing in the connected device to a certain size for each character based on the outline font data, and the character image data A procedure for creating a standard pattern for each character of the detected font, a procedure for registering the standard pattern and creating an identification dictionary, and a procedure for determining a character code and font of the character pattern using the identification dictionary; , A computer-readable recording medium on which a program for causing a computer to execute is recorded.
[0017]
(11) A procedure for detecting an outline font incorporated in a connected device, and an outline font detected from the connected device Is a font for which no identification dictionary has been created, For all characters, using the character image data, a procedure for acquiring character image data obtained by rasterizing in the connected device to a certain size for each character based on the outline font data, and the character image data A procedure for creating a standard pattern for each character of the detected font; a procedure for creating an identification dictionary for each detected font by registering the standard pattern; and a character pattern using the identification dictionary for each font A computer-readable recording medium on which a program for causing a computer to execute a procedure for determining a character code and a font is recorded.
[0018]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0019]
FIG. 1 is a block diagram showing an overall configuration of a character recognition system including a character recognition device according to an embodiment of the present invention. The character recognition system includes a character recognition device 1, a scanner 2 as an image reading device, a personal computer 3 as an image processing device, and a printer 4 as an image forming device, which are mutually connected via a network 5. It is connected so that it can communicate. The type and number of devices connected to the network 5 are not limited to the example shown in FIG.
[0020]
Next, the configuration of each device will be described, but portions having similar functions in each device will be described only for the first time in order to avoid duplication of description, and description thereof will be omitted for the second and subsequent times.
[0021]
FIG. 2 is a block diagram showing an example of the configuration of the character recognition device 1 according to the present embodiment. The character recognition device 1 is a computer, and includes a CPU 11, a ROM 12, a RAM 13, a hard disk 14, a display 15, an input device 16, a network interface 17, and a bus 18, as shown in FIG.
[0022]
The CPU 11 performs various processes of control and calculation. The ROM 12 stores various programs and data. The RAM 13 temporarily stores programs and data as a work area. The hard disk 14 stores various programs and data. The display 15 performs various displays. The input device 16 is a keyboard, a mouse, or the like, and is used for performing various inputs. The network interface 17 is an interface for connecting to a network and communicating with other devices on the network. The above-described units are connected to each other via a bus 18 for exchanging signals. In the present embodiment, the computer 1 performs a predetermined operation to be described later. A program for controlling the operation of the computer 1 for this purpose is stored in the ROM 12 or the hard disk 14.
[0023]
The scanner 2 has a function of reading an original set at a predetermined position to acquire image data (bitmap data) and outputting the image data to another device via a network.
[0024]
The personal computer 3 has the same configuration as the character recognition device 1 as shown in FIG. The personal computer 3 incorporates a font for forming a character image on an image forming apparatus such as a printer or a display as an image processing apparatus.
[0025]
The printer 4 prints out a print image (image data) based on the print job. When the print job includes a control language, the printer 4 has a function of interpreting and rasterizing the print job to create a print image. Fonts necessary for this purpose are incorporated in the printer 4.
[0026]
The network 5 includes a LAN in which computers and network devices are connected according to standards such as Ethernet (registered trademark), token ring, and FDDI, and a WAN in which LANs are connected by a dedicated line.
[0027]
Next, an outline of the operation of the character recognition device 1 in the present embodiment will be described. FIG. 3 is a flowchart showing an overall procedure of character recognition processing of the character recognition device 1 according to the present embodiment. First, the character recognition device 1 displays an input screen for a new identification dictionary creation command on the display 15 and waits until a user inputs a new creation command for an identification dictionary (NO in S101). The user operates the input device 16 to input a new identification dictionary creation command to the character recognition device 1. When the character recognition apparatus 1 receives an input of a new identification dictionary creation command from the user (YES in S101), the character recognition device 1 performs a creation process of the identification dictionary (S102).
[0028]
FIG. 4 is a flowchart showing a procedure of identification dictionary creation processing of the character recognition device 1 according to the present embodiment. In the identification dictionary creation process, the character recognition device 1 first detects a device connected to the character recognition device 1 via the network interface 17 and the network 5 (S201). When a connected device is detected on the network 5 (YES in S202), it is detected whether there is a font incorporated in the detected device (S203). For example, in the case of a personal computer, an embedded font is held by an OS (operating system), and font data is held in a specific directory used by the OS in the hard disk. In the case of a printer that supports page description languages such as Postscript (Adobe Systems Inc.), the printer has fonts. The printer interprets the page description language and selects fonts based on the font data stored on the hard disk. Rasterize. The character recognition device 1 searches a specific directory such as a hard disk of these devices and detects such font data, thereby detecting a font incorporated in the device. When a font incorporated in the detected device is detected, and the detected font has not yet created an identification dictionary (YES in S204), it is sent to the device via the network interface 17 and the network 5. Then, a character image data transfer request for requesting transfer of character image data (bitmap data) for all characters of the detected font is transmitted (S205).
[0029]
FIG. 5 shows that the character recognition device 1 detects the personal computer 3 on the network 5, further searches a predetermined directory on the hard disk of the personal computer 3 to detect font data, and detects the font data related to the detected font data on the personal computer 3. 6 is a flowchart showing the operation of the personal computer 3 when a character image data transfer request is transmitted. The personal computer 3 stands by until a character image data transfer request is received from the character recognition device 1 (NO in S301). When the character image data transfer request is received from the character recognition device 1 (YES in S301), the transfer request is received. Prepare character image data for all characters of the font. Here, there are two types of fonts: bitmap fonts that represent character shapes as a set of dots, and outline fonts that characterize outlines of characters. Bitmap fonts consist of character image data itself, whereas outline fonts include font metric data describing character width and height information and character outlines as font data. Outline data in which information for calculation is described, and character image data is obtained by rasterizing to a necessary character size based on these data. Therefore, if the font for the transfer request is a bitmap font (NO in S302), the character image data, which is the font data for each character, is transmitted as it is to the character recognition device 1, and the font for the transfer request is outlined. If it is a font (YES in S302), each character is rasterized to a certain size based on the font data (S303), and the obtained character image data is transmitted to the character recognition device 1.
[0030]
If the device related to the character image data transfer request is the printer 4, the font incorporated in the printer is an outline font. Therefore, the printer 4 receives the character image data transfer request reception procedure (S 301) in FIG. The operation is sequentially performed according to the processing procedure (S303) and the character image data transfer procedure (S304). If the character recognition device 1 itself incorporates a font as an image processing device, it treats itself as one of the connected devices, and follows the above-described font detection procedure (S203) and rasterization processing procedure (S303). It operates sequentially.
[0031]
In FIG. 4, the character recognition device 1 waits until the character image data is received from the device via the network 5 and the network interface 17 (NO in S206), and receives the character image data from the device (S206). YES), the received character image data is stored in a predetermined directory of the hard disk 14. Next, a standard pattern is created for all characters of the font using the received character image data (S207), and the created standard pattern is registered in the dictionary to create an identification dictionary for the font (S208). Here, the identification dictionary creation method differs depending on the character recognition method. In the case of the method using feature amount extraction, the feature amount is calculated by a predetermined method from the received character image data of each character, and this is registered as a standard pattern of each character. As a specific method for extracting the feature amount, there is a method of extracting the inclination of the character line, the number of loops, the line width, the character area, etc. as the extraction amount. In the case of the pattern matching method, the received character image data of each character is registered as it is as the standard pattern of each character. However, in order to deal with the positional deviation, inclination, minute size difference, etc. of the character pattern to be identified, character image data in which characters are blurred to some extent from the received character image data is created and registered as a standard pattern. There may be. The created font identification dictionary is stored in a predetermined directory of the hard disk 14.
[0032]
The character recognition device 1 continues to detect other fonts incorporated in the device (S203), and if there is one that does not create an identification dictionary (YES in S204), the above-described detection font identification dictionary creation is performed. An identification dictionary for each detected font is created by repeating the steps (S205 to S208). On the other hand, if there is no font installed in the device or if any of the built-in fonts has already created an identification dictionary (NO in S204), another connected device is detected (S201). If another connected device is detected (YES in S202), an identification dictionary for each detected font is created by repeating the above-described font detection procedure and detection font identification dictionary creation procedure (S203 to S208). To do. If no other connected device is detected (NO in S202), the identification dictionary creation process ends.
[0033]
In FIG. 3, when the identification dictionary creation process is completed, the character recognition device 1 displays an input screen for an original reading command on the display 15 and waits for an input of an original reading command from the user (NO in S103). A user operates the input device 16 to input a document reading command to the character recognition device 1. When a user inputs a document reading command (YES in S103), the character recognition device 1 transmits the document reading command to the scanner 2 via the network interface 17 and the network 5 (S104). Upon receiving a document reading command from the character recognition device 1, the scanner 2 reads a document document set at a predetermined reading position, and transmits the obtained image data to the character recognition device 1. The character recognition device 1 waits until image data is received from the scanner 2 (NO in S105). When image data is received from the scanner 2 via the network 5 and the network interface 17 (YES in S105), the received image data is received. Are stored in a predetermined directory of the hard disk 14.
[0034]
Next, a character area and a non-character area included in the received image data are determined, and the image data is separated into a character area and a non-character area (S106). In this process, by using the brightness histogram of the minute area of the image data and the frequency-resolved spectrum characteristics, the character area in the original image data and the area other than the characters such as photographs and figures are discriminated. Cut out the character area. The cut image data of the non-character area is temporarily stored in a predetermined directory of the hard disk 14 for output together with the character code data recognized later. At this time, if necessary, the image data in the non-character area may be subjected to processing such as image correction such as a smoothing filter, resolution change, and compression. In addition, for the generated image data of only the character region, image correction by filter processing may be performed so that a collapsed character or a blurred character does not occur if necessary.
[0035]
Further, pre-character recognition processing is performed on the image data of only the character area (S107). That is, since character recognition is performed using monochrome binary image data, in this process, color image data displayed in RGB, Lab, or the like is converted into monochrome binary image data in which the character has a black background as white. . Further, if necessary, the generated monochrome binary image data is subjected to processing such as noise removal of isolated points and image inclination correction.
[0036]
Next, layout analysis is performed on the monochrome binary image data obtained by the character recognition preprocessing (S108). In this process, a line structure and a column of characters are detected using a histogram of the projection image for the character region in the image, and the character region is recognized as a plurality of character blocks. Here, the character block refers to a character area that is divided by a column, a blank line, a line feed, etc. and divided into about a paragraph. A character block is also formed when the character size is different.
[0037]
Further, the character blocks recognized by the layout analysis are sequentially cut out of lines and characters (S109), and character recognition processing is performed on the cut out character image data (S110). The circumscribed rectangle data of each extracted character image is stored in a predetermined directory of the hard disk 14 as position data of each character.
[0038]
FIG. 6 is a flowchart showing a procedure of character recognition processing of the character recognition device 1 according to the present embodiment. In the character recognition processing procedure, the character recognition device 1 first selects one of the character blocks recognized in the layout analysis (S401), cuts out a line from the character block, and cuts out a character from the line ( S402).
[0039]
Next, the feature amount is extracted for the character pattern related to the extracted character (S403). The extraction of the feature amount is performed by the same method used when creating the standard pattern of each character registered in the identification dictionary in the standard pattern creation procedure (S207) of the identification dictionary creation process. Further, in the standard pattern creation procedure (S207) of the identification dictionary creation processing, when the method of pattern matching is used as the character recognition method instead of the feature amount extraction method, the character pattern itself relating to the cut-out character is the standard. Since it is compared with the pattern, the character pattern feature extraction procedure (S403) is omitted.
[0040]
Further, an identification dictionary used for identification processing is selected according to the priority order (S404), and is read from the hard disk 14 onto the RAM 13. The priority order is determined in accordance with a procedure for determining the priority order of the identification dictionary described later (S408). However, only the first time, the default priority, or a random order is used.
[0041]
Then, an identification process is performed in which the feature amount extracted from the character pattern is compared with the standard pattern of each character registered in the selected identification dictionary (S405). That is, in the identification process, the similarity between the feature amount extracted from the character pattern and the standard pattern of each character registered in the selected identification dictionary is sequentially calculated, and the similarity is equal to or higher than a predetermined threshold value. If there is a standard pattern (YES in S406), the character code of the standard pattern is determined as the character code of the character pattern, and the character code data is stored in a predetermined directory of the hard disk 14 (S407). Examples of the similarity calculation method used for the identification processing include a method based on a city block distance, a method based on a Euclidean distance, a method based on a linear identification function, a subspace method, and a Bayes identification method. If there is no standard pattern whose similarity is equal to or higher than the threshold value in the selected identification dictionary (NO in S406), the identification dictionary is sequentially changed according to the priority order until a standard pattern whose similarity is equal to or higher than the threshold value is found. The character code of the character pattern is determined by selecting again and repeating the above identification processing procedure (S404 to S406).
[0042]
Next, the font of the identification dictionary used to determine the character code is set as a font candidate for the character pattern, and the font candidate data is stored in a predetermined directory on the hard disk 14 (S408). Further, the font candidate data stored in the hard disk 14 is analyzed, and the priority order of the identification dictionary corresponding to each font is determined and updated according to the order of the frequency of the font candidates (S409). Further, the font size is calculated from the character pattern, and the font size data is stored in a predetermined directory of the hard disk 14 (S410).
[0043]
After repeating the series of character recognition processes (S402 to S411) for the character patterns of all the characters cut out from the selected character block (YES in S411), the font for the first priority identification dictionary The font of the character block is determined, and the font type data is stored in a predetermined directory on the hard disk 14 (S408). The series of character recognition processing (S401 to S412) is repeated for all character blocks recognized in the layout analysis, and the character recognition processing is terminated (YES in S413).
[0044]
In the identification processing procedure (S405 and S406), the calculation of similarity is terminated when a standard pattern having a similarity with a feature value extracted from the character pattern is equal to or greater than a predetermined threshold is found. However, the similarity with all the standard patterns of all the identification dictionaries is calculated, and the priority of the identification dictionaries according to the standard patterns in descending order of similarity is calculated in the order of updating the priority of the identification dictionaries (S409). The order may be determined. In addition, in the method for calculating the similarity to all the standard patterns in all such identification dictionaries, when two or more standard patterns having a similarity equal to or higher than a predetermined threshold are found, the spelling of the preceding and following characters You may make it employ | adopt the character code of the direction which checks and forms the word in a word dictionary. Furthermore, in the above-described character block font determination procedure (S408), the most frequently used identification dictionary font is used as the font of the character block regardless of the character code. The code may be weighted to determine the font of the character block.
[0045]
In FIG. 3, when the character recognition process is completed, the character recognition device 1 recognizes the color of the character cut out by the character recognition process and stores the obtained character color data in a predetermined directory of the hard disk 14. (S110). The character color recognition process is performed by reading color data such as RGB and Lab of pixels corresponding to the character of the original image data and obtaining an average value of all pixels corresponding to the character. The distinction between the characters and the background pixels is performed by using the black and white binary image data created by the character recognition pre-processing and setting the black pixels in the circumscribed rectangle of each character as the character pixels and the white pixels as the background pixels.
[0046]
Then, an output file is created from the character code data obtained in the character recognition procedure (S109) and the image data of the non-character region obtained in the character / non-character region separation procedure (S106) (S111). That is, the character code data and the image data of the non-character area are read from the hard disk 14 onto the RAM 13 and converted into a predetermined output file format. Further, the font type data obtained in the character recognition process (S109) and The font color and the character color data obtained in the character color recognition procedure (S110) are read from the hard disk 14, added to the character code data as character attribute data, and obtained in the character recognition processing procedure (S109). It is arranged together with the image data of the non-character area based on the character position data. The created output file is stored in a predetermined directory of the hard disk 14 or is output to other devices such as the personal computer 3 and the printer 4 via the network interface 17 and the network 5.
[0047]
In the above embodiment, the identification dictionary creation process (S102, S201 to S208) is performed every time during the character recognition process, and the identification dictionary is updated. Alternatively, the identification dictionary creation process may be performed only when there is a change in the connection environment.
[0048]
Further, in the above embodiment, an identification dictionary for each detected font is created and character recognition is performed using this. However, by default, the feature quantities of a normal identification dictionary, that is, a plurality of fonts are averaged. When the recognition rate is low, that is, when the average similarity is not more than a predetermined value in the identification process, the identification dictionary for each detected font is used. It is good also as a structure. As a result, when a special font that cannot be recognized by the default identification dictionary is used in the environment of the character recognition system, etc., the detected font identification dictionary is used only for input data relating to the font, and other input data By using a default identification dictionary, the overall recognition rate can be improved.
[0049]
Further, in the above embodiment, a separate identification dictionary is created for each detected font. However, a standard pattern obtained from character image data of all characters of all detected fonts is stored in one identification dictionary. It may be configured to register and create an identification dictionary.
[0050]
In the above embodiment, the character recognition device 1, the scanner 2, the computer 3, and the printer 4 are connected to each other via the network 5, but the character recognition device 1 and the scanner 2, or the character recognition device. 1 and the printer 4 are locally connected to each other using a serial interface such as RS-232C, USB, or IEEE1394, a parallel interface such as SCSI or IEEE1284, a wireless communication interface such as Bluetooth, IEEE802.11, HomeRF, or IrDA. A connected configuration may be used.
[0051]
The character recognition device 1 according to the present invention has a multi-functional peripheral device such as a scanner having a character recognition function, a digital copier, a facsimile machine, etc. (MFP).
[0052]
The character recognition device and the character recognition method according to the present invention can be realized by a dedicated hardware circuit for executing each of the above procedures, or by a CPU executing a predetermined program describing each of the above procedures. Can do. When the present invention is realized by the latter, the predetermined program for operating the character recognition device can be provided by a computer-readable recording medium such as a floppy disk or a CD-ROM. In this case, the program recorded on the computer-readable recording medium is usually transferred and stored in a ROM, a hard disk or the like. Further, this program may be provided as, for example, a single application software, or may be incorporated in the software of the device as one function of the character recognition device.
[0053]
【The invention's effect】
As described above, according to the present invention, an outline font embedded in a connected device used in a usage environment of a character recognition device is automatically detected, and an outline font is detected. Is a font for which no identification dictionary has been created, For all characters, a character image obtained by rasterizing in the connected device to a certain size for each character based on the outline font data is obtained from the connected device and detected using this character image data. Character recognition is performed by creating an identification dictionary for each font, so that the font can be recognized simultaneously with the character code and higher recognition than when using an identification dictionary that averages the features of multiple fonts. Rate can be achieved.
[0054]
In addition, since an identification dictionary is created only for fonts that exist in the usage environment, highly accurate character recognition can be performed without preparing identification dictionaries for many fonts in advance.
[0055]
Furthermore, since an identification dictionary for the font is automatically created even in an environment where a special font is used, highly accurate character recognition can be easily performed by adapting to the environment.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an overall configuration of a character recognition system including a character recognition device according to an embodiment of the present invention.
FIG. 2 is a block diagram showing an example of the configuration of the character recognition device 1 according to the present embodiment.
FIG. 3 is a flowchart showing an overall procedure of character recognition processing of the character recognition device 1 according to the present embodiment.
FIG. 4 is a flowchart showing a procedure of identification dictionary creation processing of the character recognition device 1 according to the present embodiment.
FIG. 5 is a flowchart showing a procedure of character image data transfer processing of the personal computer 3 in the present embodiment.
FIG. 6 is a flowchart showing a procedure of character recognition processing of the character recognition device 1 according to the present embodiment.
[Explanation of symbols]
1 ... Character recognition device,
11 ... CPU
12 ... ROM
13 ... RAM
14 ... Hard disk
15 ... Display
16 ... Input device
17 ... Network interface
18 ... Bus
2 ... Scanner,
3 ... PC
4 ... Printer,
5 ... Network.

Claims (11)

Font detection means for detecting outline fonts embedded in the connected device;
When the outline font detected by the font detection means is a font for which an identification dictionary has not been created, all the characters are rasterized in the connected device to a certain size for each character based on the outline font data. Character image data acquisition means for acquiring the obtained character image data from the connected device;
Standard pattern creating means for creating a standard pattern for each character of the detected font using the character image data obtained by the character image data obtaining means;
An identification dictionary creating means for creating an identification dictionary by registering a standard pattern created by the standard pattern creating means;
A character recognition device comprising: character identification means for determining a character code and a font of a character pattern using the identification dictionary created by the identification dictionary creation means. Font detection means for detecting outline fonts embedded in the connected device;
When the outline font detected by the font detection means is a font for which an identification dictionary has not been created, all the characters are rasterized in the connected device to a certain size for each character based on the outline font data. Character image data acquisition means for acquiring the obtained character image data from the connected device;
Standard pattern creating means for creating a standard pattern for each character of the detected font using the character image data obtained by the character image data obtaining means;
An identification dictionary creating means for creating an identification dictionary for each font detected by registering a standard pattern created by the standard pattern creating means;
A character recognition apparatus comprising: character identification means for determining a character code and a font of a character pattern using the identification dictionary for each font created by the identification dictionary creation means.   The character recognition device according to claim 2, wherein the character identification unit uses a character code of a standard pattern having a similarity equal to or higher than a predetermined threshold as the character code of the character pattern.   The character recognition device according to claim 2 or 3, wherein the character identification means determines the priority of the identification dictionary in order of identification frequency.   The character identification means, for each character block, the font of the identification dictionary with the highest identification frequency is used for each character block as a font of each character block that is divided into columns, blank lines, or line breaks. The character recognition device according to claim 2, wherein the character recognition device is determined. Detecting outline fonts embedded in the connected device;
When the detected outline font is a font for which no identification dictionary has been created , characters obtained by rasterizing in the connected device to a certain size for each character based on the outline font data for all characters Acquiring image data from the connected device;
Using the character image data to create a standard pattern for each character of the detected font;
Registering the standard pattern and creating an identification dictionary;
Determining a character code and font of a character pattern using the identification dictionary;
A character recognition method characterized by comprising: Detecting outline fonts embedded in the connected device;
When the outline font detected from the connected device is a font for which an identification dictionary has not been created, all characters are obtained by rasterizing the connected device to a certain size for each character based on the outline font data. Obtaining the obtained character image data from the connected device;
Using the character image data to create a standard pattern for each character of the detected font;
Creating an identification dictionary for each font detected by registering the standard pattern;
Determining the character code and font of the character pattern using the identification dictionary for each font. A procedure to detect outline fonts embedded in connected devices,
When the detected outline font is a font for which no identification dictionary has been created , characters obtained by rasterizing in the connected device to a certain size for each character based on the outline font data for all characters A procedure for acquiring image data from the connected device;
A procedure for creating a standard pattern for each character of the detected font using the character image data;
Registering the standard pattern to create an identification dictionary;
A procedure for determining a character code and a font of a character pattern using the identification dictionary;
A program that causes a computer to execute. A procedure to detect outline fonts embedded in connected devices,
When the outline font detected from the connected device is a font for which an identification dictionary has not been created, all characters are obtained by rasterizing the connected device to a certain size for each character based on the outline font data. To obtain the obtained character image data from the connected device,
A procedure for creating a standard pattern for each character of the detected font using the character image data;
Creating an identification dictionary for each of the fonts detected by registering the standard pattern;
A procedure for determining a character code and a font of a character pattern using the identification dictionary for each font;
A program that causes a computer to execute. A procedure to detect outline fonts embedded in connected devices,
When the detected outline font is a font for which no identification dictionary has been created , characters obtained by rasterizing in the connected device to a certain size for each character based on the outline font data for all characters A procedure for acquiring image data from the connected device;
A procedure for creating a standard pattern for each character of the detected font using the character image data;
Registering the standard pattern to create an identification dictionary;
A procedure for determining a character code and a font of a character pattern using the identification dictionary;
The computer-readable recording medium which recorded the program for making a computer perform. A procedure to detect outline fonts embedded in connected devices,
When the outline font detected from the connected device is a font for which an identification dictionary has not been created, all characters are obtained by rasterizing the connected device to a certain size for each character based on the outline font data. To obtain the obtained character image data from the connected device,
A procedure for creating a standard pattern for each character of the detected font using the character image data;
Creating an identification dictionary for each of the fonts detected by registering the standard pattern;
A procedure for determining a character code and a font of a character pattern using the identification dictionary for each font;
The computer-readable recording medium which recorded the program for making a computer perform.

Images