JP2007033964A - Braille translation device, braille translation method, braille translation program, and computer readable recording medium having program recorded therein - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provided a braille translation device, a braille translation method, a braille translation program, and a computer readable recording medium having the program recorded therein, which allow a braille data string including input errors to be easily corrected. <P>SOLUTION: The braille translation device includes; a braille storage part 103 for storing an input braille data string 10; a candidate extraction part 105 for extracting braille pattern candidates on the basis of the braille data string 10; an error detection part 104 for detecting whether braille pattern candidates are erroneous or not by detecting whether they agree with preliminarily stored dictionary data or not; and a braille corrected candidate output processing part 107 for outputting a candidate braille data string of which the translation result agree with dictionary data, out of candidate braille data strings as a corrected candidate braille data string for the braille data string 10. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a Braille translation device, a Braille translation method, a Braille translation program, and a computer-readable recording medium on which the Braille translation program is recorded, and in particular, Braille that can be used to obtain information for a visually impaired person to obtain information. The present invention relates to a translation device, a Braille translation method, a Braille translation program, and a computer-readable recording medium on which the program is recorded.

  Braille is a very important character in communication between a healthy person and a visually impaired person. However, the grammar for writing in braille is difficult and difficult for a healthy person to understand. In view of this, devices have been developed that can translate braille into ink characters or translate ink characters into braille and present them.

For example, in Patent Document 1, a user inputs braille data using an input device having an arrangement that is in the same positional relationship as the points on the convex surface of the Braille to be translated, and translates the corresponding ink characters. A display device is disclosed. As an input device in this device, for example, a numeric keypad having a maximum of six keys consisting of three vertical rows and two horizontal columns constituting Braille is used, and the user selects a point corresponding to the Braille to be translated from the numeric keypad. Key in point by point.
Japanese Patent Laid-Open No. 11-305652

  In the device disclosed in Patent Document 1, when inputting braille information, the user remembers where the dots are arranged among a total of 6 points of 3 × 2 in length that constitute one character of Braille. Must be entered. For this reason, the user may erroneously input Braille information due to misunderstanding or the like.

  However, in Patent Document 1, even if an erroneous input is made, it is necessary for the user himself / herself to determine an input error by looking at the translation result and to correct it.

  The present invention has been made to solve the above-described problems, and an object thereof is to provide a Braille translation device, a Braille translation method, and a Braille translation method that can easily correct a Braille data string including an input error. To provide a translation program and a computer-readable recording medium on which the translation program is recorded.

  A Braille translation device according to an aspect of the present invention includes Braille storage means for storing a plurality of input Braille data, and each Braille data is arranged in a matrix-like point position within the Braille cell of the corresponding Braille character. And a candidate extracting means for extracting a braille pattern candidate based on a plurality of braille data, wherein the braille pattern candidate is an input braille represented by the plurality of braille data Candidate braille corresponding to a combination of convex portions of a braille character string configured by moving at least a part of the convex portion of the character string to another point position and corresponding to a braille pattern candidate extracted by the candidate extracting means For detecting whether or not the braille pattern candidate is erroneous by detecting whether or not the translation result of the data string matches the dictionary data stored in advance And Ri detection means, among the candidate braille data string, further comprising a candidate braille data sequence translation result matches the dictionary data, and a candidate output means for outputting the corrected candidate braille data strings for a plurality of braille data.

  Preferably, the error detection means includes translation means for translating the candidate Braille data string into the ink character data string based on the dictionary data, and the error detection means is not present in the dictionary data in the candidate Braille data string. When the above braille data is detected, it is determined that the candidate braille data string is erroneous.

  Preferably, the error detection means further includes a correct / incorrect determination means for determining the correctness / incorrectness of the ink character data string translated by the translation means based on the dictionary data, and the error determination means includes the dictionary data in the ink character data string. When one or more ink character data that does not exist is detected, it is determined that the ink character string is erroneous, and the error detecting means determines that the ink character string is erroneous by the ink character correctness determining means. In this case, it is determined that the candidate braille data string is erroneous.

  Preferably, the candidate extracting means is an isolated block that is an aggregate including at least one convex portion among the convex portions of the input braille character string, and is an aggregate of convex portions that are collectively movable to adjacent flat portions. An isolated block detecting means for detecting the image and a moving means for moving the convex portion by a predetermined number of point positions in units of isolated blocks, and the pattern of the Braille character string after being moved by the moving means Extract as pattern candidates.

  Preferably, the isolated block detecting means is an aggregate of adjacent convex portions within the braille cell of the braille character constituting the input braille character string, and the convex portions that can be moved collectively to the adjacent flat portions. Are detected as isolated blocks.

  Preferably, the isolated block detecting means detects whether or not there is an empty string having no convex part when the combination of convex parts of the input braille character string is divided for each column, and there is an empty string in between. A set of convex portions and flat portions included in one or more columns that are not to be extracted is extracted as an isolated block.

  Preferably, the isolated block detecting means detects whether or not there is a blank line having no convex part when the combination of convex parts of the input Braille character string is divided for each line, and there is a blank line between them. A set of convex portions and flat portions included in one or more rows that are not to be extracted is extracted as an isolated block.

  Preferably, the apparatus further includes candidate storage means for storing candidate braille data strings that match the dictionary data as correction candidate braille data strings, and the candidate output means outputs the correction candidate braille data strings stored in the candidate storage means. To do.

  Preferably, the correction candidate braille data string further includes likelihood calculating means for calculating a likelihood index for a plurality of braille data strings, and the likelihood calculating means includes a moving direction, a moving amount, and a moving amount of the isolated block by the moving means. The likelihood index is calculated based on at least one of the movement numbers, and the candidate output means sorts and outputs the corrected candidate braille data strings in descending order of the likelihood index.

  Preferably, one or more input error patterns, a correction pattern for each input error pattern, and pattern management means for managing the likelihood index data of each correction pattern in association with each other, and information managed by the pattern management means And / or a pattern correction means for performing new addition, and the pattern correction means moves the isolated block by the movement means when the likelihood index calculated by the likelihood calculation means is equal to or greater than a predetermined threshold. The newly added partial pattern is set as the correction pattern, the partial pattern before the movement of the isolated block by the moving means is set as the input error pattern, and the likelihood index is set as the likelihood index data.

  Preferably, the apparatus further comprises pattern management means for managing one or more input error patterns in association with correction patterns for each input error pattern, and each correction pattern has a position of a convex portion of the corresponding input error pattern. , A pattern comprising a combination of convex portions moved to other point positions within a predetermined range, and the candidate extraction means is an input managed by the pattern management means among the partial patterns constituting the input braille character string A search means for searching for a partial pattern that matches the error pattern, and a replacement means for replacing the searched partial pattern with a corresponding correction pattern, and a pattern of the braille character string after replacement by the replacement means, Extract as braille pattern candidates.

  Preferably, the pattern management unit further manages the likelihood index data of the braille pattern candidates extracted by the candidate extraction unit in association with each other.

  Preferably, it further includes likelihood calculation means for calculating a likelihood index of the correction candidate Braille data string based on the likelihood index data managed by the pattern management means.

  Preferably, the candidate output unit sorts and outputs the corrected candidate Braille data strings in descending order of the likelihood calculated by the likelihood calculating unit.

  Preferably, the pattern management means further manages the likelihood index data of each correction pattern in association with each other, a pattern correction means for correcting and / or newly adding information managed by the pattern management means, and a user And a selection means for selecting one of a plurality of correction candidate braille data strings output by the candidate output means based on the instruction from the pattern correction means, wherein the pattern correction means selects the correction candidate braille selected by the selection means The value of the likelihood index data of the correction pattern corresponding to the data string is increased by a predetermined value or a predetermined ratio.

  Preferably, the candidate extracting means includes an isolated block detecting means for detecting an isolated block when a matching partial pattern is not searched by the search means, and the isolated block is at least one of the convex portions of the input braille character string. It is an aggregate that includes two convex parts, and is an aggregate that can be moved collectively to adjacent flat parts. The convex parts are moved to other point positions by a predetermined number of point positions in units of isolated blocks. A braille character string pattern that has been moved by the moving means is further extracted as a braille pattern candidate.

  Preferably, the apparatus further includes coordinate input means for inputting coordinate information and conversion means for converting the input coordinate information into a plurality of braille data.

  Preferably, data input means for receiving input of a plurality of braille data from the outside is further provided.

  Preferably, the error detection means further detects an error in a plurality of braille data, and the candidate extraction means extracts a braille pattern candidate when an error in the plurality of braille data is detected.

  The Braille translation method according to another aspect of the present invention includes a step of detecting whether or not a translation result of a plurality of input Braille data matches a dictionary data stored in advance, and a translation result of the plurality of Braille data If the dictionary data does not match, the step of extracting braille pattern candidates based on a plurality of braille data and whether the translation result of the candidate braille data string corresponding to the extracted braille pattern candidates matches the dictionary data And a step of outputting the candidate Braille data string as a correction candidate Braille data string when the translation result of the candidate Braille data string matches the dictionary data.

  A Braille translation program according to another aspect of the present invention includes a step of detecting whether or not a translation result of a plurality of input Braille data matches a dictionary data stored in advance, and a translation result of the plurality of Braille data If the dictionary data does not match, the step of extracting braille pattern candidates based on a plurality of braille data and whether the translation result of the candidate braille data string corresponding to the extracted braille pattern candidates matches the dictionary data And a step of outputting the candidate Braille data string as a correction candidate Braille data string when the translation result of the candidate Braille data string matches the dictionary data.

  A computer-readable recording medium recording a Braille translation program according to another aspect of the present invention records the Braille translation program described above.

  According to the present invention, by extracting braille pattern candidates based on a plurality of input braille data, a correction candidate braille data string for the plurality of input braille data can be output. Thus, the user can easily grasp the correct Braille data string without re-inputting a plurality of Braille data.

  Embodiments of the present invention will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals and description thereof will not be repeated.

<About configuration>
First, the configuration of the Braille translation apparatus according to the embodiment of the present invention will be described with reference to FIGS.

FIG. 1 is a functional block diagram of a braille translation device 100 according to an embodiment of the present invention.
As shown in FIG. 1, the Braille translation device 100 includes a coordinate information input unit 101 and a display unit 111. Furthermore, the Braille translation device 100 includes a Braille data conversion unit 102, a Braille data storage unit 103, an error detection unit 104, a candidate extraction unit 105, a correction candidate storage unit 106, a Braille correction candidate output processing unit 107, a likelihood calculation unit 108, A pattern correction unit 109, a pattern management unit 110, a braille-ink character correspondence dictionary storage unit 112, and a word dictionary storage unit 113 are provided.

  In the embodiment of the present invention, “Braille” is a Braille character string composed of one or more Braille characters, and “Braille character” is a convex array arranged at matrix-like point positions in the Braille cell. It is a character for visually impaired persons consisting of a combination of parts. “Ink” is a single character that is normally used by healthy individuals, and includes, for example, kana, kanji, alphabet, and numbers. A “word” is an ink character string composed of one or more ink characters, and refers to a group of words such as a word and a phrase.

First, a specific example of Braille characters will be described.
FIG. 4 is a diagram showing a configuration of a braille cell used to represent a braille character. The braille cell 30 is composed of six point positions 31 to 36 corresponding to a total of six points of 3 vertical points × 2 horizontal points. The point at the uppermost point position 31 in the left column of the braille cell 30 is “1 point”, the point at the middle point position 32 in the left column is “2 point”, and the point at the lowermost point position 33 in the left column Is “3 points”, the point at the uppermost point position 34 in the right column is “4 points”, the middle point point 35 in the right column is “5 points”, and the lowermost point in the right column The point at position 36 is called “point 6”.

  A braille character is expressed by the combination of the convex part arrange | positioned at such six point positions 31-36. Therefore, the Braille characters can express 63 combinations of convex portions. However, in 63 ways of expression, not all ink characters such as kana characters, alphabets, numbers and symbols can be expressed. Therefore, when expressing alphabets or numbers, the alphabet system or number system is distinguished by attaching a predetermined prefix.

  For example, a Braille character having a convex portion arranged at the point position 31 indicates the ink character “A”, and a Braille character having a convex portion disposed at the point position 31 and the point position 32 indicates the ink character “I”. In addition, in the case of indicating a small letter of the alphabet, the convex portion is arranged at the point position 31 after the braille character (hereinafter referred to as “foreign language mark”) in which the convex portion is arranged at the point position 35 and the point position 36. By continuing the braille character, the ink letter “a” is indicated. In addition, when an uppercase letter of the alphabet is indicated, a braille character (hereinafter referred to as an “uppercase letter”) in which a convex portion is arranged at the point position 36 is added after the foreign language mark, and a convex portion is further arranged at the point position 31. By continuing the braille character, the ink letter “A” is indicated. Further, when a numeral is shown, a convex portion is projected at the point position 31 after a Braille character (hereinafter referred to as “number mark”) in which convex portions are arranged at the point position 33, the point position 34, the point position 35, and the point position 36. The ink character “1” is indicated by continuing the braille character in which the part is arranged. Thus, one ink character can be expressed by combining a plurality of braille characters.

  Referring to FIG. 1, the Braille-Ink Character Correspondence Dictionary Storage 112 stores a plurality of Braille data, that is, conversions when converting (translating) a plurality of Braille data, that is, Braille data strings into one or more ink character data, that is, ink character data strings. For example, a conversion table that is a rule is stored. “Braille data” refers to data for expressing a combination of convex portions in corresponding Braille characters in the present embodiment. In the conversion table, for example, one ink character data and one or more braille data are stored in association with each other.

  The word dictionary storage unit 113 stores a plurality of word data as a dictionary in advance. The word data (hereinafter referred to as “word dictionary data”) stored in the word dictionary storage unit 113 may be stored in advance by the Braille translation device 100, or may be registered by the user as appropriate. It doesn't matter.

  The coordinate information input unit 101 is an interface for inputting coordinate information. For example, coordinate information is used as information for designating (selecting) a Braille convex portion from the electronic pen 211 (see FIG. 2) or a user's finger. Accepts input.

  The Braille data conversion unit 102 converts the coordinate information received by the coordinate information input unit 101 into one or more Braille data. For example, when the input of coordinate information is detected at the point positions 31, 32 and 36 of one Braille cell 30, the Braille data conversion unit 102 converts the coordinate information into Braille data “110001”.

  The braille data storage unit 103 stores the braille data string 10 converted by the braille data conversion unit 102.

  The error detection unit 104 detects whether or not an error is included in the braille data string 10. In addition, the error detection unit 104 detects whether or not an error is included in the braille data string corresponding to the braille pattern candidate extracted by the candidate extraction unit 105 described later. The error detection unit 104 includes a Braille translation unit 1041 for translating the Braille data string into the ink character data string, and an ink character correct / incorrect determination unit 1042 for determining whether the ink character data string translated by the Braille translation unit 1041 is correct. Including.

  The Braille translation unit 1041 translates the source Braille data string into the ink character data string based on the conversion table stored in the Braille-ink character correspondence dictionary storage unit 112.

  The ink character correct / incorrect determination unit 1042 determines whether the ink character data string translated by the Braille translation unit 1041 is correct, that is, whether the meaning is understood as a word or a sentence. For example, the ink character correct / incorrect determination unit 1042 performs morphological analysis on the ink character data string. Then, the individual word data after the morphological analysis is compared with a plurality of word dictionary data stored in the word dictionary storage unit 113, and whether there is word dictionary data that matches each word data. Determine. If all the word data matches the word dictionary data, the ink character correct / incorrect determination unit 1042 determines that the ink character data string translated by the Braille translation unit 1041 is correct. On the other hand, if there is even one word data that does not match the word dictionary data, the ink character correct / incorrect determination unit 1042 determines that the ink character string translated by the Braille translation unit 1041 is an error.

  The error detection unit 104 determines that the target Braille data string is an error when there is Braille data that the Braille translation unit 1041 cannot translate into ink character data. That is, when one or more braille data not associated with ink character data is detected in the target braille data string, it is determined that the braille data string is erroneous. Further, even if the target Braille data string is translated into the ink character data string by the Braille translation unit 1041, the error detecting unit 104 determines that the ink character data string is erroneous by the ink character correct / incorrect determination unit 1042. If it is determined, it is determined that the target braille data string is incorrect.

  Note that the error detection unit 104 may determine that the target Braille data string is an error only when there is Braille data that cannot be translated into ink character data by the Braille translation unit 1041.

  If no error is detected in the braille data string 10 by the error detection unit 104, the ink character data string translated by the Braille translation unit 1041 may be displayed on the display unit 111.

  The pattern management unit 110 uses an input pattern of Braille characters that is likely to cause an input error as an input error pattern, and manages correction patterns corresponding to the input error pattern in association with each other. FIGS. 5A and 5B are first and second diagrams showing examples of the data structure of the pattern management unit 110, respectively.

  Referring to FIG. 5A, the pattern management unit 110 manages one or more input error patterns 801 and a correction pattern 802 for each input error pattern 801 in association with each other in the pattern management table. For example, in the set 803, a Braille character pattern represented by Braille data “011000” is stored as an input error pattern 801, and a Braille character pattern represented by Braille data “000011” is stored as a correction pattern 802. In the set 804, a Braille character pattern represented by Braille data “100110” is stored as an input error pattern 801, and a Braille character pattern represented by Braille data “010011” is stored as a correction pattern 802. In the present embodiment, a set of patterns is stored in the table format as described above, but the present invention is not limited to the table format. Each set of patterns may be registered in advance or may be registered by the user. Further, as will be described later, a new set of patterns may be automatically registered.

  Further, as shown in FIG. 5B, in addition to the input error pattern 801 and the correction pattern 802, likelihood index data 805 may be further managed in association with each pattern set. The likelihood index data 805 is an index that relatively represents the likelihood of the new Braille data string generated by replacing the input error pattern 801 with the correction pattern 802 with respect to the Braille data string 10.

  Referring to FIG. 5B, in the pattern management table, for example, set 803 ′, the Braille character pattern represented by Braille data “011000” is replaced with the Braille character pattern represented by Braille data “000011”. The likelihood index data 805 is indicated as “100”. Here, the likelihood index data 805 is a numerical value of “0 to 100”, but is not limited to this, and may be a numerical value of a percentage, for example. For example, when a plurality of correction candidate braille data strings are registered in the correction candidate storage unit 106 to be described later, the likelihood index data 805 is output by the braille correction candidate output processing unit 107 to output these correction candidate braille data strings. Used as an indicator in determining the order. In the following description, it is assumed that the pattern management unit 110 has a data structure as shown in FIG.

  In FIGS. 5A and 5B, each of the input error pattern 801 and the correction pattern 802 is a pattern corresponding to one Braille character, but is not limited to such an example. It suffices that each set of patterns corresponds to the same number of braille characters, and a set of patterns corresponding to two or more braille characters may be managed.

  The candidate extraction unit 105 extracts braille pattern candidates based on the braille data string 10. A “braille pattern candidate” is a raised braille formed by moving at least part of the raised portion of a braille character string (hereinafter referred to as “input braille”) represented by the braille data string 10 to another point position. This is a Braille pattern corresponding to a combination of parts.

  Candidate extraction section 105 includes isolated block detection section 1051, isolated block movement section 1052, and pattern search / replacement section 1053.

  The isolated block detection unit 1051 detects the isolated block by dividing the convex portion of the input Braille into one or a plurality of groups. The “isolated block” is an aggregate including at least one convex portion among the convex portions of the input Braille, and refers to an aggregate that can be moved collectively to adjacent flat portions. The “aggregate” is an aggregate of convex portions including at least one convex portion, or a set of convex portions including at least one convex portion and flat portions (points at points where no convex portion exists). It is assumed to be a body. In addition, “movable” means that there is a flat portion on at least one of the top, bottom, left, and right of an isolated block, and even if the isolated block is shifted by a predetermined number of point positions, other convex portions are not affected (others) It does not overlap the convex part).

  The isolated block moving unit 1052 moves, for example, a predetermined number of point positions in at least one direction of up, down, left, and right in units of detected isolated blocks.

  Candidate extraction section 105 extracts the braille pattern after being moved by isolated block moving section 1052 as a braille pattern candidate.

  The pattern search / replacement unit 1053 includes a partial pattern that matches the input error pattern 801 stored in the pattern management unit 110 in the pattern of combinations of convex portions constituting the input Braille (hereinafter referred to as “input Braille pattern”). Search for. When a matching partial pattern is found, the partial pattern (input error pattern 801) is replaced with the corresponding correction pattern 802. The “partial pattern” refers to a pattern of braille characters or braille character strings included in the input braille pattern.

  The candidate extraction unit 105 extracts a braille pattern including the correction pattern replaced by the pattern search / replacement unit 1053 as a braille pattern candidate.

  The correction candidate storage unit 106 stores a braille data string corresponding to a braille pattern candidate for which no error has been detected by the error detection unit 104 as a correction candidate braille data string.

  The braille correction candidate output processing unit 107 outputs the correction candidate braille data string stored in the correction candidate storage unit 106 to the display unit 111. In the present embodiment, a combination pattern (hereinafter also referred to as “candidate pattern”) of convex portions of braille (braille character string) represented by the correction candidate braille data string is displayed on display unit 111. The braille correction candidate output processing unit 107 may display a corresponding ink character data string together with the candidate pattern.

  The likelihood calculating unit 108 calculates a likelihood index for the Braille data string 10 for each of the correction candidate Braille data strings stored in the correction candidate storage unit 106. The “likelihood index” is an amount that relatively represents the likelihood of the correction candidate Braille data string with respect to the Braille data string 10. As a likelihood index calculation method, it may be calculated based on likelihood index data 805 managed by the pattern management unit 110. Further, when the correction candidate braille data string is created by moving the isolated block by the isolated block moving unit 1052, for example, based on at least one movement information of the moving direction, moving amount, and moving number of the isolated block. Thus, the likelihood index may be calculated. For example, the likelihood index of the correction candidate Braille data string created by moving the isolated block once is higher than the likelihood index of the correction candidate Braille data string created by moving the isolated block twice Is predetermined so as to be calculated. The likelihood index may be calculated based on the number of convex portions included in the moved isolated block.

  The calculated likelihood index is used when the braille correction candidate output processing unit 107 displays the correction candidate braille data string on the display unit 111. For this purpose, for example, the correction candidate storage unit 106 stores the correction candidate braille data string and the data for calculating the likelihood index in association with each other. FIG. 6 shows an example of the data structure of the correction candidate storage unit 106.

  With reference to FIG. 6, in the present embodiment, correction candidate storage unit 106 stores correction candidate braille data string 301 and likelihood index data 302 or movement amount 303 in association with each other. The likelihood index data 302 corresponds to the likelihood index data 805 stored in the pattern management unit 110. The movement amount 303 corresponds to the number of movements of the isolated block. The likelihood calculating unit 108 calculates a likelihood index based on the likelihood index data 302 or based on the movement amount 303. The likelihood index based on the movement amount 303 may be calculated by the following calculation formula, for example.

Likelihood = (1 / movement amount) × 100
Based on the calculated likelihood index, the braille correction candidate output processing unit 107 rearranges and outputs, for example, all correction candidate braille data strings in descending order of likelihood index. Note that one correction candidate Braille data string having the highest likelihood index may be output, or a predetermined number (for example, three) of correction candidate Braille data strings having the highest likelihood index may be output. May be. Further, the user may be able to select these output methods. For example, a menu for selecting an output method may be provided to allow the user to set and change in advance, or a selection button for each output method may be displayed in a predetermined area of the display unit 111 during output. One of them may be selected. In the present embodiment, description will be made assuming that the output method is preset by the user.

  The pattern correction unit 109 appropriately corrects (updates) information in the pattern management table managed by the pattern management unit 110. For example, the timing for correcting the information in the pattern management table may be as follows. For example, when displaying a candidate pattern on the display unit 111, which of the correction candidate braille data strings presented (displayed) is a correct braille data string in the predetermined area of the display unit 111 (or the correction presented It is assumed that a button for selecting whether or not the candidate Braille data string is correct is displayed. When the user selects a correct correction candidate Braille data string, a record including the input error pattern, correction pattern, and calculated likelihood index that led to creation of the correction candidate Braille data string is stored in the pattern management table. Add new. That is, in the case of a correction candidate braille data string generated by moving an isolated block, the movement and the braille (braille character) pattern corresponding to the braille data string 10 before the movement in the braille cell including the convex portion to be moved A braille (braille character) pattern corresponding to the subsequent correction candidate braille data string is set as an input error pattern 801 and a correction pattern 802, respectively. Then, the calculated likelihood index is set in the likelihood index data 805. Further, when the likelihood index calculated by the likelihood calculating unit 108 is equal to or greater than a predetermined threshold (for example, 75), the input error pattern and the correction pattern that lead to the creation of the target correction candidate Braille data string are calculated. A record composed of the likelihood index may be newly added to the pattern management table.

  When the correct correction candidate braille data string is generated using an existing pattern set, the likelihood index data 805 of the correction pattern 802 of the set is corrected to be increased by a predetermined value or a predetermined ratio. Also good. As a result, since the likelihood index for the correction candidate braille data string corrected using the set will increase in the future, if there are a plurality of correction candidate braille data strings, the correction corrected using the set The candidate braille data string is positioned higher. Therefore, the user can easily select the correct braille data string.

  FIG. 2 is a block diagram showing a hardware configuration of Braille translation apparatus 100 according to the embodiment of the present invention.

  Referring to FIG. 2, braille translation apparatus 100 includes a CPU (Central Processing Unit) 201 for performing various operations and controls, for example, a ROM (Read Only Memory) for storing programs such as a boot-up program and various data. 202, a RAM (Random Access Memory) 203 for storing work data, the coordinate information input unit 101 and the display unit 111 described above, an operation unit 204 that is operated by the user to receive an instruction from the user, and a recording medium 210 Drive device 206 and hard disk 207 capable of reading data from the computer. These hardware are connected to the bus 208.

  The Braille translation apparatus 100 may further include a network adapter board that provides connection to a local area network (LAN).

The coordinate information input unit 101 receives an input from the electronic pen 211, for example.
The recording medium 210 stores a program (Braille translation program) for realizing the functions of the Braille translation apparatus 100 shown in FIG. 1 and further transferred to the hard disk 207. Note that such a Braille translation program may be stored in the ROM 202 in advance. Alternatively, the Braille translation program may be transmitted from the outside to the Braille translation apparatus 100 through a network (not shown) and stored in the hard disk 207. The Braille translation program is loaded into the RAM 209 at the time of execution. Note that the Braille translation program may be loaded directly into the RAM 209 from the recording medium 210 or via a network. Further, instead of the hard disk 207, the Braille translation program may be transferred to a flash memory, for example.

  The recording medium 210 may be an optical disc such as a CD-ROM (Compact Disc-ROM), an MO (Magneto-Optical disc), an MD (Mini Disc), and a DVD (Digital Versatile Disc), or an FD (flexible disc). ) And a magnetic disk such as a hard disk. Alternatively, tape media such as magnetic tape and cassette tape, card-type recording media such as IC (Integrated Circuit) cards and optical cards, and mask ROM, EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), and flash ROM Any of the semiconductor memories may be used.

  Included in the braille data conversion unit 102, error detection unit 104, candidate extraction unit 105, braille correction candidate output processing unit 107, likelihood calculation unit 108, pattern correction unit 109, error detection unit 104, and candidate extraction unit 105 described above The function of each unit is realized by the CPU 201. Note that at least one of these may be realized by hardware.

  Braille data storage unit 103 and correction candidate storage unit 106 are storage areas included in RAM 203, for example. Braille-ink-character correspondence dictionary storage unit 112, word dictionary storage unit 113, and pattern management unit 110 are storage regions included in hard disk 207, for example. Note that the data in the Braille-ink-character correspondence dictionary storage unit 112, the word dictionary storage unit 113, and the pattern management unit 110 may be stored in the hard disk 207 in advance, or read and stored from the recording medium 210, for example. May be.

  FIG. 3 shows an overview of Braille translation apparatus 100 according to the embodiment of the present invention. Referring to FIG. 3, on the surface of Braille translator 100, power button 204.1 operated to change the power-on state, menu button 204.2 for displaying various menus, and 1 This is operated when the previous candidate selection button 204.3 for displaying the previous candidate pattern, the next candidate selection button 204.4 for displaying the next candidate pattern, and the candidate pattern being displayed are correct. The operation unit 204 including the correct answer button 204.5 and the display unit 111 are provided. In the present embodiment, the coordinate information input unit 101 is provided integrally with the display unit 111 and constitutes, for example, a touch panel. Thereby, since the input coordinate information is displayed at the input location, confirmation is easy and input errors can be prevented.

  For example, eight braille cells 30 are provided in the coordinate information input unit 101. In the present embodiment, the candidate pattern is displayed in the display area corresponding to the position of the coordinate information input unit 101.

  The display unit 111 also displays a feed button 52 for advancing the braille cell 30 one by one, a return button 51 for returning one by one, and a translation button 53 for inputting a translation start instruction. The Furthermore, the display unit 111 is provided with an ink character display area 40 for displaying the translation result.

<About operation>
Next, the operation of the Braille translation apparatus 100 according to the embodiment of the present invention will be described with reference to FIGS.

(Outline of main operations)
FIG. 7 is a flowchart showing a main routine executed by CPU 201 in Braille translation apparatus 100 according to the embodiment of the present invention.

  Referring to FIG. 7, when power button 204.1 is pressed to turn on the power, the Braille translation program is activated and the process proceeds to step S2.

  In step S <b> 2, the CPU 201 accepts input of coordinate information to the coordinate information input unit 101. At this time, the CPU 201 determines which point positions 31 to 36 of the respective braille cells 30 are designated based on the received coordinate information. Information on the point position determined to be designated is temporarily recorded in the RAM 203. For example, a flag information storage area for storing flag information corresponding to each of the point positions 31 to 36 is provided for each braille cell 30 in a predetermined area of the RAM 203, corresponding to the point position determined to be selected. The flag “1” is set in the storage area to be used.

  Next, the CPU 201 determines whether or not the translation button 53 displayed on the display unit 111 has been selected (step S4). Until the translation button 53 is selected, input of coordinate information is accepted (NO in step S4).

  When the translation button 53 is selected (YES in step S4), the braille data conversion unit 102 converts the input coordinate information into one or more braille data (braille data string 10) and stores it in the braille data storage unit 103. (Step S6). For example, the Braille data conversion unit 102 converts the flag information recorded in the RAM 203 for each Braille cell 30 in the order of the dot positions 31, 32, 33, 34, 35, and 36 to convert it into the Braille data string 10.

  Next, the Braille translation unit 1041 converts the Braille data string 10 stored in the Braille data storage unit 103 in Step S6 into the ink character data string based on the conversion table stored in the Braille-inkscript correspondence dictionary storage unit 112. Translate (step S8). Then, the error detection unit 104 checks whether or not an error is included in the translation result in step S8. If an error is included, that is, if the Braille translation unit 1041 cannot translate the Braille data string 10 into the ink character data or the ink character data string (YES in Step S10), the process proceeds to Step S14. On the other hand, when there is no error, that is, when the braille data string 10 can be translated into the ink character data or the ink character data string (NO in step S10), the process proceeds to step S12.

  In step S12, the ink character correct / incorrect determination unit 1042 checks whether the translated ink character data string is strange as a word or a sentence, that is, whether the meaning is understood. Specifically, the ink character correct / incorrect determination unit 1042 performs morphological analysis on the ink character data string as described above, and individual word data matches the word dictionary data stored in the word dictionary storage unit 113. It is determined whether or not to do. When it is determined that the translated ink character data string is strange as a word or a sentence, that is, when there is word data that does not match any of the word dictionary data (YES in step S12), the process proceeds to step S14. On the other hand, if it is determined that the translated ink character data string is not a word or a sentence, that is, if all the word data matches the word dictionary data (NO in step S12), the series of processes is terminated. The

  In step S14, a correction candidate extraction process is executed. And a correction candidate output process is performed (step S16) and a series of processes are complete | finished. The correction candidate extraction process (step S14) and the correction candidate output process (step S16) will be described in detail later.

(Correction candidate extraction process)
FIG. 8 is a flowchart showing the correction candidate extraction process.

  First, the pattern search / replacement unit 1053 searches the input braille pattern for whether or not the same partial pattern as the error pattern 801 managed by the pattern management unit 110 exists (step S102). If it is determined that there is an error pattern in the input braille pattern (YES in step S102), the process proceeds to step S104. On the other hand, if it is determined that there is no error pattern in the input braille pattern (NO in step S102), the process proceeds to step S110.

  In step S104, the pattern search / replacement unit 1053 replaces the input error pattern included in the input braille pattern with the corresponding correction pattern. In this way, braille pattern candidates are extracted.

  Next, an error detection process is performed on the new braille data string corresponding to the extracted braille character string of the braille pattern (step S106). This error detection process will be described later.

  Next, the pattern search / replacement unit 1053 searches whether there is another error pattern in the input braille pattern (step S108). If it is determined that there is another error pattern (YES in step S108), the process returns to step S104. On the other hand, when it is determined that there is no other error pattern (NO in step S108), the process proceeds to step S110.

  In step S110, the isolated block detection unit 1051 performs braille square isolated block detection processing, which will be described later, on the input Braille pattern. Next, the isolated block moving unit 1052 moves one isolated block in one direction, up, down, left, and right (step S112). For example, if an error is included in the translation result, the isolated block is moved within the braille cell of the braille character in error. If erroneous word data is detected in the ink character data string, the isolated block is moved within the braille cell of the braille character string (or braille character) represented by the word data.

  Then, an error detection process is performed on a new Braille data string corresponding to the Braille character string of the Braille pattern extracted after the isolated block is moved (Step S114).

  Next, the isolated block moving unit 1052 determines whether or not to move the isolated block (step S116). Whether or not to move the isolated block is determined, for example, based on whether or not the isolated block can be moved in the direction not yet moved in step S112. Here, it is assumed that only one point position can be moved with respect to one direction, but it is also possible to move more than two point positions with respect to one direction. The same applies to step S124 and step S132 described below.

  If the isolated block moving unit 1052 determines to move the isolated block (YES in step S116), the process returns to step S112. On the other hand, when it is determined not to move the isolated block (NO in step S116), the process proceeds to step S118. Note that the order of movement in the up, down, left, and right directions in step S112 may be determined in advance or may be determined randomly each time. The same applies to step S120 and step S128 described below.

  In step S118, the isolated block detection unit 1051 executes a row isolated block detection process described later. Next, the isolated block moving unit 1052 moves one isolated block in one direction, up, down, left, and right (step S120). For example, when an error is included in the translation result, each isolated block detected in step S118 is moved. Similarly, when erroneous word data is detected in the ink character data string, the isolated blocks detected in step S118 are moved. In the latter case, it is also possible to move only the portion of the isolated block detected in step S118 that exists within the range of the braille cell corresponding to the braille (or braille character) represented by the word data in which the error is detected. Good.

  Then, an error detection process is performed on a new Braille data string corresponding to the Braille character string of the Braille pattern extracted after the isolated block is moved (Step S122). Next, the isolated block moving unit 1052 determines whether or not to move the isolated block (step S124). If it is determined that the isolated block is to be moved (YES in step S124), the process returns to step S120. On the other hand, when it is determined not to move the isolated block (NO in step S124), the process proceeds to step S126.

  In step S126, the isolated block detection unit 1051 performs a column isolated block detection process described later. Next, the isolated block moving unit 1052 moves the isolated block in any one of up, down, left and right directions (step S128). For example, when an error is included in the translation result, the isolated block including the convex portion of the Braille character in error is moved. When erroneous word data is detected in the ink character data string, the isolated block including the convex portion of the Braille character string (or Braille character) represented by the word data is moved.

  Then, an error detection process is performed on a new Braille data string corresponding to the Braille character string of the Braille pattern extracted after the isolated block is moved (Step S130). Then, the isolated block moving unit 1052 determines whether or not to move the isolated block (step S132). If it is determined that the isolated block is to be moved (YES in step S132), the process returns to step S128. On the other hand, when it is determined not to move the isolated block (NO in step S132), the process returns to the main routine.

  In the present embodiment, the isolated block detection / movement process (steps S110 to S132) is performed after the Braille pattern search / replacement process (steps S102 to S108). You may do it. In addition, although both of these processes are performed, either one may be performed. In this case, it may be determined that only one of the processes is executed in advance, or the user may select any one by operating the menu button 204.2, for example.

  In the isolated block detection / movement process, three types of isolated block detection / movement processes in braille cells, rows and columns are performed in series, but one or two types of isolated block detection / movement processes are performed. It is good also as performing. In this case, the user may be allowed to select which type of isolated block detection / movement process is performed. The order of the three types of isolated block detection / movement processing is not limited to the order shown in FIG.

  In addition, the error detection of the generated new braille data string is performed every time the pattern replacement and the isolated block move, but the error detection may not be performed every time. For example, all new braille data strings are temporarily stored in the correction candidate storage unit 106, then error detection is performed collectively, and new braille data strings in which errors are detected are deleted from the correction candidate storage unit 106. Also good.

(Error detection process)
FIG. 9 is a flowchart showing the flow of error detection processing.

  First, the Braille translation unit 1041 translates a new Braille data string into an ink character data string based on the conversion table stored in the Braille-ink character correspondence dictionary storage unit 112 (step S202). Here, the new Braille data string is a Braille data string after pattern replacement based on the Braille data string 10 (Step S104), and an isolated block movement based on the Braille data string 10 (Steps S112, S120, S128). ) Braille data string.

  Next, the Braille translation unit 1041 determines whether or not an error is included in the translation result (step S204). When it is determined that no error is included in the translation result (NO in step S204), the process proceeds to step S206. On the other hand, if it is determined that an error is included in the translation result (YES in step S204), the process proceeds to step S210.

  In step S206, the ink character correct / incorrect determination unit 1042 determines whether the ink character data string translated in step S202 is strange as a word or a sentence. If it is determined that the translated ink character data string is strange as a word or a sentence (YES in step S206), the process proceeds to step S210. On the other hand, if it is determined that the translated ink character data string is not a word or sentence (NO in step S206), the process proceeds to step S208. Since steps S204 and S206 are the same as steps S10 and S12 described above, detailed description thereof will not be repeated here.

  In step S208, the candidate extraction unit 105 registers the braille data string translated in step S202 in the correction candidate storage unit 106 as a correction candidate braille data string. When the process of step S208 ends, the error detection process ends.

  On the other hand, in step S210, the braille data string translated in step S202 is discarded without being registered as a correction candidate braille data string, and the error detection process is terminated.

(Braille square isolated block detection process)
FIG. 10 is a flowchart showing the flow of the isolated block detection process in braille cells.

  First, the isolated block detection unit 1051 determines whether there is a braille cell that has not been subjected to the isolated block detection process (step S302). If it is determined that there is an unprocessed braille cell (YES in step 302), the process proceeds to step S304. On the other hand, when it is determined that the isolated block detection process has been performed for all the braille cells (NO in step S302), the isolated block detection process in the braille cells is terminated.

  In step S304, the isolated block detection unit 1051 determines whether or not there is a convex portion that does not belong to the isolated block candidate in the braille cell determined not to perform the isolated block detection process. If it is determined that there is a convex portion that does not belong to the isolated block candidate (YES in step S304), the process proceeds to step S306. On the other hand, if it is determined that there is no convex portion that does not belong to the isolated block candidate (NO in step S304), the isolated block detection process in the braille cell is terminated, and the process returns to step S302.

  In step S306, any one convex portion that does not belong to an isolated block is set as a new isolated block candidate. Next, it is determined whether or not there is a convex portion that does not belong to the isolated block candidate at a point position adjacent to the point position of the added convex portion (step S308). If it is determined that there is a convex portion that does not belong to an isolated block candidate (YES in step S308), all the convex portions existing at adjacent point positions are added to the isolated block candidate (step S310). When the process of step S310 ends, the process returns to step S308.

  If it is determined in step S308 that there is no convex portion that does not belong to the isolated block candidate at the point position adjacent to the point position of the added convex portion (NO in step S308), the process proceeds to step S312.

  In step S312, the isolated block detection unit 1051 determines whether or not the current isolated block candidate can move in any one of up, down, left, and right directions within the target braille cell. If it is determined that it can move in any one of the up, down, left, and right directions (YES in step S312), the isolated block candidate is determined as an isolated block (step S314). When the process of step S314 ends, the process returns to step S304.

  On the other hand, if it is determined that the current isolated block candidate cannot move in any one of the upper, lower, left and right directions within the target braille cell (NO in step S312), the isolated block candidate is not regarded as an isolated block (step S316) Return to step S304.

  FIGS. 11A to 11C are diagrams for explaining the isolated block detection process in the braille cell. A case will be described in which the braille data string 10 stored in the braille data storage unit 103 represents a braille (braille character string) 620 as shown in FIG.

  Referring to FIG. 11A, the braille 620 is composed of five braille characters 621 to 625. The braille data corresponding to these braille characters 621 to 625 are “000010”, “101101”, “001011”, “110111”, and “1000011”, respectively.

  FIG. 11B is a diagram illustrating a state in which an isolated block is detected in the braille cell 30 for the braille character 622. Hereinafter, a specific description will be given with reference to the flowchart of FIG. 10 and FIG. 4.

  In step S <b> 306, the isolated block detection unit 1051 sets the convex portion at the point position 31 as a new isolated block candidate 626, and adds the convex portion to the isolated block candidate 626. Next, among the point positions 32, 34, and 35 adjacent to the point position 31 of the convex portion added to the isolated block candidate 626, the convex portion at the point position 34 that does not belong to the isolated block candidate 626 is added to the isolated block candidate 626. (Step S308 and Step S310).

  Next, it is checked whether or not there is a convex part that does not belong to the isolated block candidate 626 among the point positions 31, 32, and 35 adjacent to the point position 34 of the convex part newly added to the isolated block candidate 626 (step). S308). Here, since it does not exist, the process proceeds to step S312. In step S312, since the isolated block candidate 626 can move downward, the process proceeds to step S314, and the isolated block candidate 626 is regarded as an isolated block 630. In this manner, an isolated block 630 that is an aggregate of convex portions at the point position 31 and the point position 34 is detected.

  In step S <b> 304, the isolated block detecting unit 1051 adds the convex portion at the point position 33 as a new isolated block candidate 627 and adds the convex portion to the isolated block candidate 627. Next, among the point positions 32, 35, and 36 adjacent to the point position 33 of the convex portion added to the isolated block candidate 627, the convex portion at the point position 36 that does not belong to the isolated block candidate is added to the isolated block candidate 627. (Step S308 and Step S310).

  Next, it is checked whether or not there is a convex portion that does not belong to the isolated block candidate among the point positions 32, 33, and 35 adjacent to the point position 36 of the convex portion newly added to the isolated block candidate 627. Here, since it does not exist, the process proceeds to step S312. In step S312, since the isolated block candidate 627 can move up, the process proceeds to step S314, and the isolated block candidate 627 is regarded as an isolated block 631. In this manner, an isolated block 631 that is an aggregate of convex portions at the point position 33 and the point position 36 is detected.

  Similarly, FIG. 11C shows the result of the same processing performed on all the braille cells of the five braille characters 621 to 625 constituting the braille character string 620 shown in FIG.

  Referring to FIG. 11C, an isolated block 632 composed of a convex portion at the point position 35 is detected in the braille cell of the braille character 621. In the braille cell of the braille character 623, an isolated block 633, which is an aggregate of convex portions at the point positions 33, 35, and 36, is detected. On the other hand, no isolated block is detected in the braille cells of the braille characters 624 and 625. In the braille cell of the braille character 624, an isolated block candidate is formed by a set of five convex portions at the point positions 31, 32, 34, 35, and 36. It is because it cannot be done. Similarly, in the braille cell of the braille character 625, an isolated block candidate is formed by a set of three convex portions at the point positions 31, 35, and 36. This is because they cannot.

(Row isolation block detection processing)
FIG. 12 is a flowchart showing row isolated block detection processing.

  First, the isolated block detection unit 1051 divides the entire input braille pattern for each line, and extracts the first line, the second line, and the third line (step S402). Next, the isolated block detection unit 1051 determines how many blank lines there are (step S406). If it is determined that there is one blank line, the process proceeds to step S408. If it is determined that there are two blank lines, the process proceeds to step S416. If it is determined that there is no blank line, the process proceeds to step S424.

  In step S408, the isolated block detection unit 1051 determines which row is a blank row. If it is determined that the first row is a blank row, an isolated block is formed by the second row and the third row (step S410). That is, an aggregate of convex portions and flat portions included in the second and third rows is detected as one isolated block. If it is determined that the blank line is the second line, an isolated block is formed in each of the first line and the third line (step S412). That is, the aggregate of the convex portion and the flat portion included in the first row and the aggregate of the convex portion and the flat portion included in the third row are respectively detected as one isolated block. When it is determined that the blank line is the third line, an isolated block is formed by the first line and the second line (step S414). That is, an aggregate of convex portions and flat portions included in the first row and the second row is detected as one isolated block.

  In step S416, it is determined which line and which line is a blank line. When it is determined that the blank line is the first line and the second line, the isolated block detection unit 1051 configures an isolated block with only the third line (step S418). That is, the aggregate of the convex portion and the flat portion included in the third row is detected as one isolated block. If it is determined that the blank lines are the second and third lines, an isolated block is formed only by the first line (step S420). That is, an aggregate of convex portions and flat portions included in the first row is detected as one isolated block. If it is determined that the blank lines are the first and third lines, an isolated block is formed by only the second line (step S422). That is, the aggregate of the convex portion and the flat portion included in the second row is detected as one isolated block.

  When the processing in steps S410, 412, 414, 418, 420, and 422 is completed, this row isolated block detection processing is terminated. In step S424, the isolated row detection process is terminated without forming an isolated block.

  FIGS. 13A to 13C are diagrams for explaining row isolated block detection processing. A case will be described in which the braille data string 10 stored in the braille data storage unit 103 represents a braille (braille character string) 640 as shown in FIG.

  Referring to FIG. 13A, the braille 640 is composed of three braille characters 641 to 643. The Braille data corresponding to these Braille characters 641 to 643 are “001100”, “101000”, and “100101”, respectively.

  FIG. 13B shows a state where three lines 644, 645, and 646 are extracted by dividing the entire pattern of the Braille 640 for each line (step S402). In this case, since the second row 645 is a blank row having no protrusions, an isolated block 647 is configured by the first row 644 and an isolated block 648 is configured by the third row 646. (Step S412). The result is shown in FIG.

(Column isolated block detection processing)
FIG. 14 is a flowchart showing column isolated block detection processing.

  First, the isolated block detection unit 1051 divides the entire input braille pattern for each column, and extracts a plurality of columns (step S502). Here, since there are two columns for each Braille cell, a number of columns twice as many as the number of Braille cells constituting the input Braille pattern is extracted. Next, only empty columns having no convex portion are extracted (step S504). Then, the isolated block detection unit 1051 detects isolated blocks by dividing them with empty strings (step S506). That is, an aggregate of convex portions and flat portions included in one or more columns in which no empty column exists is detected as an isolated block. When the process of step S506 is completed, the column isolated block detection process is terminated.

  FIGS. 15A to 15C are diagrams for explaining column isolated block detection processing. A case will be described in which the braille data string 10 stored in the braille data storage unit 103 represents a braille (braille character string) 660 as shown in FIG.

  With reference to FIG. 15A, the Braille 660 is composed of six Braille characters 661 to 666. The Braille data corresponding to these Braille characters 661 to 666 are “111000”, “001011”, “000010”, “110000”, “011001”, and “011000”, respectively.

  FIG. 15B shows a state in which twelve columns 667 to 678 are extracted by dividing the entire Braille 660 pattern for each column (step S502). Here, when empty columns that do not include any convex portion are extracted, columns 668, 671, 674, and 678 are extracted (step S504). Since the isolated block detection unit 1051 forms an isolated block with one or more columns that do not include an empty column in between, an isolated block 680 is configured by only the column 667, and a column sandwiched between columns 668 and 671 that are empty columns An isolated block 681 is formed from 669 to 670. Further, an isolated block 682 is composed of columns 672 to 673 sandwiched between columns 671 and 674 which are empty columns, and an isolated block 683 is composed of columns 675 to 677 sandwiched between columns 674 and 678 which are empty columns. Is configured. The result is shown in FIG.

(Correction candidate output process)
FIG. 16 is a flowchart showing the correction candidate output process.

  First, the likelihood calculating unit 108 calculates a likelihood index of each correction candidate braille data string stored in the correction candidate storage unit 106 (step S602). Referring to FIG. 6, when likelihood index data 802 is associated with correction candidate braille data string 801, likelihood index data 802 itself is calculated as a likelihood index. When the movement amount 803 is associated with the correction candidate Braille data string 801, the likelihood index is calculated using the above-described calculation formula.

  Next, the braille correction candidate output processing unit 107 sorts the correction candidate braille data strings stored in the correction candidate storage unit 106 in descending order of likelihood index (step S604). Then, the braille correction candidate output processing unit 107 determines an output method preset by the user (step S606).

  If it is determined that there is only one output method, the correction candidate braille data string having the highest likelihood index is output (step S608). If the output method is determined to be the top N, the top N correction candidate braille data strings are output (step S610). When it is determined that the output method is all, all the correction candidate braille data strings are output in descending order of likelihood index (step S612).

  When the processes of steps S608, 610, and 612 are completed, the correction candidate braille data output process is terminated.

<Specific examples>
Next, a specific example of the operation of the Braille translation apparatus 100 according to the embodiment of the present invention will be described with reference to FIGS.

(Specific example of pattern search / replacement processing)
First, a specific example of the correction candidate braille data string extracted based on the pattern search / replacement process (steps S102 to S108 in FIG. 8) will be described.

  FIG. 17 is a diagram for explaining a specific example of the pattern search / replacement process. As shown in FIG. 17A, it is assumed that the braille data string 10 stored in the braille data storage unit 103 represents the braille 820.

  Referring to FIG. 17A, a braille 820 is composed of six braille characters 821 to 826. The braille data corresponding to these braille characters 821 to 826 are “011000”, “100000”, “111100”, “111100”, “111000”, and “100010”, respectively.

  In this case, when the Braille translation unit 1041 translates into the ink character data string, “(error) disguise” is obtained as shown in FIG. The error detection unit 104 determines that the pattern of the first braille character 821 in error is an input error. Accordingly, the pattern search / replacement unit 1053 searches the pattern management unit 110 for a pattern in which the first Braille character 821 pattern is the input error pattern 801, and extracts a correction pattern 802 managed in association with the pattern.

  In step S102, the pattern search / replacement unit 1053 searches the pattern management unit 110 for the input error pattern 801 having the same pattern as the first braille character 821. In this case, since it matches the input error pattern 801 of the set 803 ′ shown in FIG. 5B, the correction pattern 802 of the set 803 ′ is extracted and the pattern of the braille character 821 is extracted in step S104. Replace with pattern 802. FIG. 17B shows a pattern of Braille 820 ′ newly created by replacement.

  Referring to FIG. 17B, the first Braille character 821 ′ of Braille 820 ′ is represented by Braille data “000011”. When such a new Braille data string representing the Braille 820 ′ is translated into the ink character data string, “apple” is obtained. Since the ink character data string “apple” is established as a word (because it matches the word dictionary data), it is determined that the ink character data string is correct.

  Therefore, a new Braille data string representing the Braille 820 ′ is registered in the correction candidate storage unit 106 as a correction candidate Braille data string. Thereby, the registered correction candidate braille data string is presented to the user. At this time, the numerical value (100) that is the likelihood index data 805 of the set 803 'is also stored in association with the correction candidate braille data string.

  Thus, even when the braille data string 10 includes an input error, the pattern management unit 110 that manages the input error pattern 801 and the correction pattern 802 in association with each other is searched, and the input error pattern is corrected. By replacing with 802, a correction candidate braille data string can be output. Thereby, even if it is a case where a user inputs coordinate information to the wrong point position, the correct braille data sequence can be obtained easily.

(Specific example of isolated block move processing in units of rows)
FIGS. 18A to 18E are diagrams for describing a specific example of row isolated block movement processing in the embodiment of the present invention. A case will be described in which the braille data string 10 stored in the braille data storage unit 103 represents a braille (braille character string) 720 as shown in FIG.

  Referring to FIG. 18A, the braille 720 is composed of three braille characters. The braille data corresponding to these braille characters are “010010”, “001001”, and “010011”, respectively. In the ink character display area 40, a translation result “− (error)” of the braille data string representing the braille 720 is displayed. Since the translation result includes an error, the error detection unit 104 determines that an input error is included in the Braille data string 10 representing the Braille 720. Then, the candidate extraction unit 105 starts extraction processing of Braille pattern candidates.

  FIG. 18B to FIG. 18E show a procedure from detection and movement of isolated blocks in units of rows to display of a correction candidate Braille data string (of the Braille pattern).

  Since only the first row is a blank row, the isolated block detection unit 1051 detects the convex portion and the flat portion included in the second row and the third row as the isolated block 721 (step S410). FIG. 18B shows the state of the isolated block 721 detected.

  Since the isolated block 721 is an aggregate of convex portions and flat portions included in the second row and the third row, the isolated block 721 cannot move downward any further. Furthermore, the same applies to the case of moving to the right and the case of moving to the left. Therefore, in this example, it is assumed that the movement is made in any one of the upper and right directions.

  The isolated block moving unit 1052 first moves the isolated block 721 upward, for example (step S120). The state of the isolated block 721 after the movement is shown in FIG. As shown in FIG. 18C, when a newly created Braille data string is translated, a translation result of “woo” is obtained.

  Since the newly generated braille data string is translated into the ink character data string “wooru”, it is determined that the braille data string does not contain an error (NO in step S204, NO in step S206). Therefore, the newly generated Braille data string of the Braille pattern shown in FIG. 18C is registered in the correction candidate storage unit 106 as a correction candidate Braille data string (step S208). At this time, the movement amount “1” is also registered in association with the correction candidate braille data string.

  Next, the isolated block moving unit 1052 moves the isolated block 721 to the right (step S120). The state of the isolated block 721 after the movement is shown in FIG. As shown in FIG. 18D, when the newly created Braille data string is translated, “(error) (error) becomes (error)”.

  Since an error is included in the result of translating the newly generated Braille data string (YES in Step S204), it is determined that the Braille data string includes an input error. Therefore, the newly generated Braille data string of the Braille pattern shown in FIG. 18D is not registered as a correction candidate Braille data string but is discarded (Step S210).

  Thus, among the new Braille data strings created by moving the isolated block 721 shown in FIG. 18B in only one of the upper, lower, left, and right directions, it is created by moving upward. Only the braille data string is registered as a correction candidate braille data string. Therefore, the braille correction candidate output processing unit 107 outputs the braille pattern indicated by the correction candidate braille data string to the display unit 111.

  Referring to FIG. 18E, the braille pattern indicated by the correction candidate braille data string is displayed in a display area provided integrally with the coordinate information input unit 101, for example. Further, as shown in FIG. 18E, the result of translating the braille data string into the ink character data string may be output together with the ink character display area 40. Further, the difference between the Braille data string 10 including the original error and the newly created Braille data string may be displayed. For example, as shown in FIG. 18E, a message “722 moved upward” may be displayed to indicate that the isolated block 721 has been moved upward. Moreover, it is not restricted to such a form, You may display symbols, such as an arrow which shows a moving direction. Thereby, the user can easily grasp how the input is wrong.

  As described above, in this embodiment, the isolated block is moved only in one of the upper, lower, left, and right directions, but it may be moved in two directions.

  FIGS. 19A to 19E are diagrams for explaining another example of row isolated block movement processing. A case will be described in which the braille data string 10 stored in the braille data storage unit 103 represents a braille (braille character string) 730 as shown in FIG.

  Referring to FIG. 19A, the braille 730 is composed of four braille characters. The braille data corresponding to these braille characters are “000001”, “010011”, “000010”, and “011000”, respectively. In the ink character display area 40, the translation result “(error) (error) (error) (error)” of the braille data string representing the braille 730 is displayed. Since the translation result includes an error, the error detection unit 104 determines that the Braille data string 10 representing the Braille 730 includes an input error.

  FIGS. 19B to 19E also show a procedure from detection and movement of isolated blocks in units of rows to display of a correction candidate braille data string (its braille pattern).

  Referring to FIG. 19B, the isolated block detecting unit 1051 uses only the first line as a blank line, so that the convex part and the flat part included in the second and third lines are defined as isolated blocks 731. It detects (step S410).

  Referring to FIG. 19B, the isolated block 731 is an aggregate of convex portions and flat portions included in the second row and the third row, and therefore cannot move further downward. . Furthermore, the same applies to the case of moving to the right and the case of moving to the left. Therefore, in this example, it is assumed to move upward and then to the left.

  First, the isolated block moving unit 1052 moves the isolated block 731 upward (step S120). The state of the isolated block 731 after the movement is shown in FIG. As shown in FIG. 19C, when a newly created Braille data string is translated, a translation result of “(error) to (error)” is obtained.

  Since an error is included in the result of translating the newly generated Braille data string (YES in Step S204), it is determined that the Braille data string includes an input error. Accordingly, the newly generated braille data string shown in FIG. 19C is discarded without being registered as the correction candidate braille data (step S210).

  In this example, the braille pattern candidate is extracted by moving the isolated block 731 in two directions, and the process returns to step S120 again.

  In step S120, the isolated block moving unit 1052 further moves the isolated block 731 to the left from the state shown in FIG. The state of the isolated block 731 after the movement is shown in FIG. As shown in FIG. 19 (d), when a newly created Braille data string is translated, an ink character data string of “OI” can be obtained.

  Since the newly generated braille data string is translated into the ink character data string “OI”, it is determined that the braille data string does not contain an error (NO in step S204, NO in step S206). Therefore, the new braille data string shown in FIG. 19D is registered as a correction candidate braille data string (step S208).

  As described above, among the braille pattern candidates extracted by moving the isolated block 731 composed of the braille data 730 in either the upper, lower, left, or right directions, it is extracted by moving upward and further moving to the left. Only the braille data string corresponding to the candidate braille pattern is registered as the correction candidate braille data string. Thereby, the braille correction candidate output processing unit 107 outputs the braille pattern indicated by the correction candidate braille data string to the display unit 111.

  Referring to FIG. 19E, the braille pattern indicated by the correction candidate braille data string is displayed in, for example, a display area provided integrally with the coordinate information input unit 101, as in FIG. In addition, for example, a message “732 moved to the upper left” may be displayed to indicate that the isolated block 731 has been moved upward and to the left.

(Specific example of isolated block movement processing in units of columns)
20A to 20G are diagrams for describing a specific example of column isolated block movement processing according to the embodiment of the present invention. Here, it is assumed that four isolated blocks as shown in FIG. 15C are detected.

  FIG. 20A shows the translation results of the braille character strings 661 to 666. In this case, an error is included in the isolated block 682 and the isolated block 683 (a convex portion of an error braille character is included). Therefore, these isolated block 682 and isolated block 683 are to be moved.

  A braille pattern candidate extracted when one of the isolated block 682 and the isolated block 683 is moved, and an ink character data string obtained by translating the braille data string indicated by the braille pattern candidate are shown in FIG. As shown in FIG.

  FIG. 20B shows only the isolated block 682 moved to the left, FIG. 20C shows the isolated block 682 moved to the right, and FIG. 20D shows the isolated block 682. Is moved downward. Further, FIG. 20E shows the isolated block 683 moved to the left, FIG. 20F shows the isolated block 683 moved to the right, and FIG. 20G shows the isolated block. 683 is moved upward.

  As shown in FIG. 20B to FIG. 20G, when a braille data string indicating a braille pattern candidate extracted by moving one of the isolated block 682 and the isolated block 683 is translated, Only the braille pattern candidate shown in 20 (e) is translated as “carrot”. Other Braille pattern candidates include “error” in the translation result. Thereby, the Braille data string representing the Braille pattern shown in FIG. 20E is registered as a correction candidate Braille data string.

  As described above, only the braille data string of the braille pattern shown in FIG. 20E is registered as the correction candidate braille data string. Therefore, the braille pattern shown in FIG. 20 (e) is presented to the user. Presentation of the braille pattern can be performed in the same manner as in FIGS. 18 (e) and 19 (e).

  As described above, in this embodiment, only one of the isolated blocks is moved, but a plurality of isolated blocks may be moved at a time.

  FIGS. 21A and 21B are diagrams for explaining another example of the column isolated block moving process. Again, it is assumed that four isolated blocks as shown in FIG. 15C have been detected.

  FIG. 21A shows an example in which the isolated block 682 and the isolated block 683 are moved in the left direction as a set. FIG. 21B shows an example in which the isolated block 682 and the isolated block 683 are moved in the right direction as a set.

  In both FIG. 21 (a) and FIG. 21 (b), when the braille data string indicating the extracted braille pattern candidates is translated, an error is included. Accordingly, these braille data strings are discarded without being registered as correction candidate braille data strings.

  Although the case where the isolated block 682 and the isolated block 683 are moved in the same direction as a set has been described here, the isolated block 682 and the isolated block 683 may be moved in different directions. For example, braille pattern candidates may be extracted by moving the isolated block 682 to the left and moving the isolated block 683 upward at a time.

  In the above, an example of the case where the isolated block is moved in units of rows or columns has been described, but the case where the isolated block is moved in the braille cell can be similarly performed.

(Specific examples of error detection processing)
FIG. 22 is a diagram for explaining a specific example of the correct / incorrect determination process of the ink character data string by the ink character correct / incorrect determination unit 1042. As shown in FIG. 22A, it is assumed that the braille data string 10 stored in the braille data storage unit 103 represents the braille 900.

  When the Braille translation unit 1041 translates the Braille data string 900 into the ink character data string 901, the translation result is “Let's go to find out”. Thus, no error occurs in the ink character data string 901.

  Next, the ink character correct / incorrect determination unit 1042 determines whether the ink character data string 901 is correct. First, the ink character correct / incorrect determination unit 1042 breaks down the ink character string 901 into parts of speech. More specifically, for example, morphological analysis is performed on the ink character data string 901. Then, the ink character data string 901 includes word data “902”, word data 903 “ha”, word data 904 “ka”, word data 905 “ni”, word data 906 “go”. Disassembled.

  Here, the ink right / wrong determination unit 1042 determines whether there is word dictionary data that matches each of the word data 902 to 906 in the word dictionary storage unit 113. Here, since the word data 902 “IN” does not exist in the word dictionary storage unit 113, the ink character correct / incorrect determination unit 1042 determines that the word data 902 is an error.

  This is because the character data corresponding to the braille data representing the braille character that was accidentally input when the input error was included in the braille data sequence 10 based on the input coordinate information but was translated into the ink character data sequence. It is probable that no error was included in the translation result. However, even in this case, whether the ink character data string is correct or not is determined by, for example, morphological analysis of the ink character data string. That is, if there is word data that does not match the word dictionary data in the ink character data string, the word data is determined to be an error. Thereby, even when an error is not included in the translation result, it can be determined that the Braille data string 10 includes an error.

  A braille pattern candidate is extracted from the braille data string corresponding to the word data 902 determined to be an error in this way.

  For example, when a braille pattern candidate is extracted by moving an isolated block, the movement amount of the isolated block is stored in the correction candidate storage unit 106 in this embodiment. For example, it is assumed that the column isolated block detection process is performed based on the braille patterns of the braille characters 921 to 923 and is detected. In this case, as shown in FIG. 22B, the braille pattern of the braille 907 composed of the braille characters 921A to 923A moves one isolated block only once, and thus the movement amount 910 is “1”. . In addition, since the braille pattern of the braille 908 composed of the braille characters 921B to 923B has moved two isolated blocks once each, the movement amount 910 is “2”. In addition, since the Braille pattern of Braille 909 made up of Braille characters 921C to 923C has moved three isolated blocks a total of four times, the movement amount 910 is “4”.

  In this case, the likelihood index 911 of the correction candidate Braille data string of the new Braille pattern including the Braille pattern of Braille 907 is calculated as “100” by the above-described calculation formula. Further, the likelihood index 911 of the correction candidate braille data string of the new braille pattern including the braille pattern of the braille 908 is calculated as “50” by the above-described calculation formula. Further, the likelihood index of the correction candidate Braille data string of the new Braille pattern including the Braille pattern of Braille 909 is calculated as “25” by the above-described calculation formula.

  The translation results of Braille 907, 908, and 909 are “wain”, “wairo”, and “yon”, respectively.

  FIG. 23 is a diagram for describing a specific example of extracting braille pattern candidates by moving an isolated block.

  Referring to FIG. 23A, when a Braille data string corresponding to Braille 930 is translated into an ink character data string, it becomes “Oonawafu”, and no error is included in the translation result as described above. However, the ink character data string “Oonyowafu” does not match the word dictionary data stored in the word dictionary storage unit 113. In this case, for example, the isolated block 931 in units of columns including the convex portion of the braille pattern corresponding to “Nyova” can be moved to the right. Referring to FIG. 23B, when the braille data string newly generated by moving the isolated block 931 to the right is translated, the translation results are “O” and “Sakafu”. However, neither “O” nor “Sakafu” matches the word dictionary data. Next, referring to FIG. 23C, the isolated block 931 and the isolated block 932 in the column unit of the braille pattern corresponding to “Sakafu” are moved to the left. Then, the translation result will be “Osakafu”. The ink data string “Osakafu” matches the word dictionary data. Therefore, the braille data string indicating the braille 930 ′ illustrated in FIG. 23C is registered as the correction candidate braille data string.

  Thereby, even if the user accidentally translates into the ink character data string even though an input error has occurred when inputting the coordinate information that is the source of the braille data string 10, it is determined as an error. can do. Therefore, the user can easily obtain a correct Braille data string without inputting coordinate information again.

  Note that the functions of the Braille translation device 100 described in the above embodiment may be realized by a computer system using a personal computer, a PDA (Personal Digital Assistant), or the like.

  Even in this case, it can be realized by the same configuration as the hardware configuration shown in FIG. For example, a program for causing the computer system to realize the functions of the Braille translation device 100 and each dictionary data in the Braille-ink character correspondence dictionary storage unit 112 and the word dictionary storage unit 113 are stored in the recording medium 210, and further the hard disk 207. Alternatively, the program and each dictionary data may be transmitted to a computer through a network (not shown) and stored in the hard disk 207.

  This program includes a plurality of instructions for causing a computer to realize the functions of the braille translation device 100. Some of the basic functions necessary to realize these functions are provided by operating system (OS) or third-party programs running on the computer, or modules of various toolkits installed on the computer. Therefore, this program does not necessarily include all functions necessary for realizing the functions of the braille translation device 100. This program need only include instructions for controlling the Braille translation device 100 by calling an appropriate function or “tool” in a controlled manner so as to obtain a desired result. .

  Further, the Braille data string 10 may be recorded on the recording medium 210, and the above-described functions may be realized by the drive device 206 reading the Braille data string 10. Alternatively, the braille data string 10 may be transmitted to a computer through a network (not shown) and stored in the hard disk 207. The correction candidate braille data string output by the braille correction candidate output processing unit 107 may be written on the recording medium 210 by the drive device 206. Alternatively, the correction candidate data string may be transmitted to another computer through a network (not shown).

  Since the operation of the computer system is well known, the description thereof will not be repeated here.

  Moreover, the Braille translation method performed by the Braille translation apparatus of the present invention can also be provided as a program. Such a program can be recorded on an optical medium such as a CD-ROM (Compact Disc-ROM) or a computer-readable recording medium such as a memory card and provided as a program product. A program can also be provided by downloading via a network.

  The provided program product is installed in a program storage unit such as a hard disk and executed. The program product includes the program itself and a recording medium on which the program is recorded.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is a functional block diagram of the braille translation apparatus which concerns on embodiment of this invention. It is a block diagram which shows the hardware constitutions of the braille translation apparatus which concerns on embodiment of this invention. It is a general-view figure of the Braille translation apparatus concerning an embodiment of the invention. It is a figure which shows the structure of the braille cell used in order to express a braille character. It is a figure which shows the example of the data structure of a pattern management part. It is a figure which shows the example of a data structure of a correction candidate memory | storage part. It is a flowchart which shows the main routine which CPU in the braille translation apparatus of embodiment of this invention performs. It is a flowchart which shows a correction candidate extraction process. It is a flowchart which shows the flow of an error detection process. It is a flowchart which shows the flow of a braille square isolated block detection process. It is a figure for demonstrating an isolated block detection process in braille square. It is a flowchart which shows a row isolated block detection process. It is a figure for demonstrating a row isolated block detection process. It is a flowchart which shows a column isolated block detection process. It is a figure for demonstrating a column isolated block detection process. It is a flowchart which shows a correction candidate output process. It is a figure for demonstrating the specific example of a pattern search and replacement process. It is a figure for demonstrating the specific example of a row isolated block movement process in embodiment of this invention. It is a figure for demonstrating the other example of a row isolated block movement process. It is a figure for demonstrating the specific example of a column isolated block movement process in embodiment of this invention. It is a figure for demonstrating the other example of a column isolated block movement process. It is a figure for demonstrating the specific example of the correct / incorrect determination process of the ink character data sequence by an ink character correct / incorrect determination part. It is a figure for demonstrating the specific example which extracts a braille pattern candidate by moving an isolated block.

Explanation of symbols

  10 Braille data string, 30 Braille cell, 31, 32, 33, 34, 35, 36 Point position, 40 Ink display area, 51 Return button, 52 Feed button, 53 Translation button, 100 Braille translation device, 101 Coordinate information input , 102 Braille data conversion unit, 103 Braille data storage unit, 104 detection unit, 105 candidate extraction unit, 106 correction candidate storage unit, 107 Braille correction candidate output processing unit, 108 likelihood calculation unit, 109 pattern correction unit, 110 pattern Management unit, 111 display unit, 112 Braille-ink character correspondence dictionary storage unit, 113 word dictionary storage unit, 201 CPU, 202 ROM, 203 RAM, 204 operation unit, 204.1 power button, 204.2 menu button, 204. 3 Previous candidate selection button, 204.4 Next candidate selection button, 204.5 Correct answer button 206 drive unit, 207 hard disk, 208 bus, 210 recording medium, 211 an electronic pen, 1041 Braille translation unit, 1042 ink character accuracy judgment unit, 1051 an isolated block detection unit, 1052 an isolated block moving unit, 1053 pattern search and replacement unit.

Claims (22)

Braille storage means for storing a plurality of input braille data,
Each of the braille data represents a combination of convex portions arranged at matrix-like point positions in the braille cell of the corresponding braille character,
Further comprising candidate extraction means for extracting braille pattern candidates based on the plurality of braille data,
The braille pattern candidate corresponds to a combination of convex portions of the braille character string configured by moving at least a part of the convex portions of the input braille character string represented by the plurality of braille data to another point position. ,
The braille pattern candidate is erroneous by detecting whether or not the translation result of the candidate braille data string corresponding to the braille pattern candidate extracted by the candidate extraction means matches the dictionary data stored in advance. Error detection means for detecting whether or not,
A braille translation device further comprising candidate output means for outputting a candidate braille data string whose translation result matches the dictionary data among the candidate braille data strings as a corrected candidate braille data string for the plurality of braille data . The error detection means includes translation means for translating the candidate braille data string into an ink letter data string based on the dictionary data;
2. The error detection unit according to claim 1, wherein when the one or more braille data that does not exist in the dictionary data is detected in the candidate braille data string, the candidate braille data string is determined to be erroneous. Braille translator. The error detection means further includes a correctness determination means for determining the correctness of the ink character data string translated by the translation means based on the dictionary data,
The correct / incorrect determination means determines that the ink character data string is incorrect when one or more ink character data not present in the dictionary data is detected in the ink character data string,
3. The Braille translation apparatus according to claim 2, wherein the error detection unit determines that the candidate Braille data sequence is erroneous when the ink character data sequence is determined to be incorrect by the ink character correctness determination unit. The candidate extracting means includes
An isolated block detecting means for detecting an isolated block which is an aggregate including at least one convex portion of the convex portions of the input braille character string and which can be collectively moved to an adjacent flat portion; ,
Moving means for moving the convex portion by the predetermined number of point positions in units of the isolated blocks,
The braille translation device according to claim 2 or 3, wherein a pattern of the braille character string after being moved by the moving means is extracted as the braille pattern candidate.   The isolated block detection means is a set of adjacent convex portions within the braille cell of the braille characters constituting the input braille character string, and the convex portions that can be moved collectively to the adjacent flat portions. The braille translation apparatus according to claim 4, wherein an aggregate of the detected blocks is detected as the isolated block.   The isolated block detection means detects whether or not there is an empty string without a convex part when the combination of convex parts of the input Braille character string is divided for each column, and the empty string exists between them. The braille translation apparatus according to claim 4, wherein an aggregate of convex portions and flat portions included in one or more columns that are not extracted is extracted as the isolated block.   The isolated block detection means detects whether or not there is a blank line without a convex part when the combination of convex parts of the input braille character string is divided for each line, and the blank line exists between them. The braille translation apparatus according to claim 4, wherein an aggregate of convex portions and flat portions included in one or more lines that are not extracted is extracted as the isolated block. Candidate storage means for storing candidate braille data strings that match the dictionary data as the correction candidate braille data strings,
The braille translation device according to claim 2 or 3, wherein the candidate output means outputs the correction candidate braille data string stored in the candidate storage means. A likelihood calculating means for calculating a likelihood index for the plurality of braille data strings of the correction candidate braille data string;
The likelihood calculating unit calculates the likelihood index based on at least one of a moving direction, a moving amount, and a moving number of the isolated block by the moving unit,
The braille translation device according to claim 8, wherein the candidate output unit sorts and outputs the correction candidate braille data string in descending order of the likelihood index. Pattern management means for managing one or more input error patterns, correction patterns for the input error patterns, and likelihood index data of the correction patterns in association with each other;
Pattern correction means for correcting and / or newly adding information managed by the pattern management means,
When the likelihood index calculated by the likelihood calculating unit is equal to or greater than a predetermined threshold, the pattern correcting unit uses the partial pattern after the isolated block is moved by the moving unit as the corrected pattern. The Braille translation apparatus according to claim 9, wherein a new addition is performed by setting a partial pattern before movement of the isolated block by the moving means as an input error pattern and setting the likelihood index as the likelihood index data. Pattern management means for associating and managing one or more input error patterns and a correction pattern for each input error pattern;
Each of the correction patterns is a pattern composed of a combination of convex portions in which the position of the convex portion of the corresponding input error pattern is moved to another point position within a predetermined range,
The candidate extracting means includes
Search means for searching for a partial pattern that matches the input error pattern managed by the pattern management means among the partial patterns constituting the input braille character string;
Replacement means for replacing the searched partial pattern with a corresponding correction pattern,
The braille translation device according to claim 2 or 3, wherein a braille character string pattern after replacement by the replacement means is extracted as the braille pattern candidate.   The braille translation device according to claim 11, wherein the pattern management unit further manages the likelihood index data of the braille pattern candidates extracted by the candidate extraction unit in association with each other.   The braille translation device according to claim 12, further comprising likelihood calculation means for calculating a likelihood index of the correction candidate braille data string based on the likelihood index data managed by the pattern management means.   The braille translation device according to claim 13, wherein the candidate output unit sorts and outputs the corrected candidate braille data string in descending order of likelihood calculated by the likelihood calculating unit. The pattern management means further manages the likelihood index data of each of the correction patterns in association with each other,
Pattern correction means for correcting and / or newly adding information managed by the pattern management means;
Selecting means for selecting one of the plurality of correction candidate braille data strings output by the candidate output means based on an instruction from the user;
14. The Braille according to claim 13, wherein the pattern correction unit increases a value of the likelihood index data of the correction pattern corresponding to the correction candidate Braille data string selected by the selection unit by a predetermined value or a predetermined ratio. Translation device. The candidate extracting means includes
Including an isolated block detection means for detecting an isolated block when the matching partial pattern is not searched by the search means;
The isolated block is an aggregate including at least one convex portion among the convex portions of the input braille character string, and is an aggregate of convex portions that can be collectively moved to adjacent flat portions,
A moving means for moving the convex portion by a predetermined number of point positions to another point position in units of the isolated blocks;
The braille translation device according to claim 12, further comprising extracting a pattern of the braille character string after being moved by the moving means as the braille pattern candidate. Coordinate input means for inputting coordinate information;
The braille translation device according to claim 1, further comprising: conversion means for converting input coordinate information into the plurality of braille data.   The braille translation device according to claim 1, further comprising data input means for receiving input of the plurality of braille data from the outside. The error detection means further detects an error in the plurality of braille data,
The braille translation device according to claim 1, wherein the candidate extracting unit extracts the braille pattern candidates when an error is detected in the plurality of braille data. Detecting whether or not the translation result of the plurality of input braille data matches the dictionary data stored in advance;
Extracting a braille pattern candidate based on the plurality of braille data when a translation result of the plurality of braille data does not match the dictionary data;
Detecting whether or not the translation result of the candidate braille data string corresponding to the extracted braille pattern candidate matches the dictionary data;
A Braille translation method comprising: outputting the candidate Braille data string as a correction candidate Braille data string when a translation result of the candidate Braille data string matches the dictionary data. Detecting whether or not the translation result of the plurality of input braille data matches the dictionary data stored in advance;
Extracting a braille pattern candidate based on the plurality of braille data when a translation result of the plurality of braille data does not match the dictionary data;
Detecting whether or not the translation result of the candidate braille data string corresponding to the extracted braille pattern candidate matches the dictionary data;
A Braille translation program for causing a computer to execute a step of outputting the candidate Braille data string as a correction candidate Braille data string when a translation result of the candidate Braille data string matches the dictionary data.   A computer-readable recording medium on which the Braille translation program according to claim 21 is recorded.

Images