DEVICE AND METHOD FOR ELIMINATING NOISE OR HARSHNESS, SOUND DETECTION DEVICE, THEREFORE USING THE DEVICE AND VEHICLE EQUIPPED WITH THE SOUND DETECTION DEVICE.

Abstract

Device (10) for eliminating noise or noise, which exhibits: a gain detection unit (12) which determines a gain and a correction value (α) of the gain, wherein a signal-to-noise ratio (SNR) of an input signal (I) is used, wherein the gain sensing unit (12) determines the correction value (α) of the gain based on the SNR of the input signal (I), and The gain detection unit (12) determines the correction value (α) of the gain, based further on a set value (17), which is related to a relationship between the SNR of the input signal (I) and the correction value (α), and modifies the relationship between the SNR of the input signal (I) and the correction value (α), based on the set value (17), wherein the set value (17) indicates a performance of a sound detection device (50), wherein the set value (17) refers to a value that indicates a selectable situation, and where a number of selectable set values ​​(17) corresponds to a number of selectable situations; and an amplification application unit (19) which detects an output signal (o) corresponding to the input signal (I), wherein the specified amplification and the specified correction value (α) are used, wherein the output signal (o) includes an input signal (I) whose noise (N) is eliminated, and an input signal (I) whose noise (N) is not eliminated, and wherein the amplification application unit (19) is configured to determine a portion of the input signal (I) whose noise (N) has been eliminated and a portion of the original input signal (I) corresponding to the determined correction value (α).

Description

BACKGROUND 1. Technical Area

[0001] Embodiments of the present disclosure relate to a device and method for eliminating noise or sound, a sound detection device, wherein the device and a vehicle equipped with the sound detection device are used. 2. Description of the state of the art

[0002] As is well known in the field, a vehicle is a means of transport that can move an object, such as a person or cargo, to a different position while moving, for example, on a road or railway tracks. A vehicle can move primarily by the rotation of one or more wheels mounted on its main body. Examples of vehicles include three-wheeled and four-wheeled motor vehicles, two-wheeled motor vehicles such as motorcycles, motorized bicycles, construction equipment, bicycles, and trains that run on railway tracks.

[0003] A sound recognition device can be installed in a vehicle. This device can detect sounds produced by a user's voice, such as a driver or passenger. When a sound is detected, a control unit within the vehicle sends control signals corresponding to the detected sound to vehicle components, enabling them to operate accordingly. By using this sound recognition device, the user can control the vehicle's components using the sound, thus increasing driver comfort and safety.

[0004] From JP 2000 – 82 999 A, a device for eliminating noise is known, comprising a gain detection unit which determines a gain and a correction value of the gain, using a signal-to-noise ratio (SNR) of an input signal, and a gain application unit which detects an output signal corresponding to the input signal, using the determined gain and the determined correction value, wherein the output signal includes an input signal whose noise is eliminated and an input signal whose noise is not eliminated, and wherein the gain application unit is configured to determine a portion of the input signal whose noise is eliminated and a portion of the original input signal, corresponding to the determined correction value.

[0005] US Patent 2011 / 0286605A1 further discloses a speech / noise decision unit that uses a low-frequency amplitude spectrum to determine whether an input signal resembles speech. A noise spectrum estimator estimates a low-frequency noise spectrum and a high-frequency noise spectrum from the output of the speech / noise decision unit. A low-frequency processing unit and a high-frequency processing unit perform noise reduction based on the noise spectrum output by the noise spectrum estimator. OVERVIEW

[0006] Therefore, it is an object of the present invention to provide a noise-elimination device capable of improving the sound recognition rate even in noisy environments, a noise-elimination method, a sound recognition device, and a vehicle equipped with the sound recognition device. It is a further object of the present disclosure to provide a noise-elimination device capable of improving the sound recognition performance with a relatively small amount of computation, a noise-elimination method, a sound recognition device that uses the device, and a vehicle equipped with the sound recognition device.

[0007] The tasks are solved by a device with the features of claim 1, a sound recognition device with the features of claim 6, a vehicle with the features of claim 7 and a method with the features of claim 8. Advantageous further developments are found in the dependent claims.

[0008] Additional aspects of the present revelation are partly set forth in the description that follows, and partly become apparent from the description or can be learned through the practice of the revelation.

[0009] According to the embodiments of the present disclosure, a device for eliminating noise is provided, the device comprising: a gain detection unit which determines a gain and a correction value of the gain, wherein a signal-to-noise ratio (SNR) of an input signal is used, wherein the gain detection unit determines the correction value of the gain based on the SNR of the input signal, and the gain detection unit determines the correction value of the gain based on a set value, which is related to a relationship between the SNR of the input signal and the correction value and modifies the relationship between the SNR of the input signal and the correction value based on the set value, wherein the set value indicates a performance of a sound detection device, wherein the set value can refer to a valuewhich displays a selectable situation, and wherein a number of selectable set values ​​corresponds to a number of selectable situations; and a gain application unit which acquires an output signal corresponding to the input signal, wherein the specified gain and the specified correction value are used, wherein the output signal may include an input signal whose noise is eliminated and an input signal whose noise is not eliminated, and a sub-range of the input signal whose noise is eliminated and a sub-range of the input signal whose noise is not eliminated may be determined according to the specified correction value.

[0010] The correction value can be determined in such a way that the correction value increases when the SNR of the input signal increases, or that the correction value has a uniform value when the SNR of the input signal is less than a first value or greater than a second value.

[0011] The correction value can be determined in such a way that the proportion of the input signal whose noise is eliminated increases when the SNR of the input signal increases, while the proportion of the input signal whose noise is not eliminated increases when the SNR of the input signal decreases.

[0012] The device may further include a noise component estimator which estimates the noise of the input signal, using at least one of the following: a minimum-controlled recursive average (MCRA) algorithm, an improved minimum-controlled recursive average (IMCRA) algorithm, and a minimum statistics algorithm.

[0013] The device may also include an SNR estimation unit which estimates the SNR of the input signal using at least one of the following: a minimum mean squared error (MMSE), a root mean square (RMS) error, a cumulative minimum distance (CMD) and a speech presence probability (SPP).

[0014] Furthermore, according to embodiments of the present disclosure, a sound detection device is provided, comprising: an input unit that receives a sound signal in which an original signal and the noise are mixed; a converter unit that converts the sound signal into a signal in a frequency domain; a gain detection unit that determines a gain and a correction value of the gain, using a signal-to-noise ratio (SNR) of the sound signal, and detects a corrected gain obtained by applying the determined correction value to the determined gain, wherein the gain detection unit determines the correction value of the gain based on the SNR of the input signal, and the gain detection unit determines the correction value of the gain based on a further set value.which is related to a relationship between the SNR of the input signal and the correction value and modifies the relationship between the SNR of the input signal and the correction value based on the set value, wherein the set value indicates a capability of the sound detection device, wherein the set value can refer to a value indicating a selectable situation, and wherein a number of selectable set values ​​corresponds to a number of selectable situations; a gain application unit which detects an output signal by applying the corrected gain to the sound signal, wherein a portion of the input signal whose noise is eliminated is modified in the output signal, and a portion of the input signal whose noise is not eliminated is modified in the output signal according to the specified correction value; and a converter which converts the output signal.

[0015] Furthermore, according to embodiments of the present disclosure, a vehicle is provided which includes: an input unit which receives a sound signal from an occupant of the vehicle in which sound instructions and noise are mixed together; a sound recognition unit which recognizes sound instructions by: i) converting the received sound signal into a signal in a frequency domain, ii) determining a gain and a gain correction value, wherein a signal-to-noise ratio (SNR) of the signal in the frequency domain is used, the gain correction value being determined based on a set value which is related to a relationship between the SNR of the input signal and the correction value, and the relationship between the SNR of the input signal and the correction value is changed based on the set value, the set value indicating a performance level of the sound recognition unit.wherein the set value can refer to a value indicating a selectable situation, and wherein a number of selectable set values ​​corresponds to a number of selectable situations, iii) capturing an output signal by applying a corrected gain obtained by applying the specified correction value to the specified gain, and iv) converting the output signal, wherein a portion of the received sound signal whose noise has been eliminated is modified in the output signal, and a portion of the received sound signal whose noise has not been eliminated is modified in the output signal based on the specified correction value; and a control element which generates a control signal based on the detected sound instructions.

[0016] Furthermore, according to embodiments of the present disclosure, a method for eliminating noise is provided, which includes: determining a gain of a correction value of the gain, wherein an SNR of an input signal is used, wherein the correction value of the gain is determined based on a set value which belongs to a relationship between the SNR of the input signal and the correction value, and the relationship between the SNR of the input signal and the correction value is changed based on the set value, wherein the set value can refer to a value indicating a selectable situation, and wherein a number of selectable set values ​​corresponds to a number of selectable situations; acquiring a corrected gain which is obtained by applying the determined correction value to the determined gain;and capturing an output signal by applying the corrected gain to the input signal, wherein a portion of the input signal whose noise is eliminated in the output signal and a portion of the input signal whose noise is not eliminated in the output signal are modified based on the specified correction value.

[0017] Determining the gain correction value can involve determining the gain correction value based on a relationship between the SNR of the input signal and the correction value.

[0018] Determining the correction value of the gain may involve determining the correction value of the gain based on the use of a setting value which is related to a relationship between the SNR of the input signal and the correction value.

[0019] The correction value can be determined in such a way that the correction value increases when the SNR of the input signal increases, or the correction value has a uniform value when the SNR of the input signal is less than a first value or greater than a second value.

[0020] The correction value can be determined in such a way that the proportion of the input signal whose noise is eliminated increases when the SNR of the input signal increases, and the proportion of the input signal whose noise is not eliminated increases when the SNR of the input signal decreases.

[0021] The procedure may further include estimating the noise of the input signal, using at least one of the following: an MCRA algorithm, an IMCRA algorithm, and a minimal statistics algorithm.

[0022] The method may further include estimating the SNR of the input signal, using at least one of the following: an MMSE, an RMS error, a CMD and an SPP. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] These and / or other aspects of the disclosure will become apparent from the following description of the embodiments and will ultimately be assessed in conjunction with the accompanying drawings, of which: Fig. 1 a block diagram of a device to eliminate noise or sound according to embodiments of the present disclosure; Fig. 2 presents an example of the waveforms of the signals which contain noise; Fig. 3, Fig. 4 to Fig. 5 graphs are shown, which illustrate the relationship between a correction value and a signal-to-noise ratio (SNR); Fig. 6 a block diagram of a device to eliminate noise or sound according to embodiments of the present disclosure; Fig. 7 is a graph to explain high-resolution analysis and low-resolution analysis of frequencies; Fig. 8 a block diagram of a sound recognition device corresponding to embodiments of the present disclosure; Fig. 9 is a graph which shows a frequency conversion using a frequency conversion unit; Fig. 10 is a view of the internal structure of a vehicle; Fig. 11 a block diagram of a sound recognition device according to embodiments of the present disclosure, which is installed in the vehicle; Fig. 12 a block diagram of a sound recognition device according to embodiments of the present disclosure, which is installed in the vehicle; Fig. 13 a flowchart of a method for eliminating noise according to embodiments of the present disclosure; and Fig. 14 is a flowchart of a method for eliminating noise according to embodiments of the present disclosure. DETAILED DESCRIPTION

[0024] Reference is now made in detail to the embodiments of the present disclosure, examples of which are shown in the accompanying drawings, where similar reference numerals refer to similar elements therein.

[0025] The terminology used herein serves only to describe individual embodiments and is not intended to limit the invention. As used herein, the singular forms "a," "an," "one," and "the" are to include the plural forms as well, unless otherwise clearly indicated in the context. Furthermore, it is to be understood that the terms "includes" and / or "including," when used in this specification, specify the presence of the listed features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. As used herein, the term "and / or" includes any and all combinations of one or more related listed terms.

[0026] It is understood that the term "vehicle" or "vehicle-like" or any other similar term as used herein is inclusive of motor vehicles in general, such as passenger cars, including sports vehicles (SUVs), buses, trucks, various commercial vehicles, watercraft, including a variety of boats and ships, aircraft and the like, and including hybrid vehicles, electric vehicles, internal combustion engine vehicles, plug-in hybrid electric vehicles, hydrogen-powered vehicles, and other vehicles powered by alternative fuels (e.g., fuels derived from resources other than oil). As referenced herein, a hybrid vehicle is a vehicle that has two or more power sources, for example, both gasoline-powered and electric vehicles.

[0027] Additionally, it can be assumed that one or more of the following methods or aspects thereof can be executed by at least one control element. The term "control element" can refer to a hardware device that includes a memory and a processor. The memory is configured to store program instructions, and the processor is configured to execute the program instructions to perform one or more processes, which are described below. Furthermore, it can be assumed that the following methods can be executed by a device that includes a control unit, wherein the device is known in the field to be suitable for eliminating noise and / or incorporating a sound detection device.

[0028] Furthermore, the control element of the present invention can be embedded as non-transitory, computer-readable media on a computer-readable medium containing executable program instructions that are executed by a processor, a control element, or the like. Examples of computer-readable media include, but are not limited to, ROM, RAM, compact disc (CD-)ROMs, magnetic tapes, floppy disks, flash drives, smart cards, and optical data storage devices. The computer-readable recording medium can also be distributed across networked computer systems, so that the computer-readable media are stored and executed in a distributed manner, e.g., by a telematics server or a control element area network (CAN).

[0029] The following section distinguishes a multitude of elements from a single element in order to explain a device and a method for noise elimination, a sound detection device in which the device is used, and a vehicle equipped with the sound detection device. However, the elements to be described are distinguished for the sake of simplicity, and such a classification does not imply that the elements are physically separate. Furthermore, the elements to be described may be subdivided or combined.

[0030] The following describes a device for eliminating noise with regard to Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6 to Fig. 7 described.

[0031] Fig. Figure 1 is a block diagram of a device for eliminating noise according to the embodiments of the present disclosure, and Fig. Figure 2 shows an example of the waveforms of the signals that exhibit noise.

[0032] According to embodiments of the present disclosure, which are in Fig. As shown in Figure 1, a device 10 for eliminating noise can include a noise component estimator 11, a gain detection unit 12, and a gain application unit 19. With reference to Fig. 1 and Fig. 2. The device 10 can receive an input signal I (I = S+N) in which an original signal and a noise N are mixed from an external device, such as a microphone, and can output signals O in which the noise N is eliminated or attenuated compared to the received input signal I, using the noise component estimation unit 11, the gain detection unit 12 and the gain application unit 19.

[0033] The noise component estimator 11 of the device 10 for noise elimination can receive the input signal I, in which the original signal S and the noise N are mixed, from the external device and can estimate the noise (EN) of the input signal I in which the original signal S and the noise N are mixed. Specifically, the noise component estimator 11 can only estimate the EN from within the frequency components of the input signal I.

[0034] The noise component estimator 11 can estimate a noise component of the input signal I using various algorithms that a person skilled in the art might consider. For example, the noise component estimator 11 can determine the noise component of the input signal I using various algorithms, such as a minimum-controlled recursive averaging (MCRA) algorithm, an improved minimum-controlled recursive averaging (IMCRA) algorithm, and a minimum statistics algorithm. Additionally, the noise component estimator 11 can use various mathematical or statistical algorithms to estimate a noise component of the input signal I. In embodiments, the noise component estimator 11 can also estimate the noise component using a speech presence probability (SPP) that considers whether a frequency component is close to the sound.For example, the noise component estimation unit 11 can also estimate the noise, where the SPP is used in the MCRA algorithm.

[0035] In embodiments, the noise component estimator 11 can also divide the input signal I into a plurality of bands, and then separately estimate the noise component in each of the divided plurality of bands. In embodiments, the noise component estimator 11 can also estimate the noise component of the entire input signal I.

[0036] The EN, which is detected by the noise component estimation unit 11, can be transferred or sent to the gain detection unit 12.

[0037] The gain detection unit 12 can detect a gain G applied to the input signal I, using the EN. In embodiments, the gain detection unit 12 can separately detect the gain E in each of the divided bands of the input signal I. Furthermore, in embodiments, the gain detection unit 12 can also detect the gain G by calculating the gain G of the entire input signal I.

[0038] In embodiments which are in Fig. As shown in Figure 1, the gain acquisition unit 12 can include a signal-to-noise ratio (SNR) estimation unit 13, a gain estimation unit 15, a correction value determination unit 16 and a gain correction unit 18.

[0039] The SNR estimator 13 can receive the measured EN from the noise component estimator 11 and can estimate the SNR using the received EN. Here, the SNR estimator 13 can receive the EN and the input signal I from the noise component estimator 11 and the external device and can estimate the SNR using the received EN and the input signal I.

[0040] The SNR can be defined, for example, using the following equation 1. The SNR defined in equation 1 is described below. However, the SNR is not limited to the definition in equation 1 and can be defined differently depending on the designer. SNR=c log(S2N2)

[0041] S is an original signal with which the noise N is not synthesized, N is the noise, and SNR is a noise reduction ratio (SNR). c is a constant that can be applied according to the user's selection. Here, N can be the estimated noise (EN), which is estimated by the noise component estimator 11. When the SNR is defined in this way, if there is sufficient noise N in the original signal S, the SNR can have a relatively small value, and if there is little noise N in the original signal S, the SNR can have a relatively large value.

[0042] If the SNR is defined in this way, the original signal S, with which the noise N is not synthesized, should be acquired first. Accordingly, the SNR estimator 13 can acquire the EN, which is estimated by the noise component estimator 11, and the SNR SNR_EST, which is estimated using the following equation 2, and can substitute these for an original SNR. SNREST=c log(I2N2), where I=S+N

[0043] I is an input signal in which the original signal S described above and the noise N are mixed, and SNR_EST is an estimated SNR.

[0044] The SNR estimation unit 13 can capture the estimated SNR SNR_EST using equation 2 described above.

[0045] In embodiments, the SNR estimating unit 13 can estimate the SNR by using a minimum mean square error (MMSE) in which a mean square error (MSE) is minimized, can estimate the SNR by using an RMS error, or can estimate the SNR by using a cumulative minimum distance (CMD).

[0046] In embodiments, the SNR estimating unit 13 can detect the SPP or estimate the SNR using the detected SPP. Up to this point, the SNR estimating unit 13 can further include an SPP estimating unit 14, which calculates and estimates an SPP. The SPP estimating unit 14 can estimate and detect the SPP using various methods that a person skilled in the art might consider. Once the SPP has been estimated by the SPP estimating unit 14, the SNR estimating unit 13 can correct the estimated SNR SNR_EST based on the SPP. The SPP estimating unit 14 can be omitted, depending on the embodiment.

[0047] The estimated SNR SNR_EST, which is captured by the SNR estimation unit 13, can be transferred to the gain estimation unit 15 and the correction value determination unit 15. Likewise, the SPP, which is captured by the SPP estimation unit 14 of the SNR estimation unit 13, can be transferred to the gain estimation unit 15.

[0048] The gain estimation unit 15 can calculate and estimate a gain EG using the estimated SNR SNR_EST. In embodiments, the gain estimation unit 15 can also calculate and estimate the EG by further using the transmitted SPP as well as the estimated SNR SNR_EST.

[0049] The gain estimator 15 can also estimate the gain EG using a minimum mean square error short-time spectral amplitude (MMSE-STSA) estimator, a minimum mean square error logarithmic spectral amplitude (MMSE-LSA) estimator, or an optimal modified logarithmic spectral amplitude (OM-LSA) estimator, depending on the embodiment. Additionally, the gain estimator 15 can also estimate the gain EG using various methods that a person skilled in the art might consider.

[0050] Correction Value Determination Unit 16 can determine a correction value α for correcting the estimated gain EG. Specifically, Correction Value Determination Unit 16 can determine the correction value α using the SNR. The SNR used in the Correction Value Determination Unit can include the estimated SNR SNR_EST, which is transferred from the SNR Estimation Unit 13. Hereinafter, both the SNR and the estimated SNR SNR_EST can be referred to as a single SNR SNR_EST.

[0051] Fig. 3, Fig. 4 to Fig. Five graphs show the relationship between a correction value and an SNR. Fig. 3, Fig. 4 to Fig. In graph 5, the x-axis represents the SNR (SNR_EST), and the y-axis represents a correction value α to correct the estimated gain EG. The correction value α can be a specific value in the range of 0 to 1. Fig. 3, Fig. 4 to Fig. 5 is the correction value α, which corresponds to each of the points a1 to a6 on the y-axis, a value greater than 0 and less than 1. Fig. 3, Fig. 4 to Fig. In section 5, the correction value α does not have a value of 0. However, depending on the embodiment, the correction value α can also be 0. Similarly, the correction value α does not have a value of 1. However, depending on the embodiment, the correction value α can also be 1.

[0052] With reference to Fig. 3. If the SNR SNR_EST is less than a predetermined first SNR R1, the correction value determination unit 16 can determine a uniformly lower limit value a1 than the correction value α for correcting the estimated gain EG. In other words, the correction value α can be uniform with respect to the SNR SNR_EST, which is less than the first SNR R1.

[0053] Even if the SNR SNR_EST is greater than a previously defined second SNR R2, the correction value determination unit 16 can determine a uniform upper limit a2 as the correction value α to correct the estimated gain EG. In other words, if the SNR SNR_EST is greater than the second SNR R2, the correction value α can be uniform. If the SNR SNR_EST is greater than the second SNR R2, this can indicate that less noise N is present in the input signal I. Therefore, the correction value α can be determined as 1 or as a value close to 1.

[0054] With reference to Fig. 3. If the SNR SNR_EST is between the first SNR R1 and the second SNR R2, the correction value determination unit 16 can determine the correction value α in relation to a value of the SNR SNR_EST. In other words, the SNR SNR_EST and the correction value α can have a linear relationship I1 in the range of a first value R1 and a second value R2. Here, the correction value α can have a value in the range between the lower limit a1 and the upper limit a2.

[0055] With reference to Fig. 4. If the SNR SNR_EST is less than a third SNR R3, the correction value determination unit 16 can determine a more uniform lower limit a3 as the correction value α for correcting the estimated gain EG. If the SNR SNR_EST is greater than a predetermined SNR R4, the correction value determination unit 16 can determine a uniform upper limit a4 as the correction value α for correcting the estimated gain EG. If the SNR SNR_EST is between the third SNR R3 and the fourth SNR R4, the correction value determination unit 16 can determine the correction value α by applying the SNR SNR_EST to a predetermined exponential function I2.

[0056] Also with regard to Fig. 5, if the SNR SNR_EST is less than a fifth SNR R5, the correction value determination unit 16 can determine a uniform lower limit a5 as the correction value α, and if the SNR SNR_EST is greater than a sixth SNR R6, the correction value determination unit 16 can determine an upper limit a6 as the correction value α, and if the SNR SNR_EST is between the fifth SNR R5 and the sixth SNR R6, the correction value determination unit 16 can also determine the correction value α by applying the SNR SNR_EST to a previously defined log or logarithmic function I3.

[0057] Additionally, the correction value determination unit 16 can determine the correction value α to correct the estimated gain EG, using various relationships between the SNR SNR_EST and the correction value α.

[0058] The upper limit a1, a3, or a5 and the lower limit a2, a4, or a6 described above can be arbitrarily determined by a designer or engineer of the device 10 to eliminate noise, or by a user employing the device 10 for noise elimination. The upper limit a1, a3, or a5 and the lower limit a2, a4, or a6 can also be fixed values. Additionally, the upper limit a1, a3, or a5 and the lower limit a2, a4, or a6 can be variable values, depending on the embodiment. In other words, the engineer or the user can change the upper limit a1, a3, or a5 and the lower limit a2, a4, or a6, thereby altering the correction value α, which is determined according to the SNR SNR_EST.

[0059] In embodiments, the correction value determination unit 16 can determine the correction value α by further utilizing the SNR SNR_EST and a separately entered setting value 17. In this case, the correction value determination unit 16 can first determine the relationship between the SNR SNR_EST and the correction value α, corresponding to the set value 17, and can subsequently determine the correction value α by applying the entered SNR SNR_EST to the relationship between the SNR SNR_EST described above and the correction value α.

[0060] The set value 17 can refer to a value indicating a selectable situation. Therefore, the number of selectable set values ​​17 can correspond to the number of selectable situations. The set value 17 can be a value indicating the settings or performance of a sound detection device for which device 10 can be used to eliminate noise. For example, the set value 17 can indicate the sound detection device, which shows whether the noise from an output signal o is further eliminated by using another noise elimination device.

[0061] The correction value determination unit 16 can change the relationship between the correction value and the SNR according to the set value 17. For example, the correction value determination unit 16 can also change a function with respect to the relationship between the correction value and the SNR according to the set value 17 and can also change a lower limit a1, a3 or a5 or an upper limit a2, a4 or a6 of the correction value α according to the set value 17. In other words, the correction value determination unit 16 can capture different correction values ​​α which are suitable for several situations according to the set value 17.

[0062] In detail, for example, if the set value 17, which is displayed by a sound detection device that also uses another device to eliminate noise, is entered into the correction value determination unit 16, the correction value determination unit 16 can detect that the correction value α, after changing the lower limit a1, a3, or a5 described above, is relatively smaller, corresponding to the entered set value 17. If there is a lot of noise N in the input signal I, the SNR SNR_EST is low, and the output signal o is transmitted to the sound detection device, which further uses another device to eliminate noise, the correction value α can have a relatively small value. Thus, as will be described later, the proportion of an original, undisturbed input signal I in the output signal o is increased.As the proportion of the original, undisturbed input signal increases, so does its share within the input signal I, allowing multiple undisturbed original signals to be output. This can reduce errors in the sound detection of the sound detection device.

[0063] When the set value 17, which indicates that the sound detection device no longer uses another device for eliminating noise, is entered into the correction value determination unit 16, the correction value determination unit 16 can detect the correction value α after changing the lower limit a1, a3 or a5 described above, so that it is relatively smaller in relation to the entered setting value 17.

[0064] The set value 17 can be stored in a separate storage device, for example, a semiconductor storage device or a magnetic disk storage device. The correction value determination unit 16 can determine the ratio between the SNR SNR_EST and the correction value α by retrieving the set value 17 from the separate storage device.

[0065] The gain correction unit 18 can correct the gain EG transmitted by the gain estimation unit 15, using the correction value α determined by the correction value determination unit 16, and can output the corrected gain cG. The gain correction unit 16 can correct the gain using the following equation 3. cG=a(SNR,T)*G+(1.0−a(SNR,T))

[0066] cG is a corrected gain, SNR is SNR_EST, T is a setting value, and a(SNR, T) is a correction value α determined by SNR_EST and the setting value T. G is a gain EG estimated by the gain estimator 15. According to Equation 3, if the correction value α is 1 or a value close to 1, the corrected gain cG output by the gain correction unit 18 will be the same as or similar to the gain EG estimated by the gain estimator 15. If the correction value α is 0 or a value close to 0, the corrected gain cG output by the gain correction unit 18 can be 1 or a value close to 1.

[0067] The gain application unit 19 can detect the output signal o, using the corrected gain cG from the gain correction unit 18 and the input signal I. The gain application unit 19 can generate the output signal o to which the gain is applied, using the following equation 4. O=cG*I=[a*G+(1.0−a)]*I=a*G*I+(1.0−a)*I

[0068] o is an output signal and cG is a corrected gain. α is a corrected value, and G is an estimated gain EG. The correction value α can be determined by the SNR SNR_EST and the set value T. Here, a*G*I on one side, furthest to the right, is a portion of the input signal from which the noise N, corrected by the estimated gain EG, is eliminated, and (1.0-a)*I is a portion of the original input signal I that is undistorted.

[0069] According to Equation 4, the fraction of the input signal from which the noise N is eliminated and the fraction of the original input signal I can be determined according to the magnitude of the correction value α. If the correction value α is 1 or a value close to 1, the input signal from which the noise N is eliminated is output as the output signal o from the amplification application unit 19. If the correction value α is 0 or a value close to 0, the original, undisturbed input signal I is output as the output signal o from the amplification application unit 19.

[0070] With reference to Fig. 3, Fig. 4 to Fig. 5. The correction value α can be determined according to the SNR SNR_EST and the set value 17. Therefore, the proportion of the input signal from which the noise N is eliminated and the proportion of the original input signal I can be determined according to the SNR SNR_EST or the set value 17. More precisely, the proportion of the input signal from which the noise N is eliminated and the proportion of the original input signal I can be determined depending on whether there is a lot of noise N in the input signal I, or according to the settings or the capabilities of the sound detection device to which the device 10 is connected for noise elimination.

[0071] If there is less noise N in the input signal I and the SNR SNR_EST is large, the correction value α can be determined as a value close to the upper limit a2, a4, or a6. In this case, the correction value α can also be determined as 1 or a value close to 1. Then, as the correction value α is increased, the portion of the input signal from which the noise N is eliminated in the output signal o is relatively increased, and the portion of the original input signal i that is undisturbed is relatively decreased. If the SNR SNR_EST is large, the input signal to which the estimated gain EG is applied is a signal from which the noise N is estimated and which is barely disturbed, and the portion of the input signal from which the noise N is eliminated increases, so that the disturbance of the input signal I can be minimized and the optimized output signal o can be obtained.

[0072] If there is a lot of noise N in the input signal I and the SNR SNR_EST is small, the correction value α can be determined as a value close to the lower limit a1, a3, or a5. In this case, since the correction value α decreases, the proportion of the input signal from which the noise N is eliminated is proportionally reduced in the output signal o, and the proportion of the original input signal I that is undisturbed is proportionally increased. If the SNR SNR_EST is small, a lot of noise N of the input signal to which the estimated gain EG is applied is eliminated, so the distortion of an audio signal increases. Consequently, the proportion of the original input signal I that is undisturbed increases, so that the output signal o, in which the distortion is minimized, can be preserved.

[0073] If the correction value α is not applied to the estimated gain EG, there is a lot of noise N in the input signal I, and the SNR SNR_EST is small; only the input signal from which the noise N is eliminated is output as the output signal o, so the disturbance of the input signal may be increased.

[0074] However, as described above, if a suitable correction value α is applied according to the SNR SNR_EST or the settings or performance of the sound detection device, the disturbance of the input signal I can be minimized and the optimized output signal o can be obtained.

[0075] The noise component estimator 11, the gain acquisition unit 12, and the gain application unit 19, described above, can be implemented by separate processors that are physically isolated from one another, or by using a single processor. The processor can be programmed to perform one function of the noise component estimator 11, the gain acquisition unit 12, or the gain application unit 19. The processor can be implemented using one, two, or more semiconductors.

[0076] Fig. Figure 6 is a block diagram of a device for eliminating noise according to the embodiments of the present disclosure, and Fig. Figure 7 is a view to explain a high-resolution analysis algorithm and a low-resolution analysis algorithm of frequencies.

[0077] As in Fig. As shown in Figure 6, a noise elimination device 20 can include: a frequency band partitioning unit 21, a synthesis unit 22, a high-frequency noise processing unit 30, and a low-frequency noise processing unit 40. In detail, the noise elimination device 20, according to the embodiments of the present disclosure, can classify an input signal I according to a frequency band and can then eliminate the noise N by applying different methods in each frequency band.

[0078] The frequency band division unit 21 can divide the input signal I into a signal H, which has a high frequency component, and a signal L, which has a low frequency component. The input signal I can be divided into the signal H, which has the high frequency component, and the signal L, which has the low frequency component. The frequency band division unit 21 can divide the input signal I into the signal H, which has the high frequency component, and the signal L, which has the low frequency component, using a predefined reference value. For example, as shown in Fig. As shown in Figure 7, the predefined reference value is 4 kHz. In this case, the frequency band division unit 21 can divide a component of a frequency less than 4 kHz into the signal L, which has the low-frequency component, and a component of a frequency more than 4 kHz into the signal H, which has a high-frequency component. In this way, the predefined reference value can be arbitrarily determined according to the selection of the designer or the user.

[0079] The signal H, which has the high-frequency component, can be transmitted to a high-frequency noise processing unit 30, and the signal L, which has the low-frequency component, can be transmitted to a low-frequency noise processing unit 40.

[0080] The high-frequency noise processing unit 30 and the low-frequency noise processing unit 40 can eliminate the noise of a signal possessing a high-frequency component and the noise of a signal possessing a low-frequency component in the same way or by using different methods. For example, both the high-frequency noise processing unit 30 and the low-frequency noise processing unit 40 can eliminate the noise, using either a method performed by the high-frequency noise processing unit 30, which will be described later, or a method performed by the low-frequency noise processing unit 40, which will be described later.The following describes embodiments in which the high-frequency noise processing unit 30 and the low-frequency noise processing unit 40 eliminate noise using different methods. However, this does not mean that the high-frequency noise processing unit 30 and the low-frequency noise processing unit 40 can only eliminate noise according to the embodiments described.

[0081] The high-frequency noise processing unit 30 can eliminate the noise N of a signal H that has a high-frequency component. In embodiments, the high-frequency noise processing unit 30 can eliminate the noise N according to a low-resolution analysis algorithm. With reference to Fig. 7. The low-resolution analysis algorithm can be an algorithm which is set to divide a high-frequency component into a plurality of frequency bands c1 to c3, each of which has a relatively wide bandwidth, and to eliminate the noise N in each of the plurality of divided frequency bands c1 to c3.

[0082] The high-frequency noise processing unit 30 can include a first noise component estimation unit 31 and a noise elimination unit 32.

[0083] The first noise component estimator 31 can only estimate a noise component from the signal H, which contains the high-frequency component that is transferred by the frequency band partitioning unit 21 into each of the relatively wide frequency bands c1 to c3. The first noise component estimator 31 can estimate the noise component from the signal H, which contains the high-frequency component, using various algorithms that a person skilled in the art might consider. The first noise component estimator 31 can estimate the noise N using an initial signal that does not contain an original signal, such as a sound, i.e., the noise N, or an initial signal whose main component is the noise N. The first noise component estimator 31 can estimate and determine the initial signal as noise.In this case, the first noise component estimating unit 31 can calculate an average energy level from the initial signal for a predetermined period and can estimate the calculated average energy level as the noise N.

[0084] The noise elimination unit 32 can eliminate the noise N in each of the frequency bands c1 to c3 of the signal H, which has the high-frequency component transmitted by the frequency band division unit 21. The noise elimination unit 32 can eliminate the noise N from the input signal I by eliminating the initial signal, which is estimated to be the noise N, from the input signal I. The noise elimination unit 32 can eliminate the noise N by eliminating the estimated noise of the average energy level, calculated from the initial signal, from the signal H, which has the high-frequency component. The noise elimination unit 32 can eliminate the noise N from the signal H, which has the high-frequency component, by using various algorithms.For example, the noise elimination unit 32 can eliminate the noise N from the signal H, which has the high-frequency component, using spectral subtraction or a Wiener filter.

[0085] A signal o1, from which the noise is eliminated by the high-frequency noise processing unit 30, can be transferred to the synthesis unit 22 and can be synthesized with a signal o2, from which the noise, which is transferred by the low-frequency noise processing unit 40, is eliminated.

[0086] The low-frequency noise processing unit 40 can eliminate the noise N of the signal L, which has a low-frequency component. In embodiments, the low-frequency noise processing unit 40 can eliminate the noise N according to the high-resolution analysis algorithm. With reference to Fig. 7. The low-frequency noise processing unit 40 can divide the low-frequency component into a plurality of frequency bands c4 to c10, each of which becomes a relatively narrow bandwidth, according to the high-resolution analysis algorithm, and can then eliminate the noise N in each of the plurality of frequency bands c4 to c10. In other words, the low-frequency noise processing unit 40 can divide the frequency component into a plurality of frequency bands, which is relatively larger in number than that of the high-frequency noise processing units 30, and can eliminate the noise N in each of the plurality of divided frequency bands c4 to c10.

[0087] The low-frequency noise processing unit 40 can include a second noise component estimation unit 41, a gain acquisition unit 42 and a gain application unit 49.

[0088] The second noise component estimator 41 can only estimate a noise component within the frequency components of the signal L, which contains the low-frequency component. Here, the second noise component estimator 41 can estimate the noise component in each band. The second noise component estimator 41 can estimate the noise component of the signal L containing the low-frequency component, using various algorithms that a person skilled in the art might consider, such as an MCRA algorithm, an IMCRA algorithm, and a minimal statistics algorithm. Additionally, the second noise component estimator 41 can estimate the noise component of the signal L containing the low-frequency component, using various mathematical or statistical algorithms for estimating the noise signal.The second noise component estimation unit 41 can also estimate the noise component, using an SPP which takes into account whether the frequency component is close to the sound.

[0089] The gain detection unit 42 can detect a gain that is to be applied to the signal L, which has the low-frequency component, using the estimated noise. In embodiments which are described in Fig. As shown in Figure 1, the gain detection unit 42 can include an SNR estimation unit 43, a gain estimation unit 45, a correction value determination unit 46 and a gain correction unit 48.

[0090] The SNR estimator 43 can determine an estimated SNR using the estimated noise measured by the second noise component estimator 41. The SNR estimator 43 of the Fig. 6 can be the same as the SNR estimation unit 13, which is in Fig. 1 is shown.

[0091] In embodiments, the SNR estimating unit 43 can use an MMSE, an RMS error, or a CMD to estimate the SNR. The SNR estimating unit 43 can also detect the SPP or estimate the SNR using the detected SPP.

[0092] The gain estimation unit 45 can calculate and estimate a gain using the estimated SNR. In embodiments, the gain estimation unit 45 can also calculate and estimate the gain by further using the estimated SNR and the SPP.

[0093] Depending on the embodiment, the gain estimation unit 45 can use an MMSE-STSA estimator, an MMSE-LSA estimator, or an OD-LSA estimator to estimate the gain. Additionally, the gain estimation unit 15 can use various methods, which a person skilled in the art might consider, to estimate the gain.

[0094] The correction value determination unit 46 can determine a correction value to correct the estimated gain, using the SNR. Here, the SNR can include the estimated SNR transferred from the SNR estimation unit 43. The correction value determination unit 46 can determine the correction value using only the SNR or by using both the SNR and a set value 47.

[0095] The correction value determination unit 46 can determine the correction value, using the relationship between the correction value and the SNR, which, with reference to Fig. 3, Fig. 4 to Fig. 5 has been described. As in Fig. 3, Fig. 4 to Fig. As shown in Figure 5, if the measured SNR is less than a predetermined value R1, R3, or R5, or greater than a predetermined value R2, R4, or R6, the correction values ​​a1 to a6 can be uniform. The correction value and the SNR can have a relationship of a linear function I1, an exponential function I2, or a logarithmic function I3 in the range of the predetermined values ​​R1 and R2, R3 and R4, or R5 and R6. Additionally, the correction value determination unit 46 can determine the correction value to correct the estimated gain, using various relationships between the SNR and the correction value.

[0096] The correction value determination unit 46 can also determine the correction value by further using the setting value 47. In this case, the correction value determination unit 16 can first determine the relationship between the SNR and the correction value to be used according to the set value 47, and can then determine the correction value using the relationship between the SNR and the correction value as described above. Here, the set value 47 can be the same as the set value 17, which is determined with reference to Fig. 1 is described. In detail, the set value 47 can refer to a value indicating a selectable situation and can also include a value representing the settings or capabilities of a sound detection device to which device 10 can be applied for noise elimination. The relationship between the correction value and the SNR can be modified according to the set value 47. In this case, a function relating to the relationship between the correction value and the SNR can be modified according to the set value 47, and a lower limit a1, a3, or a5 or an upper limit a2, a4, or a6 of the ratio between the correction value and the SNR, which is in Fig. 3, Fig. 4 to Fig. The value shown as 5 can be changed according to the set value of 47.

[0097] The gain correction unit 48 can correct and output the gain transmitted by the gain estimation unit 45, using the correction value determined by the correction value determination unit 46. The gain correction unit 18 can correct the gain using equation 3 described above.

[0098] The amplification application unit 49 can acquire the signal o2, which is to be transmitted to the synthesis unit 22, using the gain corrected by the gain correction unit 48 and the signal L, which has the low-frequency component. The amplification application unit 49 can generate the signal o2, which is to be transmitted to the synthesis unit 22, to which the gain is applied, using equation 4 described above. Thus, the signal o2 output by the amplification unit 49 can be a signal with a high proportion of the signal L, which has the low-frequency component, or a signal with a high proportion of the signal whose noise is eliminated from the signal L, which has the low-frequency component, according to the correction value.The signal output by the amplification application unit 49 can be transmitted to the synthesis unit 22.

[0099] The synthesis unit 22 can synthesize the signal o1, output by the high-frequency noise processing unit 30, with the signal o2, output by the low-frequency noise processing unit 40, and can detect the output signal o. The output signal o can be a signal whose noise N has been eliminated, with different methods being used depending on whether the output signal o has a high frequency or a low frequency.

[0100] The frequency band division unit 21, the high-frequency noise processing unit 30, the low-frequency noise processing unit 40, and the synthesis unit 22 of the device 20, for eliminating noise as described above, can be implemented using separate processors that are physically isolated from one another, or by using a single processor. The processor can be programmed to perform a function of the frequency band division unit 21, the high-frequency noise processing unit 30, the low-frequency noise processing unit 40, or the synthesis unit 22. The processor can be implemented by one, two, or more semiconductors.

[0101] The following describes a sound detection device that uses a noise elimination device, with reference to Fig. 8 and Fig. 9 described.

[0102] Fig. Figure 8 is a block diagram of a sound recognition device, corresponding to embodiments of the present disclosure.

[0103] With reference to Fig. 8 a sound recognition device 50 can include a sound input unit 51, a frequency converter unit 52, a frequency band division unit 53, a noise elimination unit 54 and a converter 58.

[0104] The sound input unit 51 can receive a voice or sound, which is a wave generated when a human being speaks or an object vibrates. The sound input unit 51 can generate and output an electrical signal corresponding to the frequency of the voice or sound by vibrating at that frequency. The generated electrical signal can be an analog signal or a time-domain signal. The electrical signal output by the sound input unit 51 can be transmitted to the frequency conversion unit 52. If necessary, the electrical signal output by the sound input unit 51 can be transmitted to the frequency converter unit 52, using an amplifier or an analog-to-digital (A / D) converter.

[0105] Fig. Figure 9 is a graph showing the frequency conversion using a frequency converter unit.

[0106] As in Fig. As shown in Figure 9, the frequency converter unit 52 can convert an input signal J in the time domain into signals f1 to f3 in a frequency domain. The frequency converter unit 52 can convert the signal J in the time domain into the signals f1 to f3 using a fast Fourier transform (FFT). Depending on the embodiment, the frequency converter unit 52 can also be omitted.

[0107] The frequency band division unit 53 can divide the signals f1 to f3 into a signal H, which has a high-frequency component, and a signal L, which has a low-frequency component. It can transmit the signal H, which has the high-frequency component, to a high-frequency noise processing unit 55 of the noise elimination unit 54, and it can transmit the signal L, which has the low-frequency component, to a low-frequency noise processing unit 56 of the noise elimination unit 54. Depending on the embodiment, the frequency band division unit 53 can also be omitted.

[0108] The noise elimination unit 54 can include the high-frequency noise processing unit 55, the low-frequency noise processing unit 56, and the synthesis unit 57. The noise elimination unit 54 can be the noise elimination device 10, which is in Fig. Figure 1 shows, depending on the embodiment. In this case, the high-frequency noise processing unit 55 and the synthesis unit 57 can be omitted from the noise elimination unit 54, and the low-frequency noise processing unit 56 can process both the signal H, which has the high-frequency component, and the signal L, which has the low-frequency component.

[0109] The high-frequency noise processing unit 55 can eliminate the noise N of a signal H which has a high frequency component and can transfer a signal o1, from which the noise N has been eliminated, to the synthesis unit 57.

[0110] In the embodiments, the high-frequency noise processing unit 55 can eliminate the noise N of the signal H, which has the high-frequency component, according to the low-resolution analysis algorithm, as described in Fig. Figure 7 illustrates this. In this case, the high-frequency noise processing unit 55 can estimate a noise component from the signal H, which has the high-frequency component transmitted by the frequency band division unit 53, and can eliminate the noise estimated in each of the frequency bands c1 to c3 of the signal H, which has the high-frequency component. The high-frequency noise processing unit 55 can estimate the noise by calculating an average energy level from an initial signal and can eliminate the noise N from the signal H, which has the high-frequency component, according to the result of the estimate. The high-frequency noise processing unit 55 can use spectral subtraction or a Wiener filter to eliminate the noise N.

[0111] The low-frequency noise processing unit 56 can eliminate the noise N of the signal L, which has the low-frequency component, and can transmit a signal o2, from which the noise N has been eliminated, to the synthesis unit 56. In embodiments, the low-frequency noise processing unit 56 can eliminate the noise N of the signal L, which has the low-frequency component, according to the high-resolution analysis algorithm, as described in Fig. Figure 7 shows that the low-frequency noise processing unit 56 can eliminate the noise N using the noise component estimator 11 or 41, the gain acquisition unit 12 or 42, and the gain application unit 19 or 49, which are described in relation to Fig. 1 and Fig. 6 have been described. The noise component estimation unit 11 or 41, the gain acquisition unit 12 or 42 and the gain application unit 19 or 49, which are used in the low-frequency noise processing unit 56, may be the same as those described above or slightly modified according to requirements.

[0112] The synthesis unit 57 can synthesize the signal o1, which is output by the high-frequency noise processing unit 55, with the signal o2, which is output by the low-frequency noise processing unit 56, and can detect an output signal o.

[0113] Inverter 58 can invert the signal o output by synthesis unit 57 and generate a speech signal s. Inverter 58 can reverse the conversion of the signal o output to synthesis unit 57 using an inverse fast Fourier transform (IFFT).

[0114] The detected speech signal s can be transmitted to the output unit 59, such as a loudspeaker, can be output externally, or can be transmitted to a control element 61 of a device 60 to be controlled, such as a vehicle. The control element 61 can be configured by a separate microprocessor. The control element 61 can generate control instructions corresponding to the speech signal s, transmit these control instructions to a corresponding component in the device 60 to be controlled, and control the device 60 according to the user's sound instructions, which are detected by the sound recognition device 50.

[0115] The following describes a vehicle equipped with a sound detection device that uses a noise elimination device. A general four-wheeled motor vehicle is described below as an example of such a vehicle. This four-wheeled motor vehicle can include a small car, a van, a bus, or a truck capable of four-wheeled travel. The vehicle equipped with the sound detection device that uses the noise elimination device is not limited to a general four-wheeled motor vehicle.Examples of vehicles equipped with the sound recognition device may include a three-wheeled motor vehicle, a two-wheeled motor vehicle such as a motorcycle, a motorized bicycle, a construction machine, a bicycle, a train capable of traveling on railway tracks, or a ship capable of navigating a waterway.

[0116] Fig. Figure 10 is a view of an internal structure or layout of a vehicle.

[0117] As in Fig. As shown in Figure 10, a dashboard 200 can be arranged within a vehicle 100. The dashboard 200 refers to a panel or board that separates the interior of the vehicle 100 from the engine compartment and is located in front of a driver's seat 250 and a passenger seat 251, and in which various components required for driving are installed. The dashboard 200 can include an upper panel 201, a central instrument panel 220, and a gearbox 230. The upper panel 201 of the dashboard 200 can be located below a windshield 202, and an air vent 113a of an air conditioning unit 113 and a glove compartment or various indicators 140 can be installed on the upper panel 201.

[0118] A display device 110 for a vehicle, such as a navigation device, can also be installed on the dashboard 200. More specifically, the display device 110 for the vehicle can be installed at the top of the central instrument panel 220. The display device 110 for the vehicle can be recessed into the dashboard 200 and installed at the top of the central instrument panel 220, or it can be installed at the top of the central instrument panel 220 using a support unit configured from a predefined frame. One, two, or more input units 133 and 134 for receiving sound from a user, such as a driver or passenger, can be arranged on a housing 111 of the display device 110 for the vehicle. The input units 133 and 134 can be implemented using a microphone.

[0119] The central instrument panel 220 (e.g., the central console) of the dashboard 200 can be installed to be connected to the upper panel 201, and input units 221 and 222, such as physical buttons or knobs for controlling the vehicle, a radio unit 116, or a sound reproduction unit 115, such as a compact disc player or CD player, can be arranged on the central instrument panel 220 of the dashboard 200. The central instrument panel 220 can be located between the driver's seat 250 and the front passenger seat 251.

[0120] In embodiments, various components for the vehicle can be installed on the inside of the dashboard 200, including: a microprocessor for controlling an electronic device in various vehicles, including the display device 110 for the vehicle, which can be installed on the inside of the dashboard 200. Various components can include at least one of the following: at least one semiconductor chip, at least one switch, at least one integrated circuit (IC), at least one resistor, at least one volatile or non-volatile memory, and at least one printed circuit board (PCB) that performs a function of the microprocessor. The semiconductor chip, the switch, the IC, the resistor, and the volatile or non-volatile memory can be arranged on the PCB.

[0121] One, two, or more input units 131 for receiving sound from the driver or passenger can be arranged on the inside of an upper frame of the vehicle 100. The input unit 131 can be implemented by means of a microphone. The input unit 131 can be electrically connected to the microprocessor, which is located on the inside of the instrument panel 200 or the display unit 110 for the vehicle 100, using a cable. Alternatively, the input unit 131 can be electrically connected to the microprocessor, which is located on the inside of the instrument panel 200 or the display unit 110 for the vehicle 100, using a wireless communication network, such as Bluetooth or near-field communication, and can transmit the sound received by the input unit 131 to the microprocessor.

[0122] Sun visors 121 and 122 can be installed on the inside of the upper frame of the vehicle 100. One, two, or more input units 132 for receiving sound from the driver or front passenger can be arranged on the sun visors 121 and 122. The input unit 132 of the sun visors 121 and 122 can be implemented via the microphone. The input unit 132 of the sun visors 121 and 122 can be electrically connected to the microprocessor, which is arranged on the inside of the instrument panel 200 or the display unit 110 for the vehicle 100, either wired or wirelessly, and can transmit the sound signal received by the input unit 132 to the microprocessor. A locking device 112 for locking a door 117 of the vehicle 100 can also be arranged on the inside of the vehicle 100.

[0123] Fig. Figure 11 is a block diagram of a sound recognition device according to embodiments of the present disclosure, which is installed in the vehicle 100.

[0124] With reference to Fig. 11 The vehicle 100 may contain 100 different components and equipment 118, including microphones 131 to 134 installed inside the vehicle 100, or the navigation device 110, a frequency converter unit 140, a noise elimination unit 141, an inverter 145, a sound / text converter unit 146, a control element 147 and a storage unit 148.

[0125] Various components and equipment 118 in the vehicle 100 may include various devices which can be used within the vehicle 100 for driving or to provide comfort to the user, such as microphones 131 and 132, a navigation device 110, a locking device 112, a climate control device 113, a lighting device 114, the sound reproduction unit 115 and the radio device 116, as described in Fig. 11 are shown.

[0126] Microphones 133 and 134 can be installed with navigation system 110.

[0127] Microphones 131 to 134 can receive the sound from the driver or passenger and output an electrical signal corresponding to the received sound. The output electrical signal can be an analog signal. The output electrical signal can be transmitted to the frequency converter unit 140. The output electrical signal can be amplified by the amplifier or converted into a digital signal by the A / D converter before being transmitted to the frequency converter unit 140. The output electrical signal can include a signal in a time domain.

[0128] Microphones 131 to 134 can receive: the sound of the user, whether driver or passenger; the engine sound of the vehicle 100; and various types of noise, such as wind noise emitted from the air vent 113a of the climate control unit 113, or horn signals generated from outside the vehicle 100. Therefore, the electrical signal output by microphones 131 to 134 can also include various noise signals, along with signals relating to the sound of the user.

[0129] The microphones 131 and 132 can be arranged on the inside of the upper frame of the vehicle 100 or the sun visors 121 and 122, as shown in Fig. Figure 10 is shown. Additionally, the microphones 131 and 132 can be arranged in various positions inside the vehicle 100, such as on a steering wheel handle. The positions in which the microphones 131 and 132 are installed can be positions in which the sound of the driver or passenger is easily received.

[0130] Furthermore, microphones 133 and 134 may have been previously installed in the navigation system 110.

[0131] The frequency converter unit 140 can convert the signal in the time domain into the signal in the frequency domain, as described with reference to Fig. The frequency converter unit 140 is described in section 9. It can convert the signal in the time domain and the signal in the frequency domain using various methods that include an FFT. Depending on the embodiment, the frequency converter unit 140 can be omitted.

[0132] The noise elimination unit 141 performs a function of eliminating noise from the signal in the frequency domain where the user's sound and the noise within the vehicle are mixed. The noise elimination unit 141 may include a noise component estimator 142, a gain detection unit 143, and a gain application unit 144.

[0133] The noise component estimator 142 can detect the estimated noise transmitted by the microphones 131 to 134 or the frequency converter unit 140. The noise component estimator 142 can detect the estimated noise by estimating the noise component using various algorithms that a person skilled in the art might consider, such as an MCRA algorithm, an IMCRA algorithm, and a minimal statistics algorithm. In this case, the noise component estimator 142 can also estimate the noise component using the SPP.

[0134] The gain acquisition unit 143 can acquire an estimated SNR using the acquired estimated noise, can calculate and estimate a gain using the estimated SNR, can determine a correction value for correcting the estimated gain using the SNR, and can correct and output the estimated gain using the determined correction value.

[0135] The gain acquisition unit 143 can estimate the SNR using a method such as MMSE, RMS error, or CMD. The gain acquisition unit 143 can also acquire the SPP and estimate the SNR using the acquired SPP.

[0136] The gain sensing unit 143 can calculate the estimated gain using the estimated SNR. If necessary, the gain sensing unit 143 can also calculate the estimated gain using the SPP. The gain sensing unit 143 can estimate the gain using various methods that a person skilled in the art might consider, such as an MMSE-STSA estimator, an MMSE-LSA estimator, or an OM-LSA estimator.

[0137] The gain detection unit 143 can determine the correction value for correcting the estimated gain, using the estimated SNR. In this case, the gain detection unit 143 can detect the correction value using the relationship between the correction value and the SNR or a previously set value. Here, the relationship between the correction value and the SNR can include several embodiments with respect to the relationship between the correction value and the SNR, which are defined with reference to Fig. 3, Fig. 4 to Fig. 5 have been described. The set value can be a value indicating a selectable situation, and the selectable situation can include settings or the performance of the sound recognition device within the vehicle. The lower limit a1, a3 or a5 or the upper limit a2, a4 or a6 of the relationship between the correction value and the SNR, which is described in Fig. 3, Fig. 4 to Fig. The value shown as 5 can be changed according to a set value.

[0138] The correction value can be determined to be large when the SNR is high, i.e., when there is little noise, and conversely, it can be determined to be small when the SNR is low, i.e., when there is more noise. Furthermore, a correction value obtained when the sound detection device inside the vehicle detects the sound by reflecting the vehicle's driving noise (hereinafter referred to as a first correction value) can be relatively smaller than a correction value obtained when the sound detection device inside the vehicle detects the sound without reflecting the vehicle's driving noise, as in an external server or end device such as a smartphone (hereinafter referred to as a second correction value).Specifically, the first correction value can be determined to be the same as the second correction value if the SNR is large, and the first correction value can be determined to be smaller than the second correction value.

[0139] The gain detection unit 143 can correct the estimated gain using the specified correction value. The gain detection unit 143 can correct the gain according to equation 3 described above.

[0140] The amplification application unit 144 can detect an output signal by applying the corrected estimated gain, which is detected by the amplification detection unit 143, to a signal transmitted by the microphones 131 to 134 or the frequency converter unit 140. The amplification application unit 144 can detect the output signal according to equation 4 described above.

[0141] More specifically, the gain-detection unit 144 can increase the proportion of the signal whose noise has been eliminated when the detected correction value is closer to 1, and the gain-detection unit 144 can increase the proportion of an original signal when the detected correction value is closer to 0. Therefore, if the sound detection device inside the vehicle detects the sound by reflecting the vehicle's driving noise and the signal-to-noise ratio (SNR) of the sound signal is high, the correction value can be determined to be relatively small, and the gain-detection unit 144 can synthesize the original signal with the signal whose noise has been eliminated, thus increasing the proportion of the original signal.

[0142] The signal output by the amplification unit 144 can be transmitted to the inverter 145. The inverter 145 can invert the signal output by the noise elimination unit 141 using IFFT, thereby generating an audio signal with eliminated noise. The signal output by the inverter 145 can be transmitted to the control element 147 via the sound / text converter unit 146 or directly to the control element 147.

[0143] The sound / text converter unit 146 can convert sound into a text signal using various speech-to-text techniques and can transmit the converted text signal to the control element 147. If the control element 147 is capable of directly generating control instructions using the sound signal, the sound / text converter unit 146 can also be omitted.

[0144] The control unit 147 can generate corresponding control instructions. Using the sound or text signal converted by the sound / text converter unit 146, the generated control instructions can be transmitted to the corresponding components and devices to be controlled from various components and devices 118 in the vehicle, thereby controlling the components and devices to be controlled. For example, if the driver gives sound instructions for the lighting, the control unit 147 can generate control signals corresponding to the sound instructions and then transmit the generated control signals to the lighting device 114, switching on the lighting device 114.

[0145] Memory unit 148 can store various data required to generate control signals for the components and systems in the vehicle. If necessary, memory unit 148 can also store a history of the control signals stored by control element 147. This history of control signals can also be used to train the sound recognition device installed in the vehicle. Additionally, memory unit 168 can store various other data or necessary settings.

[0146] The frequency converter unit 140, the noise elimination unit 141, the converter 145, the sound / text converter unit 146, and the control element 147, as described above, can be implemented by a microprocessor installed in a specific location within the vehicle or navigation system 110. The microprocessor can be implemented as one, two, or more semiconductor chips. The frequency converter unit 140, the noise elimination unit 141, the converter 145, the sound / text converter unit 146, and the control element 147 can also be implemented by a single microprocessor or by a plurality of microprocessors that are physically separate from one another. The microprocessor can be programmed to perform the functions of the frequency converter unit 140, the noise elimination unit 141, the converter 145, the sound / text converter unit 146, and the control element 147.

[0147] Fig. Figure 12 is a block diagram of a sound recognition device according to embodiments of the present disclosure, which is installed in the vehicle.

[0148] With reference to Fig. 12, a vehicle can contain 100 different components and equipment 118 in the vehicle 100, including: microphones 131 to 134 which are installed in the vehicle 100, a frequency converter unit 150, a frequency band division unit 160, a noise elimination unit 161, a converter 165, a sound / text converter unit 166, a control element 167 and a storage unit 168.

[0149] Various components and equipment 118 in the vehicle 100 may include: microphones 131 and 132, a navigation system 110, a locking system 112, an air conditioning system 113, a lighting system 114, a sound reproduction system 115 and a radio system 116, which are used for driving the vehicle 100 or to provide convenience to a user, as described in Fig. 12 is shown.

[0150] Microphones 131 to 134 can receive the sound of the driver or passenger and can output an electrical signal corresponding to the received sound, as described in relation to Fig. 11. The output electrical signal can be an analog signal. The output electrical signal can be transmitted to the frequency converter unit 150. The output electrical signal can be amplified by an amplifier or converted into a digital signal by an A / D converter before being transmitted to the frequency converter unit 150. The output electrical signal can include a signal in a time domain. The microphones 131 to 134 can be installed in various positions on the vehicle 100, such as on an inner side of an upper frame of the vehicle 100, the sun visors 121 and 122, a steering wheel handle, or the navigation device 110.

[0151] The frequency converter unit 150 can convert the signal in the time domain into the signal in the frequency domain, as described with reference to Fig. 9 is described. The frequency converter unit 150 can convert the signal in the time domain into the signal in the frequency domain using various methods that include an FFT. Depending on the embodiment, the frequency converter unit 150 can be omitted. The frequency converter unit 150 can be implemented by a microprocessor, which is installed in a specific position in the vehicle 100 or in the navigation device 110.

[0152] The frequency band splitting unit 160 can split the signal transmitted by the microphones 131 to 134 or the frequency converter unit 150 into a signal with a high-frequency component and a signal with a low-frequency component, using a predefined reference value. This predefined reference value can be arbitrarily determined according to the designer's or user's selection. For example, the predefined reference value could be 4 kHz. The split signal containing the high-frequency component and the signal containing the low-frequency component can then be transmitted to the noise elimination unit 161.

[0153] The noise elimination unit 161 can include a high-frequency noise processing unit 162, a low-frequency noise processing unit 163 and a synthesis unit 164.

[0154] The signal which has the high-frequency component, which is output by the frequency band division unit 160, can be transmitted to the high-frequency noise processing unit 162, and the signal which has the low-frequency component can be transmitted to the low-frequency noise processing unit 163.

[0155] The high-frequency noise processing unit 162 can eliminate the noise of the signal that contains the high-frequency component. The high-frequency noise processing unit 162 can eliminate the noise using a low-resolution analysis algorithm. More precisely, the high-frequency noise processing unit 162 can divide the signal containing the high-frequency component into relatively large frequency bands (see c1 to c3 of the diagram). Fig. 7) splitting, only one noise component can be in each of the frequency bands (see c1 to c3 of the Fig. 7) estimate and can measure the noise in each of the frequency bands (see c1 to c3 of the Fig. 7) eliminate the signal that contains the high-frequency component. The high-frequency noise processing unit 162 can estimate the noise, using an initial signal in which no sound is spoken by the user, within the signals input via microphones 131 to 134, and can eliminate the estimated noise from the signals input via microphones 131 to 134. The initial signal can be configured to consist solely of noise, such as engine noise, or the main component of the initial signal can be noise. The high-frequency noise processing unit 162 can calculate an average energy level from the initial signal for a predetermined period and can eliminate the calculated average energy level from the signals input via microphones 131 to 134, thereby eliminating the noise.The high-frequency noise processing unit 162 can eliminate the noise from the signal that contains the high-frequency component, using an algorithm such as spectral subtraction or a Wiener filter. The signal, whose noise has been eliminated by the high-frequency noise processing unit 162, can be transferred to the synthesis unit 164.

[0156] The low-frequency noise processing unit 163 can eliminate the noise of the signal that contains the low-frequency component. In embodiments, the low-frequency noise processing unit 163 can eliminate the noise according to the high-resolution analysis algorithm. The low-frequency noise processing unit 163 can divide the high-frequency component into a plurality of frequency bands (see c4 to c10 of the Fig. 7), so that each of the multitude of frequency bands (see c4 to c10 of the Fig. 7) can be relatively narrow, using the high-resolution analysis algorithm, and can then analyze the noise in each of the frequency bands (see c4 to c10 of the Fig. 7) Eliminate.

[0157] The Low Frequency Noise Processing Unit 163 can estimate the noise component of a signal that possesses a low frequency component, using various algorithms considered by a person skilled in the art, such as an MCRA algorithm, an IMCRA algorithm, and a minimal statistical algorithm. The Low Frequency Noise Processing Unit 163 can estimate the noise component in each of the frequency bands. The Low Frequency Noise Processing Unit 163 can also estimate the noise component using the SPP described above.

[0158] The low-frequency noise processing unit 163 can acquire an estimated SNR using the estimated noise, calculate a gain using the estimated SNR, determine a correction value for correcting the estimated gain using the SNR, and correct the estimated gain using the determined correction value.

[0159] The Low Frequency Noise Processing Unit 163 can estimate the SNR using a method such as MMSE, RMS error, or CMD. The Low Frequency Noise Processing Unit 163 can also acquire the SPP and estimate the SNR using the acquired SPP.

[0160] The low-frequency noise processing unit 163 can determine an estimated gain using the estimated signal-to-noise ratio (SNR). The low-frequency noise processing unit 163 can determine the estimated gain using both the estimated SNR and the signal-to-pass ratio (SPP).

[0161] The low-frequency noise processing unit 163 can determine a correction value for correcting the estimated gain, using the relationship between the correction value and the SNR and a set value. The relationship between the correction value and the SNR can be given as shown in Fig. 3, Fig. 4 to Fig. Figure 5 illustrates this. For example, the correction value can be uniform within a predefined range of the SNR and can have a relationship between the SNR and a linear function I1, an exponential function I2, or a logarithmic function I3 within a different range. The set value can be used to determine the relationship between the SNR and the correction value to be used. The set value can include a value that incorporates the settings or the capabilities of a sound detection device to which device 10 can be applied for noise elimination.

[0162] The low-frequency noise processing unit 163 can correct and output the estimated gain described above, using the specified correction value. Subsequently, the low-frequency noise processing unit 163 can capture an output signal by applying the corrected gain to the signal containing the low-frequency component and then transmit the captured output signal to the synthesis unit 164. Correcting the estimated gain and applying the corrected gain to the signal containing the low-frequency component can be calculated according to equations 3 and 4.

[0163] The synthesis unit 164 can synthesize the signal output by the high-frequency noise processing unit 162 with the signal output by the low-frequency noise processing unit 163 in order to acquire a synthesized signal, and can transmit the synthesized signal to the converter 165.

[0164] The converter 165 can convert the signal output by the noise elimination unit 161 using an IFFT (Integrated Functional Transducer). This allows for the acquisition of an audio signal with its noise eliminated. The signal output by the converter 165 can be transmitted to the control element 167 via the sound / text converter unit 166 or directly to the control element 167 without passing through the sound / text converter unit 166.

[0165] The sound / text converter unit 166 can convert the sound signal into a text signal using various speech-to-text techniques and can transmit the converted text signal to the control element 167. If the control element 167 is able to generate control instructions directly using the sound signal, the sound / text converter unit 166 can also be omitted.

[0166] The control element 167 can generate control instructions according to the sound of the user, using the sound signal whose noise is eliminated or the text signal, and can transmit the generated control instructions to the corresponding components and devices to be controlled by the various components and devices 118 in the vehicle 100, thereby controlling the components and devices to be controlled.

[0167] The storage unit 168 can store various data required to generate control signals for different components and devices 118 in the vehicle 100, using the control element 167 or a history of the control signals generated by the control element 167. Additionally, the storage unit 168 can store various other data or settings.

[0168] The frequency converter unit 150, the frequency band division unit 160, the noise elimination unit 161, the converter 165, the sound / text converter unit 166 and the control element 167, which are described above, can be implemented by a microprocessor which is installed in a special position in the vehicle 100 or in the navigation device 110.

[0169] The microprocessor can be implemented using one, two, or more semiconductor chips. The frequency converter unit 150, the frequency band division unit 160, the noise elimination unit 161, the converter 165, the sound / text converter unit 166, and the control element 167 can be implemented using only one microprocessor or by using two or more microprocessors that are physically separated from each other.

[0170] The following describes a method for eliminating noise according to embodiments of the present disclosure with reference to Fig. 13 and Fig. 14 described.

[0171] The following describes the noise elimination method used in a sound detection device. However, the noise elimination method is not limited to the sound detection device. It can be used in various devices required for noise elimination. The noise detection device described below can also be a sound detection device used in a three-wheeled or four-wheeled motor vehicle, a two-wheeled vehicle such as a motorcycle, a motorized bicycle, a construction machine, a bicycle, a train capable of traveling on railway tracks, or a ship capable of navigating a waterway, as described above. However, the embodiments described in this disclosure are not limited to these.For example, a mobile phone, a personal digital assistive device, a smartphone, a tablet personal computer (PC), a notebook computer, a navigation device, or a portable terminal can also be an example of a sound detection device that uses the noise elimination method described later. Additionally, various types of equipment can be considered by a person skilled in the art as examples of a sound detection device that uses the noise elimination method described later.

[0172] Fig. Figure 13 is a flowchart of a method for eliminating noise according to embodiments of the present disclosure.

[0173] With reference to Fig. 13. First, a signal in which a sound and noise are mixed can be input via a microphone (S300). The input signal can be amplified by an amplifier or converted into a digital signal by an A / D converter. The input signal can be a signal in a time domain. In this case, the signal in the time domain can be converted into a signal in a frequency domain (S301). The conversion of the input signal to the frequency domain can be performed using an FFT. Depending on the embodiment, the conversion of the input signal to the signal in the frequency domain can be omitted.

[0174] Subsequently, a noise component of the input signal can be estimated (S302). If the input signal is divided into a multitude of frequency bands, the noise component in each of the multitude of divided frequency bands can be estimated separately.

[0175] If the noise component is estimated, an SNR can be acquired or estimated using the estimated noise component (S303). The SNR or an estimated SNR can be acquired in any of the many partitioned frequency bands. The SNR can be estimated using an MMSE, an RMS error, or a CMD. The SNR can also be estimated using an SPP.

[0176] Once the SNR is recorded, gain can be estimated using the SNR, and a correction value to be applied to the gain can be calculated (S304). Gain estimation can be performed using an MMSE-STSA estimator, an MMSE-LSA estimator, or an OD-LSA estimator. The correction value can be determined using the relationship between the correction value and the SNR and the set value, which is based on... Fig. 3, Fig. 4 to Fig. 5 have been described (S305).

[0177] The relationship between the correction value and the SNR can be set so that the correction value increases as the SNR increases. Alternatively, the relationship between the correction value and the SNR can be set so that the correction value remains constant when the SNR is within a predefined range.

[0178] The set value represents a selectable situation, and the relationship between the correction value and the SNR can be modified accordingly. Changing the relationship between the correction value and the SNR can be achieved by modifying a relationship function that represents this relationship, or by modifying at least an upper and lower limit of the selectable correction value. Here, the relationship function representing the relationship between the correction value and the SNR can take the form of a linear function, an exponential function, or a logarithmic function in a specific segment, as shown in Fig. 3, Fig. 4 to Fig. 5 is shown.

[0179] Once the gain and correction value are known, the gain can be corrected by applying the correction value to the gain, and an output signal can be obtained by applying the corrected gain to an input signal (S306). In embodiments, if the correction value is 1 or a value close to 1, the portion of the signal whose noise has been eliminated in the output signal can be further increased, and if the correction value is a value close to 0, the portion of the signal that was originally input and whose noise has not been eliminated in the output signal can be further increased.

[0180] The output signal can be inverted using an IFFT (S307). A signal with a sound corresponding to the output signal can be captured using an IFFT. The captured signal can be a signal with eliminated noise, a signal without eliminated noise, or a signal with some noise eliminated, depending on the correction value.

[0181] Fig. Figure 14 is a flowchart of a method for eliminating noise according to the embodiments of the present disclosure.

[0182] With reference to Fig. 14. First, a signal in which a sound and a noise or hiss are mixed can be input via a microphone (S310). The input signal, in which the sound and the noise are mixed, can be amplified by an amplifier or converted into a digital signal by an A / D converter.

[0183] The input signal can be a signal in a time domain. In this case, the signal in the time domain can be converted into a signal in a frequency domain (S311). The conversion of the input signal into the signal in the frequency domain can also be performed using an FFT. Depending on the embodiment, the conversion of the input signal into the signal in the frequency domain may be omitted.

[0184] The input signal can be split into a signal with a high-frequency component and a signal with a low-frequency component, depending on a predefined reference value (S312). Here, the predefined reference value can be 4 kHz. However, embodiments of the present disclosure are not limited to this. The reference value can be arbitrarily determined or changed according to the selection of the designer or the user.

[0185] Noise in the signal that has a high-frequency component and noise in the signal that has a low-frequency component can both be eliminated using the same method, or can be eliminated by using different methods.

[0186] If the noise of the signal which has the high-frequency component and the noise of the signal which has the low-frequency component are eliminated using different methods, the noise of the signal which has the low-frequency component (S313) can be eliminated by estimating a noise component (S314), estimating an SNR (S315), capturing an estimated gain of a correction value (S316) and correcting the gain and capturing an output signal (S317).

[0187] In embodiments, the noise of the signal containing the low-frequency component can be eliminated using a high-resolution analysis algorithm. When the high-resolution analysis algorithm is used to eliminate the noise of the signal containing the low-frequency component, the noise component of each of the frequency bands obtained by splitting the signal containing the low-frequency component can be estimated (S314).

[0188] If the noise component is estimated, an SNR can be determined in each of the split frequency bands using the estimated noise component (S315). The SNR can be estimated using an MMSE, an RMS error, or a CMD, and further using an SPP if necessary.

[0189] When the SNR is recorded, the gain can be estimated using the SNR, and a correction value to be applied to the gain can be calculated (S316). The gain estimation can be performed using an MMSE-STSA estimator, an MMSE-LSA estimator, or an OD-LSA estimator. The correction value can be determined by using the relationship between the correction value and the SNR and the set value, which is related to the SNR. Fig. 3, Fig. 4 to Fig. 5 has been described.

[0190] The relationship between the correction value and the SNR can be configured such that the correction value increases as the SNR increases. Alternatively, the relationship between the correction value and the SNR can be configured such that the correction value remains constant when the SNR is within a predefined range.

[0191] The relationship between the correction value and the SNR can be modified according to the set value. Changing the relationship between the correction value and the SNR can be done by modifying a relationship function that displays the relationship between the correction value and the SNR, or by modifying at least one of the following: an upper limit and a lower limit of a selectable correction value. The relationship function that displays the ratio between the correction value and the SNR can be in the form of a linear function, an exponential function, or a logarithmic function in a specific section, as described in Fig. 3, Fig. 4 to Fig. 5 is shown.

[0192] When the gain and correction value are detected, the gain can be corrected by applying the correction value to the gain, and an output signal can be detected by applying the corrected gain to the input signal (S317). The correction value can be set such that, as described above, the proportion of a signal whose noise is estimated can be further increased in the output signal when the correction value is 1 or a value close to 1, and the proportion of a signal that was originally input and whose noise is not eliminated in the output signal can be further increased when the correction value is a value close to 0.

[0193] The noise of a signal possessing a high-frequency component (S318) can be eliminated by estimating a noise component (S319), estimating the noise (S320), and capturing an output signal (S321). In embodiments, the noise of the signal possessing the high-frequency component can be eliminated using a low-resolution analysis algorithm.

[0194] When the low-resolution analysis algorithm is used to estimate the noise of the signal containing the high-frequency component, a noise component can be estimated in each of the frequency bands obtained by splitting the signal containing the high-frequency component (S319). In embodiments, an initial signal for a predetermined period or an average energy level calculated from the initial signal can be estimated as the noise.

[0195] Subsequently, the noise from the signal containing the high-frequency component can be eliminated by using the estimated noise component (S320). In this case, the noise can be eliminated in each of the frequency bands. The noise elimination can be performed using spectral subtraction or a Wiener filter. As a result, an output signal can be obtained that contains the high-frequency component with its noise eliminated (S321).

[0196] If the signal containing the low-frequency component, from which the noise has been eliminated, and the signal containing the high-frequency component, from which the noise has been eliminated, are acquired, the acquired signals can be synthesized together (S323). The synthesized signal can be converted using various conversion methods, including IFFT (S324). A signal that has a sound corresponding to the signal synthesized by IFFT can be acquired.

[0197] The noise elimination process described above can be implemented by implementing one, two, or more codes. These codes can be programmed by a microprocessor within the noise elimination device to execute the process. Alternatively, the codes for implementing the noise elimination process described above can be encoded and executed by a computer. These codes can be stored on a storage medium such as a compact disc, a semiconductor, or a magnetic disk.

[0198] As described above, in a device and method for eliminating noise, a sound recognition device which uses the device, and a vehicle which is equipped with the sound recognition device, according to the embodiments of the present invention, a sound which is produced by a speaking user can be precisely recognized with a relatively small amount of computation, even if there is a lot of noise, and thus the sound recognition performance can be improved.

[0199] Additionally, in a device and method for eliminating noise, in a sound detection device that utilizes the device, and in a method that is equipped with the sound detection device according to the embodiments of the present disclosure, the user's sound can be clearly detected even in the presence of significant noise, such as engine noise, so that the components within a vehicle can be controlled according to the user's intention, and thus the reliability of the sound detection device can be improved. Furthermore, user comfort can be enhanced, and safer vehicle operation can be achieved.

Claims

[1] Device (10) for eliminating noise or noise, which exhibits: a gain detection unit (12) which determines a gain and a correction value (α) of the gain, wherein a signal-to-noise ratio (SNR) of an input signal (I) is used, wherein the gain sensing unit (12) determines the correction value (α) of the gain based on the SNR of the input signal (I), and The gain detection unit (12) determines the correction value (α) of the gain, based further on a set value (17), which is related to a relationship between the SNR of the input signal (I) and the correction value (α), and modifies the relationship between the SNR of the input signal (I) and the correction value (α), based on the set value (17), wherein the set value (17) indicates a performance of a sound detection device (50), wherein the set value (17) refers to a value that indicates a selectable situation, and where a number of selectable set values ​​(17) corresponds to a number of selectable situations; and an amplification application unit (19) which detects an output signal (o) corresponding to the input signal (I), wherein the specified amplification and the specified correction value (α) are used, wherein the output signal (o) includes an input signal (I) whose noise (N) is eliminated, and an input signal (I) whose noise (N) is not eliminated, and wherein the amplification application unit (19) is configured to determine a portion of the input signal (I) whose noise (N) has been eliminated and a portion of the original input signal (I) corresponding to the determined correction value (α). [2] Device (10) according to claim 1, wherein the correction value (α) is determined in such a way that the correction value (α) increases when the SNR of the input signal (I) increases, or that the correction value (α) has a constant value when the SNR of the input signal (I) is less than a first value or greater than a second value. [3] Device (10) according to claim 1, wherein the correction value (α) is determined in such a way that the proportion of the input signal (I) whose noise (N) is eliminated increases when the SNR of the input signal (I) increases, while the proportion of the input signal (I) whose noise (N) is not eliminated increases when the SNR of the input signal (I) decreases. [4] Device (10) according to claim 1, which further comprises: a noise component estimator (11) which estimates noise (N) of the input signal (I), wherein at least one of the following is used: a minima-driven recursive average-forming (MCRA) algorithm, an improved minima-driven recursive average-forming (IMCRA) algorithm and a minimum statistics algorithm. [5] Device (10) according to claim 1, which further comprises: an SNR estimating unit (13) which estimates the SNR of the input signal (I), wherein at least one or one of the following is used: a minimum mean squared error (MMSE), an effective value (RMS) error, a cumulative minimum distance (CMD) and a speech presence probability (SPP). [6] Sound detection device (50) which includes: an input unit (131, 132, 133, 134) which receives a sound signal containing an original signal and Noise (N) are mixed; a converter unit (140) which converts the sound signal into a signal in a frequency domain; a gain detection unit (12) which determines a gain and a correction value (α) of the gain, using a signal-to-noise ratio (SNR) of the sound signal, and which detects a corrected gain obtained by applying the determined correction value (α) to the determined gain, wherein the gain sensing unit (12) determines the correction value (α) of the gain based on the SNR of the input signal (I), and The gain detection unit (12) determines the correction value (α) of the gain, based further on a set value (17), which is related to a relationship between the SNR of the input signal (I) and the correction value (α) and modifies the relationship between the SNR of the input signal (I) and the correction value (α), based on the set value (17), wherein the set value (17) indicates a performance of the sound detection device (50), wherein the set value (17) refers to a value indicating a selectable situation, and wherein a number of selectable set values ​​(17) corresponds to a number of selectable situations; an amplification application unit (19) which detects an output signal (o) by applying corrective amplification to the sound signal, wherein the amplification application unit (19) is configured to to determine a portion of the input signal (I) whose noise (N) has been eliminated and a portion of an original input signal (I) corresponding to the determined correction value (α); and an inverter (58) which inverts the output signal (o). [7] Vehicle which features: an input unit (131, 132, 133, 134) which receives a sound signal from a passenger of the vehicle in which sound instructions and noise or noise (N) are mixed together; a sound recognition unit that recognizes the sound instructions by: i) converting the received sound signal into a signal in a frequency domain, ii) determining a gain and a correction value (α) of the gain, wherein a signal-to-noise ratio (SNR) of the signal in the frequency domain is used, wherein the correction value (α) of the gain is determined based on a set value (17) which is related to a relationship between the SNR of the input signal (I) and the correction value \(α) and the relationship between the SNR of the input signal (I) and the correction value (α) is changed based on the set value (17), where the set value (17) indicates a performance of the sound recognition unit, wherein the set value (17) refers to a value indicating a selectable situation, and wherein a number of selectable set values ​​(17) corresponds to a number of selectable situations, iii) capturing an output signal (o) by applying a corrected gain obtained by applying the specified correction value (α) to the specified gain, and iv) inverting the output signal (o), wherein a portion of the input signal (I) whose noise (N) is eliminated and a portion of an original input signal (I) corresponding to the determined correction value (α) is determined; and a control element (61) which generates a control signal based on the detected sound instructions. [8] Methods for eliminating noise or noise which exhibits: Determining a gain and a correction value (α) of the gain, wherein a signal-to-noise ratio (SNR) of an input signal (I) is used, wherein the correction value (α) of the gain is determined based on a set value (17) which is related to a relationship between the SNR of the input signal (I) and the correction value (α), and the relationship between the SNR of the input signal (I) and the correction value (α) is changed based on the set value (17), wherein the set value (17) refers to a value indicating a selectable situation, and wherein a number of selectable set values ​​(17) corresponds to a number of selectable situations; Capturing a corrected gain obtained by applying the specified correction value (α) to the specified gain; and Capturing an output signal by applying the corrected gain to the input signal (I), wherein a gain application unit (19) takes into account a portion of the input signal (I) whose noise (N) is eliminated, and a proportion of the original input signal (I) corresponding to the specified correction value (α). [9] Method according to claim 8, wherein the correction value (α) is determined in such a way that the correction value (α) increases when the SNR of the input signal (I) increases, or the correction value (α) has a constant value when the SNR of the input signal (I) is less than a first value or greater than a second value. [10] Method according to claim 8, wherein the correction value (α) is determined in such a way that the proportion of the input signal (I) whose noise (N) is eliminated increases when the SNR of the input signal (I) increases, while the proportion of the input signal (I) whose noise (N) is not eliminated increases when the SNR of the input signal (I) decreases. [11] The method of claim 8, which further comprises: Estimating the noise (N) of the input signal (I) using at least one of the following: an Improved Minima-Driven Recursive Average-Calculating (MCRA) algorithm, an Improved Minima-Driven Recursive Average-Calculating (IMCRA) algorithm, and a minimum statistics algorithm. [12] The method of claim 8, which further comprises: Estimating the SNR of the input signal (I) using at least one of the following: a minimum determiner squared error (MMSE), an RMS error, a cumulative minimum distance (CMD) and a speech presence probability (SPP).

Images