TWI697891B - In-ear voice device - Google Patents

Abstract

An in-ear voice device, which is worn in the ear canal of a user, includes a receiving module and a detection module. The receiving module receives a biological feature transmitted from the ear canal when the user speaks. The detection module receives the biological feature and compares such biological feature with a certain feature model. When the biological feature matches a specific feature model, the detection module produces an output signal corresponding to the biological feature. Therefore, the biological feature during the user speaking is accurately received, so as to prevent the noise interference of environment. Also, by receiving the exclusive biological feature in the ear canal of the user, a unique matching effect is achieved.

Description

In-ear voice device

The invention relates to a voice device, especially an in-ear voice device.

When using personal electronic products, in order to allow the wearer to listen to the sound information provided by the electronic product or conduct voice communication without disturbing others, headphones with a microphone have become an important accessory of electronic products.

However, when the wearer uses the microphone for voice communication, the voice content will be transmitted to the microphone through the air, which makes the microphone easy to receive the wearer's voice content and noise in the environment at the same time, which affects the quality of voice communication; In the US Patent Nos. 20180098159, 20020141602, and US9036833B2, it is mentioned that the inner ear canal microphone is used to collect sound to improve the quality of voice communication.

Furthermore, the wearer can communicate with the mobile terminal device through the voice system. For example, as long as the wearer speaks a certain command to the aforementioned mobile terminal device, the system will respond to the voice signal of the wearer. Actions. At present, the voice system needs to combine the cloud server through the network to obtain the basic services and information of the cloud server. The intention is to hand over the complex and powerful voice processing supported by the computing power to the cloud server to execute. Reduce the cost of hardware configuration required for mobile terminal devices.

However, for actions such as calling and sending text messages through the address book, it is necessary to upload the address book to the cloud server to find the person to call or send the text message; therefore, the wearer’s personal privacy is still safe. There are hidden worries.

In addition, the current voice system needs to be activated when the mobile terminal device is used, and the wearer needs to speak the official key phrase to the mobile terminal device, such as hey siri or hey google now, etc., to activate the voice system; but the official The developed key phrase can be provided to anyone to use, that is to say, anyone who speaks the developed key phrase to the mobile terminal device, or someone utters the developed key phrase around the wearer in the environment, can be used The wake-up voice system, as a result, does not have an exclusive matching control method, which easily causes personal data stored in the mobile terminal device to be easily leaked, and cannot effectively protect the wearer's data privacy and use security.

In order to solve the above problems, the present invention provides an in-ear voice device that can effectively receive the biological characteristics of the wearer's speech from the wearer's ear canal, and avoid being affected by the noise of the surrounding environment; furthermore, by extracting the wearer The exclusive biological characteristics in the ear canal are compared with the specific characteristic model, which can achieve the exclusive matching effect, and then accurately generate the output signal.

An embodiment of the present invention provides an in-ear voice device for wearing in the ear canal of a wearer. The in-ear voice device includes: a receiving module for receiving the ear canal when the wearer speaks One of the biological characteristics transmitted; and a detection module, which is coupled with the receiving module, the detection module is used to receive the biological characteristics, and compare the biological characteristics with a specific feature model, when the biological characteristics and the specific characteristics When the model is matched, the detection module generates an output signal corresponding to the specific characteristic model.

Based on the above, the present invention can effectively receive the wearer’s speech by the wearer’s ear canal. Biological features can achieve high-quality reception of the wearer's speech content, so as to improve the problem that when the conventional microphone receives the wearer's voice content, it is easy to receive environmental noise at the same time, causing the problem of degraded voice communication quality.

Furthermore, the present invention extracts the exclusive biological characteristics in the ear canal of the wearer and compares them with the specific feature model, which can achieve the exclusive matching effect and accurately generate the output signal; thereby, the wearer can perform various signals through the output signal Application, in order to improve the control methods that do not have exclusive matching, cannot effectively guarantee the safety of the wearer's control and use.

In one of the embodiments, the receiving module has a first receiving unit, the first receiving unit is a microphone, and the biological feature is that the first receiving unit receives the sound transmitted through the Eustachian tube when the wearer speaks.

In one of the embodiments, the receiving module has a first receiving unit, the first receiving unit is an accelerometer, and the biological feature is that the first receiving unit receives the vibration generated by the wearer's ear bones when speaking.

In one of the embodiments, the receiving module has a second receiving unit, the second receiving unit is a microphone, and the second receiving unit receives the sound transmitted through the mouth when the wearer speaks.

Based on the above, the present invention can use the first receiving unit and the second receiving unit to cooperate with each other to improve the accuracy of the comparison of the detection module and effectively achieve the exclusive matching effect.

In one of the embodiments, the present invention further has a connection module, which is coupled to the detection module and an electronic device; when the biometric model matches the specific feature model, the comparison unit generates an output signal, and the connection module The group transmits the output signal to the electronic device, and the electronic device performs a manipulation action according to the output signal; thereby, the wearer achieves the effect of exclusively controlling the electronic device through the output signal generated by the exclusive matching.

In one of the embodiments, the present invention further has a switching module, which and the detection module Coupled, the switching module has a first mode and a second mode. In the first mode, the detection module compares the biological characteristics with the specific feature model; in the second mode, the detection module compares the biological The feature is created as a specific feature model and stored in the storage unit; in this way, the wearer is provided with a self-training learning method through the switching module to create an exclusive and personalized specific feature model to effectively improve the exclusive matching degree and protect the wearer Data privacy and security of use.

In one of the embodiments, the specific feature model is to obtain one feature value of each biological feature through signal processing of plural biological features for the same specific word, and after each feature value is sampled on average, the common feature value is obtained as the specific Feature model; thereby, it can effectively increase the probability that a specific feature model is triggered by a biological feature, thereby effectively generating an output signal.

In one of the embodiments, the present invention further has a functional module, which is coupled to the detection module, the functional module is used to obtain physiological information of the wearer, and the detection module is used to combine the physiological information with A physiological feature model comparison is used to identify the identity of the wearer; thereby, the present invention has an identity recognition function, which can effectively protect the data privacy and use safety of the wearer.

1: ear canal

2: electronic device

3: application

4: build unit

100: In-ear voice device

10: body

11: Housing space

13: installation end

12: External side

14: Sound tube

15: earplugs

20: receiving module

21: The first receiving unit

22: second receiving unit

221: Noise reduction circuit

23: Speaker unit

30: Function module

40: Detection module

41: storage unit

42: Comparison unit

50: Switch module

60: Link Module

70: Processing module

Figure 1 is a schematic cross-sectional view of the structure of the first embodiment of the present invention.

Fig. 2 is a schematic diagram of the first embodiment of the present invention being worn on the ear canal.

FIG. 3 is a schematic diagram of the structure of the first embodiment of the present invention.

Figure 4 is a diagram of a specific feature model of the present invention.

Fig. 5 is a schematic cross-sectional view of the structure of the second embodiment of the present invention.

6 is a schematic diagram (1) of the connection of the second embodiment of the present invention, showing that the electronic device is a non-intelligent mobile device.

FIG. 7 is a schematic diagram (2) of the connection of the second embodiment of the present invention, showing that the electronic device is a smart mobile device.

In order to facilitate the description of the central idea of the present invention expressed in the column of the above-mentioned invention content, specific embodiments are used to express it. The various objects in the embodiment are drawn according to the proportion, size, deformation or displacement suitable for explanation, rather than drawn according to the proportion of the actual element, which will be described first.

1 to 4, in the first embodiment of the present invention, an in-ear speech device 100 is provided, which is used to be worn in the ear canal 1 of a wearer; the in-ear speech device 100 of the present invention includes: A body 10 with an accommodation space 11 inside. The main body 10 has an outer end 12 and an installation end 13 oppositely arranged. The installation end 13 has a sound tube 14 communicating with the accommodating space 11. The outer end 12 faces the outer side of the ear canal 1 and the outer circumference of the sound tube 14 The edge is provided with an earplug 15 which is used to insert into the ear canal 1 of the wearer.

A receiving module 20, which is arranged in the accommodating space 11. The receiving module 20 is used to receive a biological characteristic transmitted in the ear canal 1 when the wearer speaks. The receiving module 20 can be a microphone and/or acceleration According to an accelerometer, the receiving module 20 has a first receiving unit 21 and a second receiving unit 22.

In one of the embodiments of the present invention, the first receiving unit 21 is an internal microphone, and the first receiving unit 21 can be arranged close to the installation end 13, such as in the accommodating space 11 or the sound tube 14, and the first The receiving unit 21 can be a bone conduction, air conduction or optical microphone, and the biological feature is that the first receiving unit 21 receives the sound transmitted by the wearer through the Eustachian tube when speaking; in another aspect of the present invention In the embodiment, the first receiving unit 21 is an accelerometer, and the biological feature is that the accelerometer receives the vibration generated by the ear bone when the wearer speaks.

Furthermore, the second receiving unit 22 is an external microphone and is arranged close to the external terminal 12. The second receiving unit 22 is used for receiving the sound transmitted through the mouth of the wearer when speaking, or receiving the surrounding sound of the wearer; in addition, the present invention further includes a speaker unit 23 coupled to the second receiving unit 22, The second receiving unit 22 is coupled to a noise reduction circuit 221. The noise reduction circuit 221 has the functions of filtering, inverting and amplifying the signal. The loudspeaker unit 23 is used to play audio signals, such as music or voice.

When the second receiving unit 22 receives external noise, the second receiving unit 22 provides an environmental signal corresponding to the external noise to the noise reduction circuit 221, after filtering the signal, the signal is reversed and amplified, and the noise cancellation signal is output to the speaker The sound unit 23 is generated by the inverse filtering and amplification of the noise cancellation signal according to the environmental signal. After being superimposed and mixed with the output sound signal of the loudspeaker unit 23, the noise cancellation signal can cancel the ambient noise from the outside and provide noise-free Sound signal or voice to the wearer.

A functional module 30 is used to obtain physiological information of the wearer, wherein the biological information can be the change of the blood vessel volume of the wearer, but is not limited to this. When the physiological information is the change of blood vessel volume, the photoplethysmography (PPG) is used to obtain the change of the diameter of the blood vessel in the ear due to the heartbeat of the wearer's ear.

A detection module 40 is arranged in the accommodating space 11 and is coupled to the receiving module 20 and the functional module 30. The receiving module 20 transmits the received biological characteristics to the detection module 40, and the detection module 40 compares the biological characteristics with a specific characteristic model. When the biological characteristics match the specific characteristic model, the detection module 40 Generate an output signal corresponding to a specific characteristic model, where the output signal is an analog or digital signal that carries information physical quantities, and the information content can be the state, audio, image, or voice that conveys phenomenon behavior or attributes, etc. The present invention does not take this as limit.

The detection module 40 has a storage unit 41 and a comparison unit 42. The storage unit 41 is provided with a specific feature model, wherein the specific feature model is a signal model formed by a specific vocabulary; the comparison unit 42 has a signal processing circuit, The comparison unit 42 is used to receive and compare the biological characteristics through the signal processing circuit Perform signal filtering and collect a specific bandwidth to generate a biometric model. The comparison unit 42 matches the biometric model with the specific feature model. When the biometric model matches the specific feature model, the detection module 40 corresponds to the specific feature The model generates an output signal.

Furthermore, the specific feature model is to process multiple biological features for the same specific vocabulary to obtain one feature value of each biological feature, and after each feature value is sampled on average, the common feature value is obtained as the specific feature model, such as As shown in FIG. 4, in the embodiment of the present invention, the specific characteristic model is a coordinate graph of cepstrum corresponding to time. Wherein, the plurality of biological characteristics can be spoken by the same wearer a plurality of specific words, or different wearers speak specific words; further explanation: the comparison unit 42 receives the biological characteristics received by the module 20 through Fourier transforms into time-frequency signals, removes unnecessary frequency ranges through filtering, performs sampling analysis on the filtered signals, and then performs characteristic analysis on the signals sampled and analyzed according to the algorithm, and the analyzed characteristics can be converted into specific neural network Feature model; among them, the biological feature model is also generated by the comparison unit 42 through the aforementioned Fourier transformation, filtering, sampling, feature analysis, and conversion into a neural network.

Therefore, the present invention accurately receives the biological characteristics of the wearer when speaking through the receiving module 20, and compares and processes through the detection module 40 to generate a uniquely matched output signal for subsequent various signal applications. In addition, when the wearer receives the biological characteristics through the first receiving unit 21, the second receiving unit 22 can also receive the specific words spoken by the mouth, so the detection module 40 can in addition to the biological characteristics of the first receiving unit 21 For comparison, the specific vocabulary received by the second receiving unit 22 can also be compared, so as to achieve a double comparison and effectively improve the matching accuracy.

In another embodiment of the present invention, the storage unit 41 of the detection module 40 is provided with a physiological characteristic model, wherein the physiological characteristic model is a biological characteristic specific to the wearer, such as a fingerprint or blood vessel volume change; a comparison unit 42 is used to compare the physiological information with the physiological characteristic model. When the physical feature model matches, the identity of the wearer is confirmed.

A switch module 50, which is disposed on the main body 10 and is coupled to the detection module 40. The switch module 50 has a first mode and a second mode. In the first mode, the detection module 40 displays biological characteristics Compare with the specific feature model; in the second mode, the detection module 40 establishes the biological feature as a specific feature model and stores it in the storage unit 41; in the embodiment of the present invention, the switching module 50 is a switch button with The wearer can switch between the first mode and the second mode through different pressing methods, for example, a short press of the switching module 50 is the first mode, and a long press of the switching module 50 is the second mode.

To further explain: when the wearer manipulates the switching module 50 in the first mode of control, the wearer can manipulate the switching module 50 to enter the second mode while using the in-ear voice device 100, and the wearer can The storage unit 41 establishes a specific characteristic model exclusive to the individual, and the specific characteristic model corresponds to an existing manipulation action.

Therefore, the wearer can use the switching module 50 to establish an exclusive and personalized specific feature model by self-training and learning, so as to effectively improve the exclusive matching degree, and thereby protect the wearer's data privacy and use security.

Please refer to FIG. 5 to FIG. 7, which is a second embodiment of the present invention. The difference from the previous embodiment is that the in-ear voice device 100 further has: a connection module 60, which is combined with a detection module 40 and an electronic device 2 coupling, where the connection module 60 can be wired or wirelessly connected to the electronic device 2. When the connection module 60 and the electronic device 2 are wirelessly connected, the connection module 60 is a Bluetooth module; the electronic device 2 can be Smart mobile devices or non-smart mobile devices (such as processing machines, home appliances, consumer electronics, automobiles and motorcycles, etc. that can receive signals).

When the biometric model matches the specific feature model, the comparison unit 42 will generate an output The output signal is transmitted to the electronic device 2 through the connection module 60, and the electronic device 2 performs a manipulation action according to the output signal; wherein, the output signal corresponding to the specific characteristic model can have a matching manipulation action, that is, When the storage unit 41 is provided with a plurality of specific feature models, it means that there are a plurality of specific vocabularies, and the storage unit 41 stores the manipulation actions associated with each specific vocabulary, and the same specific vocabulary can correspond to different electronic devices 2. There are different control actions; thereby, the wearer can achieve the effect of exclusively controlling the electronic device 2 through the output signal generated by the exclusive matching.

Please refer to Figure 6. Example 1: The specific word corresponding to the specific feature model is "turn on", the electronic device 2 is the processing machine, and the control action corresponding to the specific feature model is to start the power supply of the processing machine. When the biological characteristics of the specific word "turn on" spoken by the wearer are judged to match by the comparison unit 42, the link module 60 transmits the generated output signal to the electronic device 2 and starts the power supply of the processing machine.

Please refer to Figure 7. Example 2: The specific word corresponding to the specific feature model is "turn on", the electronic device 2 is a smart mobile device, and the control action corresponding to the specific feature model is to turn on the voice assistant in the smart mobile device Application program, therefore, when the biological characteristics of the specific word "turn on" spoken by the wearer are judged to match by the comparison unit 42, the link module 60 transmits the generated output signal to the electronic device 2 and activates Voice assistant application.

A processing module 70, which is coupled to the receiving module 20 and the connection module 60. The processing module 70 is used to perform noise filtering processing on the biological characteristics received by the receiving module 20, and connect the processed biological characteristics through the connection The module 60 is transmitted to the electronic device 2 to cooperate with the control action; wherein, the biological feature has a first biological feature and a second biological feature. The first biological feature is used to compare with a specific feature model, and the second biological model Used to cooperate with the control action.

Further explanation: When the wearer matches the specific feature model through the first biological feature Then, it can communicate with the electronic device 2 through the output signal. Then, when the wearer continues to talk, the second biological feature can be transmitted to the electronic device 2 through the connection module 60 to cooperate with the control. Action; for example: when the first biological feature matches the specific feature model, the electronic device 2 activates the voice assistant application according to the output signal and the control action; after waking up the voice assistant application, the wearer continues to speak the second creature Features: After the noise processing by the processing module 70, it can be completely and clearly transmitted to the voice assistant application.

Furthermore, when the electronic device 2 is a smart mobile device, the electronic device 2 is provided with an application program 3. The application program 3 has a establishing unit 4, the establishing unit 4 is connected to the switching module 50 by a signal, and the establishing unit 4 receives the module The biological feature received at 20 is designated as a specific feature model conforming to the manipulation action, and the detection module 40 creates the specific feature model in the storage unit 41. Further explanation: when the wearer opens the creation unit 4 of the application 3, the creation unit 4 will be connected to the switching module 50, and the switching module 50 will be manipulated into the second mode, and the wearer can create a personal in the storage unit 41 Exclusive specific feature model, and the control action corresponding to this specific feature model can be specified in the creation unit 4 of the application 3.

In summary, the effects that the present invention can achieve are as follows:

1. When the wearer puts the in-ear voice device 100 on the ear canal 1 and starts to speak, the receiving module 20 receives the biological characteristics in the ear canal 1; thus, it is not affected by the external environment by the receiving module 20 Clearly and completely receive the wearer’s speech content; and when the wearer speaks a specific vocabulary through the in-ear voice device 100, the receiving module 20 can extract the exclusive biological characteristics of the wearer’s ear canal 1 through the detection module 40 Compare the exclusive biometric model with the specific feature model to achieve the exclusive matching effect; thereby, it can effectively protect the wearer's data privacy and use security issues.

2. The present invention is further equipped with the second receiving unit 22 to assist the matching effect of the first receiving unit 21, which can effectively improve the accuracy of exclusive matching.

3. Before the wearer uses the in-ear voice device 100, he can perform body identification functions through the functional module 30, and after successful identity recognition, he can perform voice communication or signal control through the in-ear voice device 100. In order to effectively protect the wearer's data privacy and use security issues.

4. The switching module 50 of the present invention can provide the wearer to create an exclusive and personalized specific characteristic model, so as to achieve exclusive matching and control effects, so as to effectively protect the wearer's data privacy and use security.

5. The specific feature model of the present invention is formed by the statistical analysis of plural biological features, so it can compare the specific feature model with high matching degree with the biological feature model, thereby increasing the triggering probability, and generating the output signal quickly and effectively.

The above-mentioned embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. All modifications or changes that do not violate the spirit of the present invention belong to the scope of the present invention.

Claims (11)

An in-ear voice device for wearing in the ear canal of a wearer, the in-ear voice device comprising: a receiving module for receiving a biological characteristic transmitted in the ear canal when the wearer speaks; And a detection module coupled to the receiving module, the detection module having a storage unit and a comparison unit, the storage unit is provided with a specific feature model, the specific feature model is formed by a specific word A signal model, wherein the specific feature model obtains one feature value of each biological feature by signal processing the biological feature for the same specific vocabulary, and after each feature value is sampled on average, the common feature value is obtained as The specific feature model; the comparison unit has a signal processing circuit, the comparison unit is used to receive and signal processing the biological feature through the signal processing circuit to generate a biological feature model, the comparison unit The feature model is matched with the specific feature model; when the biological feature model matches the specific feature model, the detection module generates an output signal corresponding to the specific feature model. The in-ear voice device according to claim 1, wherein the receiving module has a first receiving unit, the first receiving unit is a microphone, and the biological feature is that the first receiving unit receives the wearer’s speech through the Eustachian tube The voice delivered. The in-ear voice device according to claim 1, wherein the receiving module has a first receiving unit, the first receiving unit is an accelerometer, and the biological characteristic is that when the first receiving unit receives the wearer’s ear Vibration produced by bone. The in-ear voice device according to claim 2 or 3, wherein the receiving module has a second receiving unit, the second receiving unit is a microphone, and the second receiving unit receives the speech transmitted by the wearer through the mouth The voice. The in-ear voice device according to claim 4, wherein the receiving module has a speaker unit, and the speaker unit is coupled to a noise reduction circuit. The in-ear voice device according to claim 1, further has a link module, which is coupled with the detection module and an electronic device; when the biometric model matches the specific feature model, the comparison unit The output signal is generated, the connection module transmits the output signal to the electronic device, and the electronic device performs a manipulation action according to the output signal. The in-ear voice device according to claim 6, wherein the biological characteristic has a first biological characteristic and a second biological characteristic, and the detection module forms the first biological characteristic into the biological characteristic model to be compatible with the biological characteristic Specific feature model comparison; the link module transmits the second biological feature to the electronic device, and the second biological feature cooperates with the manipulation action. For example, the in-ear voice device according to claim 7 further has a processing module, which is coupled to the receiving module and the connection module, and the processing module is used for the second creature received by the receiving module The feature performs noise filtering processing, and transmits the processed second biological feature to the electronic device through the link module to cooperate with the manipulation action. The in-ear voice device according to claim 1, further has a switching module coupled to the detection module, the switching module has a first mode and a second mode, and when in the first mode , The detection module compares the biological feature with the specific feature model; in the second mode, the detection module creates the biological feature as the specific feature model and stores it in the storage unit. The in-ear voice device described in claim 6 further has a switch module coupled to the detection module, the switch module has a first mode and a second mode; the electronic device has an application The application program has a creation unit; in the first mode, the detection module processes the biological feature into the biological feature model to compare with the specific feature model; the creation unit starts When the second mode is activated, the detection module establishes the biological feature as the specific feature model and stores it in the storage unit. The in-ear voice device described in claim 1 further has a functional module, which is coupled to the detection module, and the functional module is used to obtain physiological information of the wearer. The detection module uses The physiological information is compared with a physiological characteristic model to identify the identity of the wearer.

Images