KR20050035784A - Voice supporting web browser through conversion of html contents and method for supporting voice reproduction - Google Patents

Abstract

The present invention relates to a voice-supported web browser and a voice support method through HTML content conversion, and enables users to search a web document using a mobile terminal such as not only visual but also various user interfaces such as voice, while walking or moving on the screen. You can search the web without receiving it.

In particular, the web browser of the present invention is a page loader for downloading a document, when the document downloaded from the page loader is an HTML (Hyper Text Makeup Language) document, HTML to parse the document to generate an HTML Document Object Model (DOM) A parser, a content converter for extracting a part requiring voice support from the DOM, generating a voice recognition grammar according to the correlation of the document, and generating a VoiceXML document, parsing a document in the VoiceXML (Voice eXtensible Markup Language) document. VoiceXML parser outputs the contents of the text to TTS (Text to Speech), delivers the grammar (Grammar) and voice commands input from the user to the voice recognizer, and receives and processes the recognized words from the voice recognizer. And text to speech (TTS) and the grammar for outputting the sentences received from the VoiceXML parser as voice. It is characterized in that it comprises a speech recognizer for delivering the word recognized by the VoiceXML interpreter.

Description

Voice Supporting Web Browser Through Conversion of HTML Contents and Method for Supporting Voice Reproduction}

The present invention relates to a voice support web browser and a voice support method through HTML content conversion, and more particularly, to support voice by inserting VoiceXML into HTML through a content converter without modifying an existing web document for a personal computer (PC). By developing a browser technology, a user can use a variety of user interfaces such as voice as well as visual when searching a web document in a mobile terminal, and can search the web regardless of the screen while walking or moving.

Recently, due to the development of technology, various mobile terminals have been spread in the days when only Internet access was possible from PCs, and the desire to use the Internet in mobile terminals is increasing.

However, in the case of a mobile terminal, the resolution of the display is limited, and it is not easy to search the web and read its contents on the go.

In addition, mobile terminal users want a multimodal browser that supports a variety of input and output. HTML, a web document authoring language, is a mark-up language that is designed for visual expression, and thus, multi-modality. ) Is not supported.

In order to solve this problem, an attempt has been made to support voice in HTML documents.

For example, voice services using VoiceXML (Voice eXtensible Markup Language) are provided in various places, and most of the services are mostly used as a means for replacing an ARS (Automatic Response System) server.

This is because modification and maintenance are simple and easy by using the VoiceXML.

However, there is a problem that the ARS service is different from the existing HTML web service, and the service that enables the user to check the mail sent to the user by voice using VoiecXML in the web mail server also processes the HTML file and saves it as a VoiceXML file. There is a downside to building it.

In addition, attempts to use HTML and VoiceXML for voice services on Web servers include IBM's XHTML + VoiceXML (X + V), and Microsoft has its own specification, Speech Application Language Tags (SALT). All of them have to rewrite the web document by adding new code to the existing HTML document, and only a browser that supports the new specification can receive the service.

Accordingly, an object of the present invention is to solve the above-described problems. An object of the present invention is to add VoiceXML to an existing HTML document in real time to enable a voice output of a web document supporting only a visual interface. Voice-supported web browser and voice support method through HTML content conversion that enables users to visually and auditively search the web by voice-producing the contents of newspaper articles or bulletin boards of portal sites by configuring the browser. To provide.

In order to achieve the above object of the present invention, a voice-supported web browser through HTML (Hyper Text Makeup Language) content conversion includes a page loader for downloading a document, and when the document downloaded from the page loader is an HTML document, the document An HTML parser for generating an HTML Document Object Model (DOM) by parsing, a content converter for extracting a portion requiring voice support from the DOM, generating a voice recognition grammar according to the correlation of a document, and generating a VoiceXML document, the VoiceXML Parse the document in the document and output the contents of the document to TTS (Text to Speech), and transmit the grammar (Grammar) and the voice command input from the user to the voice recognizer required for the voice recognition, and recognized from the voice recognizer VoiceXML parser that takes over and processes words and outputs sentences received from the VoiceXML parser as voice That the word recognized by the TTS with the programmer is configured to include a speech recognizer which is transferred to the VoiceXML interpreter.

In order to achieve the object of the present invention as described above, the voice support method through HTML content conversion includes downloading a document by a page loader of a browser, and parsing the document by parsing the document if the downloaded document is an HTML document. Document object model), extracting a part to support voice from the DOM, grasping the context of the extracted document to generate a grammar for speech recognition, and generating a VoiceXML document using the speech recognition grammar. And transmitting the VoiceXML document to a VoiceXML interpreter to voicely output the contents of the document to the TTS while parsing the document.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a voice-supporting web browser through the conversion of HTML content according to the present invention.

As shown in FIG. 1, the page loader 10 of the voice-enabled web browser downloads a document and passes the downloaded document to the HTML parser 20. When the document delivered from the page loader 10 is an HTML document, the HTML parser 20 parses the delivered document to generate an HTML Document Object Model (DOM) and delivers the document from the page loader 10. If the document is a VoiceXML document, it is immediately transmitted to the VoiceXML interpreter 50 to output the voice.

 The HTML DOM generated by the HTML parser 20 is delivered to the content converter 30, and the content converter 30 extracts a portion of the HTML DOM 31 that requires voice support and relies on document correlation. The voice recognition grammar is generated to generate a VoiceXML document 32.

In addition, the HTML DOM 31 is transmitted to the renderer 40 so that the contents of the document can be displayed on the user's terminal.

The VoiceXML document 32 generated by the content converter 30 and the VoiceXML document delivered from the HTML parser 20 to the VoiceXML parser 50 are parsed while the contents of the document are transferred to the Text to Speech (TTS) 70. Voice output.

 At this time, the VoiceXML interpreter 50 transmits a grammar necessary for voice recognition and a voice command input from a user to the voice recognizer 60 from the VoiceXML document transmitted from the content converter 30, and the voice recognizer ( The word recognized from 60 is taken over and processed.

That is, the web browser is installed on the client to convert the HTML document provided in the web page in real time to the VoiceXML document to output the voice.

2 is a diagram illustrating an interface of a voice supporting web browser according to an embodiment of the present invention.

As shown in FIG. 2, if a title having a link is provided in most portal sites that support newspaper articles, bulletin boards, web mails, and the like, the user may obtain a title from the web page according to the purpose of the user terminal. Click the link with the stylus pen or the right mouse button.

Then, a menu window appears in the title having the link, and the menu window is provided with 'Read as VoiceXML', 'Go To Link', 'Go To Link, and Read With VoiceXML' menus. You will be provided with text-linked information.

When the user selects 'Read as VoiceXML' in the menu window, only the core part of the linked information corresponding to the large title is provided to the user by voice without moving the link, and when the user selects 'Move Link', The content linked to the document, such as a web page, is provided to the user, and when the 'Move Link and Read in VoiceXML' is selected, the core portion of the information linked to the selected title is simultaneously provided to the user's terminal in the document and voice. .

FIG. 3 is a diagram illustrating a portion extracted from the web browser document of FIG. 2 as a VoiceXML document. As shown in FIG. 2, only the core content linked to the large title of the web page is provided to the user as a voice or a document. .

In other words, most portal sites that support newspaper articles, bulletin boards, and web mails have a title with a link, and clicking on the title takes you to a web page with details. In addition, in the case of a link having miscellaneous information, when generating a VoiceXML document in the content converter of the speech-supporting web browser, deleting miscellaneous contents and extracting a portion to support the voice of the main core part such as a boxed portion Extract from to create a VoiceXML document. In most cases, the block containing the selected link title becomes the main block.

As such, by eliminating unnecessary parts, the user interface can be simplified and the execution speed can be improved.

4 is a flow chart of a voice support method through HTML content conversion according to the present invention. The voice support operation process of the web browser will be described in detail as follows.

The web browser downloads the document by the page loader 10 (S100), and when the document downloaded by the page loader 10 is an HTML document, the document is parsed (S110) to generate an HTML DOM ( S120).

If the document downloaded to the page loader 10 is a VoiceXML document, it is sent to the VoiceXML interpreter 50 and is outputted to the voice by the TTS 70 and provided to the user.

Subsequently, a part to support voice from the generated DOM is extracted (S130), and a correlation for the document before and after is converted into a extracted VoiceXML document to generate a grammar for speech recognition (S140).

A voiceXML document is generated using the generated voice recognition grammar (S150).

At this time, in generating the VoiceXML document, a portion corresponding to a link in the existing HTML document, that is, a sentence or word having an 'a' tag is added to the grammar, and a portion corresponding to the link when creating a sentence required for speech synthesis By inserting a special character in the TTS 70 to inform that it is a word corresponding to the voice command.

The generated VoiceXML document is transmitted to the VoiceXML parser 50, and the HTML DOM is delivered to the renderer 40 to be represented as a document according to the user's menu window selection as described in FIG. It is provided as a document on the screen.

The VoiceXML document and the grammar transmitted to the VoiceXML parser 50 are delivered to the voice recognizer 60 by the VoiceXML parser 50, and the voice recognizer 60 has a grammar passed from the VoiceXML parser 50. The recognized word is re-delivered to the VoiceXML parser 50 (S160).

In addition, the voice recognizer 60 is difficult to be mounted on the client in current technology, and may have a voice recognition server that performs only voice recognition. In this case, the VoiceXML interpreter 50 passes the grammar to the voice recognition server and passes the voice input from the client to the voice recognition server, and the voice recognition server delivers the word recognized by the grammar to the voiceXML interpreter 50 of the client.

The word recognized by the voice recognizer 60 is parsed by the VoiceXML interpreter 50 together with the VoiceXML document, and the contents of the document are output to the TTS 70 by voice and provided to the user (S170).

In addition, when generating a speech recognition grammar, insert a 'previous' and 'next' to identify the words and links between the words of the web page and the front and rear relationship of the page, so that the page can be moved by voice, Inform the user of the link movement with a specific word or sound.

For example, it inserts a certain ASCII character into a sentence and makes a beep sound that a user can know when the TTS encounters the specific character.

Although the preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the above-described embodiments, and the present invention is not limited to the above-described embodiments without departing from the spirit of the present invention as claimed in the claims. Of course, any person skilled in the art can make various modifications, and such changes are within the scope of the claims.

As described above, according to the present invention, when searching for a web document in a mobile terminal or the like, various user interfaces such as voice as well as time are possible, and there is an advantage that a web search can be performed regardless of the screen while walking or moving.

In addition, in the present invention, unlike the existing method of rewriting a document or constructing an ARS server for a voice output of an HTML web document, the client converts the document directly in real time and standardizes only necessary parts according to a user's needs. It converts to VoiceXML, which is a document form, and has an advantage in terms of cost and is easy to apply and commercialize.

1 is a block diagram of a voice support web browser through the conversion of HTML content according to the present invention,

2 is a diagram showing an interface of a voice supporting web browser according to one embodiment of the present invention;

3 is a view showing a part extracted as a VoiceXML document from the web browser document of FIG.

4 is a flowchart of a voice supporting method through HTML content conversion according to the present invention.

<Explanation of symbols for the main parts of the drawings>

10: Page Loader 20: HTML Parser

30: Content Converter 31: HTML DOM

32: VoiceXML document 40: Renderer

50: VoiceXML interpreter 60: Speech recognizer

70: TTS (Text to Speech)

Claims (8)

In a speech-enabled web browser through HTML (Hyper Text Makeup Language) content conversion, The web browser includes a page loader for downloading a document; An HTML parser for parsing the document to generate an HTML Document Object Model (DOM) when the document downloaded from the page loader is an HTML document; A content converter which extracts a portion requiring voice support from the DOM, generates a voice recognition grammar according to the correlation of the document, and generates a VoiceXML document; Parse the document from the VoiceXML document and output the content of the document to TTS (Text to Speech), and transfer the grammar required for voice recognition and voice commands input from the user to the voice recognizer, and from the voice recognizer. A VoiceXML parser that takes over and processes recognized words; And And a speech recognizer for transmitting a word received by the voiceXML parser to the VoiceXML interpreter. The method of claim 1, wherein the web browser Voice-supported web browser through HTML content conversion, characterized in that the HTML document installed on the client in real time to convert the HTML document provided by the web page to the VoiceXML document to output the voice. The method of claim 1, wherein the content converter A voice-enabled web browser using HTML content, which extracts only the core part of the downloaded HTML document, converts it into VoiceXML without outputting a link, and outputs it. The method of claim 1, wherein the content converter Voice generation by converting HTML content by inserting 'previous' and 'next' to identify the word and page of the link of the web page and the relationship between the pages when generating the voice recognition grammar. Web browser. The web browser downloading a document by a page loader, and when the downloaded document is an HTML document, parsing the document to generate an HTML DOM; Extracting a part to support voice from the DOM, and identifying a front and rear relationship in the extracted document to generate a grammar for voice recognition; Generating a VoiceXML document using the speech recognition grammar; And And transmitting the VoiceXML document to a VoiceXML interpreter, parsing the document and outputting the content of the document to the TTS. The method of claim 5, wherein the first step Voice support method by converting the HTML content, characterized in that to pass the generated HTML DOM to the renderer. The method of claim 5, wherein the step 4 And transmitting the grammar transmitted to the VoiceXML parser to a voice recognizer, receiving the recognized result again, and outputting the content of the document as a voice. The method of claim 5, wherein the step 4 Voice support method through the conversion of the HTML content, characterized in that notifying the user with a specific word or sound for the link-movable word.