WO2020189410A1

WO2020189410A1 - Speech recognition device

Info

Publication number: WO2020189410A1
Application number: PCT/JP2020/010283
Authority: WO
Inventors: 優坂西
Original assignee: 優坂西
Priority date: 2019-03-15
Filing date: 2020-03-10
Publication date: 2020-09-24

Abstract

Provided is a speech recognition device that can be used efficiently in accordance with a situation in which a conversation as the object of speech recognition takes place. The speech recognition device 1 is provided with: a plurality of microphones; a sensor 23 which determines whether the speech recognition device is in a stationary state; a microphone control unit 24 which controls whether to place each of the plurality of the microphones in on-state or off-state depending on the result of determination by the sensor; and a speech recognition unit 21 which recognizes speech that has been input to at least one microphone in on-state.

Description

Voice recognition device

The present invention relates to a voice recognition device.

With the development of voice recognition technology, the demand for devices equipped with a voice recognition function is increasing. Patent Document 1 describes a simultaneous interpretation device including two displays. In the same document, a state in which the simultaneous interpretation device is used while being placed on a stationary body such as a desk or a table is illustrated.

JP-A-2018-195276

An object of the present invention is to provide a voice recognition device that can be efficiently used according to a situation in which a conversation that is the target of voice recognition is performed.

In order to achieve the above object, the voice recognition device according to the present invention includes a plurality of microphones, a sensor for determining whether or not the voice recognition device is in a stationary state, and a plurality of sensors according to the determination result by the sensor. The voice recognition unit includes a microphone control unit that controls whether each of the microphones is turned on or off, and a voice recognition unit that recognizes a voice input to at least one microphone in the on state. The microphones are arranged on the left side surface portion, the right side surface portion, and the upper surface portion of the device.

According to the present invention, it is possible to provide a voice recognition device that can be efficiently used according to a situation in which a conversation that is the target of voice recognition is performed.

It is a front view of the voice recognition device. It is a rear view of the voice recognition device. It is a left side view of a voice recognition device. It is a right side view of the voice recognition device. It is a top view of the voice recognition device. It is a bottom view of the voice recognition device. It is a functional block diagram of a voice recognition device. It is explanatory drawing which shows the case of using the voice recognition device in the state of holding by hand. It is another explanatory drawing which shows the case where the voice recognition device is used in the state of holding by hand. It is explanatory drawing which shows the case of using with the voice recognition apparatus placed. It is another explanatory diagram which shows the case of using with the voice recognition device placed. It is a front view of the voice recognition apparatus which concerns on another embodiment. It is a rear view of the voice recognition apparatus which concerns on another embodiment. It is a left side view of the voice recognition apparatus which concerns on another embodiment. It is a right side view of the voice recognition apparatus which concerns on another embodiment. It is a top view of the voice recognition apparatus which concerns on another embodiment. It is a bottom view of the voice recognition device which concerns on another embodiment. It is a front view of the voice recognition apparatus which concerns on still another embodiment. It is a rear view of the voice recognition apparatus which concerns on still another embodiment. It is a left side view of the voice recognition apparatus which concerns on still another embodiment. It is a right side view of the voice recognition apparatus which concerns on still another embodiment. It is a top view of the voice recognition apparatus which concerns on still another Embodiment. It is a bottom view of the voice recognition device which concerns on still another embodiment. It is a perspective view of the voice recognition apparatus which concerns on still another Embodiment. It is a perspective view of the voice recognition apparatus which concerns on still another Embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the present invention is not limited to the following embodiments.

As shown in FIGS. 1 to 6, the outer shape of the voice recognition device 1 is a substantially triangular prism. This substantially triangular prism is a substantially pentahedron having two faces that are substantially triangular and three faces that are substantially rectangular. The short sides of the three faces that are substantially rectangular are also the sides of the two faces that are substantially triangular. Of the above three surfaces that are substantially rectangular, one surface is referred to as a front portion and is indicated by reference numeral 11. A display unit D visually recognized by the user of the voice recognition device 1 is arranged on the front unit 11.

The voice recognition device 1 is horizontal from the front so that the two

long sides

11a and 11b of the front portion 11 are horizontal and one long side 11a is located directly above the other long side 11b. Think about the state you saw in. In this state, the substantially triangular surfaces on the left and right are referred to as the left side surface portion 13 and the right side surface portion 14, respectively, and the substantially rectangular surface sharing the front portion 11 and the long side portion 11a is referred to as the upper surface portion 15 and is long with the front portion 11. A substantially rectangular surface that shares the side portion 11b is referred to as a lower surface portion 16.

The front surface portion 11, the upper surface portion 15, and the lower surface portion 16 of the voice recognition device 1 are substantially rectangular, and each side of the left side surface portion 13 and the right side surface portion 14 is compared with the lengths of the

long side portions

11a and 11b of the front surface portion 11. The length of the part is short. That is, the voice recognition device 1 has an elongated shape in the left-right direction.

The angle α formed by the front portion 11 and the left surface portion 13 is an acute angle (FIG. 5). Similarly, the angle formed by the front surface portion 11 and the right side surface portion 14 is also an acute angle. Further, the angle β formed by the side portion 12 and the left side surface portion 13 shared by the upper surface portion 15 and the lower surface portion 16 is an obtuse angle. Similarly, the angle formed by the side portion 12 and the right side surface portion 14 is also an obtuse angle.

The voice recognition device 1 is easy to carry and has a handy size so that it can be used even when it is held by hand. As an example, the length of the front portion 11 in the left-right direction is 120 mm, the distance between the long side portion 11a and the long side portion 11b is 30 mm, and the depth (the length from the front portion 11 to the side portion 12). Is 15 millimeters.

As shown in FIG. 5, the first microphone M1 is arranged on the upper surface portion 15. As shown in FIG. 3, a second microphone M2 is arranged on the left side surface portion 13. As shown in FIG. 4, a third microphone M3 is arranged on the right side surface portion 14. When the voice recognition device 1 is viewed horizontally from the front, the fourth microphone M4 is arranged below the display unit D in the front portion 11.

The first microphone M1, the second microphone M2, the third microphone M3, and the fourth microphone M4 are all directional microphones having directivity in front of each microphone. Compared to an omnidirectional microphone, a directional microphone can capture the voice of a speaker located farther away from the front of the directional microphone. Each microphone converts the input voice into an electrical signal when it is on. The control of whether each microphone is turned on or off will be described later.

As shown in FIG. 7, the voice recognition device 1 further includes a voice recognition unit 21, a translation unit 22, a sensor 23, and a microphone control unit 24. However, as will be described later, the translation unit 22 is not an essential element.

The voice recognition unit 21 receives an electric signal from at least one microphone in the ON state among the first to fourth microphones M1 to M4, and performs voice recognition. The voice recognition unit 21 recognizes the electric signal as the first language. This voice recognition result is displayed as a character string on the display unit D.

The translation unit 22 translates the voice recognized as the first language by the voice recognition unit 21 into a second language different from the first language. This translation is done by the translation engine. The translation engine is built in the voice recognition device 1 or is provided outside the voice recognition device 1 so as to be able to communicate with the translation unit 22. The translation result by the translation unit 22 is displayed as a character string on the display unit D.

The acceleration sensor 23 measures the acceleration of the voice recognition device 1 and determines whether or not the voice recognition device 1 is in a stationary state based on the measured acceleration. When the voice recognition device 1 is placed on a stationary body such as a desk or a table and is stationary, the acceleration sensor 23 determines that the voice recognition device 1 is in a stationary state based on the measured acceleration. On the other hand, when the voice recognition device 1 is not stationary because the user of the voice recognition device 1 holds the voice recognition device 1 in his / her hand, the acceleration sensor 23 uses the measured acceleration to make a voice. The recognition device 1 determines that it is not in a stationary state.

The microphone control unit 24 controls whether each microphone is turned on or off according to the determination result of the acceleration sensor 23. When the acceleration sensor 23 determines that the voice recognition device 1 is in the stationary state, all the first to fourth microphones M1 to M4 are controlled to be in the ON state. On the other hand, when it is determined that the voice recognition device 1 is not in the stationary state, the first microphone M1 is turned on and the second to fourth microphones M2 to M4 are controlled to be turned off.

Although not shown, the voice recognition device 1 has a computer hardware configuration of an interface device capable of communicating with a processor and an external computer, and an input used by a user of the voice recognition device 1 for input. It further includes a device and a storage device.

8 and 9 show a state in which the user P1 of the voice recognition device 1 has a conversation with the speaker P2 while holding the voice recognition device 1 in his hand. Since the user P1 holds the voice recognition device 1 in his hand, the acceleration sensor 23 determines that the voice recognition device 1 is not in a stationary state. As a result, the microphone control unit 24 is controlled so that the first microphone M1 is turned on and the second to fourth microphones M2 to M4 are turned off. The region where the first microphone M1 can acquire voice is indicated by reference numeral S1.

When the user P1 tilts the voice recognition device 1 toward the user P1 so that the display unit D can be easily seen by himself / herself, the upper surface portion 15 almost faces the speaker P2. As a result, the region S1 from which the first microphone M1 can acquire voice has a spread from the upper surface portion 15 toward the speaker P2.

The voice emitted by the speaker P2 is input to the first microphone M1 in the on state, and is recognized as the first language (English in this example) by the voice recognition unit 21. The recognition result is displayed in the area on the left side of the display unit D (FIG. 8). The voice recognized by the voice recognition unit 21 is translated into a second language (Japanese in this example) by the translation unit 22. The translation result is displayed in the area on the right side of the display unit D (FIG. 8).

As described above, since the outer shape of the voice recognition device 1 is a substantially triangular prism, the voice recognition device 1 is tilted so that the user P1 can easily see the display unit D, and at the same time, the area S1 is directed toward the speaker P2. Since it has a spread, the voice emitted by the speaker P2 can be efficiently acquired. Further, since the direction of the speaker P2 is easy for the user P1 to see, the user P1 can naturally have a conversation with the speaker P2 while using the voice recognition device 1.

Further, the voice recognition device 1 has an elongated shape in the left-right direction, and the user P1 so that the left-right direction (longitudinal direction) of the voice recognition device 1 is perpendicular to the direction in which the user P1 and the speaker P2 face each other during use. Will hold the voice recognition device 1 in his hand. Therefore, the feeling of oppression on the speaker P2 due to the use of the voice recognition device 1 can be reduced as compared with the case where the longitudinal direction of the voice recognition device 1 is parallel to the direction in which the user P1 and the speaker P2 face each other. ..

As shown in FIGS. 8 and 9, when the user P1 holds the voice recognition device 1 in his / her hand and the user P1 performs a specific operation on the voice recognition device 1, the fourth microphone M4 moves. It may be controlled to be on. Alternatively, the first microphone M1 and the fourth microphone M4 may be controlled to be in the ON state regardless of whether or not the specific operation is performed.

The voice emitted by the user P1 is captured by the fourth microphone M4 in the ON state, and can be recognized by the voice recognition unit 21. Further, the recognition result can be translated into another language by the translation unit 22. The translation result is displayed on the display unit D and can be visually recognized by the user P1. User P1 can utter the content while looking at the translation result.

10 and 11 show a state in which the user P1 of the voice recognition device 1 talks with the speakers P3 to P5 with the voice recognition device 1 placed on the table 9 as a stationary body. The voice recognition device 1 is placed on the table 9 so that the lower surface portion 16 is a contact surface with the table 9. Both the user P1 and the speakers P3 to P5 are seated. The speaker P3 sits to the right of the user P1, the speaker P4 sits facing the user P1 across the table 9, and the speaker P5 sits facing the speaker P3 across the table 9. There is.

Since the voice recognition device 1 is placed on the table 9, the acceleration sensor 23 determines that the voice recognition device 1 is in a stationary state. As a result, the microphone control unit 24 controls all the first to fourth microphones M1 to M4 to be in the ON state. In addition to the area S1 where the first microphone M1 can acquire sound, the areas where the second to fourth microphones M2 to M4 can acquire sound are shown as S2 to S4, respectively.

When the voice recognition device 1 is placed on the table 9 so that the lower surface portion 16 is a contact surface with the table 9, since the outer shape of the voice recognition device 1 is a substantially triangular prism, the display unit D is displayed for the user P1. It will be easy to see. Further, the upper surface portion 15 on which the first microphone M1 is arranged faces the speakers P4 and P5, and the right side surface portion 14 on which the third microphone M3 is arranged faces the speakers P3 and P5.

Therefore, the region S1 from which the voice can be acquired by the first microphone M1 has a spread toward the speakers P4 and P5, and the region S3 from which the voice can be acquired by the third microphone M3 becomes the speakers P3 and P5. It will have a spread toward it.

The voice emitted by the speaker P3 is input to the third microphone M3 and recognized by the voice recognition unit 21. The voice emitted by the speaker P4 is input to the first microphone M1 and recognized by the voice recognition unit 21. The voice emitted by the speaker P5 is input to at least one of the first microphone M1 and the third microphone M3, and is recognized by the voice recognition unit 21. With the voice recognition device 1 placed in front of the user P1, the voice recognition device 1 can efficiently acquire the voices emitted by the speakers P3 to P5.

Also, for user P1, speakers P3 to P5 are easier to see. That is, the user P1 can naturally have a conversation with the speakers P3 to P5 while using the voice recognition device 1.

Note that the fourth microphone M4 may be controlled so as to be turned on when a specific operation is performed on the voice recognition device 1 by the user P1.

As described above, using the voice recognition device 1, the voice recognition device 1 is placed on a stationary body such as a table 9 or a desk even when the voice recognition device 1 is held in the hand according to the situation in which a conversation is performed. Even in that case, the on / off of each of the plurality of microphones is controlled accordingly. The voice recognition device 1 can efficiently acquire the voice of the speaker by the microphone turned on. At the same time, the voice recognition result and the translation result can be easily visually recognized by the user through the display unit D.

[Other]
Although an example is shown in which the outer shape of the voice recognition device 1 is a substantially triangular prism, another shape such as a rectangular parallelepiped may be used. Further, the voice recognition device 1 is not limited to the acceleration sensor, and the voice recognition device 1 may be provided with a sensor means capable of determining whether or not the voice recognition device 1 is in a stationary state.

The voice recognition device 1 does not have to include the translation unit 22. In this case, the voice recognition device 1 can recognize the input voice as a language and store the recognition result in the storage device in the voice recognition device 1. Alternatively, the voice recognition result may be displayed on the display unit D, or may be sent to another device outside the voice recognition device 1. This voice recognition result can be used as a material for recording minutes and the like, and can also be used as a material for analyzing the productivity of conversations and meetings.

The voice recognition device 1 is small and lightweight enough to be used even when held by hand, and is equipped with a plurality of directional microphones. The plurality of directional microphones can capture and recognize the voice emitted by the participants of the target conversation or conference even if they are relatively far from the voice recognition device. In other words, there is an increased possibility that the target conversation and the voices of all the participants in the conference can be captured and recognized.

The above embodiment will be described below again.
[Part 1]
The voice recognition device 1 turns on each of the plurality of microphones, the sensor 23 for determining whether or not the voice recognition device is in a stationary state, and the plurality of microphones according to the determination result by the sensor. It includes a microphone control unit 24 that controls whether to turn off the state, and a voice recognition unit 21 that recognizes the sound input to at least one microphone in the on state.

As a result, the on / off of each microphone is controlled according to the state of the voice recognition device. As a result, it is possible to efficiently acquire the voice that is the target of voice recognition.

[Part 2]
The outer shape of the voice recognition device 1 is a substantially triangular prism having two faces having a substantially triangular shape and three faces having a substantially rectangular shape, and each of the two faces having the substantially triangular shape (left side surface portion 13 and right side surface portion 14). A microphone is arranged on each of two surfaces (front surface portion 11 and upper surface portion 15) of the three substantially rectangular surfaces (front surface portion 11, upper surface portion 15 and lower surface portion 16).

As a result, there is a high possibility that the voice acquisition area of each microphone will be expanded toward the speaker with whom the user of the voice recognition device has a conversation. This leads to efficient acquisition of audio.

[Part 3]
When the sensor determines that the voice recognition device is in a stationary state, the microphone control unit turns on all of the plurality of microphones. When the sensor determines that the voice recognition device is not in a stationary state, the microphone control unit is arranged on each of two of the three substantially rectangular surfaces (front surface portion 11 and upper surface portion 15). The microphone is turned on, and the microphones arranged on each of the two substantially triangular surfaces (left side surface portion 13 and right side surface portion 14) are turned off.

This increases the possibility that the voice can be efficiently acquired both when the voice recognition device is in the stationary state and when it is not in the stationary state.

[Part 4]
The voice recognition device 1 further includes a display unit D that displays a voice recognition result by the voice recognition unit 21. As a result, the user of the voice recognition device can visually recognize the voice recognition result through the display unit D.

[Part 5]
The voice recognition device 1 further includes a translation unit 22 that translates the voice recognized as the first language by the voice recognition unit 21 into a second language different from the first language. The display unit D further displays the translation result by the translation unit. As a result, the voice recognition result is translated into a language different from the recognized language, and the translation result can be visually recognized by the user through the display unit D.

Hereinafter, another example of the voice recognition device will be described.

12 to 17 show a voice recognition device 1A having a substantially quadrangular pyramid shape. The quadrangular pyramid stand is a three-dimensional figure obtained by cutting a quadrangular pyramid in two on a plane parallel to the bottom surface and excluding the portion including the apex of the original quadrangular pyramid. In other words, a quadrangular pyramid is a three-dimensional figure surrounded by four pyramid surfaces of a quadrangular pyramid, a bottom surface of the quadrangular pyramid, and a plane parallel to the bottom surface.

Of the two parallel flat surfaces forming the outer shape of the voice recognition device 1A, the flat surface portion having a large surface area is called a front surface portion and is indicated by reference numeral 11A. The other flat portion having a small surface area is called a back surface portion and is indicated by reference numeral 12A. Both the front portion 11A and the back portion 12A are substantially rectangular. A display unit D visually recognized by the user of the voice recognition device 1A is arranged on the front unit 11A.

The voice recognition device 1A is provided so that the two

long sides

11a ₁ and 11b _{1 of the} front portion 11A are horizontal, and one long side 11a ₁ is located above the other long side 11b _1. Consider the state of looking horizontally from the front. In this state, of the four pyramidal surface portions, the pyramidal surface portion located on the left side is called the left side surface portion of the voice recognition device 1A, and is indicated by reference numeral 13A, and the pyramid surface portion located on the right side is the right side portion of the voice recognition device 1A. It is called a face portion and is indicated by reference numeral 14A. In the same state, the pyramidal surface portion located on the upper side is referred to as the upper surface portion of the voice recognition device 1A and is indicated by reference numeral 15A, and the pyramid surface portion located on the lower side is referred to as the lower surface portion of the voice recognition device 1A and is referred to by reference numeral 16A. Shown.

The front portion 11A and the back portion 12A of the voice recognition device 1A are substantially rectangular, and the length from the front portion 11A to the back portion 12A is larger than the length of the

long side portions

11a ₁ and 11b ₁ of the front portion 11A. short. That is, the voice recognition device 1A has an elongated shape in the left-right direction.

The first microphone M1 is arranged on the upper surface portion 15A. The second microphone M2 is arranged on the left side surface portion 13A, and the third microphone M3 is arranged on the right side surface portion 14A. Further, in the front portion 11A, the fourth microphone M4 is arranged below the display portion D.

18 to 25 show another voice recognition device 1B having a substantially quadrangular pyramid shape.

Of the two parallel flat surfaces forming the outer shape of the voice recognition device 1B, the flat surface portion having a large surface area is called a front surface portion and is indicated by reference numeral 11B. Further, the other flat portion having a small surface area is called a back surface portion and is indicated by reference numeral 12B. Both the front portion 11B and the back portion 12B are substantially rectangular. A display unit D visually recognized by the user of the voice recognition device 1B is arranged on the front unit 11B.

The voice recognition device 1B is set so that the _two

long sides

11a ₂ and 11b _{2 of the} front portion 11A are horizontal, and one long side 11a ₂ is located above the other long side 11b _2. Consider the state of looking horizontally from the front. In this state, of the four pyramidal surface portions, the pyramidal surface portion located on the left side is called the left side surface portion of the voice recognition device 1A, and is indicated by reference numeral 13B, and the pyramid surface portion located on the right side is the right side portion of the voice recognition device 1B. It is called a face portion and is indicated by reference numeral 14B. In the same state, the pyramid surface portion located on the upper side is referred to as the upper surface portion of the voice recognition device 1B and is indicated by reference numeral 15B, and the cone surface portion located on the lower side is referred to as the lower surface portion of the voice recognition device 1B and is referred to by reference numeral 16B. Shown.

The front portion 11B and the back portion 12B of the voice recognition device 1B are substantially rectangular, and the length from the front portion 11B to the back portion 12B is larger than the length of the

long side portions

11a ₂ and 11b ₂ of the front portion 11B. short. That is, the voice recognition device 1B has an elongated shape in the left-right direction.

The first microphone M1 is arranged on the upper surface portion 15B. The second microphone M2 is arranged on the left side surface portion 13B, and the third microphone M3 is arranged on the right side surface portion 14B. Further, on the front portion 11B, a fourth microphone M4 is arranged below the display portion D.

As with the voice recognition device having a substantially triangular pyramid shape, the voice recognition device having a substantially quadrangular pyramid shape as described above is efficiently used according to the situation in which the conversation to be voice recognition is performed. be able to. The same effect can be obtained with a voice recognition device having a pentahedron as an outer shape including a substantially triangular prism and a voice recognition device having a hexahedron as an outer shape including a substantially quadrangular pyramid.

Although the specific embodiment of the present invention has been described, the present invention is not limited to such an embodiment, and various modifications based on the technical idea of the present invention are included in the concept of the present invention.

1 Voice recognition device 11 Front part, 13 Left side part, 14 Right side part, 15 Top part, 16 Bottom part D Display (display), M1 to M4 Microphone 21 Voice recognition part, 22 Translation part, 23 Sensor, 24 Microphone control part

Claims

It is a voice recognition device
With multiple microphones
A sensor that determines whether or not the voice recognition device is in a stationary state, and
A microphone control unit that controls whether each of the plurality of microphones is turned on or off according to the determination result by the sensor.
It includes a voice recognition unit that recognizes the voice input to at least one microphone in the ON state.
A voice recognition device in which the microphones are arranged on the left side surface portion, the right side surface portion, and the upper surface portion of the voice recognition device.
The first aspect of claim 1, wherein the outer shape of the voice recognition device is a substantially quadrangular pyramid base surrounded by four pyramid surfaces of a substantially quadrangular pyramid, a bottom surface of the substantially quadrangular pyramid, and a plane parallel to the bottom surface. Voice recognition device.
When the sensor determines that the voice recognition device is in a stationary state, the microphone control unit turns on all of the plurality of microphones.
When the sensor determines that the voice recognition device is not in a stationary state, the microphone control unit turns on the microphone arranged on the upper surface portion and turns off the microphone arranged on the left side surface portion and the right side surface portion. To be in a state
The voice recognition device according to claim 1 or 2.
The voice recognition device according to any one of claims 1 to 3, further comprising a display unit for displaying the voice recognition result by the voice recognition unit.
A translation unit for translating the voice recognized as the first language by the voice recognition unit into a second language different from the first language is further provided.
The display unit further displays the translation result by the translation unit.
The voice recognition device according to claim 4.
The voice recognition result by the voice recognition unit is displayed in the area on the left side of the display unit.
The translation result by the translation unit is displayed in the area on the right side of the display unit.
The voice recognition device according to claim 5.