Voice unlocking method for NAS device, NAS device, and system

Abstract

A voice unlocking method and apparatus for an NAS device, a system, and a storage medium, relating to the technical field of NAS. The method comprises: on the basis of acquired voice information of a user and a preset target recognition model, analyzing target feature information of the voice information, wherein the preset target recognition model comprises a preset voiceprint recognition model, and the target feature information comprises first voiceprint information (101); determining whether second voiceprint information matching the first voiceprint information is present in a preset voiceprint information set (102); and when it is determined that the second voiceprint information is present in the preset voiceprint information set, generating a target control parameter on the basis of the target feature information, wherein the target control parameter is used for controlling the NAS device to perform a target operation matching the target control parameter, and the target operation comprises an operation for unlocking the NAS device (103). Hence, the method can improve the unlocking security of the NAS device.

Description

Voice unlocking method for NAS device, NAS device and system Technical Field

[0001] The present invention relates to the field of NAS technology, and in particular to a voice unlocking method for a NAS device, a NAS device, and a system. Background Art

[0002] NAS (Network Attached Storage) is a centralized storage device. Users can store the storage resources of their mobile terminal devices on the storage medium of the NAS device, and then access the storage resources stored in the NAS device anytime and anywhere through the Internet or local area network (LAN), which greatly facilitates the user's production and life.

[0003] In actual NAS device applications, in order to ensure the security of storage resources stored by users on NAS devices, NAS devices will be locked. Existing NAS devices can only be unlocked manually by users using designated access devices. However, in practice, it is found that users manually unlock NAS devices by entering a key. However, this method of manually entering a key to unlock a NAS device has the risk of the key being forgotten, lost, or leaked. In this case, there is a risk of the storage resources in the NAS device being leaked.

[0004] It can be seen that it is particularly important to propose a technical solution to improve the security of unlocking NAS devices.

[0005] Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a voice unlocking method for a NAS device, a NAS device, and a system, which can help improve the unlocking security of the NAS device.

[0007] In order to solve the above technical problems, the first aspect of the present invention discloses a voice unlocking method for a NAS device, the method comprising:

[0008] Analyzing target feature information of the voice information based on the acquired user's voice information and a preset target recognition model, wherein the preset target recognition model includes a preset voiceprint recognition model, and the target feature information includes first voiceprint information;

[0009] and determining whether second voiceprint information matching the first voiceprint information exists in a preset voiceprint information set. When it is determined that the second voiceprint information exists in the preset voiceprint information set, generating a target control parameter based on the target feature information. The target control parameter is used to control the NAS device to perform a target operation matching the target control parameter, where the target operation includes unlocking the NAS device.

[0010] A second aspect of the present invention discloses a NAS device, comprising:

[0011] a memory storing executable program code;

[0012] a processor coupled to the memory;

[0013] The processor calls the executable program code stored in the memory to execute the voice unlocking method for the NAS device disclosed in the first aspect of the present invention.

[0014] A third aspect of the present invention discloses a computer storage medium storing computer instructions. When the computer instructions are called, they are used to execute the voice unlocking method for a NAS device disclosed in the first aspect of the present invention.

[0015] A fourth aspect of the present invention discloses a voice unlocking system for a NAS device. The voice unlocking system for a NAS device includes at least a voice unlocking device for the NAS device and a NAS device communicatively connected to the voice unlocking device for the NAS device. The voice unlocking device for the NAS device performs an unlocking operation on the NAS device according to the voice unlocking method for the NAS device disclosed in the first aspect of the present invention.

[0016] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0017] In an embodiment of the present invention, target feature information of the voice information is analyzed based on acquired user voice information and a preset target recognition model, the preset target recognition model including a preset voiceprint recognition model, and the target feature information including first voiceprint information. A determination is made as to whether second voiceprint information matching the first voiceprint information exists in the preset voiceprint information set. If the second voiceprint information exists in the preset voiceprint information set, target control parameters are generated based on the target feature information. The target control parameters are used to control the NAS device to perform a target operation matching the target control parameters, including an unlocking operation. Thus, the implementation of the embodiment of the present invention can improve the accuracy of analyzing the first voiceprint information of the voice information based on acquired user voice information and a preset voiceprint recognition model. Furthermore, if the second voiceprint information matching the first voiceprint information exists in the preset voiceprint information set, target control parameters are generated based on the first voiceprint information and / or the second voiceprint information to control the NAS device to perform a matching unlocking operation, thereby improving the accuracy of generating the target control parameters, i.e., improving the accuracy of unlocking the NAS device, and thereby improving the security of unlocking the NAS device. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] FIG1 is a flow chart of a method for voice unlocking a NAS device disclosed in an embodiment of the present invention;

[0020] FIG2 is a flow chart of another method for voice unlocking a NAS device disclosed in an embodiment of the present invention;

[0021] FIG3 is a schematic structural diagram of a NAS device disclosed in an embodiment of the present invention;

[0022] FIG4 is a schematic diagram of the structure of a voice unlocking system for a NAS device disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The present invention discloses a voice unlocking method for a NAS device, a NAS device, and a system. These methods can analyze first voiceprint information of a voice message based on acquired user voice information and a preset voiceprint recognition model, thereby improving the analysis accuracy of the first voiceprint information. Furthermore, when it is determined that second voiceprint information matching the first voiceprint information exists in a preset voiceprint information set, target control parameters are generated based on the first voiceprint information and / or the second voiceprint information to control the NAS device to execute a matching unlocking operation. This improves the accuracy of generating the target control parameters, i.e., improves the accuracy of unlocking the NAS device, thereby improving the security of unlocking the NAS device. These are described in detail below.

[0024] Example 1

[0025] Please refer to Figure 1, which is a flowchart illustrating a method for voice unlocking a NAS device disclosed in an embodiment of the present invention. The method for voice unlocking a NAS device described in Figure 1 can be applied to any NAS device or to smart devices associated with the NAS device, such as a smart device for unlocking the NAS device, including but not limited to one or more of cloud devices, edge computing devices, relay devices, base station devices, city management devices, and smart network devices. This is not limited in the present embodiment.

[0026] As shown in FIG1 , the voice unlocking method of the NAS device may include the following operations:

[0027] 101. Analyze target feature information of the voice information based on the acquired user's voice information and a preset target recognition model, where the preset target recognition model includes a preset voiceprint recognition model, and the target feature information includes first voiceprint information.

[0028] 102. Determine whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set.

[0029] In an embodiment of the present invention, as an optional implementation, the preset target recognition model further includes a preset keyword recognition model, and the target feature information further includes first keyword information. Before determining whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set, the method may further include the following operations:

[0030] A target awakening value is calculated based on the first keyword information, where the target awakening value is used to indicate a matching degree between the first keyword information and second keyword information in a pre-stored keyword set.

[0031] Determine whether the target wake-up value is greater than or equal to the preset wake-up threshold. When it is determined that the target wake-up value is greater than or equal to the preset wake-up threshold, trigger the execution of an operation of determining whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set.

[0032] In this optional embodiment, optionally, a keyword information acquisition condition can be set for the above-mentioned first keyword information. For example, after the user successfully wakes up the keyword multiple times, the average value of the voice information of the multiple wake-up keywords is taken and determined as the user's voice information, thereby further combining the preset keyword recognition model to analyze the first keyword information of the voice information, thereby further improving the analysis accuracy of the first keyword information.

[0033] It can be seen that the implementation of this optional embodiment can first analyze the first keyword information of the voice information through the preset keyword recognition model before determining whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set, thereby calculating the target wake-up value. When it is determined that the target wake-up value is greater than or equal to the preset wake-up threshold, it further triggers the execution of the operation of determining whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set. This can help further improve the feasibility and analysis accuracy of the target feature information meeting the preset wake-up conditions, as well as generate target control parameters based on the first voiceprint information and / or the second voiceprint information, improve the generation accuracy of the target control parameters, improve the accuracy of unlocking the NAS device, and thus improve the security of unlocking the NAS device.

[0034] In this optional embodiment, as another optional implementation manner, before determining that the target characteristic information meets the preset wake-up condition, the method may further include the following operations:

[0035] Analyze the user's target management authority based on the first voiceprint information and / or the second voiceprint information.

[0036] Analyze the user's target execution authority based on the first keyword information and a preset semantic analysis model.

[0037] It is determined whether the target management authority and the target execution authority match each other. When it is determined that the target management authority and the target execution authority match each other, an operation of determining whether the target characteristic information meets the preset wake-up condition is triggered.

[0038] In this optional embodiment, the above-mentioned target management authority can be temporarily modified, that is, temporarily authorized.

[0039] It can be seen that the implementation of this optional embodiment can analyze the user's target management authority based on the first voiceprint information and / or the second voiceprint information, improve the diversity of analysis methods of the user's target management authority, analyze the user's target execution authority based on the first keyword information and the preset semantic analysis model, improve the analysis accuracy of the user's target execution authority, and further improve the feasibility and analysis accuracy of the target feature information meeting the preset wake-up conditions by judging whether the target management authority and the target execution authority match, and generate target control parameters based on the first voiceprint information and / or the second voiceprint information, improve the generation accuracy of the target control parameters, and improve the accuracy of unlocking the NAS device, thereby improving the security of unlocking the NAS device and preventing users from accessing the NAS device across permissions.

[0040] 103. When it is determined that the second voiceprint information exists in the preset voiceprint information set, a target control parameter is generated according to the target feature information. The target control parameter is used to control the NAS device to perform a target operation matching the target control parameter. The target operation includes unlocking the NAS device.

[0041] In an embodiment of the present invention, optionally, the above-mentioned target operation may further include a feedback operation, that is, feeding back to the user the result of the judgment of whether the above-mentioned target characteristic information meets the preset wake-up condition, and the reason for generating the judgment result.

[0042] It can be seen that the implementation of the embodiment of the present invention can analyze the target feature information of the acquired user voice information through the preset target recognition model, thereby improving the analysis accuracy of the target feature information, and when it is determined that the target feature information meets the preset wake-up condition, generate target control parameters based on the target feature information to control the target device to perform a matching unlocking NAS device operation, thereby improving the generation accuracy of the target control parameters, that is, improving the accuracy of unlocking the NAS device, and further improving the unlocking security of the NAS device.

[0043] Example 2

[0044] Please refer to Figure 2, which is a flowchart illustrating a method for voice unlocking a NAS device disclosed in an embodiment of the present invention. The method for voice unlocking a NAS device described in Figure 2 can be applied to any NAS device or to smart devices associated with the NAS device, such as a smart device for unlocking the NAS device, including but not limited to one or more of cloud devices, edge computing devices, relay devices, base station devices, city management devices, and smart network devices. This is not limited in the present embodiment.

[0045] As shown in FIG2 , the voice unlocking method of the NAS device may include the following operations:

[0046] 201. Obtain environmental image information and environmental audio information.

[0047] 202. Determine user image information of the user in the environmental image information based on the determined user identity information and environmental image information.

[0048] 203. Analyze the user's voice information based on the user's image information and the ambient audio information.

[0049] In an embodiment of the present invention, as an optional implementation manner, the above-mentioned analysis of the user's voice information based on the user's image information and the ambient audio information may include the following operations:

[0050] At least one target detection point of the user is determined based on the user image information.

[0051] Analyze the user's activity intention information based on the dynamic activity trajectories of all target detection points.

[0052] At least one target detection audio information is determined in the ambient audio information according to the activity intention information, and the audio intention information of the target detection audio information matches the activity intention information.

[0053] For each target detection audio information, the target detection audio information is input into a preset filtering model to obtain target human voice audio information, and the filtering threshold of the preset filtering model matches the target detection audio information.

[0054] Determine the user's voice information based on all target human voice audio information.

[0055] It can be seen that the implementation of this optional embodiment can determine at least one target detection point of the user based on the user image information, and analyze the user's activity intention information based on the dynamic activity trajectories of all target detection points, thereby improving the diversity and flexibility of the analysis methods of the activity intention information on the basis of ensuring the accuracy of the analysis of the activity intention information. Furthermore, based on the activity intention information, at least one target detection audio information whose audio intention information matches the activity intention information is determined in the ambient audio information, so as to obtain the target life audio information in combination with the preset filtering model, and then determine the user's voice information, which can improve the accuracy of determining the user's voice information, thereby improving the analysis accuracy of the target feature information and the unlocking accuracy and security of the NAS device.

[0056] 204. Analyze target feature information of the voice information based on the acquired user voice information and a preset target recognition model, where the preset target recognition model includes a preset voiceprint recognition model, and the target feature information includes first voiceprint information.

[0057] 205. Determine whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set.

[0058] 206. When it is determined that the second voiceprint information exists in the preset voiceprint information set, a target control parameter is generated according to the target feature information. The target control parameter is used to control the NAS device to perform a target operation matching the target control parameter. The target operation includes unlocking the NAS device.

[0059] In the embodiment of the present invention, for the supplementary description of steps 204 to 206, please refer to the specific description of steps 101 to 103 in the first embodiment, which will not be repeated in the embodiment of the present invention.

[0060] It can be seen that the implementation of the embodiment of the present invention can determine the user image information of the user in the environmental image information based on the acquired environmental image information and environmental audio information, and further combine the environmental audio information to analyze the user's voice information with sound and emotion, thereby improving the diversity and flexibility of the user's voice information acquisition methods, and facilitating further accurate acquisition of the user's voice information in noisy environments, thereby improving the analysis accuracy of target feature information, facilitating the unlocking accuracy and security of NAS devices, and facilitating the user's NAS device unlocking experience, freeing the user's brain from having to memorize complex unlocking keys, and reducing the risk of key forgetting, loss, and leakage.

[0061] In an optional embodiment, the above-mentioned preset target recognition model includes a preset voiceprint recognition model, the above-mentioned target feature information includes the first voiceprint information, and the above-mentioned preset keyword recognition model includes a global inverse spectrum mean and variance normalization layer, a linear layer, a backbone network layer and multiple target classifiers; the global inverse spectrum mean and variance normalization layer is used to perform normalization processing on the input voice information so that the voice information is normally distributed; the linear layer is used to adjust the information format of the voice information to the target format corresponding to the preset keyword recognition model; the backbone network layer includes at least one of an RNN model, an LSTM model, a Bi-LSTM model, a TCN model, a DS-TCN model, and an MDTC model; each of the multiple target classifiers includes at least one S-shaped activation function, and the S-shaped activation function is used to predict the posterior probability of at least one keyword.

[0062] Furthermore, the preset keyword recognition model further includes a first loss function, which is:

[0063] in, m is used to represent the minimum duration frame of the keyword in the i-th voice information, N is used to represent the number of minimum duration frames in the i-th voice information, and p ij Used to express the posterior probability of predicting the jth minimum duration frame of the i-th speech information, y iIt is used to represent the true value corresponding to the posterior probability of the i-th speech information, L is used to represent the degree of inconsistency between the posterior probability and the true value, and the first loss function is used to preset the training stage of the keyword recognition model.

[0064] It can be seen that the implementation of this optional embodiment provides a model architecture of a preset voiceprint recognition model and the role of each model component, which can improve the recognition of voice information by the preset voiceprint recognition model, and thus obtain the analysis accuracy and feasibility of the voiceprint information, which is conducive to improving the unlocking accuracy and security of NAS devices.

[0065] In another optional embodiment, the above-mentioned preset target recognition model also includes a preset keyword recognition model, the above-mentioned target feature information also includes first keyword information, and the above-mentioned preset voiceprint recognition model includes at least one frame-level layer, a pooling layer, at least one segment-level transformation layer and a second loss function; the frame-level layer is used to perform at least one target processing operation on the input voice information and convert the voice information into frame-level information, and the target processing operation includes at least one of pre-emphasis, frame windowing, Fourier transform, Mel filter group and logarithmic operation, and discrete Fourier inverse transform; the pooling layer is used to perform aggregation processing on the frame-level information to obtain segment-level information; the segment-level transformation layer is used to analyze the voiceprint label corresponding to the segment-level information; the second loss function is used to analyze the degree of inconsistency between the voiceprint label and the true value corresponding to the voice information, and the second loss function is used in the training stage of the preset voiceprint recognition model.

[0066] It can be seen that the implementation of this optional embodiment provides a model architecture of a preset keyword recognition model and the role of each model component, which can improve the recognition of voice information by the preset keyword recognition model, and thus obtain the analysis accuracy and feasibility of keyword information, which is conducive to improving the unlocking accuracy and security of NAS devices.

[0067] Example 3

[0068] Please refer to Figure 3, which is a schematic diagram of the structure of a NAS device disclosed in an embodiment of the present invention. As shown in Figure 3, the NAS device may include:

[0069] The memory 401 stores executable program codes.

[0070] A processor 402 is coupled to the memory 401 .

[0071] The processor 402 calls the executable program code stored in the memory 401 to execute the steps of the voice unlocking method for the NAS device described in the first embodiment or the second embodiment of the present invention.

[0072] Example 4

[0073] An embodiment of the present invention discloses a computer storage medium storing computer instructions. When the computer instructions are called, they are used to execute the steps of the voice unlocking method for a NAS device described in the first embodiment or the second embodiment of the present invention.

[0074] Example 5

[0075] Please refer to Figure 4, which shows a voice unlocking system for a NAS device disclosed in an embodiment of the present invention. As shown in Figure 4, the voice unlocking system for the NAS device includes at least a voice unlocking device for the NAS device and a NAS device that is communicatively connected to the voice unlocking device of the NAS device. The voice unlocking device of the NAS device can be built into the NAS device, or independently communicate with the NAS device via a wired, wireless, Bluetooth / network, or other means. The voice unlocking device of the NAS device performs an unlocking operation on the NAS device according to the voice unlocking method for the NAS device described in Embodiment 1 or Embodiment 2 of the present invention.

[0076] Through the detailed description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus the necessary general hardware platform, or of course, by means of hardware. Based on this understanding, the above technical solution, in essence, or the portion that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

Claims

1. A voice unlocking method for a NAS device, characterized in that, The method comprises: Analyzing target feature information of the voice information according to the acquired voice information of the user and a preset target recognition model, wherein the preset target recognition model includes a preset voiceprint recognition model, and the target feature information includes first voiceprint information; Determine whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set; when it is determined that the second voiceprint information exists in the preset voiceprint information set, generate a target control parameter according to the target feature information, the target control parameter being used to control the NAS device to perform a target operation matching the target control parameter, the target operation including an operation of unlocking the NAS device.

2. The voice unlocking method of the NAS device according to claim 1, characterized in that, The preset target recognition model further includes a preset keyword recognition model, the target feature information further includes first keyword information, and before determining whether there is second voiceprint information matching the first voiceprint information in the preset voiceprint information set, the method further includes: Calculating a target awakening value according to the first keyword information, where the target awakening value is used to indicate a matching degree between the first keyword information and second keyword information in a pre-stored keyword set; Determine whether the target wake-up value is greater than or equal to a preset wake-up threshold. When it is determined that the target wake-up value is greater than or equal to the preset wake-up threshold, trigger the execution of the operation of determining whether there is a second voiceprint information matching the first voiceprint information in the preset voiceprint information set.

3. The voice unlocking method of the NAS device according to claim 2, wherein, Before generating the target control parameter according to the target feature information, the method further includes: analyzing the target management authority of the user according to the first voiceprint information and / or the second voiceprint information; Analyzing the target execution authority of the user according to the first keyword information and a preset semantic analysis model; Determine whether the target management authority matches the target execution authority, and when it is determined that the target management authority matches the target execution authority, trigger the execution of the The operation of collecting information and generating target control parameters.

4. The voice unlocking method of the NAS device according to claim 1, characterized in that, Before analyzing the target feature information of the voice information according to the acquired user's voice information and the preset target recognition model, the method further includes: Acquire environmental image information and environmental audio information; Determining user image information of the user in the environmental image information according to the determined identity information of the user and the environmental image information; Analyzing the user's voice information according to the user image information and the environmental audio information; And, analyzing the user's voice information according to the user image information and the environmental audio information includes: Determining at least one target detection point of the user according to the user image information; Analyzing the activity intention information of the user according to the dynamic activity trajectories of all the target detection points; According to the activity intention information, determining at least one target detection audio information in the environmental audio information, wherein the audio intention information of the target detection audio information matches the activity intention information; For each of the target detection audio information, input the target detection audio information into a preset filtering model to obtain target human voice audio information, where the filtering threshold of the preset filtering model matches the target detection audio information; Determine the voice information of the user according to all the target human voice audio information.

5. The voice unlocking method of the NAS device according to claim 2, characterized in that, The preset keyword recognition model includes a global cepstrum mean and variance normalization layer, a linear layer, a backbone network layer, and multiple target classifiers; the global cepstrum mean and variance normalization layer is used to perform normalization processing on the input voice information so that the voice information is normally distributed; the linear layer is used to adjust the information format of the voice information to the target format corresponding to the preset keyword recognition model; the backbone network layer includes at least one of an RNN model, an LSTM model, a Bi-LSTM model, a TCN model, a DS-TCN model, and an MDTC model; each of the multiple target classifiers includes at least one sigmoid activation function, and the sigmoid activation function is used to predict the posterior probability of at least one keyword; In addition, the preset keyword recognition model further includes a first loss function, and the first loss function is as follows: Among them, m is used to represent the minimum duration frame of the keyword in the i-th speech information, N is used to represent the number of the minimum duration frames in the i-th speech information, p ij is used to represent the posterior probability of predicting the j-th minimum duration frame of the i-th speech information, y i is used to represent the true value corresponding to the posterior probability of the i-th speech information, L is used to represent the degree of inconsistency between the posterior probability and the true value, and the first loss function is used in the training stage of the preset keyword recognition model.

6. The voice unlocking method of the NAS device according to claim 1, characterized in that, The preset voiceprint recognition model includes at least one frame-level layer, a pooling layer, at least one segment-level transformation layer, and a second loss function; the frame-level layer is used to perform at least one target processing operation on the input voice information and convert the voice information into frame-level information, and the target processing operation includes at least one of pre-emphasis, framing and windowing, Fourier transform, Mel filter bank and logarithm operation, and inverse discrete Fourier transform; the pooling layer is used to perform aggregation processing on the frame-level information to obtain segment-level information; the segment-level transformation layer is used to analyze the voiceprint label corresponding to the segment-level information; the second loss function is used to analyze the degree of inconsistency between the voiceprint label and the true value corresponding to the voice information, and the second loss function is used in the training stage of the preset voiceprint recognition model.

7. A NAS device, characterized in that, The device includes: A memory storing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory and executes the voice unlocking method of the NAS device according to any one of claims 1-6.

8. A computer storage medium, characterized in that, The computer storage medium stores computer instructions, which are used to execute the voice unlocking method of the NAS device according to any one of claims 1-6 when the computer instructions are called.

9. A voice unlocking system for a NAS device, characterized in that, The voice unlocking system of the NAS device at least includes a voice unlocking device and a NAS device communicatively connected to the voice unlocking device, and the voice unlocking device performs an unlocking operation on the NAS device according to the voice unlocking method of the NAS device according to any one of claims 1-6.

Images