CN103092330B - Electronic device and voice recognition method thereof - Google Patents

Abstract

A speech recognition method is applied to an electronic device. The method includes the following steps: obtaining a sound signal; converting the sound signal into a digital signal; obtaining multiple values from the digital signal, wherein each value represents a corresponding sound wave amplitude value; sequentially comparing the size of each value relative to at least one adjacent value to obtain a set of comparison result values; determining whether the set of comparison result values corresponds to at least one of the operation instructions; if so, executing the corresponding operation instruction. Through the technical solution of the present invention, after the sound information emitted by the user is converted and compared and analyzed by the signal of the present invention, it can be accurately determined whether the sound information corresponds to the existing operation instruction, so as to ensure that the correct operation instruction can be executed, and it is more user-friendly in use.

Description

Electronic device and voice recognition method thereof

technical field

The present invention relates to a voice recognition method, in particular to a voice recognition method applied to an electronic device for performing voice control; in addition, the present invention also includes an electronic device applying the method.

Background technique

With the increasing sophistication of speech recognition technology, the scope of its application is also becoming wider and wider. Taking electronic devices as an example, when combined with voice recognition technology, it can provide users with the function of controlling electronic devices through voice commands; this not only simplifies the operation of electronic devices, but also eliminates the need for users to go through various methods in the past. The inconvenience caused by the remote controller inputting commands or directly manually touching the corresponding function keys of the electronic device also improves the user's convenience in use.

Most of the voice recognition technologies commonly used at present are to directly convert the voice signal into an electrical signal, and digitize the waveform of the voice signal through an analog-to-digital converter, so as to facilitate voice recognition and judgment. However, because the sound waveforms produced by different decoders are not the same, how to effectively analyze the existing waveforms to reduce misjudgment and improve recognition during sound recognition is a subject worth studying.

Contents of the invention

The main purpose of the present invention is to provide a speech recognition method that applies the conversion of sound to numerical value to judge the speech command, so as to solve the problem of high misjudgment in the speech recognition technology adopted in the prior art.

In order to achieve the above-mentioned purpose, the present invention adopts the following technical solutions:

A voice recognition method applied to an electronic device, the method includes the following steps:

obtain an audio signal;

converting the sound signal into a digital signal;

Obtaining a plurality of numerical values from the digital signal, wherein each numerical value represents a corresponding sound wave amplitude value;

Sequentially comparing the magnitude of each of the multiple values relative to at least one adjacent value to obtain a set of comparison results;

judging whether the group of comparison result values corresponds to at least one of the operation instructions; and

When it is determined that the group of comparison result values corresponds to one of the at least one operation instruction, the corresponding operation instruction is executed.

Preferably, the sound signal is transformed into the digital signal by Fourier transformation.

Preferably, the multiple values are obtained by filtering out signal values lower than a standard value in the digital signal.

Preferably, the method further includes the following step: detecting whether a start signal is received, so as to judge whether to start to acquire the sound signal.

Preferably, in the step of judging whether the group of comparison result values corresponds to at least one of the operation instructions, by comparing at least one group of reference values corresponding to the at least one operation instruction with the group of comparison result values The coincidence or numerical difference is within a certain range to determine whether the set of comparison result values corresponds to one of the at least one operation instruction.

Preferably, each of the numerical values is compared with a single adjacent numerical value on the same side, and the corresponding group of comparison result values is obtained according to the magnitude of each of the numerical values relative to the single adjacent numerical value.

Preferably, each of the numerical values is compared with the two adjacent numerical values located on both sides of the numerical value at the same time, and the corresponding group of comparison result values is obtained according to the magnitude of each of the numerical values relative to the two adjacent numerical values .

Preferably, the following steps are further included: when it is judged that the group of comparison result values does not correspond to one of the at least one operation instruction, detecting whether a setting signal is received; if so, setting the sound signal to correspond to a new operation instruction .

The present invention also provides an electronic device, comprising:

a sound receiving module for receiving a sound signal; and

A processing module is electrically connected to the sound receiving module, and the processing module includes:

a conversion unit, used to convert the sound signal into a digital signal, and obtain a plurality of values from the digital signal, wherein each value represents a corresponding sound wave amplitude value; and

A comparison coding unit, used to sequentially compare the magnitude of each of the plurality of values relative to at least one adjacent value, so as to obtain a set of comparison result values;

The processing module judges whether the group of comparison result values corresponds to one of at least one operation instruction, so as to determine whether to execute the corresponding operation instruction.

Preferably, the processing module converts the sound signal into the digital signal through Fourier transform.

Preferably, the conversion unit filters out signal values lower than a standard value in the digital signal to obtain the plurality of values.

Preferably, an activation module is also included, configured to generate an activation signal to the processing module to notify the audio receiving module to start acquiring the audio signal.

Preferably, it also includes a storage module for storing at least one set of reference values corresponding to the at least one operation instruction, wherein when the processing module compares one of the at least one set of reference values with the set of comparison results When the values match or the numerical difference is within a certain range, it is determined that the group of comparison result values corresponds to one of the at least one operation instruction.

Preferably, each of the numerical values is compared with a single adjacent numerical value on the same side, and the corresponding group of comparison result values is obtained according to the magnitude of each of the numerical values relative to the single adjacent numerical value.

Preferably, each of the numerical values is compared with the two adjacent numerical values located on both sides of the numerical value at the same time, and the corresponding group of comparison result values is obtained according to the magnitude of each of the numerical values relative to the two adjacent numerical values .

Further, when the processing module determines that the group of comparison result values does not correspond to one of the at least one operation instruction, the processing module detects whether a setting signal is received, so as to set the sound signal to correspond to a new operation instruction.

With this design, after the signal conversion and comparison analysis of the present invention, the voice information sent by the user can accurately judge whether the voice information corresponds to the existing operation command, so as to ensure that the correct operation command can be executed, and it can be used when using more humane.

Description of drawings

1 is a schematic diagram of an electronic device of the present invention;

Fig. 2 (a) is the waveform diagram of the sound signal received by the electronic device of the present invention;

Fig. 2 (b) is the waveform diagram formed after the sound signal received by the electronic device of the present invention is converted;

Fig. 2(c) is a schematic diagram of comparing the comparison result value with the reference value by the electronic device of the present invention;

Fig. 3 is the flowchart of the first embodiment of the speech recognition method of the present invention;

Fig. 4 is the flowchart of the second embodiment of the speech recognition method of the present invention;

Detailed ways

In order to enable the examiners to better understand the technical content of the present invention, preferred embodiments are given and described as follows.

Please refer to FIG. 1 , which is a schematic diagram of an electronic device 1 of the present invention. In an embodiment of the present invention, the electronic device 1 is a remote controller for remote control of household appliances, and depending on the type of application, the electronic device 1 can also be a household appliance (such as a TV or an audio-visual player, etc.), a Portable electronic products (such as computers or mobile phones, etc.) or other electronic devices that can be operated by voice control, but not limited to this embodiment.

As shown in FIG. 1 , the electronic device 1 of the present invention includes a sound receiving module 10 , a processing module 20 and a storage module 30 . The sound receiving module 10 is used to receive the sound signal from the user. In one embodiment of the present invention, the sound receiving module 10 is a sound converter or a microphone, but other devices with similar sound receiving functions can also be used.

The processing module 20 is electrically connected to the sound receiving module 10 , and the processing module 20 includes a conversion unit 22 and a comparison and encoding unit 24 . The conversion unit 22 is used for converting the received audio signal into a digital signal, wherein the conversion unit 22 converts the audio signal into a corresponding digital signal by means of Fourier transform, such as a digital-to-analog converter.

Please refer to Figure 1 and Figure 2 (a), (b) and (c) together. Fig. 2 (a) is the waveform diagram of the sound signal received by the electronic device 1 of the present invention, Fig. 2 (b) is the waveform diagram formed after the conversion of the sound signal received by the electronic device 1 of the present invention, Fig. 2 ( C is the schematic diagram that the electronic device 1 of the present invention compares the comparison result value with the reference value.As shown in Figure 2 (a), the sound receiving module 10 of the electronic device 1 of the present invention receives the user's sent After the sound signal, the time-amplitude waveform diagram of the sound wave can be obtained as shown in the figure. Since the sound signal is too messy and difficult to identify, it needs to undergo signal conversion in order to improve the identification effect. As shown in Figure 2 (b), after The frequency-amplitude waveform diagram of the sound wave generated after Fourier transform will cause the digital signal to present multiple amplitude peaks with different heights. The technology of the present invention is aimed at the numerical processing of these amplitude peaks to achieve the effect of speech recognition.

As shown in FIG. 2( b ), the conversion unit 22 obtains a plurality of values from the converted digital signal, and in the process of obtaining these values, in order to avoid a large number of values with lower peak amplitudes from affecting the accuracy of judgment, Therefore, the conversion unit 22 first takes a set amplitude value as a standard value (as shown by the dotted line in the figure), and filters the multiple peaks obtained in the digital signal, and deletes the peaks lower than the aforementioned standard value to obtain Avoid noise interference, and finally obtain the remaining multiple values; where each value represents the corresponding sound wave amplitude value. In an embodiment of the present invention, the conversion unit 22 obtains M values from the converted digital signal, and M is an integer greater than 1.

The comparing and encoding unit 24 is used for sequentially comparing the magnitude of each numerical value relative to at least one adjacent numerical value for each numerical value among the plurality of numerical values, so as to obtain a set of comparison result values. The design of sequentially comparing each numerical value with respect to at least one adjacent numerical value mentioned here refers to comparing each numerical value among the multiple numerical values with its adjacent left and right numerical values in sequence, for example, comparing each numerical value according to The sequence is compared with a single adjacent value on the same side (left or right), or each value is compared with the adjacent values on the left and right sides at the same time; each value after the aforementioned comparison One comparison result can be generated, and after multiple values are compared, a set of comparison result values composed of multiple comparison results is obtained. In one embodiment of the present invention, it is assumed that the comparison coding unit 24 performs a pairwise comparison of the Nth value among the M values relative to the (N-1)th value adjacent to its left side, so as to Obtain a set of comparison result values, where N is an integer and 2≤N≤M. As shown in Figure 2(b), for example, if the conversion unit 22 obtains 5 values (at this time M is equal to 5, and N is any number between 2 and 5), the comparison coding unit 24 will sequentially compare For the size of the second value to the first value, then compare the size of the third value to the second value...and so on, and finally compare the size of the fifth value to the fourth value, and then Get a set of comparison result values.

The comparison results corresponding to the aforementioned numerical values may be defined to correspond to different comparison result values. When the comparison encoding unit 24 judges that the N numerical value is greater than the (N-1) numerical value, the N numerical value is recorded as a first comparison result value; when the N numerical value is judged to be equal to the (N-1) For numerical values, record the Nth numerical value as a second comparison result value; when it is judged that the Nth numerical value is smaller than the (N-1)th numerical value, record the Nth numerical value as a third comparison result value. Continuing the previous example, assume that the first comparison result is defined as 1, the second comparison result is 2, and the third comparison result is 0; in the case of obtaining 5 values, when the second value is greater than the first value, record the comparison result value of the second value as 1; when the third value is greater than the second value, record the comparison result value of the third value as 1; when the fourth value is equal to the third value, record the comparison result value of the 4th value as 0; finally, when the 5th value is smaller than the 4th value, record the comparison result value of the 5th value as 2. Therefore, the comparing and encoding unit 24 will obtain a set of comparison results consisting of the aforementioned four comparison results as “1120” after performing pairwise comparison of values. The obtained comparison result value can be used as a judgment source for subsequent instruction comparison.

In another embodiment of the present invention, if each numerical value is compared with adjacent numerical values on the left and right sides at the same time, it is assumed that the comparison coding unit 24 compares the Nth numerical value among the M numerical values with respect to its left The (N-1)th values adjacent to the side and the (N+1)th values adjacent to the right are sequentially compared to obtain a set of comparison results, where N is an integer and 2≤N ≤M-1. For example, if the conversion unit 22 obtains 5 values as shown in FIG. Compare the size of the 2nd value with respect to the 1st value and the 3rd value, then compare the size of the 3rd value with respect to the 2nd value and the 4th value, finally compare the 4th value with respect to the 3rd value and the size of the fifth value to obtain a set of comparison result values.

Among them, according to the comparison results of the aforementioned numerical values, the comparison result values corresponding to nine different results can be respectively defined, which can be roughly divided into three values that are equal (one type), any two values that are equal and matched with another value. (6 types), the three values are not equal (2 types), etc., and numbers 1 to 9 can also be used to represent the comparison result values corresponding to each type. Therefore, if judged according to the 5 numerical values obtained in Figure 2(b), assuming that the three numerical values are not equal and the second numerical value is greater than the first numerical value but less than the third numerical value, the comparison result value is 8. The comparison result value of 3 values equal to the 4th value and greater than the 2nd value is 2, and the comparison result value of the 4th value equal to the 3rd value and greater than the 5th value is 5. Therefore, after the comparing and encoding unit 24 sequentially compares each value with the adjacent values on both sides, it will obtain a set of comparison results consisting of the aforementioned three comparison results as "825", and so on. The values of the aforementioned comparison results may vary according to different identification designs, and the present invention is not limited thereto.

It should be noted that the processing module 20 of the electronic device 1 of the present invention can be a processor or its combination with software, but the processing module 20 can also be pure hardware, software, hardware or a combination of at least two or more of the foregoing. , the present invention is not limited thereto.

The storage module 30 is used for storing at least one set of reference values corresponding to the aforementioned set of comparison result values and at least one operation instruction. The user can pre-set and record at least one voice command to correspond to at least one operation command, for example, a pre-recorded voice command of “power off” or “power on” corresponds to the power off or power on operation of the electronic device 1 of the present invention. These voice commands can also be converted, compared and analyzed through the above-mentioned processing method for voice signals, so as to obtain at least one set of reference values corresponding to at least one operation command. For example, after the signal conversion and comparison, the operation command of "power off" corresponds to a set of reference values "2101", while the operation command of "power on" corresponds to another set of reference values "1121". These reference values are also stored in the storage module 30 for subsequent comparison.

Thereby, the processing module 20 judges whether it corresponds to any one of the aforementioned reference values according to the obtained group of comparison result values, so as to confirm whether the group of comparison result values corresponds to any operation instruction, and then executes the corresponding operation instruction . When the processing module 20 compares any set of reference values with the set of comparison result values, or when the numerical difference is within a certain range, it is determined that the set of comparison result values corresponds to at least one of the operation instructions, and then Execute the corresponding operation command. As shown in Figure 2(c), continuing the previous example, suppose that when the processing module 20 judges that any set of reference values is completely the same as the set of comparison result values or the difference between the two values is within 1 value, when the acquired After the group comparison result value "1120" is compared with the group reference value "2101" representing power off and the group reference value "1121" representing power on, it can be known that the group comparison result value "1120" and The set of reference values "1121" representing power-on are relatively similar with only one value difference, so the processing module 20 can judge that the set of comparison result values corresponds to the operation command of power-on, so as to execute the electronic device 1 according to the voice information. start-up operation. It should be noted that the above-mentioned difference in comparison value can also be changed according to the identification design requirements, and the present invention is not limited thereto. In an embodiment of the present invention, the storage module 20 may be a memory or other storage devices.

In addition, when the processing module 20 judges that the group of comparison result values does not correspond to any operation instruction according to the above-mentioned judgment mechanism, it means that the sound signal issued by the user does not match any operation instruction. At this time, the processing module 20 can pass Other devices or modules send out prompt information, such as preset sounds or flashes, to inform users. At this point, the user can input a setting signal through a specific button on the electronic device 1, and when the processing module 20 detects that the setting signal has been received, the received sound signal can be assigned a new corresponding setting according to the setting signal. , so that it corresponds to a new operation instruction. In this way, when the voice recognition effect is not satisfactory and the voice of the user cannot be distinguished, the user can re-record and set the currently issued voice information and the corresponding operation instructions, thereby improving the convenience of use.

The electronic device 1 of the present invention further includes an activation module 40 for generating an activation signal to the processing module 20 to notify the sound receiving module 10 to start acquiring the audio signal from the user. In an embodiment of the present invention, the activation module 40 may be an input button or replaced by other devices with similar functions, which is not limited to this embodiment.

Please refer to FIG. 3 , which is a flow chart of the first embodiment of the speech recognition method of the present invention. It should be noted that, although the electronic device 1 shown in FIG. 1 is used as an example below to illustrate the voice recognition method of the present invention, the present invention is not limited to be applicable to the electronic device 1, and any other electronic device with a similar structure can also be used. The speech recognition method of the present invention is applicable. As shown in FIG. 3 , the speech recognition method of the present invention includes steps S301 to S306. Each step of the method will be described in detail below.

Step S301: Obtain an audio signal. The electronic device 1 obtains the sound signal from the user through the sound receiving module 10 for subsequent processing.

Step S302: converting the sound signal into a digital signal. The conversion unit 22 of the processing module 20 performs Fourier transformation on the sound signal to obtain a corresponding digital signal.

Step S303: Obtain a plurality of values from the digital signal. The conversion unit 22 first filters out the amplitude peak value lower than a standard value from the converted digital signal, so as to reduce the misjudgment effect similar to noise caused by it, and then obtains a plurality of values from the remaining amplitude peak value, wherein each value represents The corresponding sound wave amplitude value.

Step S304: Sequentially compare the magnitude of each numerical value relative to at least one adjacent numerical value to obtain a set of comparison result values. The comparison encoding unit 24 of the processing module 20 compares each numerical value with at least one adjacent numerical value according to the multiple numerical values obtained above, and assigns different comparison result values according to the comparison results, thereby obtaining A set of comparison result values.

Step S305: Determine whether the group of comparison result values corresponds to one of at least one operation instruction. The processing module 20 compares the obtained set of comparison result values with at least one set of reference values corresponding to at least one operation instruction, and judges whether any set of reference values is consistent with the set of comparison result values or whether the numerical difference is within within a certain range.

Step S306: Execute the corresponding operation instruction. When the processing module 20 judges whether any set of reference values is consistent with the set of comparison result values or whether the numerical difference is within a certain range, it means that the sound signal corresponding to the set of comparison result values represents the value corresponding to the set of reference values. The operation instruction, so the processing module 20 can execute the corresponding operation instruction.

With this design, the voice information sent by the user is processed through signal conversion, filtering, comparison analysis, etc. to judge whether the voice information is consistent with the existing operation instructions, so as to determine whether to execute the corresponding operation instructions.

Please refer to FIG. 4 , which is a flowchart of a second embodiment of the speech recognition method of the present invention. As shown in FIG. 4 , the steps S401 to S406 are the same as the aforementioned steps S301 to S306 , and this embodiment further includes steps S407 , S408 to S409 .

Step S407 is also included before step S401: detecting whether a start signal is received. When the user intends to use the electronic device 1 to execute the voice recognition method of the present invention, the processing module 20 first detects whether the activation signal input by the user through the activation module 40 is received, so as to determine whether to activate the related function of voice capture. When the activation signal is detected, the processing module 20 notifies the sound receiving module 10 to start sound capture.

After the step S405, a step S408 is further included: detecting whether a setting signal is received. When it is determined through step S405 that the group of comparison result values does not correspond to any operation command, the processing module 20 detects whether a setting signal input by the user is received to determine whether the sound corresponding to the group of comparison result values The signal redefines the definition of the operation instruction.

After the step S408, a step S409 is further included: setting the sound signal to correspond to a new operation command. When the processing module 20 detects the setting signal input by the user, it redefines the sound information according to the setting signal to correspond to the new operation instruction.

Through the design of the present invention, in addition to being applied to remote controls or similar devices, users can issue corresponding operating instructions through the remote control through voice input to operate on the corresponding devices of the remote control, and the present invention can also directly It is applied to the main body of the electronic device to be operated, or indirectly controls the peripheral devices connected to the electronic device through the main body of the electronic device, and its implementation status changes according to different design requirements, so that any device can be controlled by voice.

In summary, the present invention is quite different from the prior art in terms of purpose, means and effects. However, it should be noted that the above-mentioned embodiments are only illustrative to illustrate the principles and effects of the present invention, and are not intended to limit the scope of the present invention. Any person familiar with the art can modify and change the embodiments without violating the technical principle and spirit of the present invention.

Claims (16)

1. A voice recognition method applied to an electronic device, characterized in that the method comprises the following steps: obtain an audio signal; converting the sound signal into a digital signal; Obtaining a plurality of numerical values from the digital signal, wherein each numerical value represents a corresponding sound wave amplitude value; Sequentially comparing the magnitude of each of the multiple values relative to at least one adjacent value to obtain a set of comparison results; judging whether the group of comparison result values corresponds to at least one of the operation instructions; and When it is determined that the group of comparison result values corresponds to one of the at least one operation instruction, the corresponding operation instruction is executed. 2. The method according to claim 1, wherein the audio signal is converted into the digital signal by Fourier transform. 3. The method of claim 1, wherein the plurality of values are obtained by filtering out signal values in the digital signal that are lower than a standard value. 4. The method of claim 1, further comprising the steps of: Detecting whether a starting signal is received to determine whether to start acquiring the sound signal. 5. The method according to claim 1, wherein in the step of judging whether the group of comparison result values corresponds to at least one of the operation instructions, by comparing at least one of the at least one operation instructions Whether the group reference value is consistent with the group comparison result value or the numerical difference is within a certain range, so as to determine whether the group comparison result value corresponds to one of the at least one operation instruction. 6. The method according to claim 1, wherein each of the numerical values is compared with a single adjacent numerical value on the same side, and according to the magnitude state of each of the numerical values relative to the single adjacent numerical value, to obtain The corresponding comparison result value of this group. 7. The method according to claim 1, characterized in that, comparing each of the numerical values with the two adjacent numerical values on both sides of the numerical value at the same time, according to the magnitude of each of the numerical values relative to the two adjacent numerical values , to obtain the corresponding comparison result value of this group. 8. The method of claim 1, further comprising the steps of: When it is judged that the group of comparison result values does not correspond to one of the at least one operation command, detecting whether a setting signal is received; If yes, set the sound signal to correspond to a new operation instruction. 9. An electronic device, characterized in that it comprises: a sound receiving module for receiving a sound signal; and A processing module is electrically connected to the sound receiving module, and the processing module includes: a conversion unit, used to convert the sound signal into a digital signal, and obtain a plurality of values from the digital signal, wherein each value represents a corresponding sound wave amplitude value; and A comparison coding unit, used to sequentially compare the magnitude of each of the plurality of values relative to at least one adjacent value, so as to obtain a set of comparison result values; The processing module judges whether the group of comparison result values corresponds to one of at least one operation instruction, so as to determine whether to execute the corresponding operation instruction. 10. The electronic device as claimed in claim 9, wherein the processing module converts the sound signal into the digital signal through Fourier transform. 11. The electronic device according to claim 9, wherein the conversion unit filters out signal values lower than a standard value in the digital signal to obtain the plurality of values. 12. The electronic device as claimed in claim 9, further comprising an activation module, configured to generate an activation signal to the processing module to notify the audio receiving module to start acquiring the audio signal. 13. The electronic device according to claim 9, further comprising a storage module for storing at least one set of reference values corresponding to the at least one operation instruction, wherein when the processing module compares the at least one set of reference values When one of the reference values matches the set of comparison result values or the numerical difference is within a certain range, it is determined that the set of comparison result values corresponds to one of the at least one operation instruction. 14. The electronic device according to claim 9, wherein each of the numerical values is compared with a single adjacent numerical value on the same side, and according to the magnitude of each of the numerical values relative to the single adjacent numerical value, To obtain the corresponding comparison result value of the group. 15. The electronic device according to claim 9, wherein each of the numerical values is compared with the two adjacent numerical values on both sides of the numerical value at the same time, and according to the magnitude of each of the numerical values relative to the two adjacent numerical values state to obtain the corresponding comparison result value of the group. 16. The electronic device according to claim 9, wherein when the processing module determines that the set of comparison result values does not correspond to one of the at least one operation instruction, the processing module detects whether a setting signal is received, The sound signal is set to correspond to a new operation command.