Oral dental chair lifting control method and system based on voice interaction

Abstract

The application relates to the technical field of medical equipment control, and provides a dental chair lifting control method and system based on voice interaction, which comprises the following steps: removing environmental noise components contained in original sound information to obtain intermediate sound information; identifying a working sound segment matched with preset working sound information of medical equipment from the intermediate sound information according to the preset working sound information, removing the working sound segment from the intermediate sound information, and obtaining pure sound information; determining a first similarity for each preset dental chair control sound information in a plurality of preset dental chair control sound information; taking the preset dental chair control sound information with the maximum first similarity as identified dental chair control sound information; and driving a dental chair to perform a lifting action corresponding to the dental chair control sound information when the identified dental chair control sound information meets a preset execution condition. Through the scheme, the efficiency of dental chair lifting control can be improved.

Description

Technical Field

[0001] This application relates to the field of medical device control technology, and in particular to a method and system for controlling the lifting of a dental chair based on voice interaction. Background Technology

[0002] In modern dental treatment environments, to improve treatment efficiency, ease of operation, and patient comfort, traditional dental chair operation methods, such as those relying on physical buttons or foot pedals, are gradually being replaced by more intelligent solutions. Among these, voice-based control methods and systems demonstrate significant advantages because they free the dentist's hands, enabling non-contact and precise adjustments to the dental chair's posture during treatment. However, in practical applications, especially in complex oral surgical scenarios, the acoustic conditions of the treatment environment often become exceptionally complex, posing a serious challenge to the reliability of voice control systems.

[0003] Specifically, in dental treatment, the dental chair is a fundamental and core piece of equipment. Its position and posture adjustment directly affect the ease of operation for the dentist and the comfort for the patient. Traditional dental chairs are controlled by foot pedals or physical buttons on the backrest. This requires the dentist to interrupt their instrument work and free their hands or feet to adjust the chair, which to some extent affects the continuity and efficiency of the treatment. To address this, a voice-interactive control method has been introduced. The dentist can issue specific voice commands, such as "raise," "lie down," or "rinse position," to drive the chair's height and tilt motors, achieving non-contact adjustment of the chair's posture. During routine oral examinations or simple fillings, the treatment environment is relatively quiet, the dentist's instructions are clear, and the voice recognition system can accurately capture and interpret these commands, controlling the chair to smoothly move to a preset or designated position. The entire process is smooth, effectively improving the dentist's work efficiency and allowing them to fully concentrate on the patient's oral treatment.

[0004] However, the situation changes in more complex oral surgical procedures or tooth preparation. These treatments typically require extended periods and extremely high precision. For example, during root canal treatments or implant placements, dentists need to frequently fine-tune the height of the dental chair and the patient's head position based on the depth and angle of the treatment area to achieve optimal visibility and access. In these critical procedures, the dentist's hands are completely occupied by surgical instruments, making manual adjustments impossible, which is where voice control becomes particularly important. But it is precisely in these scenarios where voice control is most crucial that the acoustic conditions of the treatment environment become most complex. During treatment, high-speed turbine handpieces, ultrasonic scalers, and powerful suction devices operate simultaneously or alternately. These devices generate continuous or intermittent noises with high decibel levels and distinct frequency characteristics, such as the sharp, high-frequency whistling sound from high-speed handpieces and the mixed noise of airflow and fluid generated by suction devices.

[0005] The loud noise from medical devices mixes with the doctor's voice commands and is picked up by the microphone of the dental chair control system. This poses a significant challenge to the backend voice recognition unit. On one hand, when a doctor issues an adjustment command, such as "raise it slightly," the command signal may be drowned out by the stronger background noise from the devices, preventing the voice recognition unit from effectively extracting the voice features and thus deeming it invalid input, resulting in command recognition failure. The doctor has to repeat the command, or even increase the volume, which not only disrupts the focused rhythm of the surgery but may also cause unnecessary disturbance to the tense patient. On the other hand, a more serious problem is that the spectral characteristics of certain device noises may coincidentally resemble the acoustic pattern of a preset control command at a specific moment. For example, the gurgling sound of a suction device at a certain frequency when drawing saliva may be incorrectly interpreted by the system as a command such as "lower" or "move." This incorrect recognition and execution will cause the dental chair to move unexpectedly without the doctor's preparation. Imagine a situation where, while a dentist is using a sharp scalpel to make an incision in the gums or a high-speed burr to precisely prepare the tooth, even a slight movement of the dental chair, such as a slight rise, fall, or tilt, could cause the instruments to deviate from their intended positions, resulting in accidental scratches or damage to the patient's oral soft tissues or teeth. This potential safety risk is absolutely unacceptable. Summary of the Invention

[0006] This application provides a method and system for controlling the height of a dental chair based on voice interaction, aiming to solve the problems of interference from medical equipment noise on voice recognition in complex dental treatment environments, and the resulting failure or misrecognition of dental chair control commands, which in turn leads to safety risks and reduced efficiency.

[0007] To achieve the above objectives, this application adopts the following technical solution: In a first aspect, a method for controlling the height of a dental chair based on voice interaction is provided, comprising: acquiring original sound information in the dental treatment environment; processing the original sound information to remove environmental noise components contained in the original sound information to obtain intermediate sound information; identifying a working sound segment matching the preset working sound information from the intermediate sound information according to the preset working sound information of the medical device, and removing the working sound segment from the intermediate sound information to obtain clean sound information; determining a first similarity for each preset dental chair control sound information among multiple preset dental chair control sound information; the first similarity being the similarity between the spectral feature information of the preset dental chair control sound information and the spectral feature information of the clean sound information; taking the preset dental chair control sound information with the largest first similarity as the identified dental chair control sound information; and driving the dental chair to perform the height adjustment action corresponding to the dental chair control sound information when the identified dental chair control sound information meets the preset execution conditions.

[0008] This technical solution effectively distinguishes between medical equipment noise and physician voice commands in the treatment environment, avoiding noise interference with voice recognition, thereby improving the recognition accuracy and execution safety of dental chair control commands. It solves the problem that existing voice control systems are prone to recognition failure or misoperation in complex noisy environments.

[0009] Further, the original sound information is processed to remove environmental noise components to obtain intermediate sound information, including: calling a preset high-pass filter; inputting the original sound information into the preset high-pass filter to remove sound information with frequencies lower than a preset frequency to obtain initial sound information; calling a preset Wiener filter model; inputting the initial sound information into the preset Wiener filter model to remove noise of a preset bandwidth from the initial sound information to obtain intermediate sound information.

[0010] This technical solution combines high-pass filtering and Wiener filtering models to more thoroughly remove environmental noise, especially low-frequency noise and broadband noise, from the original sound information, thereby obtaining purer initial sound information, providing high-quality input for subsequent speech recognition, and further improving the noise suppression effect.

[0011] Based on this, according to the preset working sound information of the medical device, the working sound segment that matches the preset working sound information is identified from the intermediate sound information, and the working sound segment is removed from the intermediate sound information to obtain clean sound information. This includes: dividing the intermediate sound information into multiple sub-intermediate sound segments; determining a second similarity for each sub-intermediate sound segment; the second similarity is the similarity between the spectral feature information of the sub-intermediate sound segment and the spectral feature information of the preset working sound information; and taking the sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold as working sound segments and removing the working sound segments from the intermediate sound information to obtain clean sound information.

[0012] This technical solution enables the accurate identification and removal of medical equipment noise segments by segmenting intermediate sound information and comparing its similarity with preset working sound information. This effectively avoids interference from medical equipment noise on the recognition of dental chair control commands, ensuring the accuracy of subsequent speech recognition.

[0013] Furthermore, the intermediate audio information is segmented into multiple sub-intermediate audio segments, including: obtaining the security level of the medical behavior corresponding to the medical device and a first correspondence; the first correspondence includes a one-to-one correspondence between multiple security level ranges and multiple segment durations; the segment duration corresponding to the security level range of the medical behavior in the first correspondence is taken as the target segment duration; the intermediate audio information is segmented into multiple sub-intermediate audio segments based on the target segment duration; the duration of the sub-intermediate audio segments is the target segment duration.

[0014] Through this technical solution, this application dynamically adjusts the segmentation duration of sound segments according to the safety level of medical procedures, so that the granularity of noise recognition matches the risk level of medical operations, thereby improving the precision and safety of noise removal and avoiding misjudgments caused by improper segmentation.

[0015] In some preferred embodiments, the medical device includes multiple medical devices, and there are multiple preset working sound information sets, each corresponding one-to-one with a specific medical device. Sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold are used as working sound segments, and these working sound segments are removed from the intermediate sound information to obtain clean sound information. This includes: using sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold as initial working sound segments; acquiring vibration information of the target medical device; the target medical device being the medical device corresponding to the target preset working sound information, and the target preset working sound information being the preset working sound information corresponding to the initial working sound segment; determining whether the target medical device is in a working state based on the vibration information of the target medical device; when the target medical device is in a working state, determining the initial working sound segment as a working sound segment and removing it from the intermediate sound information to obtain clean sound information; when the target medical device is not in a working state, determining that the initial working sound segment is not a working sound segment.

[0016] This technical solution introduces vibration information from medical devices as an auxiliary basis for judgment, further verifying whether the identified working sound segments do indeed originate from the working medical devices, effectively avoiding misjudgments caused by sound similarity, and improving the accuracy and reliability of working noise identification.

[0017] As a technological improvement, the determination of whether the target medical device is in working condition is based on the vibration information of the target medical device, including: determining whether the vibration energy within a preset frequency range is greater than a preset vibration energy threshold based on the vibration information of the target medical device; if the vibration energy within the preset frequency range is greater than the preset vibration energy threshold, the target medical device is determined to be in working condition; otherwise, the target medical device is determined not to be in working condition.

[0018] Through this technical solution, this application can more accurately determine whether a medical device is in operation by monitoring vibration energy within a specific frequency range, thereby providing a more reliable physical basis for identifying operating noise and further improving the system's accuracy in identifying noise sources.

[0019] To enhance functionality, when the vibration energy within a preset frequency range exceeds a preset vibration energy threshold, the target medical device is determined to be in a working state. This includes: determining that the target medical device is in a pre-estimated working state when the vibration energy within the preset frequency range exceeds the preset vibration energy threshold; acquiring the target medical device's current information when the target medical device is in the pre-estimated working state; determining whether the target medical device's current information is greater than the target medical device's no-load current information; and determining that the target medical device is in a working state when the target medical device's current information is greater than the target medical device's no-load current information.

[0020] This technical solution introduces current information as a secondary verification on the basis of vibration information judgment. By judging whether the current is greater than the no-load current, it can more accurately confirm whether the medical device is in actual working load state, thereby avoiding misjudgment in idling or standby state and greatly improving the accuracy of working status judgment.

[0021] As a further improvement, when the identified dental chair control sound information meets the preset execution conditions, the dental chair is driven to perform the lifting action corresponding to the dental chair control sound information, including: determining a third similarity; the third similarity is the similarity between the identified dental chair control sound information and the preset working sound information of the medical device; when the third similarity is less than the first similarity corresponding to the identified dental chair control sound information, it is determined that the identified dental chair control command meets the preset execution conditions, and the dental chair is driven to perform the lifting action corresponding to the dental chair control sound information; otherwise, it is determined that the identified dental chair control sound information does not meet the preset execution conditions.

[0022] This application introduces a similarity comparison between the identified dental chair control sound information and the medical device operating sound information as part of the execution conditions. This effectively avoids erroneous execution when the dental chair control command is highly similar to the medical device noise, and further improves the safety of dental chair control.

[0023] To improve the solution, the dental chair is driven to perform lifting and lowering actions corresponding to the dental chair control voice information, including: performing intent recognition on the dental chair control voice information to obtain intent recognition results; determining whether the lifting and lowering action corresponding to the intent recognition result is a high-risk action; the lifting and lowering distance of a high-risk action is greater than a preset lifting and lowering distance; when the lifting and lowering action corresponding to the intent recognition result is not a high-risk action, the dental chair is driven to perform the lifting and lowering action corresponding to the intent recognition result; when the lifting and lowering action corresponding to the intent recognition result is a high-risk action, a confirmation instruction is issued to the operator, and after receiving the operator's confirmation instruction, the dental chair is driven to perform the lifting and lowering action corresponding to the intent recognition result.

[0024] Through this technical solution, this application introduces a secondary confirmation mechanism for high-risk actions. For dental chair lifting actions that may lead to large displacement or potential danger, the system will actively request confirmation from the operator, thereby effectively avoiding safety accidents caused by misidentification or misoperation and greatly improving the safety of dental chair control.

[0025] Secondly, this application also discloses a voice-interactive oral chair lifting control system, comprising: an acquisition device and a processing device; the acquisition device is used to acquire original sound information in the oral treatment environment; the processing device is used to process the original sound information to remove environmental noise components contained in the original sound information to obtain intermediate sound information; the processing device is used to identify working sound segments matching the preset working sound information from the intermediate sound information according to the preset working sound information of the medical device, and remove the working sound segments from the intermediate sound information to obtain pure sound information; the processing device is used to determine a first similarity for each preset oral chair control sound information among multiple preset oral chair control sound information; the first similarity is the similarity between the spectral feature information of the preset oral chair control sound information and the spectral feature information of the pure sound information; the processing device is used to take the preset oral chair control sound information with the largest first similarity as the identified oral chair control sound information; the processing device is used to drive the oral chair to perform the action corresponding to the oral chair control sound information when the identified oral chair control sound information meets the preset execution conditions. Beneficial effects

[0026] This application discloses a voice-interactive-based method for controlling the height of a dental chair. It acquires raw sound information from the dental treatment environment and processes it in multiple stages. First, environmental noise is removed to obtain intermediate sound information. Then, based on preset working sound information from the medical device, working sound segments are accurately identified and removed to obtain pure sound information. On this basis, the similarity between the pure sound information and multiple preset dental chair control sound information is calculated, and the sound with the highest similarity is selected as the identified dental chair control sound information. Finally, when the identified dental chair control sound information meets preset execution conditions, the dental chair is driven to perform the corresponding height adjustment action. This method effectively solves the problem of severe interference from medical device noise on speech recognition in complex dental treatment environments in existing technologies, avoiding recognition failures and misoperations caused by noise drowning out commands or noise being similar to commands. Through refined noise removal and command recognition mechanisms, this application significantly improves the accuracy, reliability, and safety of dental chair voice control, allowing physicians to focus more on the patient during treatment without interrupting the operation for manual adjustments, thereby improving treatment efficiency and reducing the risks to patients that may be caused by misoperation. Attached Figure Description

[0027] Figure 1 A flowchart illustrating a voice-interactive-based method for controlling the height of a dental chair provided in this application; Figure 2 A flowchart illustrating another method for controlling the height of a dental chair based on voice interaction provided in this application; Figure 3This is a schematic diagram of a voice-interactive-based dental chair lifting control system provided in this application. Detailed Implementation

[0028] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0029] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0030] In modern dental treatment environments, to improve treatment efficiency, ease of operation, and patient comfort, traditional dental chair operation methods, such as those relying on physical buttons or foot pedals, are gradually being replaced by more intelligent solutions. Among these, voice-based control methods and systems demonstrate significant advantages because they free the dentist's hands, enabling non-contact and precise adjustments to the dental chair's posture during treatment. However, in practical applications, especially in complex oral surgical scenarios, the acoustic conditions of the treatment environment often become exceptionally complex, posing a serious challenge to the reliability of voice control systems.

[0031] In this regard, such as Figure 1 As shown, this application proposes a method for controlling the height of a dental chair based on voice interaction, including: S101. Obtain the original sound information in the oral diagnosis and treatment environment.

[0032] S102. Process the original sound information to remove the environmental noise components contained in the original sound information and obtain intermediate sound information.

[0033] S103. Based on the preset working sound information of the medical device, identify the working sound segment that matches the preset working sound information from the intermediate sound information, and remove the working sound segment from the intermediate sound information to obtain pure sound information.

[0034] S104. For each preset dental chair control sound information among multiple preset dental chair control sound information, determine the first similarity.

[0035] The first similarity is the similarity between the spectral feature information of the preset dental chair control sound information and the spectral feature information of the pure sound information.

[0036] S105. Use the preset dental chair control sound information with the highest first similarity as the identified dental chair control sound information.

[0037] S106. When the identified dental chair control sound information meets the preset execution conditions, drive the dental chair to perform the lifting action corresponding to the dental chair control sound information.

[0038] This application ensures the purity of dental chair control sound information by performing multi-stage noise removal and working sound filtering on the original sound information. Combined with similarity matching and preset execution condition judgment, it significantly improves the accuracy and security of voice command recognition, effectively avoiding misidentification and misoperation caused by environmental noise and medical equipment working sound interference. Thus, it realizes precise and reliable voice control of dental chair in complex oral treatment environments.

[0039] To better understand the technical solution proposed in this application, some key terms and implementation environments involved will be explained first.

[0040] "Oral care environment" refers to the place where medical activities such as oral examination, treatment, and surgery are carried out. It usually includes dental chairs, medical equipment (such as high-speed turbine handpieces, ultrasonic scalers, suction devices, etc.), lighting equipment, medical staff, and patients.

[0041] "Raw audio information" refers to the unprocessed raw audio data collected in the dental treatment environment through microphones and other sound pickup devices, which may contain various sound components such as the doctor's voice instructions, the working noise of medical equipment, and environmental background noise.

[0042] "Preset operating sound information" refers to pre-recorded or stored sound characteristic information representing the sound emitted by a specific medical device during operation, such as the whistling sound of a high-speed turbine handpiece or the vibration sound of an ultrasonic scaler. This information is used to identify and remove the operating noise of the medical device in subsequent processing.

[0043] "Preset dental chair control voice information" refers to pre-recorded or stored speech feature information that represents specific dental chair control commands (such as "raise," "lower," etc.). This information is used to match with clean voice information to identify the dentist's dental chair control commands.

[0044] "Spectral feature information" refers to the representation of a sound signal in the frequency domain, such as the frequency and energy distribution features extracted by methods like Fourier transform. These features can effectively characterize the uniqueness of different sounds.

[0045] "Similarity" refers to the degree of matching between the spectral features of two sound signals. It is usually obtained by calculating the correlation coefficient, Euclidean distance, or cosine similarity. The higher the similarity, the closer the two sound signals are.

[0046] The implementation environment of this application typically includes one or more microphones for acquiring raw sound information, a processing unit (e.g., an embedded processor, a digital signal processor, or a general-purpose computer) for executing algorithms such as sound processing, feature extraction, similarity calculation, and instruction recognition, and a dental chair control module for receiving instructions and driving the dental chair to perform corresponding lifting and lowering actions.

[0047] The core of the voice-interactive-based dental chair lifting control method proposed in this application lies in ensuring that the dentist's dental chair control commands can be accurately identified in complex treatment environments through a series of refined sound processing and recognition steps, and that the dental chair can be safely driven to perform the corresponding actions.

[0048] First, the method involves "acquiring raw sound information from the dental treatment environment." In practice, this can be achieved by deploying one or more high-sensitivity microphones within the dental treatment environment. For example, microphones can be mounted near the headrest of the dental chair, on the bracket of a treatment light, or integrated into a headset worn by the dentist. These microphones can capture all sounds within the treatment area in real time and convert them into raw, digitized sound information. The raw sound information can be mono or multi-channel data, and the sampling rate and bit depth can be configured according to actual needs. For example, a sampling rate of 44.1 kHz and a bit depth of 16 bits can be used to ensure the integrity of the sound information.

[0049] Secondly, the acquired raw sound information is processed to "remove environmental noise components from the raw sound information to obtain intermediate sound information." This step aims to initially purify the sound signal and eliminate non-specific noise prevalent in the background. For example, statistical model-based methods, such as spectral subtraction, can be used to suppress noise by estimating the spectrum of the noise and subtracting it from the spectrum of the raw sound signal. Another approach is to use adaptive filtering techniques, such as the Least Mean Square (LMS) algorithm or the Recursive Least Squares (RLS) algorithm, to estimate and eliminate noise using an adaptive filter. Furthermore, deep learning models, such as denoising models based on Convolutional Neural Networks (CNNs) or Recurrent Neural Networks (RNNs), can be used to perform end-to-end noise suppression on the raw sound information, resulting in cleaner intermediate sound information.

[0050] Next, based on the preset operating sound information of the medical device, the method "identifies operating sound segments that match the preset operating sound information from the intermediate sound information, and removes the operating sound segments from the intermediate sound information to obtain pure sound information." This step is one of the key innovations of this application, aiming to solve the interference of medical device operating noise on voice command recognition. For example, typical sound samples of various commonly used medical devices (such as high-speed turbine handpieces, ultrasonic scalers, suction devices, etc.) during operation can be collected and stored in advance, and their spectral feature information can be extracted as preset operating sound information. During processing, the intermediate sound information can be divided into several short frames, and spectral features can be extracted for each short frame, and similarity calculation can be performed with the spectral features of the preset operating sound information. When the similarity between a short frame and a preset operating sound information exceeds a preset threshold, the short frame is considered to contain medical device operating noise and is marked as an operating sound segment. Subsequently, these identified operating sound segments will be removed from the intermediate sound information, for example, by setting the energy of these segments to zero or performing more complex masking processing, thereby obtaining pure sound information without medical device operating noise.

[0051] Subsequently, for each preset dental chair control voice message among multiple preset dental chair control voice messages, "a first similarity is determined; the first similarity is the similarity between the spectral feature information of the preset dental chair control voice message and the spectral feature information of the clean voice message." In this step, the system pre-records and stores a series of standard dental chair control command voice samples, such as "dental chair raise," "dental chair lower," and "dental chair lie flat," and extracts their spectral feature information as preset dental chair control voice messages. Then, the clean voice information obtained after multiple noise reduction processes is compared with each preset dental chair control voice message for similarity. For example, Mel-frequency cepstral coefficients (MFCCs) can be used as spectral features, and dynamic time warping (DTW) algorithms or cosine similarity can be used to calculate the similarity between the clean voice information and each preset dental chair control voice message. In this way, the degree of matching between the clean voice information and each potential dental chair control command can be quantified.

[0052] Furthermore, this method uses the preset dental chair control sound with the highest similarity as the identified dental chair control sound. After calculating the first similarity between the clean sound information and all preset dental chair control sound information, the system selects the preset dental chair control sound with the highest similarity as the final identified dental chair control command. For example, if the "raise dental chair" command has the highest similarity to the clean sound information, then the system will use "raise dental chair" as the identified dental chair control sound. This "maximum similarity" decision-making mechanism ensures that the system can select the command that best matches the physician's intention.

[0053] Finally, when the identified dental chair control sound information "meets the preset execution conditions," the dental chair is driven to perform the lifting and lowering action corresponding to the dental chair control sound information. This step is a crucial step in ensuring the safety of the dental chair control.

[0054] This application further proposes a more refined noise processing method to more effectively remove environmental noise components from the original sound information, thereby obtaining higher quality intermediate sound information.

[0055] In response, this application further proposes processing the original sound information to remove environmental noise components and obtain intermediate sound information, including: Call the preset high-pass filter; input the original sound information into the preset high-pass filter to remove sound information with frequencies lower than the preset frequency from the original sound information to obtain the initial sound information; call the preset Wiener filter model; input the initial sound information into the preset Wiener filter model to remove noise with a preset bandwidth from the initial sound information to obtain the intermediate sound information.

[0056] Specifically, calling the preset high-pass filter means that the system has pre-stored and prepared a high-pass filter. This filter is designed to allow sound signals above a certain frequency to pass through, while attenuating or blocking sound signals below that frequency. The preset frequency is the cutoff frequency of the high-pass filter. Sound information below this frequency will be effectively filtered out. Its purpose is to remove low-frequency noise commonly found in the dental treatment environment, such as the low-frequency hum of equipment or low-frequency background noise in the environment.

[0057] Furthermore, calling the preset Wiener filter model means that the system pre-stores and prepares a Wiener filter model. This model is a linear filter based on the minimum mean square error criterion, commonly used to recover the original signal from noisy signals. The Wiener filter model can adaptively adjust the filter parameters according to the statistical characteristics of the signal and noise to achieve the best noise suppression effect. The noise with the preset bandwidth can be understood as noise components with a specific frequency range present in the initial sound information, such as broadband white noise or other background noise without specific frequencies. Its purpose is to further refine the removal of residual noise components in the initial sound information that are difficult to remove by simple high-pass filtering, thereby obtaining a purer intermediate sound information.

[0058] This application's solution effectively addresses the removal of various noise components from raw sound information by employing a phased, multi-stage filtering strategy. First, by invoking a preset high-pass filter and setting a preset frequency, low-frequency noise commonly found in dental clinic environments, such as low-frequency humming sounds generated by air conditioning, ventilation systems, or certain medical equipment, can be specifically removed. This initial filtering significantly reduces low-frequency noise interference in the sound information, laying the foundation for subsequent refined processing. Second, after removing low-frequency noise to obtain the initial sound information, a preset Wiener filter model is further invoked. The Wiener filter model adaptively suppresses noise within a preset bandwidth based on the statistical characteristics of the initial sound information and noise, effectively removing broadband noise or other background noise of non-specific frequencies. It is precisely this combination of high-pass and Wiener filtering that allows environmental noise components in the raw sound information to be removed more comprehensively and thoroughly, resulting in higher-quality intermediate sound information.

[0059] Through the above technical solution, this application can achieve more effective and thorough removal of environmental noise components from the original sound information in the oral treatment environment. Compared with single noise processing methods, combining high-pass filters and Wiener filtering models can process different types of noise (such as low-frequency noise and broadband noise) step by step, significantly improving the efficiency and effectiveness of noise suppression. As a result, the obtained intermediate sound information has a higher signal-to-noise ratio and purity, providing a more reliable and clear input for subsequent recognition of working sound segments and dental chair control sound information. This effectively improves the accuracy and stability of the voice-interactive dental chair lifting control method, reduces the false recognition rate, and ensures the safety and smoothness of the treatment process.

[0060] This application further proposes a method for identifying working sound segments matching the preset working sound information from intermediate sound information based on the preset working sound information of a medical device, and removing the working sound segments from the intermediate sound information to obtain clean sound information. This process includes: The intermediate sound information is divided into multiple sub-intermediate sound segments; for each sub-intermediate sound segment, a second similarity is determined; the second similarity is the similarity between the spectral feature information of the sub-intermediate sound segment and the spectral feature information of the preset working sound information; the sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold are taken as working sound segments, and the working sound segments are removed from the intermediate sound information to obtain clean sound information.

[0061] Specifically, segmenting intermediate audio information into multiple sub-intermediate audio segments refers to dividing the intermediate audio information, after preliminary noise reduction processing, into several shorter, independent audio segments according to a preset time length or specific rules. The purpose is to discretize the continuous audio stream so that each small segment can be analyzed and processed independently, improving the granularity and accuracy of working audio segment recognition.

[0062] The second similarity determination, for each of the multiple sub-intermediate sound segments, refers to calculating the degree of similarity between each segmented sub-intermediate sound segment and the preset working sound information. The second similarity can be understood as quantifying the similarity by comparing the spectral feature information of the two. Spectral feature information is the representation of sound in the frequency domain, reflecting key attributes such as timbre and frequency distribution. For example, methods such as Mel-frequency cepstral coefficients (MFCC), linear predictive coding (LPC) coefficients, and short-time Fourier transform (STFT) can be used to extract the spectral features of sound. Similarity calculation methods may include, but are not limited to, cosine similarity, the reciprocal of Euclidean distance, and dynamic time warping (DTW), with the aim of objectively evaluating the degree of matching between the sub-intermediate sound segments and the known working sound of medical equipment.

[0063] In practical applications, sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold are designated as working sound segments. These working sound segments are then removed from the intermediate sound information to obtain clean sound information. This involves calculating the similarity between each sub-intermediate sound segment and the preset working sound information, and then setting a preset first similarity threshold. Only when the second similarity of a sub-intermediate sound segment exceeds this threshold is it identified as containing the working sound of the medical device, and it is removed from the intermediate sound information. The setting of the preset first similarity threshold needs to comprehensively consider the characteristics of the working sound of the medical device, the residual environmental noise, and the recognition accuracy requirements to ensure that the working sound is effectively removed while avoiding the accidental deletion of non-working sounds. In this way, the working sound component in the intermediate sound information can be accurately located and removed, thereby obtaining clean sound information that does not contain the working sound of the medical device.

[0064] The proposed solution meticulously segments intermediate sound information into multiple sub-intermediate sound segments, enabling the recognition of working sounds to move beyond holistic assessment and allow for refined analysis of local areas within the sound stream. By calculating the spectral similarity between each sub-intermediate sound segment and preset working sound information, and setting a preset first similarity threshold, segments highly matching the working sounds of medical equipment can be effectively filtered out. This segment-based similarity matching and threshold judgment mechanism can more accurately identify the start and end of the working sound and its temporal distribution, even if the working sound is intermittent or interwoven with other sounds, it can be precisely captured and removed. This ensures the accuracy and completeness of removing working sound segments from intermediate sound information, avoiding residual or misjudged working sounds due to insufficient overall recognition, and providing a cleaner input for the subsequent recognition of dental chair control sound information.

[0065] Through the above technical solution, this application can significantly improve the accuracy and thoroughness of removing medical device operating sounds from the sound environment of oral treatment. Compared with the potential problems of insufficient overall recognition or incomplete removal in basic solutions, this application introduces additional technical features such as sound segment segmentation, segment-by-segment similarity calculation, and threshold screening, making the recognition and removal process of operating sounds more refined and intelligent. This not only effectively avoids interference from medical device operating sounds on the recognition of dental chair control voice commands, ensuring the quality of pure sound information, but also improves the reliability and safety of voice-interactive control of dental chair lifting and lowering movements, providing a more stable and efficient voice control experience for the oral treatment process.

[0066] This application further proposes a step for segmenting intermediate audio information into multiple sub-intermediate audio segments, including: Obtain the security level and first correspondence of the medical behavior corresponding to the medical device; the first correspondence includes a one-to-one correspondence between multiple security level ranges and multiple segment durations; take the segment duration corresponding to the security level range of the medical behavior in the first correspondence as the target segment duration; divide the intermediate sound information into multiple sub-intermediate sound segments based on the target segment duration; the duration of the sub-intermediate sound segments is the target segment duration.

[0067] Specifically, the safety level of a medical procedure refers to the risk assessment level associated with a specific operation performed by medical equipment during oral treatment. For example, procedures such as tooth extraction and root canal treatment may be assigned a higher safety level, while procedures such as teeth cleaning and examinations may have a lower safety level. This safety level can be preset and categorized based on medical standards, expert experience, or historical data. The first correspondence can be understood as a mapping table or rule set, which establishes a relationship between different safety level ranges and corresponding segment durations. For example, when the safety level is "high," the corresponding segment duration can be set to a shorter time to improve the precision of recognition; when the safety level is "low," the corresponding segment duration can be set to a longer time to reduce processing overhead. This correspondence aims to dynamically adjust the segmentation granularity of sound information according to the risk level of the medical procedure.

[0068] In practical applications, the target segment duration is determined by searching and matching the first correspondence based on the safety level of the current medical procedure, and is used to segment intermediate audio information. For example, if the safety level of the current medical procedure is determined to be "medium risk," and the segment duration corresponding to "medium risk" in the first correspondence is 500 milliseconds, then the target segment duration is set to 500 milliseconds. Consequently, the intermediate audio information will be segmented into a series of continuous sub-intermediate audio segments based on this target segment duration, with the duration of each sub-intermediate audio segment being consistent with the target segment duration.

[0069] This application's solution achieves adaptive adjustment of the segmentation granularity of intermediate sound information by introducing a correspondence between the safety level of medical actions and the segment duration. After obtaining the safety level of the medical action currently being performed by the medical device, the system dynamically determines a target segment duration that matches that safety level based on a preset first correspondence. For example, for high-risk medical actions, the system selects a shorter segment duration for segmentation, ensuring that each sub-intermediate sound segment contains more refined sound information. This helps to more accurately identify potential working sound segments and avoids introducing irrelevant noise or confusing information due to excessively long segments. Conversely, for low-risk medical actions, the system can select a longer segment duration to reduce the computational load of segmentation operations and improve processing efficiency. This dynamic segmentation mechanism based on safety levels allows sound information processing to better adapt to the needs of different medical scenarios, improving the robustness and accuracy of working sound segment recognition.

[0070] The aforementioned technical solution intelligently adjusts the granularity of intermediate audio information segmentation based on the actual risk level of the medical procedure. This not only avoids the problems of insufficient recognition accuracy or wasted computing resources that may result from fixed segment lengths, but also effectively improves the accuracy and efficiency of recognizing working audio segments through fine-grained or coarse-grained processing. Especially in high-risk medical scenarios, using shorter segment lengths for segmentation allows for faster and more accurate capture of subtle working audio features, thus providing a more reliable foundation for subsequent extraction of clean audio information and further enhancing the security and reliability of the entire voice interaction system.

[0071] This application further proposes an optimization scheme, wherein the medical device includes multiple medical devices, and there are multiple preset working sound information, each of which corresponds one-to-one with a single medical device. Sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold are used as working sound segments, and these working sound segments are removed from the intermediate sound information to obtain clean sound information, including: The sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold are used as the initial working sound segments; the vibration information of the target medical device is obtained; the target medical device is the medical device corresponding to the target preset working sound information, and the target preset working sound information is the preset working sound information corresponding to the initial working sound segment; based on the vibration information of the target medical device, it is determined whether the target medical device is in a working state; when the target medical device is in a working state, the initial working sound segment is determined to be a working sound segment, and the working sound segment is removed from the intermediate sound information to obtain pure sound information; when the target medical device is not in a working state, the initial working sound segment is determined not to be a working sound segment.

[0072] Specifically, the aforementioned medical equipment can be understood as various devices used in the oral healthcare environment, such as dental drills, ultrasonic scalers, and light-curing machines. These medical devices typically contain one or more medical devices, each of which produces a specific sound when operating. Therefore, there can be multiple preset operating sound information entries, and each preset operating sound information entry corresponds one-to-one with a specific medical device, so that the system can distinguish the operating sounds of different devices. Specifically, using sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold as initial operating sound segments means that segments that may belong to the operating sounds of medical devices are initially screened through similarity matching of sound spectrum features.

[0073] Furthermore, acquiring vibration information of the target medical device refers to monitoring its physical vibration state in real time by installing vibration sensors on the target medical device. Vibration information is a reliable physical basis for determining whether a medical device is truly in a working state. A target medical device refers to a medical device associated with preset working sound information corresponding to the currently identified initial working sound segment. For example, if the initial working sound segment matches the preset information of "dental drill working sound," then the target medical device is a dental drill. In practical applications, determining whether a target medical device is in a working state based on its vibration information aims to introduce physical-level verification to improve the accuracy of working sound segment identification.

[0074] Only when vibration data indicates that the target medical device is indeed in operation can the initial working sound segment be definitively confirmed as a genuine working sound segment. Conversely, if the vibration data indicates that the target medical device is not in operation, even if the sound spectrum similarity is high, the initial working sound segment should be determined not to be a genuine working sound segment, thus avoiding the erroneous removal of non-working sounds. Therefore, when the target medical device is in operation, the initial working sound segment is definitively identified as a working sound segment and removed from the intermediate sound data to obtain purer sound information. When the target medical device is not in operation, even if the sound features match, the initial working sound segment will not be incorrectly identified as a working sound segment and removed.

[0075] The solution proposed in this application effectively solves the problem of misjudgment that may occur when relying solely on sound spectrum features to identify working sound segments by introducing verification of the physical working state of the medical device. Specifically, when the system initially identifies a certain sub-intermediate sound segment as possibly being the working sound of the medical device through sound spectrum similarity, it does not immediately confirm it as a working sound segment.

[0076] Instead, the system further acquires vibration information of the target medical device corresponding to the sound segment. Since medical devices typically generate specific mechanical vibrations during operation, monitoring this vibration information allows for accurate determination of whether the device is truly operational. Only when the vibration information confirms that the target medical device is indeed operational is the sound segment ultimately identified as operational and removed from the intermediate sound information. This dual verification mechanism, combining sound feature matching with physical state confirmation, significantly improves the accuracy and robustness of operational sound segment identification, avoiding misidentification caused by environmental noise or similar sounds in non-operational states.

[0077] Through the above technical solution, this application can more accurately identify and remove the operating sounds of medical devices in the oral treatment environment. Compared with recognition methods that rely solely on sound spectrum characteristics, introducing vibration information from medical devices as an auxiliary judgment criterion greatly reduces the false recognition rate and ensures the quality of pure sound information. Therefore, subsequent recognition of dental chair control sound information will be based on purer input, thereby improving the accuracy and reliability of voice-interactive control of dental chair lifting and lowering movements, effectively avoiding misoperation or recognition failure caused by background device sound interference, and enhancing user experience and treatment safety.

[0078] This application further proposes a specific method for determining whether a target medical device is in operation based on vibration information of the target medical device, in order to improve the accuracy and robustness of the determination.

[0079] Specifically, such as Figure 2 As shown, based on the vibration information of the target medical device, it is determined whether the target medical device is in a working state, including: S201. Determine whether the vibration energy within the preset frequency range is greater than the preset vibration energy threshold based on the vibration information of the target medical device.

[0080] S202. When the vibration energy within the preset frequency range is greater than the preset vibration energy threshold, the target medical device is determined to be in working condition; otherwise, the target medical device is determined not to be in working condition.

[0081] The vibration information of the target medical device can be understood as physical vibration data collected in real time by vibration sensors (such as accelerometers, piezoelectric sensors, etc.) installed on the medical device. This data is typically a time-domain signal, reflecting the mechanical vibration characteristics generated by the device during operation. The preset frequency range refers to a frequency interval predetermined based on the operating characteristics of a specific medical device and the frequency distribution of its main noise sources. For example, the operating vibration of a high-speed rotating dental handpiece may be concentrated in a higher frequency range; while the vibration frequency of a low-speed suction pump may be lower.

[0082] By limiting the frequency range, background vibrations or environmental noise unrelated to the device's operation can be effectively filtered out. Vibration energy refers to the intensity or power of a vibration signal within a specific frequency range; it can be the root mean square value, power spectral density integral, or other energy measures. The preset vibration energy threshold is a critical value determined experimentally or empirically based on the vibration energy distribution of the medical device in its non-operating and normal operating states. When the vibration energy within the preset frequency range exceeds this threshold, the device is considered to be actually operating, rather than in standby, idling, or under external interference.

[0083] This application's solution addresses the potential for misjudgment based solely on general vibration information by performing a refined analysis of the vibration information of the target medical device. Specifically, when performing its core functions, the medical device generates mechanical vibrations with specific frequency characteristics and energy levels. By pre-setting a frequency range consistent with the device's operating characteristics and monitoring vibration energy within this range, the actual operating vibration of the device can be effectively distinguished from environmental noise, accidental collisions, or weak vibrations during non-operational states. When the vibration energy within the preset frequency range exceeds a preset vibration energy threshold, this strongly indicates that the internal mechanical components of the device are actively operating, thus accurately determining that the target medical device is in operation. Conversely, if the vibration energy does not reach the threshold, it indicates that the device is not in an effective operating state, thereby avoiding subsequent voice control errors caused by misjudgment.

[0084] The above technical solution significantly improves the accuracy and reliability of determining whether a target medical device is in operation. Compared to relying solely on generalized vibration information, this solution effectively filters out irrelevant noise and interference by introducing a preset frequency range and vibration energy threshold, resulting in a more accurate assessment of the device's operational status. This helps avoid misinterpreting vibrations in a non-operating state as those in an operating state, thereby reducing voice command recognition errors caused by misinterpreting operating sound segments and further enhancing the overall stability and safety of the voice-interactive dental chair height control method.

[0085] This application further proposes a step for determining that the target medical device is in a working state when the vibration energy within a preset frequency range is greater than a preset vibration energy threshold, including: When the vibration energy within the preset frequency range exceeds the preset vibration energy threshold, the target medical device is determined to be in the estimated working state. When the target medical device is in the estimated working state, the current information of the target medical device is acquired. It is determined whether the current information of the target medical device is greater than the no-load current information of the target medical device. When the current information of the target medical device is greater than the no-load current information of the target medical device, the target medical device is determined to be in the working state.

[0086] Specifically, when the vibration energy of the target medical device meets preset conditions, it is first identified as being in a "preliminary operating state." The "preliminary operating state" refers to an intermediate state based on preliminary vibration detection, suggesting the device may be operating, but not yet definitively confirmed. To improve the accuracy of this assessment, further acquisition of the target medical device's current information is necessary. Current information refers to the actual current consumed by the medical device during operation, which can be monitored in real-time using a current sensor integrated into the device's power supply line. The no-load current information of the target medical device refers to the minimum current consumption when powered on but not performing any actual tasks, such as the current during standby or when only maintaining basic circuit operation. By comparing the real-time acquired current information with the no-load current information, it is possible to more accurately determine whether the medical device is truly under workload.

[0087] This application's solution addresses the potential misjudgment caused by relying solely on vibration information by introducing the assessment of the target medical device's current information. When the vibration energy of the target medical device reaches a preset threshold, although it is initially judged that it may be in a working state, it is not immediately confirmed as working; instead, it is placed in a "preliminary working state." Under this "preliminary working state," the system further acquires the medical device's current information. If the current information of the medical device is greater than its no-load current information, it indicates that the device is not only vibrating but is also actually consuming electrical energy to perform its work, thus accurately determining that the target medical device is in its true working state. Conversely, if the vibration energy meets the condition but the current information does not exceed the no-load current, it indicates that the vibration may not be caused by actual operation, thereby avoiding misjudgment.

[0088] Through the above technical solution, this application can more accurately determine whether a medical device is in actual working condition. This dual confirmation mechanism, combining vibration and current information, effectively avoids misjudgments caused by vibration signals from non-working states, such as environmental interference, equipment inertia, or occasional vibrations. Therefore, it can more accurately identify and remove medical device operating sound segments in the dental treatment environment, thereby obtaining purer voice information, significantly improving the accuracy and reliability of subsequent dental chair control sound information recognition, and reducing the risk of misoperation.

[0089] This application further proposes that, when the identified dental chair control sound information meets preset execution conditions, the dental chair is driven to perform a lifting action corresponding to the dental chair control sound information, specifically including: A third similarity is determined; wherein the third similarity is defined as the similarity between the identified dental chair control sound information and the preset working sound information of the medical device; when the third similarity is less than the first similarity corresponding to the identified dental chair control sound information, it is determined that the identified dental chair control command meets the preset execution conditions, and the dental chair is driven to perform the lifting action corresponding to the dental chair control sound information; otherwise, it is determined that the identified dental chair control sound information does not meet the preset execution conditions.

[0090] Specifically, the third similarity measure aims to assess the degree of similarity between the identified dental chair control sounds and the operating sounds of medical devices that may exist in the current treatment environment. This similarity can be obtained by comparing the spectral characteristics of the two, for example, by using algorithms such as cosine similarity, Euclidean distance, or dynamic time warping (DTW). Its purpose is to assess whether the identified dental chair control sounds may be interfered with or misjudged by the operating sounds of medical devices.

[0091] This application's solution effectively solves the aforementioned problems of misidentification and misoperation by introducing a third similarity score and comparing it with the first similarity score corresponding to the identified dental chair control sound information. Specifically, the first similarity score reflects the degree of matching between the identified dental chair control sound information and the clean sound information (i.e., the sound after removing environmental noise and the sound of the medical device operating), and is the main basis for judging the validity of voice commands. The third similarity score provides an additional security verification mechanism, which evaluates the similarity between the identified dental chair control sound information and the preset operating sound information of the medical device.

[0092] When the third similarity score is less than the first similarity score, it means that the identified dental chair control voice information matches the preset dental chair control commands more closely, while having a lower similarity to the working voice of the medical equipment. This eliminates misjudgments caused by interference from the working voice of the medical equipment. Therefore, the system can more accurately determine the true intent of the voice commands, avoiding erroneous dental chair raising and lowering actions due to sound confusion while the medical equipment is operating.

[0093] Through the above technical solution, this application can significantly improve the safety and reliability of the voice-interactive dental chair lifting control system. This solution, by introducing a judgment on the similarity between the identified dental chair control voice information and the operating voice of medical equipment, effectively avoids misidentification and misoperation caused by the similarity in the frequency spectrum between the operating voice of medical equipment and the dental chair control voice information in complex treatment environments. Especially when high-noise medical equipment such as high-speed turbine handpieces and ultrasonic scalers are operating, this solution provides an additional safety barrier, ensuring that the dental chair only performs lifting actions when it receives clear and interference-free control commands, thereby protecting the safety of patients and medical staff and improving the system's intelligence and user-friendliness.

[0094] This application further proposes the following steps for driving the dental chair to perform lifting and lowering actions corresponding to the aforementioned dental chair control sound information: The system performs intent recognition on the control sound information of the dental chair to obtain the intent recognition result; it then determines whether the lifting action corresponding to the intent recognition result is a high-risk action; if the lifting distance of a high-risk action is greater than the preset lifting distance; if the lifting action corresponding to the intent recognition result is not a high-risk action, the system drives the dental chair to perform the lifting action corresponding to the intent recognition result; if the lifting action corresponding to the intent recognition result is a high-risk action, the system sends a confirmation instruction to the operator, and upon receiving the operator's confirmation instruction, drives the dental chair to perform the lifting action corresponding to the intent recognition result.

[0095] Specifically, intent recognition of dental chair control voice information refers to converting the recognized dental chair control voice information into specific, executable commands for raising or lowering the dental chair using natural language processing (NLP) technology or a pre-trained voice command parsing model. For example, the command might include "raise," "lower," "adjust to the highest position," or "adjust to the lowest position." The intent recognition result typically includes the action type (e.g., raise or lower) and the action magnitude (e.g., a specific distance value or target location).

[0096] High-risk actions can be understood as lifting or lowering movements that may significantly impact the patient or the treatment environment. Specifically, the lifting or lowering distance for high-risk actions is set to be greater than a preset distance. This preset distance can be flexibly set based on various factors such as clinical experience, dental chair model, and patient safety standards. For example, it can be set to 30% or 50% of the total lifting and lowering travel of the dental chair, or a fixed distance value, such as 20 centimeters. When the actual lifting or lowering distance corresponding to the intent recognition result exceeds this preset distance, it is judged as a high-risk action by the system.

[0097] In practical applications, when the lifting or lowering motion corresponding to the intent recognition result is a high-risk action, the system will issue a confirmation instruction to the operator. This confirmation instruction can be implemented in various ways, such as displaying a warning message on the dental chair control panel, issuing a voice announcement saying "High-risk operation detected, please confirm," or displaying a confirmation dialog box on the connected monitor. After receiving the confirmation instruction, the operator needs to provide a confirmation command through a specific method (such as pressing a confirmation button or issuing a confirmation voice command). Only after receiving the operator's confirmation command will the dental chair be driven to perform the high-risk lifting or lowering motion; otherwise, the action will be paused or canceled.

[0098] This application's solution effectively addresses the potential safety hazards of directly executing commands by introducing intent recognition, high-risk action judgment, and operator confirmation mechanisms. Specifically, firstly, intent recognition is performed on the dental chair control voice information to ensure that voice commands are accurately interpreted into specific lifting and lowering movements and their amplitudes. Secondly, by judging whether the lifting and lowering movements corresponding to the intent recognition results are high-risk movements, the system can intelligently identify operations that may have a significant impact on the patient or the treatment process. Because the lifting and lowering distance for high-risk movements is set to be greater than a preset distance, the system can distinguish between routine operations and potentially dangerous operations. Finally, for movements identified as high-risk, the system does not execute them immediately but instead sends a confirmation instruction to the operator and awaits their confirmation. This dual confirmation mechanism, which adds human intervention based on the system's judgment, greatly improves operational safety and avoids adverse consequences caused by misidentification or operational errors.

[0099] Through the above technical solution, this application can significantly improve the safety of the voice-interactive dental chair height control system. By conducting risk assessments for height adjustment movements and incorporating operator confirmation, it effectively avoids large-scale, high-risk height adjustments caused by voice recognition errors, environmental interference, or unintentional operator commands, thereby reducing the risk of harm to patients or damage to medical equipment. This human-machine collaborative decision-making model, while ensuring ease of operation, provides a more reliable and safer method for controlling dental chairs in the dental treatment environment.

[0100] This application also discloses a voice-interactive oral chair lifting control system, comprising: an acquisition device and a processing device; the acquisition device is used to acquire original sound information in the oral treatment environment; the processing device is used to process the original sound information to remove environmental noise components contained in the original sound information to obtain intermediate sound information; the processing device is used to identify working sound segments matching the preset working sound information from the intermediate sound information according to the preset working sound information of the medical device, and remove the working sound segments from the intermediate sound information to obtain pure sound information; the processing device is used to determine a first similarity for each preset oral chair control sound information among multiple preset oral chair control sound information; the first similarity is the similarity between the spectral feature information of the preset oral chair control sound information and the spectral feature information of the pure sound information; the processing device is used to select the preset oral chair control sound information with the largest first similarity as the identified oral chair control sound information; the processing device is used to drive the oral chair to perform the action corresponding to the oral chair control sound information when the identified oral chair control sound information meets the preset execution conditions.

[0101] In practice, the acquisition device can be one or more microphones configured to pick up raw sound information from the dental treatment environment. For example, the acquisition device could be a unidirectional microphone mounted near the headrest of the dental chair, primarily used to capture sound from the direction of the dentist.

[0102] The processing unit can be an embedded processor integrated within the dental chair control unit, directly receiving and processing the raw sound information transmitted from the acquisition device in real time. Alternatively, the processing unit can be a standalone computing unit, such as an industrial PC or server, communicating with the acquisition device and dental chair control module via a network interface for offline or semi-real-time sound information processing. Furthermore, the processing unit can include a dedicated digital signal processor or artificial intelligence acceleration chip to improve the efficiency and accuracy of sound processing and recognition. The output of the processing unit is used to drive the dental chair to perform corresponding actions.

[0103] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for controlling the height of a dental chair based on voice interaction, characterized in that, include: Acquire raw sound information from the oral treatment environment; The original sound information is processed to remove the environmental noise components contained in the original sound information to obtain intermediate sound information; Based on the preset working sound information of the medical device, a working sound segment that matches the preset working sound information is identified from the intermediate sound information, and the working sound segment is removed from the intermediate sound information to obtain pure sound information; For each preset dental chair control sound information among multiple preset dental chair control sound information, a first similarity is determined; the first similarity is the similarity between the spectral feature information of the preset dental chair control sound information and the spectral feature information of the clean sound information. The preset dental chair control sound information with the highest first similarity is used as the identified dental chair control sound information; When the identified dental chair control sound information meets the preset execution conditions, the dental chair is driven to perform the lifting and lowering action corresponding to the dental chair control sound information.

2. The method for controlling the height of a dental chair based on voice interaction according to claim 1, characterized in that, The original sound information is processed to remove environmental noise components to obtain intermediate sound information, including: Invoke the preset high-pass filter; The original sound information is input into a preset high-pass filter to remove sound information with frequencies lower than the preset frequency from the original sound information, thus obtaining the initial sound information; Invoke the preset Wiener filter model; The initial sound information is input into a preset Wiener filter model to remove noise of a preset bandwidth from the initial sound information to obtain intermediate sound information.

3. The method for controlling the height of a dental chair based on voice interaction according to claim 1, characterized in that, Based on the preset operating sound information of the medical device, a working sound segment matching the preset operating sound information is identified from the intermediate sound information, and the working sound segment is removed from the intermediate sound information to obtain clean sound information, including: The intermediate audio information is divided into multiple sub-intermediate audio segments; For each of the multiple sub-intermediate sound segments, a second similarity is determined; the second similarity is the similarity between the spectral feature information of the sub-intermediate sound segment and the spectral feature information of the preset working sound information. Sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold are taken as working sound segments, and the working sound segments are removed from the intermediate sound information to obtain clean sound information.

4. The method for controlling the height of a dental chair based on voice interaction according to claim 3, characterized in that, The intermediate audio information is divided into multiple sub-intermediate audio segments, including: Obtain the security level and primary correspondence of medical behaviors corresponding to medical devices; the primary correspondence includes a one-to-one correspondence between multiple security level ranges and multiple segment durations. The segment duration corresponding to the safety level range of the medical behavior in the first correspondence is taken as the target segment duration. The intermediate audio information is divided into multiple sub-intermediate audio segments based on the target segment duration; the duration of each sub-intermediate audio segment is the same as the target segment duration.

5. The method for controlling the height of a dental chair based on voice interaction according to claim 3, characterized in that, The medical device includes multiple medical devices, and there are multiple preset working sound information sets. Each preset working sound information set corresponds one-to-one with a specific medical device. Sub-intermediate sound segments with a second similarity greater than a preset first similarity threshold are used as working sound segments. These working sound segments are then removed from the intermediate sound information to obtain clean sound information, including: Sub-intermediate audio segments with a second similarity greater than a preset first similarity threshold are used as initial working audio segments; Acquire vibration information of the target medical device; the target medical device is the medical device corresponding to the target preset working sound information, and the target preset working sound information is the preset working sound information corresponding to the initial working sound segment; Based on the vibration information of the target medical device, determine whether the target medical device is in working condition; When the target medical device is in working condition, the initial working sound segment is identified as the working sound segment, and the working sound segment is removed from the intermediate sound information to obtain pure sound information; When the target medical device is not in operation, the initial working sound segment is determined to be a non-working sound segment.

6. The method for controlling the height of a dental chair based on voice interaction according to claim 5, characterized in that, Based on the vibration information of the target medical device, determine whether the target medical device is in a working state, including: Based on the vibration information of the target medical device, determine whether the vibration energy within the preset frequency range is greater than the preset vibration energy threshold. If the vibration energy within the preset frequency range is greater than the preset vibration energy threshold, the target medical device is determined to be in working condition; otherwise, the target medical device is determined to be in non-working condition.

7. The method for controlling the height of a dental chair based on voice interaction according to claim 6, characterized in that, When the vibration energy within a preset frequency range exceeds a preset vibration energy threshold, the target medical device is determined to be in a working state, including: When the vibration energy within the preset frequency range exceeds the preset vibration energy threshold, the target medical device is determined to be in the estimated working state. When the target medical device is in the estimated working state, acquire the current information of the target medical device; Determine whether the current information of the target medical device is greater than the no-load current information of the target medical device; When the current information of the target medical device is greater than the no-load current information of the target medical device, it is determined that the target medical device is in working condition.

8. The method for controlling the height of a dental chair based on voice interaction according to claim 1, characterized in that, When the identified dental chair control sound information meets the preset execution conditions, the dental chair is driven to perform a lifting action corresponding to the dental chair control sound information, including: The third similarity is determined; the third similarity is the similarity between the identified dental chair control sound information and the preset operating sound information of the medical device. If the third similarity is less than the first similarity corresponding to the identified dental chair control sound information, it is determined that the identified dental chair control command meets the preset execution conditions, and the dental chair is driven to perform the lifting action corresponding to the dental chair control sound information; otherwise, it is determined that the identified dental chair control sound information does not meet the preset execution conditions.

9. A method for controlling the height of a dental chair based on voice interaction according to claim 8, characterized in that, Driving the dental chair to perform a lifting action corresponding to the dental chair control sound information includes: Intent recognition is performed on the voice information controlling the dental chair to obtain the intent recognition result; Determine whether the lifting motion corresponding to the intent recognition result is a high-risk motion; the lifting distance of a high-risk motion is greater than the preset lifting distance. When the lifting or lowering action corresponding to the intent recognition result is not a high-risk action. Drive the dental chair to perform the lifting and lowering actions corresponding to the intent recognition results; When the lifting action corresponding to the intent recognition result is a high-risk action, a confirmation instruction is issued to the operator. After receiving the confirmation instruction from the operator, the dental chair is driven to perform the lifting action corresponding to the intent recognition result.

10. A voice-interactive-based dental chair lifting control system, characterized in that, include: Acquisition device and processing device; Acquisition device for acquiring raw sound information in the oral treatment environment; A processing device is used to process the original sound information to remove the environmental noise components contained in the original sound information and obtain intermediate sound information; The processing device is used to identify working sound segments that match the preset working sound information from intermediate sound information based on the preset working sound information of the medical device, and remove the working sound segments from the intermediate sound information to obtain pure sound information; The processing device is used to determine a first similarity for each of a plurality of preset dental chair control sound information; the first similarity is the similarity between the spectral feature information of the preset dental chair control sound information and the spectral feature information of the clean sound information; The processing device is used to take the preset dental chair control sound information with the highest first similarity as the identified dental chair control sound information; The processing device is used to drive the dental chair to perform an action corresponding to the dental chair control sound information when the identified dental chair control sound information meets the preset execution conditions.

Images