An ultrasonic tooth cleaning and periodontal examination electric toothbrush

Abstract

The application discloses an ultrasonic tooth cleaning and periodontal detection electric toothbrush, and belongs to the cross field of oral care and periodontal ultrasonic diagnosis. The application integrates daily cleaning and periodontal screening in the same handheld electric toothbrush, realizes a "cleaning-detection-feedback" closed loop, and reduces the probability that periodontal abnormalities are ignored in a home scene. Based on the recognition potential of high-frequency ultrasonic waves on soft tissues, inflammation and dental calculus, the application can output tissue state grades in a non-invasive manner in a consumer-level scene, and avoid discomfort and operation dependence caused by traditional manual probing.

Description

Technical Field

[0001] This invention belongs to the cross-technical field of oral care instruments, periodontal ultrasound diagnostic equipment and intelligent health monitoring equipment, and specifically relates to an electric toothbrush for ultrasonic teeth cleaning and periodontal testing. Background Technology

[0002] Currently, periodontal clinical examination still relies on manual periodontal probing and two-dimensional imaging. Publicly available literature shows that the results of manual probing are affected by factors such as probe morphology, probe tip thickness, insertion angle, pressure applied, and the degree of inflammation, resulting in limited repeatability. Excessive probing pressure can also lead to false-positive bleeding. The literature also points out that probing is invasive, causing poor patient comfort, and carries the risk of cross-contamination of bacteria into healthy grooves during periodontal pocket examination (Renaud et al., 2023; Rochefort, Denis and Renaud, 2025). Furthermore, while two-dimensional periapical and apical radiographs can assess alveolar bone changes, they are insufficient to directly display soft tissue, early inflammation, and the true periodontal pocket contour, and are inadequate in reflecting lingual / palatal bone defects and soft tissue conditions (Figueredo et al., 2023; Discepoli et al., 2025; Estrade et al., 2026).

[0003] Recent studies on periodontal ultrasound have shown that high-frequency ultrasound can visualize structures such as enamel, cementoenamel junction (CEJ), alveolar ridge, keratinized gingiva, connective tissue, and gingival sulcus / periodontal pocket in real time without ionizing radiation. Systematic reviews indicate that periodontal ultrasound is suitable for imaging superficial tissues with a depth less than 10 mm and dimensions in the sub-millimeter range, with optimal operating frequencies generally between 15 MHz and 40 MHz. When the frequency exceeds 20 MHz, it enters the high-resolution imaging range, with resolutions exceeding 100 μm (Estrade et al., 2026). 20 MHz and 40 MHz devices have been used for periodontal tissue measurement, CEJ localization, gingival thickness assessment, and alveolar bone-related parameter evaluation, demonstrating high reliability or good correlation with clinical / imaging reference methods (Chifor et al., 2015; Figueredo et al., 2023).

[0004] Further research indicates that high-frequency intraoral ultrasound can not only measure periodontal pockets but also classify tissues based on echogenicity and texture differences. Estrade et al. (2026) found significant differences in the average grayscale / echoic characteristics of different periodontal tissues, with inflamed tissues generally exhibiting hypoechoicity, while highly mineralized tissues such as enamel exhibit hyperechoicity. This study also observed that deep inflammation and supragingival / subgingival calculus could be identified in real-time images (Estrade et al., 2026). This suggests that introducing high-frequency detection units into home-use devices and jointly analyzing echogenicity, shadows, boundary continuity, and texture could provide an engineering basis for achieving "post-cleaning screening" or "screening while brushing."

[0005] From the perspective of oral care products, publicly available reviews and ADA data indicate that powered toothbrushes have established a mature product path in terms of safety and plaque / gingivitis control. ADA data states that for a powered toothbrush to obtain its Seal rating, it needs to pass tests related to electrical safety, bristle retention, mechanical strength, and chemical tolerance, and provide clinical evidence demonstrating its safety for use on oral hard tissues, soft tissues, and restorations, as well as its effectiveness in removing plaque and reducing gingivitis (American Dental Association, 2025). Regarding ultrasonic toothbrushes, publicly available reviews suggest that their core technology typically integrates piezoelectric elements within the brush head, using the inverse piezoelectric effect to drive the bristles or sound field, allowing microstreaming in the liquid to work in conjunction with the mechanical movement of the bristles for cleaning. Existing ultrasonic / sonic toothbrushes are generally superior to manual toothbrushes in maintaining oral health, but they do not necessarily have an overwhelming advantage compared to other powered brush types (Digel et al., 2020; Ng et al., 2020). Summary of the Invention

[0006] To address the technical problems mentioned above, the present invention provides an ultrasonic cleaning and periodontal detection electric toothbrush, comprising: a handle, a replaceable brush head assembly, a cleaning drive module, an ultrasonic detection module, a main control module, a power supply module, a pressure / attitude sensing module, a display / communication module, and a waterproof isolation structure; The workflow of the main control module includes: Based on user instructions or preset procedures, an interlocking switch is performed between cleaning mode and detection mode; In response to entering the detection mode, at least one valid frame is obtained by joint gating based on the contact pressure collected by the pressure / attitude sensing module, the parallel deviation angle between the longitudinal axis of the detection window and the tooth long axis, the vertical deviation angle between the surface normal of the detection window and the local gingival surface normal, and the coupling quality index. Based on the valid frame, the ultrasonic detection module is controlled to emit short pulses at the MHz level to the gingival margin, gingival sulcus and superficial tissue of the tooth being tested, and to receive the echo signal to reconstruct the B-mode image; Based on the B-mode image, extract structural parameters and state scores; Based on structural parameters and status scores, the organization status level is output.

[0007] Preferably, the main control module performs interlock switching according to the detection mode instruction. The interlock switching includes: shutting down or derating the cleaning drive module, then performing detection transducer excitation parameter loading, simulation front-end gain initialization, pressure and attitude self-check, and coupling status confirmation, before allowing echo acquisition.

[0008] Preferably, the main control module obtains a valid frame by joint gating based on the contact pressure, parallel deviation angle, vertical deviation angle, and coupling quality index. The joint gating includes determining the current frame as a valid frame only when the contact pressure is within a preset pressure window, the parallel deviation angle is less than a preset parallel angle threshold, the vertical deviation angle is less than a preset vertical angle threshold, and the coupling quality index is greater than a preset coupling threshold.

[0009] Preferably, the main control module performs B-mode image reconstruction using a delay-sum beamforming method based on the echo signal corresponding to the valid frame. The B-mode image reconstruction includes: delay-summing of multi-channel echoes, Hilbert envelope detection, logarithmic compression, and grayscale mapping.

[0010] Preferably, the main control module performs geometric registration of the superficial reflection line, cementoenamel junction candidate area, soft tissue boundary, and pocket bottom candidate area in the image domain based on the B-mode image, and extracts structural parameters, including the estimated periodontal pocket depth and the shadow attenuation ratio.

[0011] Preferably, the main control module outputs an organization status level based on structural parameters and status scores, wherein the organization status level includes: Level A, routine maintenance status; Level B, recommended local enhanced cleaning and retesting status; and Level C, recommended further professional inspection status.

[0012] Preferably, the main control module is gating the organization status level based on the overall confidence level. When the overall confidence level is lower than a preset confidence threshold, the organization status level is not output, and a prompt to re-detect or adjust the contact status is output.

[0013] Preferably, the main control module obtains the comprehensive confidence score by weighted fusion based on the average value of the effective frame gating criterion, the image quality score, the artifact intensity index, and the network output confidence score.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention integrates daily brushing and periodontal screening into a single handheld electric toothbrush, achieving a closed-loop "cleaning-detection-feedback" process and reducing the probability of periodontal abnormalities being overlooked in home settings. Based on the potential of high-frequency ultrasound to identify soft tissue, inflammation, and tartar, this invention can output tissue status levels non-invasively in consumer-grade settings, avoiding the discomfort and operational dependence of traditional manual probing. Simultaneously, it employs a dual-frequency architecture where the cleaning drive module and the detection ultrasound module operate at different frequency bands, avoiding confusion between cleaning acoustics and detection ultrasound, thus ensuring clear product function definition. Through joint gating of pressure / posture / coupling conditions, the effective frame ratio and repeatability of home self-examination are improved. Furthermore, by recording daily brushing data and periodontal status data over time, personalized gingival health curves can be generated, providing auxiliary data support for long-term follow-up and re-examination decisions. Attached Figure Description

[0015] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the toothbrush structure according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the system functional modules and signal chain according to an embodiment of the present invention; Figure 3 This is a flowchart illustrating the detection process according to an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] Example 1 like Figure 1 As shown, this invention provides a handheld electric toothbrush device, which includes a handle, a replaceable brush head assembly, a cleaning drive module, a detection ultrasonic module, a main control module, a power supply module, a pressure / attitude sensing module, a display / communication module, and a waterproof isolation structure. Figure 2As shown, the handle houses a battery, main control circuit board, power drive circuit, and wireless communication module. The front of the brush head assembly forms a bristle cleaning area, with a high-frequency ultrasonic detection window located on one side or around the perimeter of this area. The cleaning drive module and the ultrasonic detection module operate at different frequency bands; the cleaning drive module outputs kHz-level vibrations, while the ultrasonic detection module outputs high-frequency pulses in the range of 15MHz to 40MHz, achieving a clearly defined dual-frequency system. Figure 1 As shown, this embodiment employs a slender neck and a small probe area at the front end to allow entry into the mouth and scanning near the long axis of the tooth.

[0020] The brush head assembly preferably adopts a modular and detachable structure, comprising at least: First, a bristle base for routine plaque removal from the tooth surface, gingival margin, and proximal areas; Second, a detection window, which can be positioned on the front, side, or near the bristle tufts of the brush head and is acoustically connected to the transducer; Third, a guide section, which defines the contact angle and contact depth during detection; Fourth, a coupling medium output section, which can be a micropore, a thin film cavity, or a hydrophilic coating, used to form stable acoustic coupling between the detection window and the gingival surface. To reduce the obstruction of the detection window by brushing residue, the detection window can be slightly raised relative to the bristles or positioned on the side edge of the brush head to form an independent detection surface.

[0021] In a preferred embodiment, a piezoelectric element or a microarray transducer is disposed at the front end of the brush head, and a biocompatible acoustic membrane is used to cover the detection window; a damping isolation layer is disposed between the brush bristle vibration area and the detection window to suppress the coupling of mechanical vibration to the high-frequency detection channel during cleaning mode. To accommodate anterior teeth, posterior teeth, and lingual sites, the brush head can also be designed in three specifications: A-type straight head, B-type curved head, or C-type narrow head.

[0022] In cleaning mode, the main control module drives the cleaning drive unit. This unit can be a piezoelectric drive, a magnetostrictive drive, an eccentric motor drive, or a combination thereof. A preferred solution employs a piezoelectric / acoustic drive, causing the bristles to generate kHz-level high-frequency micro-vibrations, which directly contact the bristles to remove plaque, while simultaneously generating localized microfluidics and shearing effects within the saliva / toothpaste film.

[0023] In detection mode, the cleaning drive unit stops or switches to background mode, and the main control module switches to the detection transducer path. The detection transducer emits short pulses at the MHz level to the gingival margin, gingival sulcus, and superficial tooth tissue, and receives the echo signals. The analog front end amplifies, performs time gain compensation (TGC), filters, and performs analog-to-digital conversion on the echoes, which are then sent to the algorithm module for tissue boundary recognition, periodontal pocket estimation, hypoechoic area recognition, strong echo with shadow recognition, and risk classification. To improve imaging consistency, during detection, the device uses pressure sensor, IMU, and position sensor information to determine whether the conditions of "contact pressure within a safe window," "probe surface approximately parallel to the long axis of the tooth," and "detection window approximately perpendicular to the gingival surface" are met. Only when the thresholds are met are valid frames acquired.

[0024] Furthermore, the main control module is not only used for mode switching, but also for completing acquisition timing control, parameter adaptive configuration, effective frame gating, image reconstruction, feature extraction, network inference, confidence arbitration, and state output mapping. The main control module preferably includes a mode control unit, an excitation timing unit, an acquisition control unit, a signal preprocessing unit, an imaging reconstruction unit, a feature analysis unit, a network inference unit, a result arbitration unit, and a display and communication unit. Each unit can be integrated into the same MCU, DSP, FPGA, SoC, or a combination thereof, where the MCU is responsible for the state machine and peripheral scheduling, the DSP or FPGA is responsible for high-speed signal processing and beamforming, and the SoC or embedded AI accelerator is responsible for network model inference.

[0025] The mode control unit performs interlocked switching between cleaning mode and detection mode based on user button presses, preset brushing procedures, or mobile terminal commands. To avoid crosstalk between kHz-level cleaning vibrations and MHz-level detection channels, the main control module, upon entering detection mode, first disables or degrades the cleaning drive, then performs transducer excitation parameter loading, analog front-end gain initialization, pressure and attitude self-checks, and coupling status confirmation before allowing echo acquisition. This interlocked switching process can be described as follows: in, Indicates the detection mode is enabled; This indicates a detection trigger command; Indicates the cleaning drive status; Indicates the security gating status.

[0026] To improve the consistency of detection results in home scenarios, the main control module performs joint gating of pressure, attitude, and coupling adequacy, saving valid frames only when preset conditions are met. The pressure gating function, attitude gating function, and coupling gating function can be written as follows: Based on this, the joint gating function is written as: Therefore, the determination of the valid frame of the k-th frame is written as: in, This represents the contact pressure during the acquisition of the k-th frame. This indicates the parallel deviation angle between the longitudinal axis of the detection window and the long axis of the tooth. This indicates the vertical deviation angle between the normal to the surface of the detection window and the normal to the local gingival surface. This represents the coupling quality index for the k-th frame. The coupling quality index is a weighted average of the front surface echo energy, short-time signal-to-noise ratio, and interface continuity index. The main control module performs joint gating of pressure, attitude, and coupling adequacy, only gating when... and When the k-th frame is determined to be a valid frame and saved, it is otherwise determined to be an invalid frame and discarded. At that time, the first The frame enters the subsequent image reconstruction and feature analysis process; when At this time, the main control module controls the current frame not to be stored and triggers prompts for re-lamination, decompression, or angle adjustment.

[0027] Once the main control module completes the mode switching and determines that the current frame meets the valid acquisition conditions, it controls the detection transducer to perform ultrasonic excitation and echo acquisition, and sends the obtained valid frames to the analog front end, image reconstruction unit and algorithm analysis unit.

[0028] Based on the differences in tissue echogenicity in user-provided literature, this embodiment converts the detection results into a set of states that are understandable to the user. Preferred states include: State A "Normal / Maintenance," indicating that the estimated periodontal pocket depth is no greater than a set threshold and the local tissue is within the normal echogenicity range; State B "Suspected Inflammation," indicating the presence of a persistent hypoechoic area, blurred boundaries, or suspected soft tissue swelling, suggesting enhanced local cleaning and retesting; State C "Deep Risk / Caloculation," indicating the detection of deep hypoechoic areas, strong echoes with posterior shadows, calculus-like reflections, or a risk of deep periodontal pockets, prompting the user to seek medical attention as soon as possible. The above thresholds can be calibrated based on training data and regional populations.

[0029] like Figure 3 As shown, the above signal chain processing procedure, image reconstruction method, structural parameter calculation method, network model structure, and result fusion and output method are as follows: S1. Acquire the raw echo and reconstruct the B-mode image.

[0030] S2. Geometric registration is completed based on surface reflection, CEJ candidate points, alveolar ridge candidate lines, and soft tissue boundaries.

[0031] S3 extracts features such as local echo intensity, texture entropy, edge continuity, shadow length, relative depth, and temporal series stability.

[0032] S4 outputs the estimated periodontal pocket depth (PDestimate), inflammation probability (inflammationscore), and calculus probability (calculusscore).

[0033] S5. Map the results to device icons, light colors, and mobile reports. The algorithm can employ a rule engine, support vector machine, random forest, convolutional neural network, or a combination thereof.

[0034] This embodiment further introduces a quality control layer to remove or prompt for rescanning of low-confidence images caused by artifacts, insufficient coupling, overvoltage, and angle anomalies.

[0035] Furthermore, in this embodiment, the signal chain and algorithm flow are executed according to a fixed link: "pulse excitation and echo acquisition—analog front-end preprocessing—B-mode image reconstruction—geometric registration and structural parameter extraction—multi-task network inference—quality control and status output." The signal chain consists of a detection transducer, an analog front-end, an analog-to-digital conversion unit, a main control processor, and a display / communication module; the algorithm chain consists of an image reconstruction unit, a geometric registration unit, a structural segmentation unit, a parameter regression unit, a status scoring unit, a confidence assessment unit, and a fusion decision unit.

[0036] (1) Pulse excitation and multi-channel echo acquisition In detection mode, the main control module controls the detection transducer to emit a central frequency to the gingival margin, gingival sulcus, and superficial tooth tissue of the tested tooth. short pulses, Set at to Within the range. The transmitted signal is output by the excitation drive stage, and the detection transducer converts it into an acoustic pulse; the tissue echo is then converted into an electrical signal by the receiving unit of the detection transducer before entering the analog front end.

[0037] The excitation signal is written as: in, To excite the amplitude, For pulse envelope window function, The initial phase is used. For pulse-echo ultrasound, transmitting a short pulse and receiving the scattered echo, and treating the echo as a convolutional superposition of the system response and the scattered reflection sequence, is a common and standard modeling method.

[0038] No. The discrete echo of the receiving channel is written as: in, Indicates the first The system impulse response of the channel, This represents the equivalent reflection sequence from the tissue scatterer. Indicates the propagation delay. Indicates noise and interference terms. This indicates the length of the system impulse response.

[0039] (2) Simulate front-end preprocessing The analog front end sequentially performs low-noise amplification, variable gain amplification, time-gain compensation, bandpass filtering, and analog-to-digital conversion. To compensate for echo attenuation caused by increased propagation depth, the main control module adjusts the output according to the depth... Control gain compensation function After compensation and filtering, only when the first... Only when a frame meets the aforementioned contact condition determination and valid frame gating requirements can the corresponding multi-channel preprocessing sequence enter the subsequent image reconstruction and inference chain.

[0040] The time gain compensation function is written as: Then the time gain compensation i The channel signal is written as: in, Indicates sampling point The corresponding depth, The gain coefficient obtained from system calibration.

[0041] Then, the main control module performs bandpass filtering on the compensated signal to obtain the preprocessed echo sequence: in, For the impulse response of the bandpass filter, This represents the length of the bandpass filter.

[0042] Based on the effective frame gating results, the first The valid input sequence corresponding to the frame is written as: in, For the first Frame validity determination result. When hour, Proceed to the subsequent image reconstruction and algorithm analysis process; when When this happens, the current frame will not proceed to subsequent processing.

[0043] (3) Pattern Image Reconstruction For the pre-processed multi-channel echoes, the main control module uses a delayed summation beamforming method to reconstruct the B-mode image. Conventional B-mode imaging consists of summing the data from the focused receiving channels, Hilbert envelope detection, and logarithmic compression.

[0044] For imaging point r, the first i The focusing delay corresponding to the channel is denoted as The dynamic aperture weighting coefficient is denoted as The beamforming output is then written as: in, N Number of receive channels For the pre-processed first Channel echo sequence.

[0045] The radio frequency signal obtained by beamforming is then subjected to envelope detection, logarithmic compression, and grayscale mapping to form a B-mode image, the expression of which is written as: in, Represents the Hilbert transform. Represents the imaging point The envelope amplitude, This represents the grayscale intensity after logarithmic compression. For reference amplitude, To prevent extremely small positive numbers with a denominator of zero, Hilbert envelope detection and logarithmic compression are common processing steps in ultrasound B-mode image display. The image reconstruction unit outputs... This image serves as the input for subsequent geometric registration and network inference.

[0046] (4) Geometric registration and structural parameter extraction The main control module performs geometric registration of the superficial reflection line, cementoenamel junction candidate area, soft tissue boundary, and pocket bottom candidate area in the image domain, and calculates structural parameters. For the pocket depth parameter, the Euclidean distance between the pocket opening reference point and the bottom reference point is calculated along the local normal, and converted to the actual distance using calibration coefficients, as follows: in, Indicates the reference point for the opening of the trench bag. Indicates the reference point at the bottom of the trench bag. This is the calibration coefficient from the pixel size to the actual distance.

[0047] For strong reflection accompanied by shadowing characteristics, the shadow attenuation ratio is defined as: in, This represents the average gray level of a local area in front of a strong reflector. The average gray level of the shadow area behind it. To prevent extremely small positive numbers with a denominator of zero, The larger the value, the more significant the energy attenuation behind the image. The posterior acoustic shadow in an ultrasound image is essentially a reduction in the energy and grayscale of the echo behind a high-attenuation structure. The physical distance between two points in an image can usually be obtained by multiplying the Euclidean distance between pixels by the pixel scale factor.

[0048] (5) Multi-task network inference model To perform boundary segmentation, parameter regression, state scoring, and confidence estimation under limited computing power, the main control module employs a lightweight U-shaped encoder-decoder multi-task network. The network structure consists of an encoder, decoder, segmentation head, parameter regression head, state scoring head, and confidence head. The encoder is responsible for extracting multi-scale context and edge features, while the decoder is responsible for restoring spatial resolution and outputting structural representations. Let the input B-mode image be... The encoder is denoted as The decoder is denoted as Then the shared features and the decoded features are written as follows: The original structure of U-Net consists of a shrinking path and an expanding path, which is suitable for segmentation tasks; in multi-task learning, it is also a common practice to jointly optimize the loss of multiple tasks in the form of a weighted sum.

[0049] Based on this, the segmentation output, parametric regression output, state score output, and confidence output are respectively written as: in, This is a structural segmentation diagram. For continuous parameter regression results, This is a state score vector, containing low-echo state score, strong-reflection with shadow state score, and boundary integrity score. Output the confidence level of the current detection result. The Sigmoid function has the following expression: The Sigmoid function compresses the output to... arrive Intervals are suitable for expressing probabilistic segmentation results, state scores, and confidence levels.

[0050] The network employs a multi-task joint loss for parameter optimization during the training phase. The loss function is written as: in, To divide the loss, For continuous parameter regression loss, For state scoring loss, Loss is constrained by confidence level. The loss weight is determined by calibration.

[0051] (6) Quality control, integrated decision-making and status output The main control module does not directly output disease diagnosis conclusions. Instead, it integrates structural parameters, state scores, image quality indicators, and acquisition quality indicators to output tissue state parameters and their levels related to cleaning, maintenance, and retesting prompts. Using quality scores and confidence levels for result gating and suppressing output or triggering verification when reliability is low is a common reliability control approach in medical imaging AI.

[0052] Let the normalized trench depth parameter, low echo state score, shadow attenuation parameter, boundary continuity score, and quality score be respectively... , , , and The overall state value is then written as: in, .

[0053] Let the average value of the gating criterion for valid frames within the current detection segment be: The overall confidence level is then written as: in, , Rate the image quality. As an index of artifact intensity, Output confidence level for the network. to The fusion weights are obtained after calibration.

[0054] The main control module first performs result gating based on the overall confidence level. When When [the situation is as described], no grade result is output; only a prompt of "Re-detect" or "Adjust contact status" is output. Then, based on With threshold , Perform hierarchical mapping: Among them, Level A indicates routine maintenance, Level B indicates a recommendation for enhanced local cleaning and retesting, and Level C indicates a recommendation for further professional inspection. This forms a closed-loop signal chain and algorithm flow of "effective frame gating—image reconstruction—structural segmentation—parameter regression—state fusion—level output". The multi-task model output, image quality assessment, and prediction uncertainty are jointly used to control the reliability of the final output, which is consistent with existing medical image quality control and uncertainty gating practices.

[0055] (7) Safety and hygiene structure Considering that this device is used for both brushing teeth and testing, this embodiment preferably incorporates the following safety / hygiene designs: First, the testing window is covered with a disposable protective film, or a sterilizable and replaceable testing cap is used; second, a micro-liquid reservoir for the coupling medium is installed inside the device, quantitatively dispensing the water-based coupling fluid only during testing to avoid long-term residue; third, an acoustic isolation layer and a waterproof layer are installed inside the brush head, and the handle achieves the waterproof rating commonly found in home oral care devices; fourth, when the pressure exceeds the gingival safety threshold, the testing automatically stops and an alarm is issued.

[0056] Example 2 This embodiment provides a household integrated electric toothbrush. The handle is approximately 150 mm to 180 mm long and 20 mm to 35 mm wide at its maximum. The width of the brush head tip is preferably less than 10 mm to facilitate access to the posterior teeth and gingival margin areas. The handle houses a lithium battery, a main control PCB, a cleaning drive circuit, an ultrasonic excitation / receiving circuit, and a Bluetooth module. The brush head includes a bristle area and a detection window. A 20 MHz to 40 MHz high-frequency piezoelectric transducer is fixed behind the detection window, and the bristle area is driven by a kHz-level acoustic / mechanical drive module. After completing a regular brushing session, the user presses and holds the mode button to enter "detection mode." At this time, the motor stops, and the detection window slowly slides along the gingival margin. A pressure sensor detects whether the contact pressure is within a preset range, for example, 0.05 N to 0.30 N; if the threshold is exceeded, the device vibrates to alert the user and pauses data collection. The system generates an estimated periodontal pocket value, an inflammation probability value, and a tartar probability value for each tooth position and displays a color status map on the handle or a mobile device. This application can be used on both humans and animals.

[0057] Example 3 In this embodiment, the handle is the same as in Embodiment 2, but two replaceable brush heads are used. The first brush head is a daily cleaning brush head, retaining only the bristles and the cleaning drive coupling structure. The second brush head is a detection brush head, with no large area of ​​bristles at the front end, only narrow bristles, a detection window, and a thin cavity for the coupling medium. After brushing, the user replaces the detection brush head for rapid screening of the entire mouth or specific quadrants. This design helps reduce detection window contamination, lowers the difficulty of coupling design, and facilitates the formation of subordinate protection alongside the integrated brush head during the application process.

[0058] Example 4 This embodiment adds higher computing power SoC and wireless network functionality to Embodiment 2, supporting the uploading of raw frames or feature data to a tablet terminal. Dentists or hygienists can view B-mode images, CEJ locations, periodontal pocket contours, hypoechoic areas, and hyperechoic areas with shadows on the tablet, and correct and annotate the algorithm results to continuously optimize the model. This version is suitable for community screening, postoperative follow-up, or orthodontic / implant maintenance scenarios.

[0059] An acoustic cavity is housed within the brush head housing, with a detection window at its front. A piezoelectric transducer is mounted behind the detection window, followed by a backing layer and a matching layer. The bristle base is located to one side of the detection window and is separated from the acoustic cavity by a damping isolation layer. A hydrophilic reservoir or microchannel is positioned near the detection window to form a thin coupling layer during detection. A waterproof sealing ring and electrical connection terminals are located at the connection between the handle and the brush head. This structure allows the brush head to simultaneously meet the requirements of waterproofing, teeth cleaning, detection, and replaceability / maintenance.

[0060] The detection algorithm implementation steps are as follows: Step 1: After brushing your teeth, the user selects "Detection Mode". The system prompts that a coupling liquid film is formed on the surface of the brush head.

[0061] Step 2: The system uses the IMU and pressure sensor to confirm whether the brush head posture and contact pressure meet the acquisition conditions.

[0062] Step 3: The high-frequency transducer transmits pulses and receives echoes, and the analog front end performs amplification, filtering, TGC and A / D conversion.

[0063] Step 4: Reconstruct the superficial periodontal ultrasound image and extract surface reflection lines, CEJ candidate points, alveolar ridge candidate lines, soft tissue boundaries, and shadow features.

[0064] Step 5: The algorithm calculates the estimated periodontal pocket, inflammation score, and calculus score, and outputs the confidence level.

[0065] Step 6: When the confidence level is lower than the threshold, the system prompts for a rescan; when the risk is higher than the threshold, the system saves the site and sends a "recommend medical treatment" reminder.

[0066] Step 7: The test results are stored in association with the date, tooth position, and user ID for longitudinal comparison.

[0067] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. An ultrasonic electric toothbrush for teeth cleaning and periodontal testing, characterized in that, include: Handle, replaceable brush head assembly, cleaning drive module, ultrasonic detection module, main control module, power supply module, pressure / attitude sensing module, display / communication module, and waterproof isolation structure; The workflow of the main control module includes: Based on user instructions or preset procedures, an interlocking switch is performed between cleaning mode and detection mode; In response to entering the detection mode, at least one valid frame is obtained by joint gating based on the contact pressure collected by the pressure / attitude sensing module, the parallel deviation angle between the longitudinal axis of the detection window and the tooth long axis, the vertical deviation angle between the surface normal of the detection window and the local gingival surface normal, and the coupling quality index. Based on the valid frame, the ultrasonic detection module is controlled to emit short pulses at the MHz level to the gingival margin, gingival sulcus and superficial tissue of the tooth being tested, and to receive the echo signal to reconstruct the B-mode image; Based on the B-mode image, extract structural parameters and state scores; Based on structural parameters and status scores, the organization status level is output.

2. The ultrasonic scaling and periodontal testing electric toothbrush according to claim 1, characterized in that, The main control module performs interlock switching according to the detection mode command. The interlock switching includes: shutting down or derating the cleaning drive module, then loading the detection transducer excitation parameters, initializing the analog front-end gain, performing pressure and attitude self-checks, and confirming the coupling status before allowing echo acquisition.

3. The ultrasonic scaling and periodontal testing electric toothbrush according to claim 1, characterized in that, The main control module obtains valid frames by joint gating based on contact pressure, parallel deviation angle, vertical deviation angle, and coupling quality index. The joint gating includes determining the current frame as a valid frame only when the contact pressure is within a preset pressure window, the parallel deviation angle is less than a preset parallel angle threshold, the vertical deviation angle is less than a preset vertical angle threshold, and the coupling quality index is greater than a preset coupling threshold.

4. The ultrasonic scaling and periodontal testing electric toothbrush according to claim 1, characterized in that, The main control module performs B-mode image reconstruction using a delay-summing beamforming method based on the echo signal corresponding to the valid frame. The B-mode image reconstruction includes: delay-summing of multi-channel echoes, Hilbert envelope detection, logarithmic compression, and grayscale mapping.

5. The ultrasonic scaling and periodontal testing electric toothbrush according to claim 1, characterized in that, The main control module performs geometric registration of the superficial reflection line, cementoenamel junction candidate area, soft tissue boundary, and pocket bottom candidate area in the image domain based on the B-mode image, and extracts structural parameters, including the estimated periodontal pocket depth and the shadow attenuation ratio.

6. The ultrasonic scaling and periodontal testing electric toothbrush according to claim 1, characterized in that, The main control module outputs the organization status level based on structural parameters and status scores. The organization status levels include: Level A, routine maintenance status; Level B, local enhanced cleaning and retesting recommended status; and Level C, further professional inspection recommended status.

7. The ultrasonic scaling and periodontal testing electric toothbrush according to claim 6, characterized in that, The main control module is gating the organization status level based on the overall confidence level. When the overall confidence level is lower than the preset confidence threshold, the organization status level is not output, and a prompt to re-detect or adjust the contact status is output.

8. The ultrasonic scaling and periodontal testing electric toothbrush according to claim 7, characterized in that, The main control module obtains the comprehensive confidence score by weighted fusion based on the average value of the effective frame gating criteria, the image quality score, the artifact intensity index, and the network output confidence score.

Images