JP2000259616A - Device for collating user defined word dictionary - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a collation device capable of effectively utilizing plural user dictionaries. SOLUTION: This user defined word dictionary collation device storing plural user dictionaries storing at least plural word data and capable of collating word data with character code data and executing processing based on the collated result is provided with a text analysis part 2 for storing user defined word data in a user dictionary table to be a data storing area in the previously determined priority order of user dictionaries, collating word data stored in the user dictionary table with character code data inputted from a text input part 1 and executing processing.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a user-defined word dictionary matching device. In particular, in a device that performs text-to-speech processing on text data and pronounces it as speech, it is intended to effectively apply a user-defined word dictionary used when converting text data, for example, word by word.

[0002]

2. Description of the Related Art Conventionally, text data (data in which characters are converted into predetermined character codes, including data in which kanji and kana character strings are mixed) is subjected to speech synthesis processing, converted into speech synthesis data, and converted into speaker data. (Hereinafter referred to as speech synthesis technology). The speech synthesis technology is performed by a device called a text-to-speech synthesis engine that converts text data into speech synthesis data.

Here, a process of converting text data into speech synthesis data will be considered. The text-to-speech engine collates the input text data with word data recorded in a certain unit (generally a word unit, henceforth referred to as word data), and separates each word. The word data includes a system word data dictionary (hereinafter, referred to as a system dictionary) that stores general registered word data, and a user-defined word data dictionary (hereinafter, referred to as a user dictionary) that stores registered word data uniquely registered by a user. Can be divided into two. Here, when performing collation with text data, word data in the user dictionary is preferentially collated with word data in the system dictionary to generate speech synthesis data. In particular, registered word data registered in a user dictionary may differ from the original character reading method, and it is important to use the user dictionary efficiently in terms of accuracy.

[0004]

However, the conventional text-to-speech synthesis engine does not consider using a plurality of user dictionaries, so that only one user dictionary can be specified. For this reason, for example, even if a user creates a user dictionary in which personal names are registered and a user dictionary in which place names are registered, only one of them can be registered. Therefore, in order to effectively use both word data, there is no other choice but to combine the independently created word data into one and register it as a user dictionary. However, in this case, since the number of word data registered in one dictionary increases, for example, when it is desired to correct the word data of a certain person, it is necessary to spend a very long time searching for the word data. did not become. Further, there is a problem that the capacity of a memory used when editing word data must be increased in accordance with the number of word data.

[0005] Therefore, it has been desired to realize a collation device that can effectively use a plurality of user dictionaries.

[0006]

A user-defined word dictionary matching device according to the present invention stores user-defined word data in a user dictionary table, which is a data storage area, in a predetermined priority order of the user-defined word dictionary. There is provided a user-defined word data storage means, and a collation processing means for collating and processing the user-defined word data stored in the user dictionary table with the input character code data. In the present invention, the user-defined word data storage means determines in advance the order of priority for collation of a plurality of user-defined word dictionaries, and stores them in a user dictionary table, which is a data storage area. Then, the matching processing unit performs a process by comparing the user-defined word data stored in the user dictionary table with the input character code data.

Further, the user-defined word dictionary matching device according to the present invention merges user-defined word data in a plurality of user-defined word dictionaries, sorts them in descending order of the number of words, and stores the data in a user dictionary table as a data storage area. A user-defined word data storage means, and a collation processing means for collating and processing the user-defined word data stored in the user dictionary table with the input character code data. In the present invention, the user-defined word data storage unit merges user-defined word data in a plurality of user-defined word dictionaries and sorts them in descending order of the number of words. Then, the matching processing unit performs a process by comparing the user-defined word data stored in the user dictionary table with the input character code data.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiment 1 FIG. FIG. 1 shows the first embodiment of the present invention.
It is a block diagram of the text-to-speech synthesis engine to which the user-defined word dictionary matching device according to the embodiment is applied. In the figure, 1 is a text input unit. Text input unit 1
For example, text data from a clipboard, a text file, or the like, which is a part for temporarily storing data, is input to. Text data is data obtained by encoding characters, control characters, and the like. Here, the character code includes a full-width code represented by a 2-byte data amount and a half-width code represented by a 1-byte data amount. If the text data includes a half-width code, the text input unit 1 converts all the half-width code to a full-width code corresponding to the half-width code. 2 is a text analysis unit. The text analysis unit 2 divides the text data input to the text input unit 1 into text data in word units.
At that time, the word data stored in the user dictionary 3 and the system dictionary 4 are referred to. Then, the reading data corresponding to the matched word data is transmitted. What is reading data?
How to pronounce the word data (reading kana, accent, etc.)
Are recorded as data.

The voice data generation unit 5 includes a text analysis unit 2
The voice data is generated by referring to the voice data dictionary 6 based on the read data transmitted from the. The audio output unit 7
The audio data generated by the audio data generation unit 5 is transmitted to an audio output device such as a speaker, and sounds. Also, instead of transmitting this to the audio output device, the audio data may be stored in a storage unit or the like as a file.

Here, at least the text input unit 1, text analysis unit 2, and speech data generation unit 5 of the speech synthesis engine are actually constituted by a computer and a program for operating the computer. Therefore, the operations of the text input unit 1, the text analysis unit 2, and the voice data generation unit 5 are performed based on this program. The program is stored in a recording medium or the like, and the computer reads and executes the program.

In this embodiment, by setting a plurality of user dictionaries and merging the word data in the speech synthesis engine, it is not necessary to merge the user dictionaries which can normally be set only once. It is one that allows multiple settings. In addition, by setting the priority, for example, it is possible to set the user dictionary of the personal name with priority for the text data having many personal names, so that the efficiency can be improved.

FIG. 2 is a diagram showing the operation of the speech synthesis engine of FIG. The operation of the speech synthesis engine according to the present embodiment centering on the text analysis unit 2 will be described with reference to the drawings. First, in a text input process S1, text data from a text file, a clipboard, or the like is input to the text input unit 1. Next, in the user dictionary application process S2, the text analysis unit 2 stores the registered word data of each user dictionary in the user dictionary table based on a predetermined priority (in order of higher priority). In this embodiment, an application program interface (hereinafter, referred to as an API) or a system call is used to set the priority of the collation. The system call is used as a means for using a service (a file system operation, an input / output operation, etc.) provided by an operation system (OS). The API is a more sophisticated system call. It is also sometimes used as a system call. For example, a programmer or the like sets a user dictionary name to be used and its priority using an API. At the time of collation, the text analysis unit 2 calls the API, and stores the registered word data of each user dictionary in the user dictionary table in priority order based on the API.

In the word division processing S3, the text analysis unit 2
Refers to the user dictionary table and the system dictionary table, and divides the text data. The division of the text data is performed by collating with the word data stored in the user dictionary table and the system dictionary table. . Here, the user dictionary table and the system dictionary table hold reading data corresponding to each word data. The text analysis unit 2 transmits the read data corresponding to the recognized word data.

In the voice data generation process S4, voice synthesis data is generated based on the read data to be transmitted and the voice data recorded in the voice data dictionary 6. In the voice output processing S5, the generated voice synthesized data is output to a voice output device such as a speaker via the voice output unit 7 to generate sound, or the voice synthesized data is stored as a file.

As described above, according to the first embodiment,
Since the user dictionary is prioritized using the API and the system call to divide the text data, the voice synthesis data can be created by referring to a plurality of user dictionaries. Therefore, it is possible to use a plurality of dictionaries without merging the user dictionaries in advance,
Since a user dictionary for each use or the like can be created, for example, dictionary data can be easily corrected and time can be reduced.

Embodiment 2 In the first embodiment described above, the priorities of the user dictionaries are assigned using an API or a system call, but the present invention is not limited to this. For example, a setting file such as a registry or a resource used in various operation systems may be provided, and the setting may be performed by describing the file. Then, the text analysis unit 2 stores the word data of each user dictionary in the user dictionary table in the order of priority based on the settings described in the file. For example, in the case of a registry, the name of a user dictionary to be used is used as a key and the priority is described as a value in a registry key. In this way, if the setting is made by describing the setting in the setting file, the priority order of the user dictionary can be set externally as compared with the setting using the API, so that it is more versatile and can be easily modified.

Embodiment 3 In the above-described embodiment, a plurality of user dictionaries are assigned priorities and stored in the user dictionary table, and each user dictionary is referred to. However, the present invention is not limited to this. For example, a plurality of user dictionaries are kept as they are, and when storing a table, a plurality of user dictionaries are merged, sorted in the order of data having the largest number of characters, stored, and collated with text data. Therefore, the matching with the text data is performed from the data having the largest number of characters (the data having the largest data size in the case of all double-byte character codes).

Normally, the probability that the character code string in the text data matches certain word data is highest in the sentence represented by the text data, where the word meaning by the word data is included. Moreover, the probability increases as the number of characters increases. Therefore, by performing collation from data having a large number of characters, the accuracy of division of text data can be improved, and speech synthesis processing can be performed more efficiently.

Embodiment 4 In the above-described embodiment, the application of the user dictionary of the speech synthesis engine has been described. However, the present invention is not limited to this. For example, conversion of an input character code such as automatic kanji conversion of a word processor or the like is performed. It can also be applied for the purpose of doing so.

[0020]

As described above, according to the present invention, the user-defined word data storage means preliminarily determines the collation priority order of a plurality of user-defined word dictionaries and stores them in the user dictionary table which is a data storage area. Since the matching processing means performs processing by comparing the user-defined word data stored in the user dictionary table with the input character code data, the user-defined word dictionaries need not be merged in advance. However, since a plurality of user-defined word dictionaries can be used and user-defined word dictionaries for different purposes can be created, for example, user-defined word data can be easily corrected and time can be reduced.

Further, according to the present invention, the user-defined word data storage means merges the user-defined word data in the plurality of user-defined word dictionaries and sorts them in descending order of the number of words. When the number of characters is large, the probability of matching is reduced, but when the number of characters is large, the matching is performed in another sense. The accuracy of the collation result can be improved by utilizing the fact that there is little possibility.

[Brief description of the drawings]

FIG. 1 is a block diagram of a text-to-speech synthesis engine to which a user-defined word dictionary matching device according to a first embodiment of the present invention is applied.

FIG. 2 is a diagram showing the operation of the speech synthesis engine of FIG. 1;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Text input part 2 Text analysis part 3 User-defined word data dictionary 4 System word data dictionary 5 Voice data generation part 6 Voice data dictionary 7 Voice output part

Claims (6)

[Claims] 1. A user-defined dictionary that stores at least a plurality of user-defined word dictionaries storing a plurality of user-defined word data, compares the user-defined word data with character code data, and performs processing based on the result. In the word dictionary matching device, in a user dictionary table that is a data storage area, a user-defined word data storage unit that stores the user-defined word data in a predetermined priority order of the user-defined word dictionary; A user-defined word dictionary matching device, comprising: matching processing means for comparing stored user-defined word data with input character code data to perform processing. 2. The user-defined word dictionary matching device according to claim 1, wherein the priority order is described and set by an application program interface or a system call. 3. The user-defined word data storage unit stores the user-defined word data based on the setting information file, wherein the priority order is described and set in a setting information file. Item 1. The user-defined word dictionary matching device according to item 1. 4. A user-defined dictionary that stores at least a plurality of user-defined word dictionaries storing a plurality of user-defined word data, compares the user-defined word data with character code data, and performs processing based on the result. In the word dictionary matching device, a user-defined word data storage unit that merges user-defined word data in the plurality of user-defined word dictionaries, sorts the data in descending order of the number of words, and stores the data in a user dictionary table that is a data storage area. A user-defined word dictionary matching device, comprising: matching processing means for checking and processing user-defined word data stored in a user dictionary table with input character code data. 5. The apparatus according to claim 1, further comprising a voice conversion unit configured to convert reading data indicating how to read a word represented by the user-defined word data into voice data, wherein the matching processing unit includes the user-defined word data and the character code data. 5. The user-defined word dictionary matching device according to claim 1, wherein the reading data corresponding to the user-defined word data that matches the character code data is output to the voice conversion unit. 6. 6. The user-defined word dictionary matching device according to claim 5, further comprising a voice output unit that converts the voice data into a sound.