JPH0442363A - Portable voice recognizing electronic dictionary - Google Patents

Abstract

PURPOSE:To quickly correct a misrecognized word which is inputted in voices by providing a correcting function that works with the voices inputted for each single sound or monosyllable. CONSTITUTION:The voice signals inputted through a microphone are processed by a voice input means 501 and converted into the digital signals. These digital signals are stored in a voice data store means 502 and then recognized by a voice recognizing means 503 to be converted into the phonetic character code strings. A translation means 504 extracts the equivalents of all homonyms. Then various control switches 100 and 104 - 107 are controlled by a control means 506 in relation to other functions. Furthermore the means 506 owns a function to correct the recognizing results for each single sound or monosyllable. As a result, the Japanese equivalents of a desired foreign language can be obtained with a comparatively short interruption even in a conversation.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a portable speech recognition electronic dictionary that has a function of correcting single sounds or monosyllables based on voice input.

[Conventional technology 1] Many small portable electronic dictionaries have been commercialized in the past, but all of them use a manual key input method, and input and correction of words are performed manually using the keys. .

Furthermore, although voice recognition technology has been realized to a certain extent on large-scale computers, there is no such technology available on the market as a consumer product yet.

[Problem to be solved by the invention] For example, when having a conversation in a foreign language, it is a common experience to have trouble remembering certain words. In such cases, with conventional electronic dictionaries, the user inputs the desired word manually using the keypad to obtain the desired foreign word. If you do so, either type it again from the beginning, or point to the wrong character in some way and retype the correct character.
This procedure was necessary. Therefore, during a conversation,
It was impossible to use it naturally.

In view of this problem, the present invention has a correction function by voice input in units of single sounds or single syllables, so that even during a conversation, the desired foreign language interpretation can be obtained without much interruption.

The purpose is to provide a portable voice recognition electronic dictionary.

[Means for Solving the Problems] The portable speech recognition electronic dictionary of the present invention has the following features: 1. A voice input means, a voice data storage means capable of accumulating one word's worth of voice data, and a voice input means capable of accumulating voice data for one word; It has a speech recognition means for recognizing, a translation means for translating the recognized word into another language, a display means for displaying the translation result, and a control means for controlling various functions, and each of the above means is configured as one. It is characterized by

2. The control means is characterized in that it has a function of correcting recognition results in units of single sounds or single syllables.

3. The above correction function is characterized in that correct single sounds or single syllables can be input by voice.

[Operation] The analog voice signal for one word inputted to the human voice means is converted into a digital signal and stored in the voice data storage means. The voice recognition means reads the voice data stored in the voice data storage means, performs voice recognition processing, and converts it into a phonetic character code string. The data converted into a phonetic character code string is searched by the built-in bilingual dictionary by the translation means, all homophones are extracted, and the result is displayed on the display means together with the recognition results obtained by the speech recognition means. . At this time, if there is an error in the recognition result recognized by the speech recognition means, the correction is made by various switches connected to the control means, and the desired translation result is obtained again by the translation means. When making the above corrections, the corrections are made for each single sound or single syllable, and especially when inserting or changing, the voice input means, voice data storage means, and voice recognition means are used after indicating the position to be inserted or changed. This can be done using the voice input method.

[Example] An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is an external view of a portable speech recognition electronic dictionary of the present invention. For example, let's say you're having a conversation with a foreigner in English and you want to know the word "constraint." While pressing the voice input switch 100, say ``Seiyaku°'' and then release the voice input switch 100, the voice signal input from the microphone 101 will be recognized by voice recognition processing, and the voice will be pronounced as ``sweet yaku''. The phonetic character recognition result is displayed on the input window 103A. Also, through the translation process, all homophones of "Seiyaku" are extracted, and the translation results are displayed on the output window of 103B. At that time, for example, if there are many homophones or many parallel translations, and they cannot all be displayed on the display means 103, 104 left scroll switch, 105 right scroll switch, 108 upper scroll switch, 109
Necessary information is obtained by scrolling the displayed contents using the scroll switch.

Incidentally, the above is a case where the input voice is accurately recognized. Next, an example will be described regarding a case where erroneous recognition occurs.

FIG. 2 shows a case where the voice input of "Seiyaku" is mistakenly recognized as "Seiya<". As shown in FIG. 2(A) on the display means 103, "'Seya<°°" is displayed in the input window 103A, and since the corresponding word cannot be found by the translation means, "?" is displayed in the output window 103B. °° is displayed. In this case, if ``°°'' is inserted in ``Seya<°°'', it will be a correct input, so move the 201 cursor to 105
Use the right scroll switch to move to the right, and then move to Figure 2 (
When you reach the position indicated by B), press the 107 insertion switch. Next, as shown in Figure 2 (C), 100
Press the voice input switch and input the single sound or single syllable to be inserted. In this case, pronounce it as "i" and
Release the voice power switch. The voice recognition means recognizes it as ``'', and the display on the 103-octave window changes from ``Saya<° to Paseiya<'', as shown in FIG. 2 (CD).

The translation means immediately extracts all the homophones of °seiya<pa, and displays the bilingual results as shown in FIG. 2(D).

FIG. 3 shows a case where the voice input of "Seiyaku" is incorrectly recognized as "Seiyaku". Figure 3 (A
), the 103A input window is “pampered”.
is displayed. In this case, if ゛゛°° is changed to ゛゛゛゛゜°, then the input is correct, so move the 201 cursor to 105
Move it using the right scroll switch, and when it comes to the position shown in FIG. 3(B), press the 100 voice manual switch and input voice as "I". Voice recognition means °゛°
゛ is recognized, and the input window 103A correctly displays ゛ or <°゛, and the desired translation is obtained as shown in Fig. 3(C).

FIG. 4 shows a case where the voice input of "Seiyaku" has an extra -single added to it, "Seiyaku". As shown in FIG. 4(A), 9° is displayed in the output window 103A because the corresponding word is not found in the translation output stage. In this case, if “°seeiya<°°゛” is deleted, it will be a correct input, so move the 201 cursor to 10
5 Move rightward using the right scroll switch and when you reach the position shown in FIG. 2(B), press the delete switch 106. Then, the display of 1 to 103A people's counters changes from ゛seiya<°゛ to °゛seiya (pa), and the translation means extracts all the corresponding homonyms and displays them as shown in Figure 4 (C). Get the desired bilingual results.

FIG. 5 is a functional block diagram of the portable speech recognition electronic dictionary of the present invention. A voice signal inputted from a microphone 101 is processed by a voice input means 501 and converted into a digital signal, and then stored in a voice data storage means 502, and is recognized by a voice recognition means 503 into a phonogram code. It is converted into a string, and the translation means extracts the parallel translations of all homophones. The various control switches described above are all controlled by a control means 506 in cooperation with other functions.

[Effects of the Invention] As described above, the portable voice recognition electronic dictionary of the present invention has a correction function based on voice input for each single sound or single syllable, so even if a manually inputted word is misidentified as ME,
It is possible to fix it very quickly. Particularly in the case of Japanese, the number of phonetic characters is large, and it takes time to select them manually, making it impractical in situations such as conversation, but the portable voice recognition electronic dictionary of the present invention has a correction function using voice input. is quite practical.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a perspective view of an embodiment of a portable voice recognition electronic dictionary of the present invention. FIGS. 2(A) to 2(D) are operation explanatory diagrams of one embodiment of the portable speech recognition electronic dictionary of the present invention when insertion and correction are performed. 3(A) to 3(C) are operation explanatory diagrams of an embodiment of the portable voice recognition electronic dictionary of the present invention when changes and corrections are made. , an operation explanatory diagram of an embodiment of the portable voice recognition electronic dictionary of the present invention when deleting and modifying. FIG. 5 is a functional block diagram of the portable voice recognition electronic dictionary of the present invention. Voice input switch Microphone power switch display means Input output line, left screen, right screen 100 ・ ・ 101 ・ 102 ・ ・ ・ 103 ・ ・ 130A ・ ・ 103 B ・ 104 ・ 105 ・ Window switch 107 ・ 108 ・ 109 ・ 201・ 502 ・ 504 ・ 505 ・ 506 ・ Delete switch Insert switch ・ Up scroll switch ・ Down scroll switch ・ Cursor ・ Voice input means ・ Voice data storage means ・ Voice recognition means ・ Translation means ・ Display means ・ Control means application Person Seiko Epson Co., Ltd. agent Patent attorney Kizobe Suzuki (1 person) + A) C day] (A) (C) Figure 3

Claims (1)

[Claims] 1. Voice input means, voice data storage means capable of accumulating voice data for one word, voice recognition means for recognizing the accumulated voice data as a word, and converting the recognized word into another language. 1. A portable speech recognition electronic dictionary, comprising: a translation means for translating into text; a display means for displaying the translation results; and a control means for controlling various functions. 2. The portable speech recognition electronic dictionary according to claim 1, wherein the control means according to claim 1 has a function of correcting the recognition result in units of single sounds or single syllables. 3. The portable speech recognition electronic dictionary as claimed in claim 1 and claim 2, wherein the recognition result correction function as claimed in claim 2 allows inputting a correct single sound or monosyllable by voice.