JPS60205594A - Recognition results display system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] <Technical Field of the Invention> The present invention recognizes speech uttered in units of phrases in units of syllables, and creates and displays a candidate string in units of phrases based on combinations of recognized syllable candidates. 0 ↓ <Technical background of the invention and its problems> Generally speaking, speech uttered in phrases can be defended in units of syllables. With voice input devices, unlike input from a keyboard, input includes so-called 6-likelihood.However, the current level of recognition devices. , it is rare that the uttered phrase is input as is.As a result, a process is performed to develop it into a string of syllable candidates and then a string of phrase candidates, which are then displayed on the screen in order of ``most likely'' candidates. At this time, the display displays the most likely phrase candidate, and if the displayed phrase is different from the phrase that was uttered, the operator operates the keyboard etc. The next candidate is displayed on the screen.

However, with this conventional method of displaying recognition results, if the uttered phrase is located at the bottom of the candidate column, it takes time to find the phrase, and if the phrase is not in the candidate column, the phrase is not found in the candidate column. , there were problems such as even if you went to the end to search for candidates, it would be in vain.

<Objective of the Invention> The present invention was made with the purpose of eliminating the above-mentioned conventional problems, and in order to achieve this purpose, the recognition result display method of the present invention displays the speech uttered in units of phrases. A speech input device that recognizes syllables and creates and displays candidate strings for each phrase has been designed to display multiple phrase candidate strings on the screen using a menu-only method, which involves uncertainty in the recognition results. has been done.

<Embodiment of the Invention> Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an example of the configuration of a voice input device in which the present invention is implemented.

In FIG. 1, l is a microphone that picks up voice input, and the voice detected by this microphone 1 is input to a monosyllable recognition section 2. This monosyllable recognition unit 2 is conventionally known, and classifies speech in units of phrases input through the microphone 1 into units of syllables, and extracts features for each monosyllable. On the other hand, the memory 8 stores a standard pattern for each monosyllable, and the monosyllable recognition unit 2 performs a matching calculation process between the characteristic pattern of the input speech and the standard pattern, and as a result of this matching calculation process, the most approximate The selected candidate is selected as the first candidate, and the sequentially approximated ones are selected as the next candidates, and the results are stored as a syllable lattice in the memory 4 together with distance difference information indicating the degree of approximation (certainty).

The content recognized by the monosyllable recognition unit 2 and stored in the memory 4 as a syllable lattice is input to the candidate sequence creation unit 5 and is highly accurate using distance difference information indicating the degree of approximation (likelihood). Clause candidates (kana character strings) are created in order and stored in the clause candidate memory 6.

The plurality of candidate columns created in the candidate column creation section 5 and stored in the memory 6 are configured to be displayed as a menu on the screen of the display device 7 according to the present invention, as will be described later.

Note that β is an input device having kana keys, function keys, etc., 9 is a kana-kanji conversion processing unit, and 1o is a controller (CPU) that controls each of the above devices.Next, the operation of the device configured as described above will be explained below. This will be explained according to the processing flow shown in FIG.

The speech uttered in units of phrases is detected by the microphone 1 and recognized by the monosyllable recognition unit 2 in units of monosyllables by acoustic analysis (nO to n3), and the recognition results are input and stored in the syllable lattice memory 4.

For example, a syllable lattice as shown in Table 1 is formed as a monosyllable recognition result for the input speech "/ko/</mi/n/ha/".

In Table 1 recognition, since ambiguity is included, the words are output as syllable candidates as shown in Table 1 above.

Recognized by the monosyllable recognition unit 2, the syllable latenos J
Each syllable unit candidate stored in the memory 4 at -1 is input to the candidate string creation section 5.

The candidate sequence creation unit 5 considers the combination of syllable/unit recognition results stored in the syllable lattice memory 4 and develops it into a phrase candidate sequence (step n4), and stores the created phrase candidate sequence in the order of accuracy. In the example shown in Table 1 above, 24 candidate columns are created and stored in the memory 6 as shown in Table 2. At this time, the phrase of the input speech that was uttered is the 12th phrase in the candidate string.

Table 2 Conventionally, the highest phrase candidate of such a plurality of candidate columns was displayed on the screen, and the next phrase candidate was displayed in order according to an operator's instruction. Also, (at this time, the display was always limited to one phrase candidate.

In the method of the present invention, along with the input sentence displayed on the screen 71 of the display device 7 as shown in FIG.
Step n5). That is, following the input sentence, a part of the clause candidate string is displayed in a multi-window (clause menu) 72.

The operator determines whether or not there is a phrase that has been uttered in the displayed one phrase menu "72" (step n6).
If there is no phrase that has been uttered, the display is switched to the next clause menu by using, for example, the cursor key on the input device 8 (step n7).

If there is a uttered phrase in the ``0 phrase menu'' displayed in this way, by indicating the number displayed together with the ``6 phrase menu'' (step n8
), the uttered phrase is determined. In the example shown in FIG. 8, it is the second of the third "Phrase Menu".

After determining the phrase, it is converted into a kanji-mixed sentence by the kana-kanji conversion processing section 9, for example, by operating a conversion key, and the result is displayed (step n9).

Furthermore, even if the uttered phrase is not in the phrase candidate column, this can be known with a few operations.

Furthermore, according to the present invention, the lower the uttered phrase is in the phrase candidate string, the shorter the time required to determine the phrase.

In the above embodiment, as shown in FIG.
Since we are performing step 2, there may be cases where the ``1 Bunsetsu Menu'' extends off the screen, but in this case, you can scroll the screen upwards to ensure space for the ``Bunsetsu Menu'' to be displayed. good.

Furthermore, in the above embodiment, a method was described in which a "one phrase menu" is displayed together with the input text, but the present invention is not limited to this, and the "one phrase menu" is always allocated to the phrase menu display area #72. After the phrase is determined, the determined phrase may be displayed in the input sentence. <Effect of the Invention> As described above, according to the present invention, the phrase is uttered in units of phrases. In a voice input device that recognizes speech in units of syllables and creates and displays candidate sequences in units of phrases, multiple phrase candidate sequences with uncertain recognition results are displayed on the screen using a menu-only method. Since the phrase is configured so as to force the user to speak the phrase, when the number of phrase candidates increases, the time required to determine the phrase to be uttered can be shortened. It is possible to check with a few operations whether or not the phrase that has yet to be uttered is included in the candidate string.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an example of the configuration of a voice input device embodying the present invention, and FIG. 2 shows an explanation of its operation. +11iSQi4n
:fKIIYIrtJ-M key de f1-Ji- FIG. 1... Microphone, 2... Monosyllable recognition unit. 3... Standard pattern memory, 4... Syllable lattice memory, 5... Candidate sequence creation diagram,
6... Clause candidate memory, 7... Display device, 8...
- Input device, 71... Display screen, 72... Clause menu display screen. Agent Patent attorney Aihiko Fukushi (and 2 others) Figure 1

Claims (1)

[Claims] 1. In a speech input device that recognizes speech uttered in units of phrases in units of syllables and creates and displays a candidate string in units of phrases, A recognition result display method characterized in that a phrase candidate string is displayed on a screen using a menu.