Hearing aid with sensitive word reminding function

Abstract

The invention relates to a hearing aid, in particular to a hearing aid with a sensitive word reminding function, for a multi-microphone hearing aid, the hearing aid comprises a signal acquisition and beam forming module, a multi-beam sensitive word detection module, a direction reminding module and a user personalized setting module; the signal acquisition and beam forming module is used for generating audio beams in a plurality of directions by utilizing a beam forming technology according to original audio signals acquired by a plurality of microphones of the hearing aid at the same time; the multi-beam sensitive word detection module is used for simultaneously detecting whether the audio beams in multiple directions contain sensitive words or not, and when the audio beams contain the sensitive words, recording a source direction and generating a detection signal at the same time; the direction reminding module generates a second reminding signal according to the source direction of the sensitive word and the detection signal, reminds the user to pay attention to the sensitive word and enables the user to perceive the source direction of the sensitive word; according to the technical scheme provided by the invention, the defect that a user cannot be helped to accurately detect sensitive information in a complex environment in the prior art can be effectively overcome.

Description

Technical Field

[0001] The present invention relates to a hearing aid, and more particularly to a hearing aid with a sensitive word reminder function. Background Art

[0002] Hearing aids are the main devices for solving hearing loss problems. They collect, process, and amplify external sound signals to help users improve their auditory experience. However, the development of traditional hearing aid technologies has mainly focused on aspects such as sound amplification, noise reduction, and frequency compensation, with relatively limited intelligence levels and difficulty in meeting the personalized usage needs of users.

[0003] Currently, the hearing aids on the market mainly process sound signals in the environment, but there are still some deficiencies:

[0004] 1) The problem of speech loss in noisy environments

[0005] In noisy environments, in order to improve the auditory comfort of users, hearing aids usually adopt noise reduction technologies. However, such technologies may also weaken speech signals, resulting in users missing important information, such as someone calling their name.

[0006] 2) The limitations of directional functions

[0007] Some hearing aids have directional functions that can amplify sounds from a specific direction (such as the front) and suppress sounds from other directions. However, this function may lead to the missed detection of sensitive information in an environment with multi-directional sound sources, such as someone calling the user's name from behind.

[0008] 3) The lack of targeted reminder functions

[0009] The main functions of existing hearing aids focus on enhancing sound and improving speech clarity, but they lack the ability of active recognition and reminder. Users may not be able to pay attention to important information related to themselves in a timely manner.

[0010] In summary, adding a sensitive word detection and reminder function to hearing aids can make up for the deficiencies of existing hearing aids. For example, users can preset sensitive words to be concerned about (such as their own names). The hearing aid detects in real time and sends a reminder to the user when a sensitive word is recognized, thereby enhancing the user's information perception ability in complex environments. This improvement not only solves the problem of missed detection of sensitive information but also significantly improves the intelligence level of hearing aids, bringing a better usage experience to users. Summary of the Invention

[0011] (I) Technical problems to be solved

[0012] In view of the above-mentioned disadvantages of the prior art, the present invention provides a hearing aid with a sensitive word reminder function, which can effectively overcome the defect that the prior art cannot help users accurately detect sensitive information in complex environments.

[0013] (II) Technical Solution

[0014] To achieve the above objectives, the present invention is realized through the following technical solutions:

[0015] A hearing aid with a sensitive word reminder function. For a single-microphone hearing aid, it includes a signal acquisition module, a sensitive word detection module, a sensitive word reminder module, and a personalized setting module;

[0016] The signal acquisition module directly transmits the original audio signal collected by the single microphone of the hearing aid to the sensitive word detection module;

[0017] The sensitive word detection module performs real-time processing on the original audio signal to detect whether it contains sensitive words. When the original audio signal contains sensitive words, it generates a detection signal and sends it to the sensitive word reminder module;

[0018] The sensitive word reminder module generates a first reminder signal according to the detection signal to remind the user to pay attention to the sensitive word;

[0019] The personalized setting module allows the user to preset a list of sensitive words and select a reminder method;

[0020] For a multi-microphone hearing aid, it includes a signal acquisition and beamforming module, a multi-beam sensitive word detection module, a direction reminder module, and a user personalized setting module;

[0021] The signal acquisition and beamforming module generates audio beams in multiple directions using beamforming technology based on the original audio signals simultaneously collected by the multiple microphones of the hearing aid, and transmits them to the multi-beam sensitive word detection module;

[0022] The multi-beam sensitive word detection module simultaneously detects whether the audio beams in multiple directions contain sensitive words. When the audio beam contains sensitive words, it records the source direction and generates a detection signal at the same time, and sends it to the direction reminder module;

[0023] The direction reminder module generates a second reminder signal according to the sensitive word source direction and the detection signal to remind the user to pay attention to the sensitive word and enable the user to perceive the sensitive word source direction;

[0024] The user personalized setting module allows the user to preset a list of sensitive words, select a reminder method, and at the same time allows the user to adjust the priority of the beam direction to optimize the sensitive word detection effect in a specific scenario.

[0025] Preferably, the sensitive word detection module detects whether the original audio signal contains sensitive words, and the multi-beam sensitive word detection module simultaneously detects whether the audio beams in multiple directions contain sensitive words, including:

[0026] S1. Perform frame segmentation and splicing on the audio signal to obtain multiple sub-audio signals;

[0027] S2. Use the sensitive word detection model to detect sensitive words in multiple sub-audio signals;

[0028] S3. Implement interval sampling through the output of the sensitive word detection model, and input the sampling result into the secondary confirmation model to obtain the final sensitive word detection result.

[0029] Preferably, in S1, performing frame segmentation and splicing on the audio signal to obtain multiple sub-audio signals includes:

[0030] Perform frame segmentation on the audio signal according to a frame duration of 10 ms and a frame shift of 10 ms;

[0031] Obtain multiple consecutive sub-audio signals with a frame length of 21 by splicing the first 10 frames and the last 10 frames.

[0032] Preferably, in S2, using the sensitive word detection model to detect sensitive words in multiple sub-audio signals includes:

[0033] Adopt the Viterbi algorithm to identify possible wake-up paths from the sub-audio signals, calculate the scores of each wake-up path, and select the optimal path from multiple wake-up paths through dynamic programming technology to detect sensitive words in the sub-audio signals.

[0034] Preferably, in S3, implementing interval sampling through the output of the sensitive word detection model and inputting the sampling result into the secondary confirmation model to obtain the final sensitive word detection result includes:

[0035] Perform interval sampling through the output of the sensitive word detection model according to a frame interval of 20, and input the sampling result into the secondary confirmation model to obtain the final sensitive word detection result.

[0036] Preferably, the sensitive word reminder module generates a sound effect reminder or a voice reminder according to the detection signal sent by the sensitive word detection module to remind the user to pay attention to sensitive words.

[0037] Preferably, the direction reminder module generates a sound effect reminder or a voice reminder according to the sensitive word source direction and the detection signal sent by the multi-beam sensitive word detection module to remind the user to pay attention to sensitive words, and simulates the sense of direction through different channels to enable the user to perceive the sensitive word source direction.

[0038] (III) Beneficial effects

[0039] Compared with the prior art, a hearing aid with a sensitive word reminder function provided by the present invention has the following beneficial effects:

[0040] 1) Improve the user's ability to capture sensitive information

[0041] By setting a sensitive word detection system in the hearing aid, the present invention can capture and remind the user in real time regardless of the volume and source direction of the sensitive word, effectively making up for the deficiency of traditional hearing aids in information capture;

[0042] 2) Improve the detection accuracy in a noisy environment

[0043] The present invention adopts beamforming technology and multi-beam sensitive word detection technology, enabling the hearing aid to process audio signals from different directions respectively, ensuring accurate detection of sensitive words in a complex sound field;

[0044] 3) Enhance the user's perception ability of sensitive information

[0045] The present invention helps the user quickly perceive sensitive words through various prompting methods (such as sound effect prompts, voice prompts), and can be personalized according to the user's needs, greatly improving the user experience;

[0046] 4) Have high versatility and flexibility

[0047] The sensitive word detection system in the present invention can meet the requirements of various scenarios through simple settings, and can be applied to both single-microphone and multi-microphone hearing aids, and supports users to customize the sensitive word list and prompting methods, with high versatility and flexibility;

[0048] 5) Improve the intelligent level of the hearing aid

[0049] The present invention realizes the transformation of the hearing aid from "passive listening" to "active reminder" through the sensitive word detection and reminder function, enabling it to have a certain intelligent interaction ability, providing a new direction for the function expansion of the hearing aid;

[0050] 6) Enhance user safety and life convenience

[0051] The present invention helps the user pay attention to key voice information (such as their own name, emergency reminder, etc.) in a timely manner through accurate sensitive word detection and instant reminder function, significantly improving safety and life convenience. Description of the Drawings

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0053] Figure 1 It is a schematic diagram of the hardware of the hearing aid in the present invention;

[0054] Figure 2 It is a schematic flowchart of the sensitive word detection by the sensitive word detection module and the multi-beam sensitive word detection module in the present invention. Detailed implementation manners

[0055] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0056] A hearing aid with a sensitive word reminder function, as Figure 1 shown, for a single-microphone hearing aid, it includes a signal acquisition module, a sensitive word detection module, a sensitive word reminder module, and a personalized setting module;

[0057] The signal acquisition module directly transmits the original audio signal collected by the single microphone of the hearing aid to the sensitive word detection module;

[0058] The sensitive word detection module performs real-time processing on the original audio signal to detect whether it contains sensitive words. When the original audio signal contains sensitive words, it generates a detection signal and sends it to the sensitive word reminder module;

[0059] The sensitive word reminder module generates a first reminder signal according to the detection signal to remind the user to pay attention to the sensitive word;

[0060] The personalized setting module allows the user to preset a list of sensitive words and select a reminder method.

[0061] The sensitive word reminder module generates a sound effect reminder (similar to the mobile phone notification sound, emitting a preset reminder sound when a sensitive word is detected) or a voice reminder (prompting the content of the sensitive word to the user through pre-recorded or synthesized voice, such as "someone called your name") according to the detection signal sent by the sensitive word detection module to remind the user to pay attention to the sensitive word.

[0062] In the technical solution of this application, the sensitive word detection system operates independently of the hearing aid system. Since the hearing aid system usually includes algorithms related to hearing aid fitting such as WDRC (Wide Dynamic Range Compression) and frequency band compensation, it may interfere with the detection of sensitive words. Therefore, for a single-microphone hearing aid, the sensitive word detection system directly performs sensitive word detection on the original audio signal collected by the single microphone of the hearing aid, which can ensure the generality of sensitive word detection and avoid the impact of the hearing aid function on sensitive word detection.

[0063] A hearing aid with a sensitive word reminder function, as Figure 1 shown, for a multi-microphone hearing aid, it includes a signal acquisition and beamforming module, a multi-beam sensitive word detection module, a direction reminder module, and a user personalized setting module;

[0064] The signal acquisition and beamforming module generates audio beams in multiple directions using beamforming technology based on the original audio signals simultaneously collected by the multi-microphones of the hearing aid, and transmits them to the multi-beam sensitive word detection module;

[0065] The multi-beam sensitive word detection module simultaneously detects whether the audio beams in multiple directions contain sensitive words. When an audio beam contains a sensitive word, it records the source direction and generates a detection signal at the same time, and sends it to the direction reminder module;

[0066] The direction reminder module generates a second reminder signal based on the sensitive word source direction and the detection signal to remind the user to pay attention to the sensitive word and enable the user to perceive the sensitive word source direction;

[0067] The user personalized setting module allows the user to preset a list of sensitive words, select a reminder method, and at the same time allows the user to adjust the priority of the beam direction to optimize the sensitive word detection effect in a specific scenario.

[0068] The direction reminder module generates a sound effect reminder or a voice reminder based on the sensitive word source direction and the detection signal sent by the multi-beam sensitive word detection module to remind the user to pay attention to the sensitive word, and simulates the sense of direction through different channels to enable the user to perceive the sensitive word source direction.

[0069] In the technical solution of this application, the array advantages of multiple microphones are fully utilized. Audio beams are formed in the front, rear, and other directions of the user using beamforming technology, or multiple audio beams in multiple directions are generated simultaneously. Each beam signal is processed by an independent sensitive word detection system. Even in a noisy environment, sensitive words can be effectively detected. In addition, this design can accurately locate the source direction of the sensitive word and provide more reliable reminder information for the user.

[0070] The sensitive word detection module detects whether the original audio signal contains sensitive words, and the multi-beam sensitive word detection module simultaneously detects whether the audio beams in multiple directions contain sensitive words, such as Figure 2 shown, including:

[0071] S1. Perform frame segmentation and splicing processing on the audio signal to obtain multiple sub-audio signals;

[0072] S2. Use the sensitive word detection model to detect sensitive words in multiple sub-audio signals;

[0073] S3. Implement interval sampling through the output of the sensitive word detection model, and input the sampling result into the secondary confirmation model (since the sensitive word detection model using the sensitive word detection method based on the Viterbi algorithm may have the risk of false wake-up, so when detecting the sensitive word candidate path, a secondary confirmation strategy needs to be adopted to further improve the sensitive word detection accuracy), and obtain the final sensitive word detection result.

[0074] 1) In S1, perform frame segmentation and splicing processing on the audio signal to obtain multiple sub-audio signals, such as Figure 2 shown, including:

[0075] Perform frame segmentation on the audio signal according to a frame duration of 10 ms and a frame shift of 10 ms;

[0076] Obtain multiple consecutive sub-audio signals with a frame length of 21 by splicing the first 10 frames and the last 10 frames.

[0077] 2) In S2, use the sensitive word detection model to detect sensitive words in multiple sub-audio signals, including:

[0078] Adopt the Viterbi algorithm to identify possible wake-up paths from the sub-audio signals, calculate the score of each wake-up path, and select the optimal path from multiple wake-up paths through dynamic programming technology to detect sensitive words in the sub-audio signals.

[0079] 3) In S3, implement interval sampling through the output of the sensitive word detection model, and input the sampling result into the secondary confirmation model to obtain the final sensitive word detection result, such as Figure 2 shown, including:

[0080] Perform interval sampling according to a frame interval of 20 through the output of the sensitive word detection model, input the sampling result into the secondary confirmation model, and obtain the final sensitive word detection result.

[0081] In the above technical solution, by directly using the output of the sensitive word detection model (instead of the original audio signal) as the input of the secondary confirmation model, it is possible to avoid recalculation and waste of processing power, and directly utilize the effective information extracted by the primary model (sensitive word detection model), thereby improving the sensitive word detection efficiency.

[0082] Meanwhile, interval sampling is achieved through the output of the sensitive word detection model, and the sampling result is input into the secondary confirmation model to obtain the final sensitive word detection result. This sampling method reduces the amount of data that the secondary confirmation model needs to process, while ensuring sufficient feature information, reducing the computational complexity, and improving the overall response speed. Compared with the traditional method, the amount of computation is significantly reduced, which is especially suitable for low-power devices, enabling the sensitive word detection system to operate stably in low-power devices (such as hearing aids). The optimized secondary confirmation model can detect sensitive words in real time while avoiding excessive power consumption, improving the performance of the overall system, and is especially suitable for scenarios with extremely high power consumption requirements.

[0083] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the present invention in each embodiment.

Claims

1. A hearing aid with a sensitive word reminder function, characterized in that: For a single-microphone hearing aid, it includes a signal acquisition module, a sensitive word detection module, a sensitive word reminder module, and a personalized setting module; The signal acquisition module directly transmits the original audio signal collected by the single microphone of the hearing aid to the sensitive word detection module; The sensitive word detection module processes the original audio signal in real time to detect whether it contains sensitive words. When the original audio signal contains sensitive words, a detection signal is generated and sent to the sensitive word reminder module; A sensitive word reminder module generates a first reminder signal according to the detection signal to remind the user to pay attention to the sensitive words; Personalization setting module, allowing users to preset sensitive word lists and select prompt methods; For multi-microphone hearing aids, it includes a signal acquisition and beamforming module, a multi-beam sensitive word detection module, a direction reminder module, and a user personalized setting module; The signal acquisition and beamforming module uses beamforming technology to generate audio beams in multiple directions based on the original audio signals collected simultaneously by the multiple microphones of the hearing aid, and transmits them to the multi-beam sensitive word detection module; The multi-beam sensitive word detection module simultaneously detects whether the audio beams in multiple directions contain sensitive words. When the audio beam contains sensitive words, it records the source direction and generates a detection signal, which is sent to the direction reminder module. A direction reminder module generates a second reminder signal according to the direction of the sensitive word source and the detection signal, reminding the user to pay attention to the sensitive word and allowing the user to perceive the direction of the sensitive word source; The user personalization setting module allows users to preset a list of sensitive words and select a prompt method. It also allows users to adjust the priority of the beam direction to optimize the sensitive word detection effect in specific scenarios.

2. The hearing aid with sensitive word reminder function according to claim 1, characterized in that: The sensitive word detection module detects whether the original audio signal contains sensitive words, and the multi-beam sensitive word detection module simultaneously detects whether the audio beams in multiple directions contain sensitive words, including: S1, performing frame division and splicing processing on the audio signal to obtain multiple sub-audio signals; S2. Detect sensitive words on multiple sub-audio signals using a sensitive word detection model; S3. Interval sampling is implemented through the output of the sensitive word detection model, and the sampling results are input into the secondary confirmation model to obtain the final sensitive word detection results.

3. The hearing aid with sensitive word reminder function according to claim 2, characterized in that: In S1, the audio signal is framed and spliced ​​to obtain multiple sub-audio signals, including: The audio signal is processed by dividing the frame into frames according to the frame length of 10ms and the frame shift of 10ms; By splicing the previous and next 10 frames, multiple continuous sub-audio signals with a frame length of 21 are obtained.

4. The hearing aid with sensitive word reminder function according to claim 3, characterized in that: In S2, a sensitive word detection model is used to detect sensitive words on multiple sub-audio signals, including: The Viterbi algorithm is used to identify possible wake-up paths from the sub-audio signal, and the score of each wake-up path is calculated. The optimal path is selected from multiple wake-up paths through dynamic programming technology to perform sensitive word detection on the sub-audio signal.

5. The hearing aid with sensitive word reminder function according to claim 4, characterized in that: In S3, interval sampling is implemented through the output of the sensitive word detection model, and the sampling results are input into the secondary confirmation model to obtain the final sensitive word detection results, including: With a frame interval of 20, interval sampling is achieved through the output of the sensitive word detection model, and the sampling results are input into the secondary confirmation model to obtain the final sensitive word detection results.

6. The hearing aid with sensitive word reminder function according to claim 1, characterized in that: The sensitive word reminder module generates a sound effect prompt or a voice prompt according to the detection signal sent by the sensitive word detection module to remind the user to pay attention to the sensitive words.

7. The hearing aid with sensitive word reminder function according to claim 1, characterized in that: The direction reminder module detects sensitive words according to the sensitive words sent by the multi-beam sensitive word detection module. The source direction and detection signal generate sound prompts or voice prompts to remind users to pay attention to sensitive words. And through different audio channels, the sense of direction is simulated to enable users to perceive the direction from which sensitive words come.

Images