JP3402480B2 - Document reading device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document reading device for reading a document input in characters. 2. Description of the Related Art In recent years, a document reading device has been widely used as a device for converting an input document into voice and reading it out. For example, in the case of a Japanese text-to-speech reading device, for example, if a user inputs a document saying "There was a large flower clock in front of the station." There was. " [0003] However, in the above-mentioned conventional document reading device, there is a case where the word cannot be uttered correctly unless all the notations of the words in the document are standard notations. For example, for the document "There was a big clock in front of the station", "Hana" and "Clock" are interpreted as different words, so "Hanato Kei" is uttered. Resulting in. Therefore, the user had to check in advance for any spelling in the document, and if so, had to correct it to the standard notation. An object of the present invention is to solve the above-mentioned conventional problems, and an object of the present invention is to provide a document reading apparatus which can make a correct utterance even if there is a fluctuation of a notation in a document. According to the present invention, a rendaku compound word dictionary, an alternative notation dictionary and a compound word interpretation judging unit are added to a conventional document reading apparatus in order to achieve the above object. is there. According to the present invention, when a series of nouns appears in a document, it is determined whether or not the continuous part of the noun can be interpreted as a fluctuation of the notation of a compound word. For example, by giving priority to the interpretation of compound words, it is not necessary to register all spelling variations such as "Hana Clock" in the word dictionary, and the user checks whether such swaying variations exist in the document. Since there is no necessity, correct vocalization can be performed even if there are fluctuations in notation in the document. An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a functional block diagram of a document reading device according to an embodiment of the present invention. In FIG. 1, reference numeral 1 denotes an input means for inputting a document to be read and start of reading of the document, 2 a document storage unit for storing the input document, 3 word notation and reading and grammatical information and A word dictionary for storing information relating to accents in association with each other, 4 is a word-to-word connection table describing the possibility of connection between words, 5 is a word dictionary 3 and a word-to-word connection cable 4
, A morphological analysis unit that performs a morphological analysis on the document stored in the document storage unit 2 with reference to an analysis result storage unit that stores the analysis result output from the morphological analysis unit 5, and utterances out of compound words Then, a rendaku compound dictionary containing only the compound words that cause the phenomenon of rendaku, and the notation, reading, and readings that are read without rendaku separated by the constituent words of the compound words are stored. 8 is a word of a certain noun. If there is an alternative notation, an alternative notation dictionary that stores the pronunciation, the notation, and the alternative notation in association with each other. Reference numeral 9 refers to the analysis result storage unit 6, and if there is a portion where nouns are continuous, that portion is rendaku. A compound word interpretation judging unit for judging by referring to the alternative notation dictionary 8 whether it can be interpreted as the contents registered in the compound word dictionary 7, a speech unit 10 for uttering a voice, and an analysis result storage unit 6
An utterance driving unit 12 for outputting a voice waveform to be output to the utterance unit 10 with reference to the contents stored in the input unit 1 stores the document input from the input unit 1 in the document storage unit 2,
When the user instructs to read the document, the contents of the document storage unit 2 are passed to the morphological analysis unit 5, the analysis result is stored in the analysis result storage unit 6, and the analysis result is sent to the compound word interpretation determination unit 9. Let me judge whether there is shaking of the notation inside,
The document reading control unit corrects the analysis result, if any, and passes the corrected analysis result to the utterance unit 10 to actually utter. The above configuration can be considered as adding a rendaku compound word dictionary 7, an alternative notation dictionary 8, and a compound word interpretation determining unit 9 to the conventional document reading device. FIG. 2 is a circuit block diagram of the machine translation apparatus according to the present embodiment, wherein 21 is a keyboard, 22 is a central processing unit (hereinafter abbreviated as CPU), 23 is a speaker, and 24 is a speaker.
Is a read only memory (hereinafter abbreviated as ROM), 25 is a random access memory (hereinafter RAM).
Abbreviated. ). In FIG. 2, the input unit 1 shown in FIG. 1 uses a keyboard 21 to operate a word dictionary 3, an inter-word connection table 4, a rendaku compound word dictionary 7, and an alternative notation dictionary 8.
Is stored in the document storage unit 2 and the analysis result storage unit 6 by the ROM 24.
A RAM 25, a morphological analysis unit 5, a compound word interpretation determination unit 9, an utterance drive unit 11, and a document reading control unit 12
2 executes a program stored in the ROM 24 while exchanging data with the ROM 24 and the RAM 25, so that the utterance unit 10
Each has been realized. The operation of the document reading apparatus according to the embodiment of the present invention constructed as described above will be described below with reference to the flowchart of FIG. First, step 3
In step 1, the user inputs a document from the input unit 1. At this time, the document reading control section 12 writes the input document into the document storage section 2. Here, a description is given assuming that the sentence “There was a large flower clock in front of the station” is input. FIG. 4 shows the contents of the document storage unit 2 at this time. Next, in step 32, the user gives an instruction to start reading the document from the input means 1. Then, the document reading control section 12 passes the contents of the document storage section 2 to the morphological analysis section 5. In step 33, the morphological analysis unit 5 performs a morphological analysis on the received document with reference to the word dictionary 3 and the inter-word connection table 4, and writes the result in the analysis result storage unit 6. FIG. 5 shows the contents of the word dictionary 3, and FIG. 6 shows the contents of the inter-word connection table 4. FIG. 7 shows the contents of the analysis result storage unit 6 at this time. In step 34, the document reading control section 12
The compound word interpretation judging unit 9 checks whether or not the contents of the analysis result storage unit 6 include fluctuation of the notation of the compound word, and if so, updates the contents of the analysis result storage unit 6. Issue a command. Then, the compound word company determination unit 9 refers to the contents of the analysis result storage unit 6 and checks whether or not there is a portion where nouns are consecutive in the analysis result. If not, step 3
Proceed to 4. In the above example, “hana” and “clock” are noun continuations. In step 35, the compound word interpretation judging section 9
However, the rendaku compound word dictionary 7 is searched for a compound word in which “the reading of the noun continuation part” and “the reading when the compound word is read without rendaku” are equal. If not, step 34
Proceed to. In the above example, the compound word having “reading without rendaku” is registered in the rendaku compound word dictionary. In addition,
FIG. 8 shows the contents of the rendaku compound word dictionary. In step 36, the compound word interpretation judging section 9
Is replaced by another notation that can be considered with reference to the alternative notation dictionary 8 to determine whether or not the noun continuation portion can be a matched compound notation in step 35. If the matching notation cannot be obtained in step 35, the process proceeds to step 34. In the above example, "Hana" is entered in the alternative notation dictionary 8.
It is shown that the notation is an alternative notation of the notation of "flower", so by replacing "hana" with the notation of "flower", the noun continuation of "hana" and "clock" becomes "flower clock" Can be determined to be a fluctuation of the notation of the compound word ". In this case, the process proceeds to step 37. FIG. 9 shows the contents of the alternative notation dictionary. In step 37, the compound word interpretation judging section 9 corrects the contents of the analysis result storage section 6 for the portion judged to be the sway of the notation, and returns to step 34. FIG. 10 shows the contents of the analysis result storage unit 6 at the time of correction. In step 34, it is checked again whether or not a noun continuous section exists in the analysis result storage section 6. If there are no more noun continuations, the process proceeds to step 38. In step 38, the utterance drive unit 11 refers to the contents of the analysis result storage unit 6 to utter a speech waveform to be passed to the utterance unit 10. In step 39, the voice generator 10 actually outputs the voice waveform generated by the voice driver 11 as voice. This makes it appear to the user that the document is being read. According to the present invention, as is apparent from the above embodiment, a rendaku compound word dictionary, an alternative notation dictionary, and a compound word interpretation judging unit are added to a conventional document reading device. Appears in a document, checks whether the noun sequence can be interpreted as a fluctuation in the notation of the compound word, and if so, gives priority to the interpretation of the compound word, so that it looks like a "hana clock" It is not necessary to register all the notational fluctuations in the word dictionary, and the user does not need to check whether such notational fluctuations exist in the document. Can do it.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a functional block diagram of a document reading device according to an embodiment of the present invention. FIG. 2 is a circuit block diagram of the device. FIG. 3 is a flowchart showing an outline of the operation of the apparatus. FIG. 4 is a list showing contents of a document storage unit in the apparatus. FIG. 5 is a list showing the contents of a word dictionary in the device. FIG. 6 is a list showing the contents of an inter-word connection table in the device. FIG. 7 is a list showing the contents of an analysis result storage unit in the apparatus. FIG. 8 is a list showing the contents of a rendaku compound word dictionary in the device. FIG. 9 is a list showing the contents of an alternative notation dictionary. FIG. 10 is a list showing the contents of an analysis result storage unit. [Description of Signs] 1 input means 2 document storage unit 3 word dictionary 4 inter-word connection table 5 morphological analysis unit 6 analysis result storage unit 7 rendaku compound word dictionary 8 alternative notation dictionary 9 compound word interpretation judgment unit 10 utterance unit 11 utterance drive Part 12 Document reading control part

Claims (1)

(57) [Claims] [Claim 1] Input means for inputting a document to be read and instructing start of reading of the document, a document storage unit for storing the input document, word notation and reading. A word dictionary that stores grammatical information and information about accents in association with each other, an inter-word connection table describing whether or not words can be connected, and the document storage unit with reference to the word dictionary and the inter-word connection table. A morphological analysis unit that performs morphological analysis on the document stored in the morphological analysis unit, an analysis result storage unit that stores the analysis result output from the morphological analysis unit, and only a compound word that, when it occurs, causes a phenomenon called rendaku. A rendaku compound word dictionary in which readings that are collected and read without performing rendaku and notation are separated and stored in compound word units,
An alternative notation dictionary in which, when an alternative notation exists for a word of a business card, the pronunciation, the notation, and the alternative notation are stored in association with each other;
If there is a part where nouns are continuous by referring to the analysis result storage unit, it is determined whether the part can be interpreted as the content registered in the rendaku compound word dictionary by referring to the alternative notation dictionary. A word interpretation determining unit, an uttering unit for uttering a voice, an utterance driving unit that outputs a voice waveform to be output to the uttering unit with reference to the content stored in the analysis result storage unit, and The input document is stored in the document storage unit, and when the reading of the document is instructed from the input unit, the contents of the document storage unit are transferred to the morphological analysis unit, and the analysis result is stored in the analysis result storage unit. Further, let the compound word interpretation determination unit determine whether there is a sway of the notation in the analysis result, correct the analysis result, if any, and pass the corrected analysis result to the speech unit. Actually speak Document reading apparatus having a document reading control unit for.