JP3825234B2 - Hypertext relay method and apparatus for voice browsing, and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a hypertext relay method and apparatus and recording medium for voice browsing, and more particularly to a hypertext relay method and apparatus and recording medium for voice browsing in which a link to be read is designated by voice input from a browser. .
[0002]
[Prior art]
In recent years, a framework called WWW (World Wide Web) that realizes operations of request, reception, and display by a hypertext display device based on hypertext via a computer network such as the Internet has become widespread.
In the hypertext, a URL (uniform resource locator) indicating the location of other hypertext and its label can be described. For the label, text or an image is used.
[0003]
The URL on the hypertext and its label are collectively called a link. The hypertext display device is called a browser and functions as follows via a computer network and a hypertext relay device (proxy server).
First, when hypertext is displayed on the browser, the browser highlights the link portion with a different color or the like. The user performs a click operation on such a link portion with a pointing device such as a mouse, and selects hypertext specified in the link. At this time, the browser requests the relay device to transmit the hypertext selected by the user. The relay device identifies the location of the corresponding hypertext from the computer network based on the transmission request from the browser, and transmits the corresponding hypertext from the computer network to the requesting browser. The browser displays the hypertext transmitted by the relay device to the user.
[0004]
With the widespread use of the WWW, hypertext specifications have also become diversified and highly functional, and browsers as display devices have also responded to this.
Under such circumstances, a method of browsing using a voice uttered by a human being has been invented. This is referred to as audio browsing. Japanese Patent Application Laid-Open No. 10-154063 (Japanese Patent Application No. 8-312016) discloses a hypertext relay method and apparatus for adding a command that enables voice browsing to existing hypertext. Japanese Patent Application Laid-Open No. 10-222342 (Japanese Patent Application No. 9-24024) discloses a hypertext voice control method and apparatus for adding / extending a voice input function to a browser. Audio browsing is realized by both of the inventions as follows.
[0005]
First, a hypertext with a browser is requested from a hypertext relay device for voice browsing. Its relay device imparts a voice command such as the following in hypertext key strike has been requested.
・ "Read link"-> (T1, ..., Ti, ..., Tn)
・ "Jump to Wij"-> Jump (URLi)
However, Ti and URLi are a text part and a URL part of a certain link Li, respectively. Wij is a certain word included in Ti.
[0006]
Next, the relay device assigns the voice command as described above to the hypertext and transmits it to the requesting browser, and the browser displays the hypertext to the user. At this time, if the browser has a voice input / output function, the user can perform the following operations.
・ Speech input called “Read Link”
Read the text of all hypertext links.
・ Enter the voice input "Jump to",
Call the hypertext specified by the link containing "~".
[0007]
[Problems to be solved by the invention]
It is an object of the present invention to solve the problem of voice browsing malfunction prevention and the enhancement of interactive voice browsing.
First, invoking hypertext by voice input has a problem of causing malfunction depending on the expression of the text of the link. This occurs when each link has the same word in its text. For example, assume that the text of a hypertext link is as follows.
[0008]
“Health condition of the Prime Minister”
“A smart life in the Edo period”
"Proposal for a healthy diet"
Here, even if the user performs a voice input of “jump to health” in order to browse the hypertext represented by “advice to healthy eating”, the hypertext represented by “health condition of the prime minister”. Text may be browsed. Even if the voice input “jump to life” is made, the same applies. The user needs to say “jump to advice”.
[0009]
Next, from the viewpoint of realizing a voice dialogue, the conventional operation for voice browsing is problematic in that it is unnatural or incomplete. In other words, in the conventional voice browsing, the voice is used only as a command to the browser, and unlike a human-to-human dialogue, it is unilateral and not reciprocal. As a voice output, the conventional technology has a function of reading out the text of the entire link portion. In order to further enhance the reciprocity of this function, the present invention adds and expands a function of reading out only the text of a link that matches the word that the user has input by voice.
[0010]
[Means for Solving the Problems]
The present invention adds / extends the function of reading out only the text of a link that matches the word input by the user in order to prevent malfunction of the voice browsing and to enhance the interactive voice browsing.
First, a hypertext HT with a browser is requested from a hypertext relay apparatus for voice browsing. The relay device extracts link portions L1,..., Ln from the HT. For each Li, a text part Ti and a URL part URLi are divided, and voice recognition words (Wi1,..., Wim) for voice recognition are further specified from Ti. The following voice command is created that reads out only the text Ti including these words Wij.
“Read Wij”-> Read (Ti)
Next, the relay device gives the above voice command to the hypertext and transmits it to the requesting browser, and the browser displays the hypertext to the user. At this time, if the browser has a voice input / output function, the user can perform the following operations.
Make a voice input saying "Read aloud"
Read the text of the link containing "~".
[0011]
DETAILED DESCRIPTION OF THE INVENTION
FIG. 1 shows a principle configuration diagram of the present invention.
The user requests the relay device to obtain hypertext by voice through the voice browser, and the relay device reads the hypertext from the computer network and transmits it to the voice browser, and the browser displays the hypertext. Further, the user instructs the voice browser to read out the link by voice.
[0012]
FIG. 2 shows a configuration diagram of a hypertext browsing system to which the present invention is applied.
The hypertext browsing system includes a computer network 1 such as the Internet, a relay device 2, and a voice browser 3. The computer network 1 and the relay device 2 are connected via a communication line 4, and the relay device 2 and the voice browser 3 are connected via a communication line 5. It is connected.
[0013]
The relay device 2 includes a hypertext reading unit 21, a link extraction unit 22, a speech recognition word specifying unit 23, a voice command creating unit 24, a hypertext adding unit 25, a request source browser transmitting unit 26, and a buffer 27.
The voice browser 3 includes a browser 31, a voice recognition device 32, a command interpretation execution device 33, a voice input device 34, a voice synthesis device 35, and a voice output device 36.
Operations of the relay device 2 and the voice browser 3 will be described.
(1) When the hypertext is requested from the voice browser 3, the hypertext reading unit 21 reads the hypertext from the computer network 1 connected via the communication line 4 and stores it in the buffer 27.
(2) The link extraction unit 22 extracts a link portion for the hypertext in the buffer 27.
(3) The voice recognition specifying unit 23 divides the link extracted by the link extraction unit 22 into a text part and a URL part, and further specifies a voice recognition word for voice recognition from the text part. (See Japanese Patent Laid-Open No. 10-154063)
(4) the voice command creation part 24, for every speech recognition word specified by the speech recognition identification unit 23, creates a voice command read the text of the link containing the speech recognition word.
(5) The hypertext giving unit 25 gives the voice command created by the voice command creating unit 24 to the hypertext.
(6) The request source browser transmission unit 26 transmits the hypertext to which the voice command is given to the browser 31 via the communication line 5.
(7) The browser 31 receives the hypertext to which the voice command is given, displays it on the display device, outputs it through the voice synthesizer 35 and the voice output device 36, and outputs it to the voice recognition device 32. .
(8) The voice input device 34 inputs command voice from the user's microphone or the like, and outputs voice data to the voice recognition device 32.
(9) The voice recognition device 32 recognizes the voice data, selects the closest voice command received from the browser 31, and outputs it to the command interpretation execution device 33. (10) The command interpretation execution device 33 interprets the voice command received from the voice recognition device 32 and controls the operation of the browser 31 based on the result.
(11) The browser 31 transmits a voice command to the relay device 2.
(12) The relay device 2 receives the voice command, executes the command, reads the text from the computer network, and transmits it to the browser 31.
[0014]
A specific example of the hypertext relay method of the present invention will be described with reference to the flowchart of FIG.
First, (a) the browser requests the following hypertext from the hypertext relay apparatus for voice browsing.
This is <the health condition of the Prime Minister>.
This is the “popular lifestyle of the Edo period”.
[0015]
This is <advice for a healthy diet>.
However, the text enclosed in <> indicates a link.
Next, (b) the relay device acquires the hypertext from the computer network and extracts the link from the hypertext as follows.
<Health condition of the Prime Minister>
<Smart life of the Edo period>
<Advice to a healthy diet>
Each link is divided into a text part and a URL part, and (c) a speech recognition word for speech recognition is specified from the text part as follows.
[0016]
Cabinet, ..., health, condition, Edo, ..., commoner, life, health, ... life, ..., advice
(d) For all these words, create the following voice command that reads only the text of the link that matches each word.
Read "Cabinet"-> Read out ("Health condition of the Prime Minister")
...
Read “Health”-> Read (“Health condition of the Prime Minister”)
Read "Status"-> Read ("Prime Minister's health")
Read "Edo"-> Read aloud ("Smart life of the Edo period")
...
Read out "People"-> Read out ("Smart life of the Edo period")
Read "Life"-> Read out ("Smart life in the Edo period")
Read "Health"-> Read ("Healthy eating advice")
...
Read "Life"-> Read ("Advice to a healthy diet")
...
Read "advice"-> read out ("advice for a healthy diet")
(e) The relay device gives the above voice command to the hypertext, and (f) transmits it to the browser that is the request source, and the browser displays the hypertext to the user. At this time, if the browser is a voice browser having a voice input / output function, the user can perform the following operations, for example.
[0017]
The voice input “Read advice” is made, and the text “advice to a healthy diet” is read out by the speech synthesis function.
The relay device 2 includes a computer having a CPU, a memory, and the like, and a user terminal and a recording medium used by a user who is an access subject. The recording medium is a machine reading such as a CD-ROM, a magnetic disk, and a semiconductor memory. A program that is a possible recording medium and that executes the text relay method recorded therein is read by a computer, controls the operation of the computer, and implements each component in the above-described embodiment on the computer.
[0018]
【The invention's effect】
In the present invention, the text can be read out by voice input using a word included in the text of the link. Thereby, it can be confirmed by the present invention whether or not the user can use the current word for voice browsing. For example, assume that the text of a hypertext link is as follows.
“Health condition of the Prime Minister”
“A smart life in the Edo period”
"Proposal for a healthy diet"
Here, in order to browse the hypertext represented by “advice to a healthy diet”, the user reads out the text of the link by voice input “read aloud health”. Thus, for example, when it is read out as “the health status of the prime minister”, it can be understood that “jump to health” may browse hypertext represented by “the health status of the prime minister”. There is an effect that erroneous browsing can be avoided in advance by using the function provided by the present invention by the user. In addition, it enables speech browsing in a form closer to human-to-human dialogue by answering voice input with voice output according to the content.
[Brief description of the drawings]
FIG. 1 is a principle configuration diagram of the present invention.
FIG. 2 is a configuration diagram of a hypertext browsing system to which the present invention is applied.
FIG. 3 is a flowchart for explaining a specific example of the hypertext relay method of the present invention.
[Explanation of symbols]
1 Computer network 2 Relay device 3 Voice browser 4 5 Communication line
21 Hypertext reading part
22 Link extractor
23 Speech recognition word identification section
24 Voice command generator
25 Hypertext adding section
26 Requester browser transmitter
31 Browser
32 Voice recognition device
33 Command interpretation execution device
34 Voice input device
35 Speech synthesizer
36 Audio output device

Claims (3)

Extract the link part from the requested hypertext, divide each link into a text part and a URL part, identify a plurality of speech recognition words for speech recognition from those texts, For a voice recognition word, when a voice input is made to read out the voice recognition word , a voice command that reads the text of the link that includes the voice recognition word is created, and the voice command is assigned to the hypertext. A text relay method for voice browsing, characterized in that it is transmitted to a browser that is a request source. A link extraction unit that extracts a link portion from the requested hypertext, and for each link, the link portion is divided into a text portion and a URL portion, and a plurality of speech recognition words for speech recognition are further identified from the text. A voice recognition word specifying unit and a voice command creation unit for creating a voice command for reading out the text of a link including the voice recognition word when there is a voice input for reading out the voice recognition word for all of the voice recognition words And a hypertext giving unit for giving the voice command to the hypertext, and a request source browser sending unit for sending the hypertext to which the voice command is given to the browser that is the hypertext request source. A text relay device for browsing. A process of extracting a link part from the requested hypertext, dividing each link into a text part and a URL part, and further specifying a plurality of speech recognition words for speech recognition from the text, For all voice recognition words, when there is a voice input for reading out the voice recognition word , a voice command that reads the text of the link including the voice recognition word is created, and the voice command is given to the hypertext; A recording medium storing a program for causing a computer to execute a text relay method for voice browsing, which includes processing for transmitting hypertext to which a voice command is given to a browser that is a request source of the hypertext .

Images