JP2001014305A - Method and device for electronic document processing, and recording medium where electronic document processing program is recorded - Google Patents

Abstract

PROBLEM TO BE SOLVED: To read aloud an arbitrary document in the form of an electronic document by voice synthesis with high precision without any feeling of incompatibility. SOLUTION: The document processor when receiving a tag file as a tagged document (S1) derives property information for a reading-aloud process from the tag in the tag file and embeds the property information to generate a file for the reading-aloud process (S2). Then the document processor performs a process matched with a voice synthesis engine (S3) by using the generated file for the voice reading-aloud process and performs a process (S4) corresponding to operation that a user does by using a user interface.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an electronic document processing method and an electronic document processing apparatus for processing an electronic document, and a recording medium on which an electronic document processing program is recorded.

[0002]

2. Description of the Related Art Conventionally, WWW (World Wide Web) has been provided as an application service for providing hypertext information in a window format on the Internet.

[0003] The WWW is a system that executes document processing for creating, publishing, or sharing a document and showing the way of a new style document. However, from the viewpoint of practical use of documents, advanced document processing beyond WWW, such as classification and summarization of documents based on the contents of the documents, is required. For such advanced document processing, mechanical processing of the contents of the document is indispensable.

However, mechanical processing of the contents of a document is still difficult for the following reasons. First, HTML (Hyper Text Markup Language), which is a language for describing hypertext, specifies the expression of a document, but hardly specifies the content of the document. Second, hypertext networks constructed between documents are not always easy for readers of the document to understand the contents of the document. Third, although the author of a document generally works without the convenience of the reader in mind, the convenience of the reader of the document is not coordinated with the convenience of the author.

[0005] As described above, WWW is a system showing the way of a new document. However, since the document is not mechanically processed, advanced document processing cannot be performed.
In other words, in order to perform advanced document processing, it is necessary to process the document mechanically.

Therefore, a system for supporting mechanical processing of documents has been developed based on the results of natural language research, with the goal of mechanical processing of documents. As a document processing based on natural language research, mechanical document processing using tags attached to a document has been proposed on the assumption that attribute information about the internal structure of the document by the author of the document, so-called tags are added. I have.

[0007] By the way, a user uses an information search system such as a so-called search engine to search for desired information from a vast amount of information provided via the Internet. This information search system is a system that searches for information based on a specified keyword and provides the searched information to a user. The user selects desired information from the provided information.

[0008] In the information retrieval system, information can be easily retrieved as described above. However, the user reads the retrieved and provided information, understands the outline, and checks whether the information is the desired information. It is necessary to determine whether or not. This operation imposes a heavy burden on the user, especially when the amount of provided information is large. Therefore, recently, a so-called automatic summarization creating system, which is a system for automatically summarizing text information, that is, the contents of a document, has attracted attention.

The automatic summarizing system is a system for summarizing the original information, that is, by reducing the length and complexity of the text information while retaining the meaning of the document. The user can understand the outline of the document by reading the summary created by the automatic summary creation system.

Normally, the automatic summarization creating system treats sentences and words in a text as one unit, and assigns a degree of importance based on some information to order them. Then, the automatic summarizing system collects sentences and words ordered in a higher order to create a summarization.

[0011]

By the way, with the spread of computers and the progress of networking in recent years, high-performance document processing has been demanded. Among them, a function of synthesizing a document by voice and reading it out is demanded. Have been.

[0012] Speech synthesis originally generates speech mechanically based on the results of speech analysis and simulation of a human speech generation mechanism, and digitally controls elements or phonemes of individual languages. To assemble.

However, in speech synthesis, when reading an arbitrary document, it is not possible to read the document in consideration of the breaks in the document, and it is not possible to perform natural reading. Further, in speech synthesis, it is necessary for a user to appropriately select a speech synthesis engine to be used according to a language. Furthermore, in speech synthesis, the accuracy with which words that are prone to reading errors, such as technical terms and difficult words, are read accurately depends on the dictionary used.

The present invention has been made in view of such circumstances, and has an electronic document processing method, an electronic document processing apparatus, and an electronic document processing method capable of reading an arbitrary document by speech synthesis with high accuracy and without a sense of incongruity. An object is to provide a recording medium on which a document processing program is recorded.

[0015]

According to the present invention, there is provided an electronic document processing method for processing an electronic document, comprising the steps of: It is characterized by comprising a voice reading file generating step of generating a reading file.

The electronic document processing method according to the present invention generates a voice reading file based on the electronic document, and reads the electronic document.

The electronic document processing method according to the present invention for achieving the above-mentioned object has a structure in which a plurality of elements are hierarchized in the electronic document processing method for processing an electronic document. It is characterized by comprising a document input step of inputting an electronic document to which tag information shown is given in advance, and a document reading step of reading out the electronic document by voice synthesis based on the tag information.

In the electronic document processing method according to the present invention, an electronic document to which tag information indicating an internal structure in which a plurality of elements are hierarchized is added in advance is input, and the electronic document is converted based on the tag information. Read directly.

Further, the electronic document processing apparatus according to the present invention, which achieves the above-mentioned object, is a digital document processing apparatus for processing an electronic document. It is characterized by comprising a voice reading file generating means for generating.

The electronic document processing apparatus according to the present invention generates a text-to-speech file based on the digital document, and reads the digital document using the text-to-speech file.

Further, an electronic document processing apparatus according to the present invention for achieving the above-mentioned object has an internal structure in which a plurality of elements are hierarchized in an electronic document processing apparatus for processing an electronic document. It is characterized by comprising a document input unit for inputting an electronic document to which tag information is given in advance, and a document reading unit for reading the electronic document by voice synthesis based on the tag information.

Such an electronic document processing apparatus according to the present invention inputs an electronic document to which tag information indicating an internal structure in which a plurality of elements are hierarchized is added in advance, and a tag assigned to the electronic document. Read electronic documents directly based on information.

Further, a recording medium on which an electronic document processing program according to the present invention for achieving the above object is recorded,
In a recording medium on which a computer-controllable electronic document processing program for processing an electronic document is recorded, the electronic document processing program generates a speech-to-speech file for speech synthesis and speech based on the electronic document. It is characterized by having a file generation step.

[0024] The recording medium in which the electronic document processing program according to the present invention is recorded provides an electronic document processing program which generates a file for speech reading based on the electronic document and reads the electronic document.

Further, a recording medium on which an electronic document processing program according to the present invention for achieving the above-mentioned object is recorded is a computer-controlled electronic document processing program for processing an electronic document. The processing program has an internal structure in which a plurality of elements are hierarchized, and a document inputting step of inputting an electronic document to which tag information indicating the internal structure is added in advance, and audio processing of the electronic document based on the tag information And a document reading-out step for reading out by combining.

The recording medium on which the electronic document processing program according to the present invention is recorded receives an electronic document to which tag information indicating an internal structure in which a plurality of elements are hierarchized is added in advance, An electronic document processing program for reading out an electronic document directly based on the electronic document.

[0027]

Embodiments of the present invention will be described below in detail with reference to the drawings.

A document processing apparatus shown as an embodiment of the present invention has a function of synthesizing a given electronic document and a summary sentence created from the electronic document by using a speech synthesis engine and reading out the speech. In the following description, an electronic document is simply referred to as a document.

As shown in FIG. 1, the document processing apparatus includes a main body 10 having a control unit 11 and an interface 12;
An input unit 20 that supplies information input by a user to the main body 10, a receiving unit 21 that receives an external signal and supplies the information to the main body 10, and a communication unit that performs communication processing between the server 24 and the main body 10. 22, an audio output unit 30 for outputting information output from the main unit 10 as audio, a display unit 31 for displaying information output from the main unit 10, and recording and / or reproducing information on a recording medium 33. Recording / playback unit 32
And Hard Disk Drive (HD)
D) 34.

The main body 10 has a control unit 11 and an interface 12, and constitutes a main part of the document processing apparatus.

The control unit 11 is a CPU (Central Processing Unit) 1 that executes processing in the document processing apparatus.
3 and RAM (Random Access Me
mory) 14 and a ROM (Read O
nly Memory) 15.

The CPU 13 performs control for executing the program in accordance with, for example, the program recorded in the ROM 15 or the hard disk. The RAM 14 temporarily stores programs and data necessary for the CPU 13 to execute various processes as necessary.

The interface 12 includes an input unit 20, a receiving unit 21, a communication unit 22, a display unit 31, and a recording / reproducing unit 32.
And a hard disk drive 34. Under the control of the control unit 11, the interface 12 inputs data supplied via the input unit 20, the reception unit 21 and the communication unit 22, outputs data to the display unit 31, and outputs data to the recording / reproduction unit 32. For data input / output, adjust the data input / output timing and convert the data format.

The input section 20 is a section for receiving a user input to the document processing apparatus. This input unit 20
For example, it is composed of a keyboard and a mouse. The user can use the input unit 20 to input a keyword using a keyboard or select and input an element of a document displayed on the display unit 31 using a mouse, for example. Note that the element is an element constituting a document, and includes, for example, a document, a sentence, and a word.

The receiving section 21 receives data transmitted from the outside to the document processing apparatus via, for example, a communication line. The receiving unit 21 receives a plurality of documents that are electronic documents and an electronic document processing program for processing these documents. The data received by the receiving unit 21 is supplied to the main unit 10.

The communication section 22 is composed of, for example, a modem and a terminal adapter, and is connected to the Internet 23 via a telephone line. A server 24 storing data such as documents is connected to the Internet 23.
The communication unit 22 is connected to the server 24 via the Internet 23.
And data can be received from the server 24. The data received by the communication unit 22 is supplied to the main unit 10.

The audio output unit 30 is composed of, for example, a speaker, and receives an electric audio signal obtained by synthesizing a voice by a voice synthesizing engine or the like and various other audio signals via the interface 12 and converts the audio signal into audio. And output.

The display unit 31 inputs and displays character information and image information via the interface 12. The display unit 31 is, for example, a cathode ray tube (CR).
T) and Liquid Crystal Display (LC)
D), for example, one or more windows are displayed, and characters, graphics, and the like are displayed on these windows.

The recording / reproducing unit 32 records and / or reproduces data on a removable recording medium 33 such as a floppy disk, an optical disk, or a magneto-optical disk under the control of the control unit 11. Also, the recording medium 3
3 records an electronic document processing program for processing a document and a document to be processed.

The hard disk drive 34 records and / or reproduces data on a hard disk which is a large-capacity magnetic recording medium.

Such a document processing apparatus receives a desired document and displays it on the display unit 31 as follows.

In the document processing apparatus, first, the user operates the input unit 20 to start a program for performing communication via the Internet 23, and inputs a URL (Uniform Resource Locator) of the server 24 (search engine). , The control unit 11 controls the communication unit 22 and the server 2
Access 4

In response, the server 24 outputs the data of the search screen to the communication unit 22 of the document processing device via the Internet 23. CPU1 in the document processing device
3 outputs the data to the display unit 31 via the interface 12 and causes the display unit 31 to display the data.

In the document processing apparatus, when the user inputs a keyword or the like on this search screen using the input unit 20 and instructs a search, the communication unit 22 transmits the keyword to the Internet 23.
, A search command is transmitted to the server 24 as a search engine.

Upon receiving the search command, the server 24 executes the search command and transmits the obtained search result to the communication unit 22 via the Internet 23. In the document processing apparatus, the control unit 11 controls the communication unit 22 and
, And a part of the search result is displayed on the display unit 31.

Specifically, when the user instructs a search by inputting a keyword such as “TCP” using the input unit 20, the document processing apparatus includes the keyword “TCP” from the server 24. Various information is transmitted, and the following documents are displayed on the display unit 31.

"TCP / IP (Transmission Control Protocol)
It is no exaggeration to say that the history of the (Internet Protocol) is the history of computer networks in North America, or even the world. And the history of TCP / IP cannot be told without ARPANET. ARPANET has the official name Advance
d Research Project Agency Network, the US Department of Defense's DOD (Departme
nt of Defense) Defense Advanced Research Projects Agency (DARPA: Defense A)
advanced research project agency) sponsored and built for experimental and research packet switching networks. 1969 Four universities on the west coast of North America,
ARPANET departed from a very small network of research institute host computers connected by a 50kbps line. At the time, ENIAC, the world's first computer in 1945
Was developed at the University of Pennsylvania and became the first IC in 1964
Was implemented as a theoretical element, and a mainframe general-purpose computer series that formed the history of the third generation of computers was developed. Given this historical background, such a project that anticipated the future of computer communications could be said to have been unique to the United States. "

In this document, its internal structure is described by attribute information by tagging described later. Document processing in the document processing device is performed with reference to a tag attached to the document. In this embodiment, a semantic / pragmatic tag is added to a document to enable a mechanical understanding of the content of the document between multiple languages, in addition to a syntactic tag indicating the structure of the document. .

As a syntactic tagging, there is a tagging that describes a tree-like internal structure of a document. That is, in the present embodiment, as shown in FIG. 2, the internal structure by this tagging, each element such as a document, a sentence, and a vocabulary element, a normal link, a reference / referenced link, and the like are added to the document in advance as tags. Has been granted. In the figure, a white circle “○” is a document element such as a vocabulary, a segment, a sentence, that is, an element.
Is the vocabulary element corresponding to the lowest level word in the document. A solid line indicates a normal link (normal lin) indicating a connection between elements of the document such as a word, a phrase, a section, and a sentence.
k). A broken line is a reference link indicating a dependency relationship between a reference and a referenced. The internal structure of a document is the document (documen
t), subdivision, paragraph (paragr)
aph), sentence (sentence), subsentential segment (subsentential segment), ..., vocabulary element. Of these, subdivisions and paragraphs are optional.

On the other hand, the semantic / pragmatic tagging includes tagging of a syntactic structure indicating a dependency, for example, a referent of a pronoun, or semantic like a meaning of a polysemy. There is something that describes the information. Tagging in the present embodiment is performed in HTML.
XML (eXte) similar to (Hyper Text Markup Language)
nsible Markup Language).

Here, an example of the internal structure of the tagged document is shown below, but the tagging of the document is not limited to this method. In the following, examples of English and Japanese documents are shown, but it should be noted that the description of the internal structure by tagging is applicable to other languages as well.

For example, "Time flies like an arrow."
For the sentence, <sentence><noun phrase meaning = “Time0”> time </ noun phrase><verbphrase><verb meaning = “fly1”> flies </ verb><adjective verb phrase><adjective verb meaning = “Like0”> like <
/ Adjective verb><nounphrase> an <noun word meaning = "arrow0"> arrow </ noun></ noun phrase></ adjective verb phrase></ verb phrase>. Can be.

Here, <sentence>, <noun>, <noun phrase>,
<Verb>, <verb phrase>, <adjective verb>, <adjective verb phrase>
Is a preposition phrase or postposition phrase / adjective phrase / adjective phrase / adjective phrase / including sentence, noun, noun phrase, verb, verb phrase, and adjective, respectively.
It represents the syntactic structure of a sentence such as an adjective verb. Tags are arranged corresponding to immediately before the front end and immediately after the end of the element. The tag placed immediately after the end of the element indicates that it is the end of the element by the symbol "/". Elements are syntactic constructs, ie
Show phrases, clauses and sentences. Note that the word sense = “ti
"me0" has multiple meanings of the word "time",
This means that it is the 0th meaning of the plural meanings. Specifically, “time” includes a noun and a verb, but here, “time” indicates that it is a noun.
In addition to this example, for example, the word “orange” has at least the meaning of a plant name, a color, and a fruit, and these can also be distinguished by their meanings.

In a document processing apparatus using such a document, as shown in FIG.
01 can indicate a syntactic structure. Window 1
In 01, the vocabulary element is displayed on the right half 103, and the internal structure of the sentence is displayed on the left half 102. In this window 101, not only a document described in Japanese but also a document described in an arbitrary language such as English can display a syntactic structure.

More specifically, in the right half 103 of the window 101, the following tagged document “City where the B meeting of Mr. A has finished, The newspaper has stated on paper that it will voluntarily regulate its photo reporting. " An example of tagging this document follows.

<Document><sentence><adjective verb phrase relation = “place”><nounphrase><adjective verb phrase place = “C city”><adjective verb phrase relation = “subject”><noun phrase identifier =
“B meeting”><adjective verb phrase relation = “own”><person name identifier = “Mr. A”> Mr. A </ person name></ adjective verb phrase><organization name identifier = “B meeting”> B meeting < / Organization name><Nounphrase></ Adjective verb phrase> Finished </ Adjective verb phrase><Place name identifier = "C city">
C city </ place name></ noun phrase>, </ adjective verb phrase><adjective verb phrase relation = "subject"><noun phrase identifier = "newspaper" syntactic = "parallel"><nounphrase><adjective verb Phrases> Some </ adjective verb phrases> Popular newspaper </ noun phrase> and <noun> general paper </ noun></ noun phrase></ adjective verb phrase><adjective verb phrase> Relationship = "object"><Adjective verb phrase relation = “contents” subject = “newspaper”><adjective verb phrase relation =
"Object"><nounphrase><adjective verb phrase><noun co-reference = "B-kai"> so </ noun></ adjective verb phrase> photo coverage </ noun phrase></ adjective verb phrase> The self-regulated </ adjective verb phrase> policy was clarified as </ adjective verb phrase><adjective verb phrase relation = "position"></ adjective verb phrase> on paper. </ Text></text>

In this document, "part of popular paper and general paper" is indicated by the tag "syntax" as "parallel". The definition of parallel is to share a dependency relationship. If nothing is specified, for example, <noun phrase relation = “x”><noun> A <//
Noun><Noun> B </ Noun></ Noun phrase> indicates that A has a dependency on B.

The relation = “x” represents a relation attribute.
This relationship attribute describes the syntactic, semantic, and rhetorical relationships. Grammar functions such as subjects, objects, and indirect objects, subject roles such as actors, actors, and beneficiaries, and rhetorical relations such as reasons and results are described by the relation attributes. The relation attribute is represented in the form of relation = ***. In this embodiment, relational attributes are described for relatively easy grammatical functions such as subjects, objects, and indirect objects.

In this document, for example, “A
For proper nouns such as "Mr.", "B-kai", and "C city",
Attributes are described by tags such as a place name, a person name, and an organization name.
These words to which tags such as place names, personal names, and organization names are given are proper nouns.

The document processing apparatus can receive the document tagged in this way. The document processing device is a CP
When the voice reading program among the electronic document processing programs recorded in the ROM 15 or the hard disk is activated by U13, the document is read through a series of steps as shown in FIG. First, a simplified description of each step will be given, and then a detailed description of each step will be given using a specific document example.

First, as shown in the figure, the document processing apparatus receives a tagged document in step S1. It is assumed that a tag required for performing speech synthesis is added to this document, as described later.
Also, the document processing device receives the tagged document,
It is also possible to create a document by newly adding a tag necessary for performing speech synthesis to the document. Further, the document processing apparatus may receive an untagged document, perform tagging on the document including a tag necessary for performing speech synthesis, and create a tag file. In the following, a tagged document received or created and prepared in this manner is referred to as a tag file.

Subsequently, in step S2, the document processing apparatus generates a voice reading file based on the tag file under the control of the CPU 13. As will be described later, the voice reading file is generated by deriving attribute information for reading from a tag in the tag file and embedding the attribute information.

Subsequently, in step S3, the document processing device performs a process suitable for the speech synthesis engine using the speech reading file under the control of the CPU 13. The speech synthesis engine may be configured by hardware or may be realized by software. When the speech synthesis engine is implemented by software, the application program is stored in the ROM 15 or the hard disk of the document processing apparatus in advance.

Then, in step S4, the document processing apparatus performs a process in accordance with an operation performed by a user using a user interface described later.

By performing such processing, the document processing apparatus can synthesize a given document by speech and read it out. Each of these steps will be described in detail below.

First, the reception or creation of a tagged document in step S1 will be described. The document processing device is, for example, as described above, the server 2 shown in FIG.
4 and receives the document as a result of the search based on the keyword or the like. Further, the document processing apparatus receives the tagged document, and newly adds a tag necessary for performing speech synthesis to the document to create the document. Further, the document processing apparatus can receive a document that has not been tagged, tag the document with a tag necessary for performing speech synthesis, and create a tag file.

Here, it is assumed that a tag file in which a document in Japanese or English as shown in FIG. 5 or FIG. 6 is tagged is received or created. That is, the original document of the tag file shown in FIG. 5 is the following Japanese document.

"[Nice aging] / 8 Cancer metastasis can be suppressed! Cancer has been the leading cause of death in Japan for more than 10 years.
The mortality rate is increasing with age. When thinking about the health of the elderly, you cannot avoid cancer.
Characterizing cancer is cell proliferation and metastasis. There are two types of human cells: "cancer genes", which act as accelerators in automobiles and rapidly grow cancer, and "cancer suppressing genes" that act as brakes. There is no problem if both are balanced. Loss of normal regulatory function and nonbreaking mutations in the cell initiate cancer growth. In the elderly, the mutation accumulates over the years, increasing the proportion of cells with cancerous conditions, leading to more cancers. By the way, without the other characteristic, metastasis, cancer doesn't have to be so afraid. This is because a complete cure is possible just by resection. Here is the importance of suppressing metastasis. This metastasis does not occur simply by increasing the number of cancer cells. Cancer cells dissolve proteins (proteins) between cells, create their own way, and enter blood vessels and lymph vessels. In recent years, it has been elucidated that complicated movements such as searching for a new “dwelling house” while circulating are performed. "

When receiving the Japanese document, the document processing apparatus displays the document in a window 110 displayed on the display unit 31, as shown in FIG. Window 1
Reference numeral 10 denotes a document name display section 111 on which a document name is displayed.
A keyword input section 112 for inputting a keyword, an abstract creation execution button 113 as an execution button for creating an abstract sentence of a document as described later, and a reading execution button 114 as an execution button for executing voice reading aloud.
And the like, and a display area 130 where a document is displayed. At the right end of the display area 130, there are provided a scroll bar 131 and buttons 132 and 133 for moving the scroll bar 131 up and down. Move directly up and down or press buttons 132 and 133 to scroll
By moving 1 vertically, the display content displayed in the display area 130 can be scrolled in the vertical direction.

On the other hand, the original document of the tag file shown in FIG. 6 is the following English document.

"During its centennial year, The Wall
Street Journal will report events of the past cen
tury that stand as milestones of American business
history. THREE COMPUTERS THAT CHANGED the face of
personal computing were launched in 1977. That ye
ar the Apple II, Commodore Pet and Tandy TRS came
to market.The computers were crude by today's sta
ndards.Apple II owners, for example, had to use t
heir television sets as screens and storeddata on
audiocassettes. "

When receiving the English document, the document processing apparatus displays the document in a window 140 displayed on the display unit 31, as shown in FIG. Window 14
Reference numeral 0 denotes a document name display section 141 for displaying a document name, a keyword input section 142 for inputting a keyword, and a summary creation execution button 143 as an execution button for creating a summary of a document, similarly to the window 110. In addition, the display area 150 is divided into a display area 150 in which a read-out execution button 144 or the like, which is an execution button for executing voice reading-out, is displayed, and a display area 160 in which a document is displayed. At the right end of the display area 160, a scroll bar 161 and buttons 162 and 162 for moving the scroll bar 161 up and down are provided.
163 is provided, and the user moves the scroll bar 161 up and down directly by using, for example, a mouse or the like of the input unit 20, or moves the scroll bar 161 up and down by pressing the buttons 162 and 163.
The display content displayed at 0 can be scrolled in the vertical direction.

The Japanese or English document shown in FIG. 5 or FIG. 6 is configured as a tag file as shown in FIG. 7 or FIG. 8, respectively.

Here, the tag file shown in FIG. 7 has the heading “[Nicely Aged] /
8 cancer metastases can be suppressed! ? "
In the same paragraph (B), the last paragraph, “This metastasis does not occur just by increasing the number of cancer cells. The cancer cells dissolve proteins and the like between cells and create their own path,
Get into blood vessels and lymph vessels. In recent years, it has been elucidated that complicated movements such as searching for a new “dwelling house” while circulating are performed. "
The remaining paragraphs have been omitted. in this case,
The actual tag file is composed of one file from the heading to the last paragraph.

<Heading> in the heading portion shown in FIG. 9A indicates that this portion is a heading.
In addition, a tag or the like indicating that the related attribute is “condition” or “means” is added to the last paragraph shown in FIG. Further, in the last paragraph shown in FIG. 3B, an example of a tag necessary for performing the above-described speech synthesis is shown.

First, as a tag required for performing speech synthesis, there is a tag such as "protein" which is added when information indicating a reading pseudonym is given to an original document. In other words, in this case, in order to prevent the text from being read aloud in duplicate with “protein tan”, the reading attribute information of pronunciation = “null” is described.
A tag is added to prohibit the reading of "(protein)". In addition, the tag indicates information indicating that the tag has a special function.

Also, tags required for performing speech synthesis are attached to technical terms such as “lymphatic vessels” and difficult-to-learn portions such as “dwelling house” that may be erroneously read aloud. There is something. That is, in this case, in order to prevent the pronunciation of “Rinpakuda” or “Sumie”, pronunciation = “Rinpaku” and pronunciation =
Reading attribute information indicating a reading pseudonym “Sumika” is described.

On the other hand, the tag file shown in FIG. 8 is provided with a tag indicating that it is a supplementary sentence or a tag indicating that a plurality of sentences are continuously formed as one sentence. In addition, as a tag necessary for performing speech synthesis in this tag file, pronunciation attribute information of pronunciation = “two” is described for Roman numeral “II”. This is described in order to prevent "II" from being read out as "two" when it is desired to be read out as "two".

Further, for example, when a citation is included in a document, although not shown, such a tag file
A tag indicating that the sentence is a quote is attached. Further, when there is a question sentence in the document, for example, a tag indicating that the sentence is a question sentence is added to the tag file, though not shown.

In step S1 shown in FIG. 4, the document processing apparatus receives or creates a document to which a tag necessary for performing speech synthesis has been added.

Next, the generation of the voice reading file in step S2 will be described. The document processing device derives attribute information for reading out from the tag in the tag file, and embeds this attribute information,
Generate a speech file.

More specifically, the document processing apparatus finds tags indicating the start positions of paragraphs, sentences and phrases of the document, and embeds attribute information for reading out in correspondence with these tags. Further, as described later, when a document summary is created, the document processing apparatus finds the start position of a portion included in the summary from the document, and embeds attribute information for increasing the volume when reading out. It can also be emphasized that the part is included in the summary sentence.

The document processing apparatus generates a speech reading file as shown in FIG. 9 or FIG. 10 from the tag file shown in FIG. 7 or FIG. The file for speech reading shown in FIG. 9A corresponds to the excerpt of the heading part shown in FIG. 7A earlier, and the file for speech reading shown in FIG. This corresponds to the excerpt of the last paragraph previously shown in FIG. It goes without saying that the actual voice reading file is composed of one file from the heading to the last paragraph.

The attribute information Com = Lang = *** is embedded in the text-to-speech file shown in FIG. 9A corresponding to the start position of the document. This attribute information indicates the language describing the document. Here, the attribute information is Com = Lang = JPN, indicating that the language describing the document is Japanese. In the document processing apparatus, by referring to the attribute information, it is possible to select an appropriate speech synthesis engine corresponding to the language for each document.

Further, the voice reading files shown in FIGS. 9A and 9B include Com = begin_p and Com = begin_s.
And attribute information Com = begin_ph is embedded.
These pieces of attribute information indicate the start positions of paragraphs, sentences, and phrases of the document, respectively. The document processing device detects the start positions of at least two of these paragraphs, sentences, and phrases based on the tags in the tag file described above. In the text-to-speech file, for example, for portions where tags representing syntactic structures of the same level appear continuously, such as <adjective verb phrase><nounphrase> in the above-described tag file, the corresponding numbers are used. Com = begin_ph is not embedded,
It is put together and one Com = begin_ph is embedded.

[0086] Further, the voice reading file includes Co
According to m = begin_p, Com = begin_s and Com = begin_ph,
Attribute information of Pau = 500, Pau = 100, and Pau = 50 are respectively embedded. Each of these attribute information,
It shows that there are pause periods of 500 ms, 100 ms, and 50 ms when reading out. That is, at the start position of the paragraph, sentence and phrase of the document,
Documents are read aloud by the speech synthesis engine with pause periods of 500 ms, 100 ms, and 50 ms, respectively. In addition, these attribute information is Com = begin_p, Com
= begin_s and Com = begin_ph are embedded. Therefore, for example, a portion in which tags representing syntactic structures of the same level appear consecutively, such as <adjective verb phrase><nounphrase> in a tag file, is regarded as one phrase, and a corresponding number of Pau = 50 is not embedded, but one Pau = 50 is embedded. Also, for example, in portions where tags representing different levels of syntactic structure appear consecutively, such as <paragraph><sentence><nounphrase> in the tag file, corresponding Pau = *** is embedded. . For this reason, when reading out such a portion, the document processing apparatus reads out, for example, a pause period of 650 milliseconds obtained by adding the pause periods of paragraphs, sentences, and phrases of the document. As described above, the document processing apparatus provides a pause period corresponding to a paragraph, a sentence, and a phrase, for example, so that the length becomes shorter in the order of the paragraph, the sentence, and the phrase. Speaking can be performed without discomfort taking into account. Note that this pause period is 500 milliseconds at the start position of the paragraph, sentence, and phrase of the document, respectively.
The time does not need to be 00 ms and 50 ms, and can be changed as appropriate.

Further, in the voice reading file shown in FIG. 13B, the pronunciation attribute information of pronunciation = “null” described in the tag file corresponds to
“(Tanpaku)” has been removed, and “lymphatic vessels” and “dwellers” have been replaced by “Rinpkan” in accordance with the reading attribute information of pronunciation = “Rinpaku” and pronunciation = “Sumika”, respectively. , "Sumika". By embedding such reading attribute information, the document processing apparatus does not make a reading error due to a defect in the dictionary referred to by the speech synthesis engine.

Further, the voice reading file includes a tag based on a tag indicating a quote included in the document.
Attribute information for designating only this quoted sentence to use another speech synthesis engine may be embedded.

Further, attribute information for raising the intonation of the ending of the sentence may be embedded in the voice reading file based on the tag indicating the question sentence.

Furthermore, if necessary, attribute information for converting a so-called "Dana-tona" style into a "Dan-Masuna" style can be embedded in the voice reading file. In this case, the document processing apparatus does not embed such attribute information in the text-to-speech file, but converts the style of "Dana-tona" to the style of "Dan-ma-tona" and converts the text-to-speech file. You may make it generate | occur | produce.

On the other hand, the text-to-speech file shown in FIG. 10 has attribute information Com = Lang = ENG embedded therein corresponding to the start position of the document, and the language describing the document is English. Is shown.

[0092] In addition, the file for voice reading aloud includes Com =
The attribute information of Vol = *** is embedded. This attribute information indicates the volume at the time of reading out. For example, Com = Vol =
0 indicates that the document is read out at the default volume of the document processing apparatus. Com = Vol = 80 indicates that the default volume is read out at a volume increased by 80%. Any Com = Vol = *** is valid until the next Com = Vol = ***.

Further, in the voice reading file, "II" is replaced by "two" in accordance with the reading attribute information of pronunciation = "two" described in the tag file.

The document processing apparatus generates such a text-to-speech file through a series of steps shown in FIG.

First, in step S11, the document processing apparatus analyzes the received or created tag file by the CPU 13 in step S11. Here, the document processing apparatus determines the language in which the document is described, and searches for the starting position of the paragraph, sentence, and phrase of the document and the reading attribute information based on the tag.

Subsequently, the document processing device proceeds to step S12
, The CPU 13 embeds Com = Lang = *** at the start position of the document according to the language in which the document is described.

Subsequently, the document processing device proceeds to step S13
, The CPU 13 replaces the start positions of the paragraphs, sentences and phrases of the document with the attribute information in the voice reading file. That is, the document processing apparatus replaces <paragraph>, <sentence>, and <*** phrase> in the tag file with,
Replace with Com = begin_p, Com = begin_s and Com = begin_ph.

Subsequently, the document processing device proceeds to step S14.
, The same level of syntactic structure appears and the same Com = begin _ ***
Into one Com = begin _ ***.

Subsequently, the document processing device proceeds to step S15
, Pau = *** is embedded by the CPU 13 in correspondence with Com = begin _ ***. That is, the document processing device
Embed Pau = 500 just before om = begin_p, embed Pau = 100 just before Com = begin_s, and Pau = 5 just before Com = begin_ph
Embed 0.

Then, the document processing device proceeds to step S16.
, The CPU 13 replaces the correct reading based on the reading attribute information. That is, the document processing apparatus, based on pronunciation attribute information of pronunciation = “null”,
“(Tanpaku)” is removed, and “lymphatic vessel” and “dweller” are changed to “Rinpkan” and “Sumika” based on the reading attribute information of pronunciation = “Rinpaku” and pronunciation = “Sumika”, respectively. ".

In step S2 shown in FIG. 4, the document processing apparatus automatically performs the processing shown in FIG. 11 to automatically generate a file for reading out voice. The document processing device stores the generated voice reading file in the RAM.
14 is stored.

Next, the processing using the voice reading file in step S3 in FIG. 4 will be described. The document processing device uses the read-aloud file to
Under the control of the CPU 13, processing suitable for the speech synthesis engine stored in advance in the M15 or the hard disk or the like is performed.

More specifically, the document processing apparatus selects a speech synthesis engine to be used based on the attribute information Com = Lang = *** embedded in the speech reading file. The speech synthesis engine is provided with an identifier according to a language, a type of male voice / female voice, and the like, and the information is recorded on the hard disk as, for example, an initialization file. The document processing device refers to the initialization file and selects the speech synthesis engine of the identifier corresponding to the language.

Further, the document processing device converts Com = begin _ *** embedded in the voice reading file into a format suitable for the voice synthesis engine. For example, the document processing apparatus marks Com = begin_p with a number in the 100s, such as Mark = 100, and sets Com = begin_s to 10 such as Mark = 1000.
Mark with 00's number, Com = begin_ph to Mark = 1
Mark with a number in the 10000s range, such as 0000.

Further, in the text-to-speech file, since the attribute information of the volume is expressed as a percentage of the increase with respect to the default volume, such as Vol = ***, the document processing apparatus stores the attribute information in this attribute information. Based on the information, the percentage information is converted into absolute value information.

In step S3 shown in FIG. 4, the document processing apparatus performs processing using such a text-to-speech file so that the text-to-speech file can be read by the text-to-speech engine. To a different format.

Next, the operation using the user interface in step S4 in FIG. 4 will be described. The document processing apparatus activates the speech synthesis engine when the user operates the mouse or the like of the input unit 20 and presses the read-out execution button 114 or the read-out execution button 144 shown in FIG. 5 or FIG. Then, the document processing apparatus displays a user interface window 170 as shown in FIG.

User interface window 17
0 has a play button 171 for reading a document, a stop button 172 for stopping reading, and a pause button 173 for temporarily stopping reading as shown in FIG. The user interface window 170 includes a cue button 174, a rewind button 175, and a fast-forward button 176 for cueing, rewinding, and fast-forwarding in sentence units, and a cueing, rewinding, and fast-forwarding in paragraph units. Cue button 177, rewind button 178 and fast forward button 179
And a cue button 180, a rewind button 181 and a fast-forward button 182 for cueing, rewinding and fast-forwarding in phrase units. Further, the user interface window 170 has selection switches 183 and 184 for selecting whether to read the entire sentence or a summary sentence created as described later. In addition,
Although not shown here, the user interface window 170 is, for example, a button for increasing or decreasing the volume, a button for increasing or decreasing the reading speed, a male voice /
A button or the like for changing a voice such as a female voice may be provided.

The document processing apparatus performs a reading operation by the speech synthesis engine when the user operates / presses / selects these various buttons / switches by operating the mouse or the like of the input unit 20, for example. For example, the document processing apparatus starts reading the document by pressing the play button 171 by the user, and jumps to the start position of the currently read sentence by pressing the cue button 174 during reading. And read again. Further, the document processing apparatus can make such a jump in mark units at the time of reading aloud by the marking performed in step S3 in FIG. That is, when the user presses the rewind button 178 or the fast forward button 179 using, for example, a mouse or the like of the input / output unit 20, the document processing apparatus outputs a paragraph having a number in the 100s, such as Mark = 100. Jumps by identifying only the mark indicating the start position of. Similarly, when the user presses the rewind button 175, the fast-forward button 176, the rewind button 181, and the fast-forward button 182 using a mouse or the like of the input / output unit 20, for example,
= 1000, Mark = 10000, 1000s, 1000s
Jump is performed by identifying only the mark indicating the start position of the sentence or phrase, which is the zeroth number. In this way, the document processing apparatus can respond to a request that a user wants to repeatedly reproduce a desired portion in a document, for example, by performing a jump in paragraphs, sentences, and phrases in reading.

In step S4, the document processing device determines
When the user performs an operation using such a user interface, the document is read out by the speech synthesis engine. The read information is output from the audio output unit 30.

In this way, the document processing apparatus can read out a desired document by the speech synthesis engine without feeling uncomfortable.

Next, a description will be given of a reading process when a summary sentence of a document is created. First, here, a process of summarizing the tagged documents to generate a summary sentence will be described with reference to FIGS.

In the document processing apparatus, when a summary of a document is created, the user operates the input unit 20 to execute the automatic summary creation mode while the document is displayed on the display unit 31. Command. That is, the document processing apparatus drives the hard disk drive 34 under the control of the CPU 13 to activate an automatic summary sentence creation program among the electronic document processing programs stored in the hard disk. In the document processing apparatus, the CPU 13 controls the display unit 31 to display an initial screen for an automatic summary sentence creating program as shown in FIG. here,
A window 190 displayed on the display unit 31 includes a document name display unit 191 for displaying the name of the document, a keyword input unit 192 for inputting a keyword, and a summary creation execution button for executing a summary sentence of the document. A display area 200 on which buttons 193 and the like are displayed, a display area 210 on which a document is displayed, and a display area 22 on which a summary of the document is displayed
0.

In the document name display section 191 of the display area 200, the document name of the document displayed in the display area 210 is displayed. In addition, a keyword for creating a summary of a document is input to the keyword input unit 192 using, for example, the keyboard of the input unit 20. The summary creation execution button 193 is an execution button for starting execution of a summary creation process of the document displayed in the display area 210 by being pressed using, for example, the mouse of the input unit 20.

In the display area 210, a document is displayed.
At the right end of the display area 210, a scroll bar 211,
Buttons 212 and 213 for moving the scroll bar 211 up and down are provided. The user can directly move the scroll bar 211 up and down by using, for example, a mouse of the input unit 20 or scroll by pressing the buttons 212 and 213. By moving the bar 211 up and down, the display content displayed in the display area 210 can be scrolled in the vertical direction. By operating the input unit 20, the user can select and summarize a part of the document displayed in the display area 210, or can summarize the entire document.

In display area 220, a summary is displayed. In the figure, since a summary has not been created yet, nothing is displayed in this display area 220. The user can change the display range (size) of the display area 220 by operating the input unit 20. Specifically, the user can enlarge the display range (size) of the display area 220 shown in FIG. 14, for example, as shown in FIG.

The document processing device allows the user to operate the input unit 2 for example.
When the user presses the summary creation execution button 193 with the mouse 0 or the like to turn it on, the process shown in FIG. 15 is executed under the control of the CPU 13 to create a summary sentence.

The process of creating a summary sentence from a document is executed based on tagging relating to the internal structure of the document. In the document processing apparatus, the size of the display area 220 of the window 190 can be changed as shown in FIG. In the document processing apparatus, under the control of the CPU 13, after the window 190 is newly drawn on the display unit 31 or the size of the display area 220 is changed, the digest creation execution button 193 is operated. In some cases, a process of creating a summary sentence from a document at least a part of which is displayed in the display area 210 of the window 190 so as to match the display area 220 is executed.

First, as shown in FIG. 15, the document processing apparatus performs a process called active diffusion under the control of the CPU 13 in step S21. In the present embodiment, the document is summarized by employing the central activity value obtained by the activity diffusion as the importance. That is, in a document tagged with an internal structure, by performing active diffusion, a central activation value corresponding to the tagging with respect to the internal structure can be given to each element.

Here, the active diffusion is a process for giving a high central activity value to an element related to an element having a high central activity value. In other words, the active diffusion means that the central activity values become equal between the element expressed in anaphora (coreference) and its antecedent, and otherwise, each central activity value converges to the same value. . Since the central activity value is determined according to tagging relating to the internal structure of the document, it can be used for analysis of the document in consideration of the internal structure.

The document processing apparatus executes active diffusion by going through a series of steps shown in FIG.

First, as shown in the figure, the document processing apparatus initializes each element under the control of the CPU 13 in step S41. The document processing device assigns the initial value of the central activity value to all the elements except the vocabulary element and the vocabulary element. For example, the document processing apparatus sets the initial value of the central activity value to “1” for all elements except the vocabulary element.
And “0” is assigned to the vocabulary element. In addition, the document processing apparatus can reflect the bias of the initial value in the central activity value obtained as a result of activity diffusion by assigning a non-uniform value to the initial value of the central activity value of each element in advance. For example, the document processing apparatus can obtain a central activity value reflecting the interest of the user by setting a high initial value of the central activity value for an element of interest to the user.

Regarding the reference / referenced link, which is a link having a dependency relationship by reference / reference between elements, and the normal link, which is another link, the end point activation value of the end point of the link connecting the elements. Is set to “0”. The document processing apparatus stores the initial value of the end point activation value thus assigned, for example, in the RAM 14.

Here, an example of a connection structure between elements is shown in FIG. In the figure, an element _Ei and an element _Ej are shown as a part of the structure of the element and the link that constitute the document. Element E
_i and element E _j are central activation values e _i and e _j , respectively.
And are connected by a link L _ij . The end point of the link L _ij connected to the element E _i is T _ij , and the end point of the link L _ij connected to the element E _j is T _ji . Element E _i
, In addition to the elements E _j, which is connected by a link L _ij, the link L _ik, L _il and L _im element E _k (not shown) by, respectively connected to the E _l and E _m. Element E _j is connected to element E _i connected by link L _ji
Besides, they are connected respectively to the link L _uk, L _jq and L elements E _p (not shown) by _jr, E _q and E _r of.

[0125] Then, the document processing device, in step S42 in FIG. 16, under the control of the CPU 13, performs initialization of a counter for counting the elements E _i of a document. That is, the document processing apparatus sets the counter value i of the counter for counting the elements to “1”. Thus, the counter will be that refers to the first element E _1.

Subsequently, the document processing device proceeds to step S43.
, Under the control of the CPU 13, a link process for calculating a new central activation value is executed for the element referred to by the counter. This link processing will be further described later.

Subsequently, the document processing device proceeds to step S44.
, Under the control of the CPU 13, it is determined whether or not the calculation of a new central activation value has been completed for all elements in the document.

Here, when the document processing apparatus determines that the calculation of the new central activation value has been completed for all the elements in the document, the document processing apparatus shifts the processing to step S45. If it is determined that the calculation of the new central activity value has not been completed for all the elements, the process proceeds to step S47.

More specifically, the document processing device is a CPU 13
Under the control of, it is determined whether or not the counter value i of the counter has reached the total number of elements included in the document. If the document processing apparatus determines that the counter value i of the counter has reached the total number of elements included in the document, it is determined that all elements have been calculated, and the process proceeds to step S45. . On the other hand, when the document processing apparatus determines that the counter value i of the counter has not reached the total number of elements included in the document, it is determined that the calculation has not been completed for all elements, and the process proceeds to step S47. Transition.

The document processing apparatus sets the counter value i
Determines that the total number of elements included in the document has not been reached, in step S47
Under the control of 3, the count value i of the counter is incremented by “1” and the count value of the counter is incremented by “i +
As a result, the counter refers to the (i + 1) th element, that is, the next element. Then, the document processing apparatus shifts the processing to step S43, calculates the endpoint activation value, and follows the calculation. A series of steps is performed for the next (i + 1) th element.

When the document processing device determines that the counter value i of the counter has reached the total number of elements included in the document, the document processing device determines in step S45 that the CPU 1
Under the control of 3, the average value is calculated for the change in the central activity value of all the elements included in the document, that is, the change in the newly calculated central activity value from the original central activity value.

Under the control of the CPU 13, the document processing apparatus reads, for example, the original center activity value stored in the RAM 14 and the newly calculated center activity value for all elements included in the document. The document processing apparatus divides the sum of the respective changes of the newly calculated central activity value from the original central activity value by the total number of elements included in the document, thereby obtaining the change in the central activity value of all the elements. Calculate the average value. The document processing device calculates the average value of the change in the central activity value of all the elements calculated in this way,
For example, it is stored in the RAM 14.

Then, the document processing apparatus proceeds to step S46.
In step S45 under the control of the CPU 13,
It is determined whether or not the average value of the change in the central activity values of all the elements calculated in the above is within a preset threshold value. Then, when the document processing device determines that the change is within the threshold, the document processing apparatus ends the series of steps.
On the other hand, when the document processing device determines that the change is not within the threshold, the process proceeds to step S42, where the count value i of the counter is set to “1” and the central activation value of the element of the document is set. Execute the series of steps to be calculated again.
In the document processing apparatus, each time the loop of steps S42 to S46 is repeated, the amount of change gradually decreases.

The document processing apparatus can perform active diffusion in this way. Next, a link process executed in step S43 to perform the active diffusion will be described with reference to FIG. The flowchart shown in the figure, but shows the processing for one element E _i, this processing is to be performed for all elements.

[0135] First, the document processing apparatus, as shown in the figure, in step S51, under the control of the CPU 13, the counter for counting the links one element E _i and one end of a document is connected Perform initialization. That is, the document processing apparatus sets the count value j of the counter for counting links to “1”. This counter will refer to the first link L _ij connected to element E _i .

Subsequently, the document processing device proceeds to step S52.
, Under the control of the CPU 13, the elements E _i
It is determined whether or not the link L _ij is a normal link by referring to the tag of the relation attribute for the link L _ij connecting the link L _ij and E _j . The document processing device uses the link L _ij
Is a normal link indicating the relationship between the vocabulary element corresponding to the word, the sentence element corresponding to the sentence, the paragraph element corresponding to the paragraph, and the like, and a reference link indicating the dependency relationship by reference / reference. Judge. When the document processing device determines that the link L _ij is a normal link, the document processing device shifts the processing to step S53 and executes the link L _ij
If it is determined that is a reference link, step S5
The processing shifts to step 4.

If the document processing device determines that the link L _ij is a normal link, it proceeds to step S53.
An end point T _ij connected to the normal link L _ij of the element E _i
To calculate a new end point activity value.

In step S53, step S52
It is clear from the determination in that the link _Lij is a normal link. The new endpoint activation value t _ij of the endpoint T _ij connected to the normal link L _ij of the element E _i is:
Of the end point activation values of the element E _j, the end point activation values t _jp , t _jq , t _{jr of} all the end points T _jp , T _jq , T _jr connected to the links other than the link L _ij , and the element E _i is the link L _ij
And the central activity value e _j of the element E _j connected by the following _formula, and the value obtained by this addition is divided by the total number of elements included in the document.

Under the control of the CPU 13, the document processing device reads necessary endpoint activation values and central activation values from the RAM 14, for example. The document processing device calculates a new endpoint activity value of the endpoint connected to the normal link as described above for the endpoint activity value and the center activity value that have been read.
Then, the document processing apparatus stores the new endpoint activation value calculated in this way in, for example, the RAM 14.

[0140] On the other hand, the document processing device link if L _ij is determined not to be normal link, in step S54, the end point T which is connected to the reference link element E _i
A process for calculating the endpoint activity value of _ij is performed.

In step S54, step S52
It is clear from the determination in that the link _Lij is a reference link. Point activation values t _ij endpoint T _ij that is connected to the reference link L _ij of the element E _i, of the end-point activation value of the element E _j, all endpoints T _uk which is connected to the link, excluding the link L _ij, T _jq, point activation value t _uk of t _jr, is obtained by adding t _jq, and t _jr, a central activation value e _j of the element E _j of the element E _i is connected by a link L _ij.

Under the control of the CPU 13, the document processing device reads necessary endpoint activation values and central activation values from, for example, the endpoint activation values and the central activation values stored in the RAM 14. The document processing apparatus calculates a new endpoint activity value connected to the reference link as described above using the read endpoint activity value and the central activity value. Then, the document processing apparatus stores the calculated endpoint activation value in the RAM 1
4 is stored.

The processing of the normal link in step S53 and the processing of the reference link in step S54 go from step S52 to step S55, and are referred to by the count value i as shown in a loop returning to step S52 via step S57. Element E
_This is executed for all the links L _ij connected to _i .
In step S57, the are increments the count value j for counting the links connected to the element E _i.

The document processing apparatus performs these steps S53
Alternatively, after performing the processing of step S54, step S55
, Under the control of the CPU 13, the elements E _i
It is determined whether or not the endpoint activation values have been calculated for all the links connected to. If the document processing apparatus determines that the end point activation values have been calculated for all the links, the process proceeds to step S56, and determines that the end point activation values have not been calculated for all the links. In this case, the process proceeds to step S57.

[0145] Here, the document processing device, when it is determined that the end-point activation value for all links is calculated, in step S56, under the control of the CPU 13, the central activity value e of the element E _i Perform _i update.

[0146] new value of central activation value e _i of the element E _i, i.e. updated value is currently a central activation value e _i of the element E _i, the new end-point activation values of all of the end points of the element E _i It is obtained by taking the sum e _i '= e _i + _{ t _j '. Here, the prime “′” means a new value. Thus, the new central activity value is obtained by adding the element's original central activity value to the sum of the new endpoint activity values for the endpoints of the element.

Under the control of the CPU 13, the document processing device reads necessary endpoint activation values from, for example, endpoint activation values and center activation values stored in the RAM 14. The document processing device performs the calculations as described above and its elements E _i
Is calculated. Then, the document processing apparatus stores the calculated new central activation value e _{i in} , for example, the RAM 14.
To memorize.

In this way, the document processing device calculates a new central activation value for each element in the document. Then, the document processing apparatus executes the active diffusion in step S21 in FIG.

Subsequently, in step S22 in FIG. 15, the document processing apparatus, under the control of the CPU 13, sets the window 1 previously displayed on the display unit 31 shown in FIG.
Size of the display area 220 of 90, i.e., sets the maximum number of characters that can be displayed on the display area 220 and W _s. Further, the document processing apparatus initializes the digest S under the control of the CPU 13 and sets the initial value S ₀ = “”. this is,
Indicates that no strings are present in the summary. Document processing device, thus set, the initial value S ₀ of the maximum number of characters W _s and summary S can be displayed in the display area 220, is stored in, for example, the RAM 14.

Subsequently, the document processing device proceeds to step S23.
, Under the control of the CPU 13, the count value i of a counter that counts the sequential creation of the skeleton of the summary sentence is set to “1”. That is, the document processing apparatus sets i = 1 for the count value. The document processing apparatus stores the count value i set in this way in, for example, the RAM 14.

Subsequently, the document processing device proceeds to step S24.
, Under the control of the CPU 13, for the count value i of the counter, the skeleton of the sentence having the ith highest average central activity value is extracted from the text to be summarized. Here, the average central activity value is the average of the central activity values of the elements constituting one sentence. The document processing apparatus reads, for example, the digest S _i-1 stored in the RAM 14 and adds the extracted character string of the skeleton of the sentence to the digest S _i-1 to obtain the digest S _i . Then, the document processing device, a summary S _i obtained in this way, is stored in, for example, the RAM 14. At the same time, the document processing device, creates a list l _i the central activation value order of the elements that are not included in the backbone of the sentence, and stores the list l _i for example the RAM 14.

That is, in step S24, the document processing device selects sentences in descending order of the average central activity value by using the result of the activity diffusion under the control of the CPU 13, and sets the skeleton of the selected sentence. Is extracted. The skeleton of the sentence is
It consists of required elements extracted from the sentence. Required elements can be element heads (hea
d) and subject, object, indirect object, owner (posessor), cause (c
an element having a relation attribute of “ause”, “condition” or “comparison”, and an element directly included in the coordination structure when the related element having the coordination structure is an essential element. The document processing apparatus connects the essential elements of the sentence to generate a sentence skeleton,
Add to the summary.

Subsequently, the document processing device proceeds to step S25.
Under the control of the CPU 13, the length of the digest S _i , that is, the number of characters is displayed in the display area 22 of the window 190.
0 often determines whether or not than the maximum number of characters W _s of.

If the document processing apparatus determines that the number of characters of the digest S _i is greater than the maximum character count W _s , the process proceeds to step S 30, where the digest S _i-1 is finalized under the control of the CPU 13. This is set as a brief summary, and a series of processing ends. In this case, since the summary S _i = S ₀ = “” is output, the summary is not displayed in the display area 220.

On the other hand, when the document processing apparatus determines that the number of characters of the digest S _i is not larger than the maximum number of characters W _s ,
The process proceeds to step S26, and under the control of the CPU 13, the central activity value of the sentence having the (i + 1) -th highest average central activity value and the central activity value among the elements of the list l _i created in step S24. Is compared to the central activity value of the element with the higher. Then, the document processing apparatus determines that the central activity value of the sentence having the (i + 1) -th highest average central activity value is the list l _i
If it is determined that the central activity value of the element having the highest central activity value is higher than the central activity value of the element having the highest
The processing shifts to 28. On the other hand, the document processing device
If it is determined that the central activity value of the sentence having the first highest average central activity value is not higher than the central activity value of the element having the highest central activity value among the elements of the list l _i ,
The process moves to step S27.

The document processing apparatus determines that the central activity value of the sentence having the (i + 1) -th highest average central activity value is not higher than the central activity value of the element having the highest central activity value among the elements of list l _i . In this case, in step S27, under the control of the CPU 13, the count value i of the counter is incremented by "1", and in step S24
Return processing to

The document processing apparatus determines that the central activity value of the sentence having the (i + 1) -th highest average central activity value is higher than the central activity value of the element having the highest central activity value among the elements of the list l _i. If so, in step S28, under the control of the CPU 13, the element e having the highest central activation value among the list l _i elements is summarized S.
to generate a SS _i in addition to the _i, further, to remove the element e from the list l _i. Then, the document processing apparatus stores the summary SS _i generated in this manner in, for example, the RAM 14.

Subsequently, the document processing device proceeds to step S29.
In, under the control of the CPU 13, the number of characters in summary SS _i it is determined whether or not more than the maximum number of characters W _s of the display area 220 of the window 190. Document processing device, when the number of summary SS _i is determined to not more than the maximum number of characters W _s repeats the processing from step S26.
On the other hand, when the document processing apparatus determines that the number of characters of the digest SS _i is larger than the maximum number of characters W _s , the process proceeds to step S31.
, Under the control of the CPU 13, the summary S _i is set as the final summary sentence, displayed in the display area 220, and the series of processing is terminated. In this way, the document processing apparatus generates a summary so as not more than the maximum number of characters W _s.

By performing such a series of processing, the document processing apparatus can summarize a tagged document and create a summary sentence. When the document processing apparatus summarizes, for example, the document shown in FIG. 13, the document processing apparatus creates a summary sentence as shown in FIG. 19 and displays it in the display area 220 of the display range.

That is, the document processing apparatus states, "The history of TCP / IP cannot be described without ARPANET.
In 1969, ARPANET departed from a small network connecting host computers of four universities and research institutes on the west coast of North America with 50 kbps lines. At that time, a general-purpose computer series of mainframes was developed in 1964. Given this historical background, such a project that anticipated the future of computer communications could be said to have been unique to the United States. Is created and displayed in the display area 220.

In the document processing apparatus, instead of reading the entire sentence of the document, the user reads the summary sentence to understand the outline of the sentence and determines whether or not the sentence is the desired information. can do.

In the document processing apparatus, as a method for assigning importance to elements in a document, it is not always necessary to use the active diffusion as described above. May be weighted by the tf * idf method, and the total weight of words appearing in the document may be used as the importance of the document. See “K.Ze
chner, Fast generation of abstracts from general d
omain text corporaby extracting relevant sentence
s, In Proc. of the 16th International Conference o
n Computational Linguistics, pp. 986-989, 1996 ". In addition, a method of assigning importance may be other than these methods. Further, the keyword input unit 192 of the display area 200 may be used. By inputting a keyword into, the importance can be set based on the keyword.

The document processing apparatus can enlarge the display range of the display area 220 of the window 190 displayed on the display unit 31 as shown in FIG.
When the display range of the display area 220 is changed in a state where the created summary text is displayed in the display area 220, the information amount of the summary text can be changed according to the display range. In this case, the document processing device performs the processing shown in FIG.

That is, as shown in the figure, the document processing apparatus responds to the user operating the input unit 20 under the control of the CPU 13 in step S61,
Display area 2 of window 190 displayed on display unit 31
It waits until the display range of 20 is changed.

Then, the document processing device operates the display area 220.
Is changed, the process proceeds to step S62, and the display range of the display area 220 is measured under the control of the CPU 13.

The following steps S63 to S65
The process performed in step S22 is the same as the process performed after step S22 in FIG. 15, and a summary corresponding to the display range of the display area 220 is created, and the process ends.

That is, the document processing apparatus performs step S6.
3, the display area 22 is controlled under the control of the CPU 13.
The total number of characters of the summary sentence displayed in the display area 220 is determined based on the measurement result of the display range of 0 and the character size specified in advance.

Subsequently, the document processing device proceeds to step S64.
, Under the control of the CPU 13, sentences or words are selected from the RAM 14 in descending order of importance so that the created summary does not exceed the number of characters determined in step S63.

Then, the document processing device proceeds to step S65.
In step S64 under the control of the CPU 13,
A summary sentence is created by joining the sentences or words selected in the above, and is displayed in the display area 220 of the display unit 31.

By performing such processing, the document processing apparatus can newly create a summary sentence according to the display range of the display area 220. For example, when the user expands the display range of the display area 220 by dragging the mouse of the input unit 20 with the mouse, the document processing apparatus newly creates a more detailed summary sentence, and as shown in FIG. The summary is displayed in the display area 220 of the window 190.

That is, the document processing apparatus states, "The history of TCP / IP cannot be described without ARPANET.
Is an experimental and research packet-switched network sponsored by the U.S. Department of Defense DOD's Defense Advanced Research Projects Agency. In 1969, ARPANET departed from a very small network connecting host computers of four universities and research institutes on the west coast of North America with 50 kbps lines. At that time it was the world's first computer in 1945
ENIAC was developed at the University of Pennsylvania, and in 1964 the first mainframe general-purpose computer series that implemented ICs as theoretical elements was developed. Given this historical background, such a project that anticipated the future of computer communications could be said to have been unique to the United States. Is created, and the display area 2
20 is displayed.

As described above, in the document processing apparatus, when the displayed summary is too simple to grasp the outline of the document, the user can increase the display range of the display area 220 to increase the display range. A more detailed summary sentence having the information amount of can be referred to.

When the document processing apparatus creates the digest of the document in this way, when the CPU 13 activates the voice reading program among the electronic document processing programs recorded in the ROM 15 or the hard disk, the CPU 13 executes the process shown in FIG.
By going through a series of steps as shown in (1), a document or an abstract can be read aloud. Here,
First, the document shown in FIG. 6 will be described as an example.

First, the document processing apparatus receives a tagged document in step S71 as shown in FIG. As described above, this document is provided with a tag necessary for performing speech synthesis, and is configured as a tag file shown in FIG. Further, the document processing apparatus can receive a tagged document and create a document by newly adding a tag necessary for performing speech synthesis to the document. Further, the document processing apparatus may receive an untagged document, perform tagging on the document including a tag necessary for performing speech synthesis, and create a tag file. This step corresponds to step S1 in FIG.

Subsequently, the document processing device proceeds to step S72.
In step (1), under the control of the CPU 13, a digest of the document is created by the method described above. Here, since the document serving as the source of the summary is tagged as shown in step S71, the tag corresponding to the document is also added to the created summary.

Subsequently, the document processing device proceeds to step S73.
Then, under the control of the CPU 13, a speech reading file for all contents of the document is generated based on the tag file. The voice reading file is generated by deriving attribute information for reading from the tag in the tag file and embedding the attribute information.

At this time, as described above, the attribute information indicating the volume at the time of reading aloud Com = Vol = *** is embedded in the voice reading file. Here, the document processing apparatus embeds the attribute information Com = Vol = 80 in element units for the start position of the part included in the summary sentence created in step S72 in the entire contents of the document.
For the start position of the other part, attribute information of Com = Vol = 0 is embedded. That is, the document processing apparatus sets the default volume to 8 for the part included in the summary sentence.
Read aloud at 0% increased volume. Note that the sound volume does not need to be increased by 80% from the default sound volume, and can be changed as appropriate. As described above, the document processing apparatus can emphasize the part included in the summary sentence even when reading out, and can call the user's attention. This step corresponds to step S2 in FIG.

Subsequently, the document processing device proceeds to step S74.
Under the control of the CPU 13, a process suitable for the speech synthesis engine stored in advance in the ROM 15, the hard disk, or the like is performed using the speech reading file. This step corresponds to step S3 in FIG.

Then, the document processing apparatus proceeds to step S75.
, Processing is performed according to the operation performed by the user using the above-described user interface. This step corresponds to step S4 in FIG. The document processing device reads out the summary sentence created in step S72 by, for example, selecting the selection switch 184 of the user interface window 170 shown in FIG. 12 using the mouse or the like of the input unit 20 by the user. Can be targeted. In this case, the document processing apparatus can start reading out the summary sentence, for example, when the user presses the play button 171 using the mouse or the like of the input unit 20. In addition, the user can use the selection switch 18 by, for example,
When the user selects 3 and presses the play button 171, the reading of the document is started as described above. At this time, based on the attribute information embedded in the voice reading file in step S73, the document processing device reads the portion included in the summary sentence while increasing the volume.

By performing such processing, the document processing apparatus can read out a given document or a prepared summary sentence. Further, when reading out a given document, the document processing apparatus can change the reading style according to the created summary.

As described above, the document processing apparatus automatically generates a text-to-speech file from a given document, and reads a document and a summary sentence from the document using a suitable speech synthesis engine. Can be.

The present invention is not limited to the above-described embodiment, and it goes without saying that, for example, tagging a document or a file for reading out aloud is not limited to the above. .

Further, in the above-described embodiment, the description has been made assuming that the document is transmitted from the outside to the communication unit 22 via the telephone line. However, the present invention is not limited to this. For example, the present invention can be applied to a case where a document is transmitted via a satellite or the like. In addition, the document may be read from the recording medium 33 in the recording / reproducing unit 32 or the document may be recorded in the ROM 15 in advance.

Further, in the above-described embodiment,
Although the voice reading file is generated from the received or created tag file, the voice reading file may be directly read based on the tag file without generating such a voice reading file.

In this case, after receiving or creating the tag file, the document processing apparatus uses the speech synthesis engine to generate a paragraph, a sentence, and a phrase based on the tag indicating the paragraph, sentence, and phrase attached to the tag file. , And a predetermined pause period is provided at the start position of these paragraphs, sentences and phrases to read out. As described above, the tag file is provided with attribute information for prohibiting reading aloud and attribute information indicating a reading kana or pronunciation, and the document processing device removes the portion where reading is prohibited and , And perform reading aloud by replacing with correct reading or pronunciation. In addition, the document processing apparatus operates the above-described user interface during reading, and the unit of the paragraph, the sentence, and the phrase based on the tag indicating the paragraph, the sentence, and the phrase attached to the tag file. Can be used to perform cueing, fast forward or rewind when reading out.

Thus, the document processing apparatus can directly read out a document based on the tag file without generating a voice reading out file.

Further, in the present invention, it is possible to easily realize, as the recording medium 33, a disk-shaped recording medium or a tape-shaped recording medium in which the above-mentioned electronic document processing program is written.

Further, in the above-described embodiment, the mouse of the input unit 20 is exemplified as a device for operating various windows displayed on the display unit 31, but the present invention is not limited to this. Needless to say.
For example, a tablet, a light pen, or the like can be used as such a device.

Further, in the above-described embodiment,
Although Japanese and English documents have been illustrated, it is understood that the present invention is applicable to any language.

As described above, it goes without saying that the present invention can be appropriately changed without departing from the spirit of the present invention.

[0191]

As described above in detail, the electronic document processing method according to the present invention includes a voice reading file generating step of generating a voice reading file for voice synthesis based on an electronic document. Prepare.

Therefore, the electronic document processing method according to the present invention makes it possible to read out an electronic document by generating a voice reading file based on the electronic document.

In the electronic document processing method according to the present invention, there is provided a document input step of inputting an electronic document having an internal structure in which a plurality of elements are hierarchized, and to which tag information indicating the internal structure is added in advance. And a text-to-speech process in which an electronic document is voice-synthesized and read out based on tag information.

Therefore, in the electronic document processing method according to the present invention, an electronic document to which tag information indicating an internal structure in which a plurality of elements are hierarchized is added in advance is input, and the electronic document is directly converted based on the tag information. Enables reading aloud.

Further, the electronic document processing apparatus according to the present invention is provided with a voice reading file generating means for generating a voice reading file for voice synthesis based on the electronic document.

Therefore, the electronic document processing device according to the present invention can generate a text-to-speech file based on an electronic document, and can read a digital document using this text-to-speech file.

Further, the electronic document processing apparatus according to the present invention has a document input means for inputting an electronic document having an internal structure in which a plurality of elements are hierarchized, and to which tag information indicating the internal structure is added in advance. And a text-to-speech means for reading out an electronic document by voice synthesis based on the tag information.

Therefore, the electronic document processing apparatus according to the present invention inputs an electronic document to which tag information indicating an internal structure in which a plurality of elements are hierarchized is added in advance, and the tag information added to this electronic document is input. The electronic document can be read aloud directly based on.

An electronic document processing program in a recording medium on which an electronic document processing program according to the present invention is recorded is a speech reading file for generating a speech reading file for speech synthesis and reading based on an electronic document. The method includes a generation step.

Therefore, the recording medium on which the electronic document processing program according to the present invention is recorded can provide an electronic document processing program which generates a file for reading out voice and reads out the electronic document. Therefore, the device provided with the electronic document processing program can read out the electronic document.

Further, the electronic document processing program in the recording medium on which the electronic document processing program according to the present invention is recorded has an internal structure in which a plurality of elements are hierarchized,
The method includes a document input step of inputting an electronic document to which tag information indicating an internal structure is added in advance, and a document reading step of reading out an electronic document by voice synthesis based on the tag information.

Therefore, the recording medium on which the electronic document processing program according to the present invention is recorded inputs an electronic document in which tag information indicating an internal structure in which a plurality of elements are hierarchized is added in advance, and the tag information is used as the tag information. An electronic document processing program that reads an electronic document directly based on the electronic document can be provided. Therefore, an apparatus provided with the electronic document processing program can input an electronic document and read it directly.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a document processing apparatus shown as an embodiment of the present invention.

FIG. 2 is a diagram showing an internal structure of a document.

FIG. 3 is a view for explaining display contents of a display unit, and is a view showing a window in which an internal structure of a document is displayed by a tag.

FIG. 4 is a flowchart illustrating a series of processes when reading out a document.

FIG. 5 is a diagram illustrating an example of a received or created Japanese document, and is a diagram illustrating a window displaying the document.

FIG. 6 is a diagram illustrating an example of a received or created English document, and is a diagram illustrating a window displaying the document.

FIG. 7 is a diagram showing a tag file which is a tagged Japanese document shown in FIG. 5;

FIG. 8 is a diagram showing a tag file which is the tagged English document shown in FIG. 6;

FIG. 9 is a diagram showing a voice reading file generated from the tag file shown in FIG. 7;

FIG. 10 is a diagram showing a voice reading file generated from the tag file shown in FIG. 8;

FIG. 11 is a flowchart illustrating a series of processes when generating a voice reading file.

FIG. 12 is a diagram showing a window for a user interface.

FIG. 13 is a diagram illustrating a window displaying a document.

14 is a diagram illustrating a window displaying a document, and is a diagram illustrating a state where a display area for displaying a summary sentence is larger than the display area illustrated in FIG. 13;

FIG. 15 is a flowchart illustrating a series of processes when creating a summary sentence.

FIG. 16 is a flowchart illustrating a series of processes when performing active diffusion.

FIG. 17 is a diagram showing a connection structure of elements for explaining an active diffusion process.

FIG. 18 is a flowchart illustrating a series of processes when performing link processing of active spread.

FIG. 19 is a diagram showing a window displaying a document and its summary.

FIG. 20 is a flowchart illustrating a series of processes when a new summary is created by changing the display range of the display area for displaying the summary.

FIG. 21 is a diagram showing a window displaying a document and its summary sentence, and showing a state in which the summary sentence is displayed in the window shown in FIG. 14;

FIG. 22 is a flowchart illustrating a series of processing when a summary sentence is created and a document is read aloud.

[Explanation of symbols]

10 body, 11 control unit, 12 interface, 13 CPU, 14 RAM, 15 ROM,
20 input section, 21 receiving section, 22 communication section, 3
0 audio output unit, 31 display unit, 32 recording / reproducing unit, 33 recording medium, 34 HDD

Claims (61)

[Claims] 1. An electronic document processing method for processing an electronic document, comprising: a voice reading file generating step of generating a voice reading file for voice synthesis and reading based on the electronic document. Electronic document processing method. 2. The electronic document processing according to claim 1, further comprising a processing step of performing a process suitable for a voice synthesis engine using the voice reading file generated in the voice reading file generating step. Method. 3. The electronic document processing according to claim 1, wherein the electronic document has an internal structure in which a plurality of elements are hierarchized, and tag information indicating the internal structure is added in advance. Method. 4. The electronic document is provided with tag information indicating at least a paragraph, a sentence, and a phrase among a plurality of elements constituting the electronic document. 4. The electronic document processing method according to claim 3, wherein a paragraph, a sentence, and a phrase constituting the electronic document are identified based on tag information indicating a sentence and a phrase. 5. The electronic document processing method according to claim 3, wherein the electronic document is provided with tag information necessary for performing speech synthesis. 6. The electronic document processing method according to claim 5, wherein the tag information necessary for performing the speech synthesis includes attribute information for prohibiting reading out. 7. The electronic document processing method according to claim 5, wherein the tag information necessary for performing the speech synthesis includes attribute information indicating a pronunciation or pronunciation. 8. The electronic document processing device according to claim 1, wherein in the step of generating a file for reading out aloud, the file for reading out aloud is generated by adding attribute information indicating a language describing the electronic document. Method. 9. The method according to claim 9, wherein in the step of generating a file for reading out aloud, a paragraph,
2. The electronic document processing method according to claim 1, wherein the voice reading file is generated by adding attribute information indicating a start position of a sentence and a phrase. 10. In the step of generating a text-to-speech file, if the attribute information indicating the same type of syntactic structure among the attribute information indicating the start position of the paragraph, sentence, and phrase is continuously duplicated, 10. The electronic document processing method according to claim 9, wherein the attribute information of the electronic document is singly collected. 11. The speech reading file generating step includes the step of adding attribute information indicating that a pause period is provided in correspondence with the attribute information indicating the start position of the paragraph, sentence, and phrase to convert the speech reading file. The electronic document processing method according to claim 9, wherein the electronic document is generated. 12. The electronic document processing method according to claim 1, wherein in the step of generating a file for reading out aloud, the part for which reading out is prohibited is removed to generate the file for reading out aloud. 13. The electronic document processing method according to claim 1, wherein in the voice reading file generating step, the voice reading file is generated by substituting correct reading or pronunciation. 14. The electronic document processing method according to claim 1, wherein in the voice reading file generating step, the voice reading file is generated by adding attribute information indicating a volume of voice reading. 15. A speech synthesis engine according to claim 2, wherein in the processing step, a speech synthesis engine is selected based on attribute information indicating a language describing the electronic document attached to the speech reading file. Electronic document processing method. 16. The electronic document according to claim 2, wherein, in the processing step, an absolute value of a reading volume is obtained based on attribute information indicating a reading volume given to the reading file. Processing method. 17. The speech synthesis engine according to claim 1, wherein said plurality of elements constituting said electronic document include a paragraph, 3. The electronic document processing method according to claim 2, wherein cueing, fast-forwarding, or rewinding at the time of reading is performed in units of sentences and phrases. 18. An electronic document processing method for processing an electronic document, wherein a document input for inputting an electronic document in which a plurality of elements has a hierarchical internal structure and tag information indicating the internal structure is added in advance. An electronic document processing method, comprising: a process; and a text-to-speech process of reading out the electronic document by voice synthesis based on the tag information. 19. In the document input step, an electronic document to which tag information indicating at least a paragraph, a sentence, and a phrase is input among a plurality of elements constituting the electronic document, 19. The electronic document processing method according to claim 18, wherein a pause is provided at a start position of the paragraph, the sentence, and the phrase based on the tag information indicating the paragraph, the sentence, and the phrase, and the reading is performed. 20. The electronic document has tag information indicating at least a paragraph, a sentence, and a phrase among a plurality of elements constituting the electronic document. In the document reading step, the paragraph, the sentence, and the tag are displayed. 19. The electronic document processing method according to claim 18, wherein paragraphs, sentences, and phrases constituting the electronic document are identified based on tag information indicating phrases. 21. The electronic document processing method according to claim 18, wherein the electronic document is provided with tag information necessary for performing speech synthesis. 22. The electronic document processing method according to claim 21, wherein the tag information necessary for performing the speech synthesis includes attribute information for prohibiting reading out. 23. The electronic document processing method according to claim 21, wherein the tag information necessary for performing the speech synthesis includes attribute information indicating a reading kana or pronunciation. 24. The electronic document processing method according to claim 18, wherein in the document reading step, a part whose reading is prohibited is removed to read the electronic document. 25. The electronic document processing method according to claim 18, wherein in the document reading step, the electronic document is read aloud by replacing with correct reading or pronunciation. 26. In the reading step, based on tag information indicating a paragraph, a sentence, and a phrase out of a plurality of elements constituting the electronic document, cueing at the time of reading out in units of a paragraph, a sentence, and a phrase. 19. The electronic document processing method according to claim 18, wherein fast forward or rewind is performed. 27. An electronic document processing apparatus for processing an electronic document, comprising: a voice reading file generating unit that generates a voice reading file for voice synthesis and reading based on the electronic document. Electronic document processing device. 28. The electronic document processing according to claim 27, further comprising processing means for performing processing suitable for a speech synthesis engine using the speech reading file generated by the speech reading file generating means. apparatus. 29. The electronic document processing according to claim 27, wherein the electronic document has an internal structure in which a plurality of elements are hierarchized, and tag information indicating the internal structure is added in advance. apparatus. 30. The electronic document has tag information indicating at least a paragraph, a sentence, and a phrase among a plurality of elements constituting the electronic document. 30. The electronic document processing apparatus according to claim 29, wherein a paragraph, a sentence, and a phrase constituting the electronic document are identified based on tag information indicating a sentence and a phrase. 31. The electronic document processing apparatus according to claim 29, wherein the electronic document is provided with tag information necessary for performing speech synthesis. 32. The electronic document processing apparatus according to claim 31, wherein the tag information necessary for performing the speech synthesis includes attribute information for prohibiting reading out. 33. The electronic document processing apparatus according to claim 31, wherein the tag information necessary for performing the speech synthesis includes attribute information indicating a reading kana or a pronunciation. 34. The electronic document processing apparatus according to claim 27, wherein said voice reading file generating means generates the voice reading file by adding attribute information indicating a language describing the electronic document. apparatus. 35. The reading-aloud file generating means, wherein a plurality of elements constituting the electronic document include a paragraph,
28. The electronic document processing apparatus according to claim 27, wherein the voice reading file is generated by adding attribute information indicating a start position of a sentence and a phrase. 36. When the attribute information indicating the same type of syntactic structure among the attribute information indicating the start position of the paragraph, sentence, and phrase is continuously duplicated, 36. The electronic document processing apparatus according to claim 35, wherein the attribute information of the electronic document is collectively collected. 37. The text-to-speech file generating means adds attribute information indicating that a pause period is to be provided corresponding to the attribute information indicating the start position of the paragraph, sentence or phrase, and generates the text-to-speech file. The electronic document processing apparatus according to claim 35, wherein the electronic document processing apparatus generates the electronic document. 38. The electronic document processing apparatus according to claim 27, wherein the voice reading file generating means generates the voice reading file by removing a portion where reading is prohibited. 39. The electronic document processing apparatus according to claim 27, wherein the voice reading file generating means generates the voice reading file by replacing the reading with correct reading or pronunciation. 40. The electronic document processing apparatus according to claim 27, wherein the voice reading file generating means generates the voice reading file by adding attribute information indicating a reading volume. 41. The speech processing system according to claim 28, wherein said processing means selects a speech synthesis engine based on attribute information, which is given to said speech reading file and indicates a language describing said electronic document. Electronic document processing device. 42. The electronic document according to claim 28, wherein said processing means obtains an absolute value of a reading volume based on attribute information indicating a reading volume given to the voice reading file. Processing equipment. 43. The text-to-speech engine, based on the text-to-speech file to which attribute information indicating a start position of a paragraph, a sentence and a phrase among a plurality of elements constituting the electronic document is added. 29. The electronic document processing apparatus according to claim 28, wherein cueing, fast-forwarding, or rewinding at the time of reading is performed in units of sentences and phrases. 44. An electronic document processing apparatus for processing an electronic document, comprising: a document input device for inputting an electronic document in which a plurality of elements have a hierarchical internal structure and tag information indicating the internal structure is added in advance. An electronic document processing apparatus comprising: means for reading out the electronic document based on the tag information; 45. The document input unit inputs, from among a plurality of elements constituting the electronic document, an electronic document to which tag information indicating at least a paragraph, a sentence, and a phrase is added, The electronic document processing apparatus according to claim 44, wherein a pause is provided at a start position of the paragraph, the sentence, and the phrase based on the tag information indicating the paragraph, the sentence, and the phrase, and the reading is performed. 46. The electronic document has tag information indicating at least a paragraph, a sentence, and a phrase among a plurality of elements constituting the electronic document, and the document reading means reads the paragraph, sentence, and phrase. Paragraphs constituting the electronic document based on tag information indicating phrases,
The electronic document processing apparatus according to claim 44, wherein a sentence and a phrase are identified. 47. The electronic document processing apparatus according to claim 44, wherein the electronic document is provided with tag information necessary for performing speech synthesis. 48. The electronic document processing apparatus according to claim 47, wherein the tag information necessary for performing the speech synthesis includes attribute information for prohibiting reading out. 49. The electronic document processing apparatus according to claim 47, wherein the tag information necessary for performing the speech synthesis includes attribute information indicating a pronunciation or pronunciation. 50. The electronic document processing device according to claim 44, wherein said document reading means reads out the electronic document by removing a portion for which reading is prohibited. 51. The electronic document processing apparatus according to claim 44, wherein the document reading means reads the electronic document by replacing the reading with correct reading or pronunciation. 52. The document reading means, based on tag information indicating a paragraph, a sentence, and a phrase among a plurality of elements constituting the electronic document, finds a head at the time of reading in a unit of a paragraph, a sentence, and a phrase. The electronic document processing apparatus according to claim 44, wherein fast forward or rewind is performed. 53. A recording medium on which a computer-controllable electronic document processing program for processing an electronic document is recorded, wherein the electronic document processing program is used for reading aloud a voice for synthesizing and reading aloud based on the electronic document. A recording medium on which an electronic document processing program is recorded, comprising a voice reading file generating step of generating a file. 54. The electronic document processing program further comprises a processing step of performing processing suitable for a speech synthesis engine using the speech reading file generated in the speech reading file generating step. Term 53
A recording medium on which the electronic document processing program described above is recorded. 55. The electronic document processing according to claim 53, wherein the electronic document has an internal structure in which a plurality of elements are hierarchized, and tag information indicating the internal structure is added in advance. A recording medium on which a program is recorded. 56. The electronic document, to which tag information indicating at least a paragraph, a sentence, and a phrase among a plurality of elements constituting the electronic document is added. 56. The recording medium according to claim 55, wherein a paragraph, a sentence, and a phrase constituting said electronic document are identified based on tag information indicating a sentence and a phrase. 57. A recording medium according to claim 55, wherein said electronic document is provided with tag information necessary for performing speech synthesis. 58. A recording medium on which a computer-controllable electronic document processing program for processing an electronic document is recorded, wherein said electronic document processing program has an internal structure in which a plurality of elements are hierarchized, Electronic document processing, comprising: a document inputting step of inputting an electronic document to which tag information indicating the following is given in advance; and a text-to-speech step of voice-synthesizing and reading out the electronic document based on the tag information. A recording medium on which a program is recorded. 59. In the document inputting step, an electronic document to which tag information indicating at least a paragraph, a sentence and a phrase is added among a plurality of elements constituting the electronic document is input. The recording according to claim 58, wherein a pause is provided at a start position of the paragraph, the sentence, and the phrase based on the tag information indicating the paragraph, the sentence, and the phrase. Medium. 60. The electronic document has tag information indicating at least a paragraph, a sentence, and a phrase among a plurality of elements constituting the electronic document. In the document reading step, the paragraph, sentence, and 59. The recording medium according to claim 58, wherein a paragraph, a sentence, and a phrase constituting the electronic document are identified based on tag information indicating a phrase. 61. A recording medium according to claim 58, wherein said electronic document is provided with tag information necessary for performing speech synthesis.