Dietary behavior monitoring method, earphone and storage medium

Abstract

The invention relates to the technical field of diet monitoring, in particular to a diet behavior monitoring method, an earphone and a storage medium. The method comprises the steps of firstly performing vibration sensing on a target object to obtain a sensing data stream; performing noise frequency band screening on the sensing data stream to obtain a target sensing signal; sending the target sensing signal to a target terminal, and performing feature extraction on the target sensing signal through the target terminal to obtain a target behavior feature of the target object; and behavior pattern recognition is performed on the target behavior characteristics in the target terminal to determine a dietary behavior pattern of the target object, and dietary behavior analysis is performed based on the dietary behavior pattern and the target behavior characteristics to obtain a dietary behavior monitoring result. In this way, on the premise that daily wearing habits of the user are changed as little as possible, non-inductive, continuous and accurate digital capture can be carried out on tiny diet actions of the human body.

Description

Technical Field

[0001] This application relates to the field of dietary monitoring technology, and in particular to a method for monitoring dietary behavior, as well as an earphone and a storage medium. Background Technology

[0002] Currently, in the field of dietary monitoring technology, there is an irreconcilable obstacle between the invasiveness of monitoring methods and the continuity and accuracy of data collection.

[0003] Existing dietary monitoring methods rely on smartphone cameras or microphones, attempting to circumvent the complexities of hardware integration through non-contact sensing. However, this approach exposes users to privacy risks related to image exposure and voice monitoring. Furthermore, in actual dining scenarios, variations in ambient light, background noise interference, and limitations in shooting angles make it difficult for such solutions to generate a stable and reliable data stream, hindering their usability as everyday tools. While dedicated sensor devices worn on the wrist, neck, or head are physically close to the eating organs, their design often fails to conform to ergonomic principles, adding extra burden to users' existing habits and resulting in a degraded user experience.

[0004] Based on this, how can we achieve seamless, continuous, and accurate digital capture of subtle eating movements of the human body while changing users' daily wearing habits as little as possible? Summary of the Invention

[0005] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a method for monitoring eating behavior, as well as headphones and a storage medium, which can achieve imperceptible, continuous, and accurate digital capture of subtle eating movements of the human body with minimal changes to the user's daily wearing habits.

[0006] The dietary behavior monitoring method according to the first aspect of this application, applied to a dietary behavior monitoring headset, includes: Vibration sensing is performed on the target object to obtain a sensing data stream; The target sensing signal is obtained by filtering out noise frequency bands from the sensing data stream. The target perception signal is sent to the target terminal so that the target terminal can extract features from the target perception signal to obtain the target behavior features of the target object; In the target terminal, behavioral pattern recognition is performed on the target behavioral characteristics to determine the dietary behavior pattern of the target object. Based on the dietary behavior pattern and the target behavioral characteristics, dietary behavior analysis is performed to obtain dietary behavior monitoring results.

[0007] According to some embodiments of this application, the dietary behavior monitoring headphones use a six-axis inertial measurement unit as a vibration sensor, and the vibration sensing of the target object to obtain a sensing data stream includes: Obtain the target sampling sequence; Based on the target sampling sequence, the linear vibration generated by the jaw movement of the target object is captured by the six-axis inertial measurement unit to obtain the linear vibration sensing component; Based on the target sampling sequence, the rotation component of the target object's jaw opening and closing is captured by the six-axis inertial measurement unit to obtain the opening and closing rotation sensing component; By integrating the linear vibration sensing component and the opening / closing rotation sensing component, the vibration sensing data corresponding to the target sampling step sequence of the sensing data stream is obtained.

[0008] According to some embodiments of this application, the integration of the linear vibration sensing component and the opening / closing rotation sensing component to obtain the vibration sensing data corresponding to the target sampling step sequence in the sensing data stream includes: Obtain the characterization conditions for linear vibration of chewing and the characterization conditions for chewing opening and closing rotation; The amplitude of the linear vibration sensing component is calculated based on each target sampling step sequence to obtain the corresponding linear vibration amplitude. If the linear vibration amplitude satisfies the chewing linear vibration characterization condition, the target sampling step sequence corresponding to the linear vibration amplitude is determined as the key verification step sequence; Based on the opening and closing rotation sensing components of each key verification step, the rotation angle is calculated to obtain the corresponding opening and closing rotation angle. If the opening and closing rotation angle of the key verification step satisfies the chewing opening and closing rotation characterization condition, the linear vibration sensing component and the opening and closing rotation sensing component corresponding to the key verification step are integrated to obtain the vibration sensing data of the sensing data stream corresponding to each key verification step.

[0009] According to some embodiments of this application, the step of obtaining the target sampling sequence includes: Obtain computational resource constraints and power consumption limits; The chewing vibration spectrum of the target object was measured to obtain a vibration test spectrum sample. Spectral analysis was performed on the vibration test spectrum sample to obtain the test spectrum analysis results; The sampling rate is calculated based on the test spectrum analysis results to obtain the theoretical sampling rate; Based on the computational resource constraints and the power consumption constraints, the theoretical sampling rate is adjusted to the target vibration sampling rate; Based on the target vibration sampling rate, the target sampling sequence is set.

[0010] According to some embodiments of this application, the dietary behavior monitoring headphones further include an electromyography sensor, and the step of sensing vibrations in the target object to obtain a sensing data stream further includes: The vibration sensor detects vibration in the target object, resulting in a vibration sensing data stream; wherein the sampling rate of the vibration sensing data stream is the target vibration sampling rate. The electromyography (EMG) sensor is used to detect the electromyography of the target object, thereby obtaining an EMG data stream; wherein the sampling rate of the EMG data stream is the target EMG sampling rate. Based on the target vibration sampling rate and the target electromyography sampling rate, the sampling ratio relationship is determined; Based on the sampling ratio relationship, the vibration sensing data stream and the electromyography sensing data stream are aligned; The aligned vibration sensing data stream and electromyography sensing data stream are identified as the sensing data stream.

[0011] According to some embodiments of this application, the step of filtering out noise frequency bands in the sensing data stream to obtain the target sensing signal includes: Based on the vibration sensing data corresponding to each target sampling step sequence in the sensing data stream, a target vibration sensing spectrum is generated. Spectral analysis is performed on the vibration sensing spectrum of the target to obtain the target spectral analysis results; Based on the target spectrum analysis results, a filtering and denoising operation is performed on the sensing data stream to obtain the target sensing signal.

[0012] According to some embodiments of this application, the step of performing filtering and denoising operations on the sensing data stream based on the target spectrum analysis results to obtain the target sensing signal includes: Obtain the target vibration sampling rate corresponding to the target sampling step sequence; A low-pass filter cutoff frequency is set based on the target vibration sampling rate, and the sensing data stream is subjected to low-pass filtering based on the low-pass filter cutoff frequency to obtain a first intermediate data stream. Based on the target spectrum analysis results, the target interference frequency is determined; Based on the target interference frequency, preset notch filter parameters are set, and the first intermediate data stream is processed point by point through the notch filter with the set parameters to obtain the second intermediate data stream. Random noise removal processing is performed on the second intermediate data stream to obtain the target sensing signal.

[0013] The dietary behavior monitoring method according to a second aspect embodiment of this application, applied to a target terminal, includes: Target perception signal is received from the diet behavior monitoring earphone; wherein, the target perception signal is obtained in the diet behavior monitoring earphone through the following steps: vibration sensing is performed on the target object to obtain a perception data stream, and noise frequency band is filtered out from the perception data stream to obtain the target perception signal; Feature extraction is performed on the target perception signal to obtain the target behavior features of the target object; Behavioral pattern recognition is performed on the target behavioral characteristics to determine the dietary behavior pattern of the target object. Based on the dietary behavior pattern and the target behavioral characteristics, dietary behavior analysis is performed to obtain dietary behavior monitoring results.

[0014] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the dietary behavior monitoring method as described in any one of the embodiments of the first aspect of this application.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium storing a program that is executed by a processor to implement the dietary behavior monitoring method as described in any one of the embodiments of the first aspect of this application.

[0016] The dietary behavior monitoring method, earphone, and storage medium according to the embodiments of this application have at least the following beneficial effects: The dietary behavior monitoring method according to a first aspect embodiment of this application includes: This application requires first using dietary behavior monitoring headphones to sense vibrations in a target object, obtaining a sensing data stream; then filtering out noise frequencies from the sensing data stream to obtain the target sensing signal; finally, sending the target sensing signal to a target terminal, where feature extraction is performed to obtain the target object's behavioral characteristics; next, behavioral pattern recognition is performed on the target behavioral characteristics in the target terminal to determine the target object's dietary behavior pattern; and finally, dietary behavior analysis is performed based on the dietary behavior pattern and target behavioral characteristics to obtain the dietary behavior monitoring results. In this way, it is possible to achieve imperceptible, continuous, and accurate digital capture of subtle dietary movements with minimal changes to the user's daily wearing habits.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 A schematic flowchart of a dietary behavior monitoring method provided in an embodiment of this application; Figure 2 Another flowchart illustrating the dietary behavior monitoring method provided in this application embodiment; Figure 3 Another flowchart illustrating the dietary behavior monitoring method provided in this application embodiment; Figure 4 Another flowchart illustrating the dietary behavior monitoring method provided in this application embodiment; Figure 5 Another flowchart illustrating the dietary behavior monitoring method provided in this application embodiment; Figure 6 Another flowchart illustrating the dietary behavior monitoring method provided in this application embodiment; Figure 7 Another flowchart illustrating the dietary behavior monitoring method provided in this application embodiment; Figure 8 This is a schematic diagram of the hardware structure of the eating behavior monitoring earphone provided in an embodiment of this application; Figure 9 This is a schematic diagram of another hardware structure of the dietary behavior monitoring earphone provided in this application embodiment. Detailed Implementation

[0019] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0020] In the description of this application, "several" means one or more, "more than" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. The use of "first" and "second" in the description is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features.

[0021] In the description of this application, it should be understood that the orientation descriptions, such as up, down, left, right, front, and back, are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0022] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0023] In the description of this application, it should be noted that, unless otherwise explicitly defined, terms such as "setting," "installation," and "connection" should be interpreted broadly. Those skilled in the art can reasonably determine the specific meaning of the above terms in this application based on the specific content of the technical solution. Furthermore, the identification of specific steps in the following text does not imply a limitation on the order of steps or execution logic. The execution order and logic between each step should be understood and inferred from the content described in the embodiments.

[0024] Currently, in the field of dietary monitoring technology, there is an irreconcilable obstacle between the invasiveness of monitoring methods and the continuity and accuracy of data collection.

[0025] On the one hand, existing dietary monitoring methods rely on smartphone cameras or microphones, attempting to avoid the complexity of hardware integration through non-contact sensing. However, this approach exposes users to privacy risks such as image exposure and voice monitoring. Furthermore, in actual dining scenarios, changes in ambient light, background noise interference, and limitations in shooting angles make it difficult for such solutions to form a stable and reliable data stream, thus hindering their ability to become truly usable tools in daily life.

[0026] On the other hand, while dedicated sensor devices worn on the wrist, neck, or head are physically close to the organs of ingestion, their design is often difficult to conform to ergonomics, adding extra burden to users' existing wearing habits and resulting in a decline in user experience.

[0027] On the other hand, even smart headphones suffer from the limitation of "functional generalization" in their built-in sensors and algorithm architecture. That is, traditional smart headphones are optimized to recognize the inertial characteristics of macroscopic movements such as running and jumping, but completely ignore the complex kinematic nature of chewing, a specific physiological activity with micro-amplitude, periodicity, and multi-degree-of-freedom coupling. This leads to a mismatch in the configuration of sensor range, bandwidth, and resolution, causing the subtle vibrations of the jaw to be submerged in the system noise at the analog-to-digital conversion front end.

[0028] The aforementioned existing technologies all point to a pressing technical challenge: how to achieve seamless, continuous, accurate, and privacy-secure digital capture of subtle eating movements while minimizing changes to users' daily wearing habits.

[0029] The core technical challenge this solution addresses is accurately capturing and identifying the faint signals generated during chewing in the specific location of the ear. The jaw movement during chewing is minimal, resulting in extremely weak vibrations and displacements transmitted to the ear. Simultaneously, everyday actions such as head turning, walking, and talking generate acceleration interference ranging from a few grams to tens of grams, leading to signal strength differences of tens or even hundreds of times. Extracting these faint chewing signals from such strong background noise is the first technical hurdle. This requires a sensor with sufficiently high resolution and sensitivity, while also ensuring good contact between the sensor and the ear canal through structural design—neither too loose, resulting in signal loss, nor too tight, generating additional noise.

[0030] Secondly, regarding data processing, chewing is not an isolated action; it's interspersed throughout various daily activities. The algorithm must be able to distinguish between different behaviors such as chewing, speaking, coughing, and walking in real time. Due to the small size and limited battery capacity of the earphones, it's impossible to transmit all the raw data to the phone for processing. Signal preprocessing and feature extraction must be performed inside the earphones. This requires real-time computation with extremely low power consumption, placing high demands on the performance of the microprocessor and the efficiency of the algorithm. Furthermore, different people have significantly different chewing habits, dental conditions, and ear canal structures, resulting in varying signal characteristics. The algorithm needs to be adaptive, capable of building a personalized recognition model for each user, but it cannot store large amounts of data or perform complex training on the device to avoid increasing power consumption and latency.

[0031] Secondly, there's the balance between overall system power consumption and user experience. To ensure signal integrity, the sensor needs to sample at a high frequency, generating a large amount of data. If all data is transmitted wirelessly, the battery will be quickly depleted. Therefore, the data must be compressed at the headset end to extract key features before transmission. However, this presents a new trade-off: excessive compression results in lost details and affects recognition accuracy; retaining too much increases power consumption. Simultaneously, Bluetooth connectivity needs to remain stable, but maintaining a high-speed connection continuously also consumes significant power. The system needs to dynamically switch between deep sleep and rapid response based on the user's actual usage, ensuring the device can be worn all day while providing timely feedback during use. These challenges are interconnected and require collaborative optimization of hardware, algorithms, communication protocols, and user experience design to create a truly practical product.

[0032] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a method for monitoring eating behavior, as well as headphones and a storage medium, which can achieve imperceptible, continuous, and accurate digital capture of subtle eating movements of the human body with minimal changes to the user's daily wearing habits.

[0033] The following explanation is based on the accompanying drawings.

[0034] Reference Figure 1 The dietary behavior monitoring method according to the embodiments of this application, applied to dietary behavior monitoring headphones, may include: Step S101: Vibration sensing is performed on the target object to obtain a sensing data stream; Step S102: Noise frequency bands are filtered out from the sensing data stream to obtain the target sensing signal; Step S103: Send the target perception signal to the target terminal so that the target terminal can extract features from the target perception signal to obtain the target behavior features of the target object. Step S104: In the target terminal, perform behavioral pattern recognition based on the target behavioral characteristics to determine the dietary behavior pattern of the target object. Based on the dietary behavior pattern and the target behavioral characteristics, perform dietary behavior analysis to obtain dietary behavior monitoring results.

[0035] The dietary behavior monitoring method according to the embodiments of this application, when applied to a target terminal, may include: Step S105: Receive target perception signals from the diet behavior monitoring headphones, and extract features from the target perception signals to obtain the target behavior features of the target object; Step S106: Perform behavioral pattern recognition on the target behavioral characteristics to determine the dietary behavior pattern of the target object. Based on the dietary behavior pattern and the target behavioral characteristics, perform dietary behavior analysis to obtain dietary behavior monitoring results.

[0036] This application requires first sensing vibrations in the target object to obtain a sensing data stream; then filtering out noise frequencies in the sensing data stream to obtain the target sensing signal; sending the target sensing signal to a target terminal for feature extraction to obtain the target object's behavioral characteristics; finally, performing behavioral pattern recognition on the target behavioral characteristics in the target terminal to determine the target object's eating behavior pattern; and then analyzing the eating behavior based on the eating behavior pattern and target behavioral characteristics to obtain the eating behavior monitoring results. In this way, it is possible to achieve imperceptible, continuous, and accurate digital capture of subtle eating movements of the human body with minimal changes to the user's daily wearing habits.

[0037] In step S101 of some embodiments, vibration sensing is performed on the target object to obtain a sensing data stream; It should be noted that the dietary behavior monitoring method in this application first addresses the core issue of how to imperceptibly capture dietary movements in everyday wearable devices by starting with the key step of "vibration sensing of the target object to obtain a sensing data stream." By integrating a high-precision vibration sensor, such as a six-axis inertial measurement unit, into the dietary behavior monitoring earphone, this sensor can capture in real time the minute vibrations generated by chewing movements in the ear. These vibration signals differ significantly in frequency, amplitude, and pattern from signals generated by other daily head movements, such as speaking, providing a foundation for subsequent signal processing. The vibration sensing step not only ensures the continuity of data acquisition but also avoids the privacy issues that may arise from traditional camera or microphone monitoring methods, as the vibration signal itself does not involve any personally identifiable information.

[0038] Reference Figure 2 According to some embodiments of this application, a six-axis inertial measurement unit is used as a vibration sensor in the diet behavior monitoring headphones. Step S101, which involves vibration sensing of the target object to obtain a sensing data stream, may include: Step S201: Obtain the target sampling sequence; Step S202: Based on the target sampling sequence, the linear vibration generated by the jaw movement of the target object is captured by a six-axis inertial measurement unit to obtain the linear vibration sensing component; Step S203: Based on the target sampling sequence, the rotation component of the target object's jaw opening and closing is captured by a six-axis inertial measurement unit to obtain the opening and closing rotation sensing component. Step S204: Integrate the linear vibration sensing component and the opening / closing rotation sensing component to obtain the vibration sensing data corresponding to the target sampling sequence in the sensing data stream.

[0039] In the embodiments of this application, the diet behavior monitoring headphones utilize a six-axis inertial measurement unit (IMU) as a vibration sensor to perceive vibrations in the target object. The diet behavior monitoring headphones select a six-axis IMU as the vibration sensor to perceive vibrations in the target object. Chewing involves not only linear movements of the jaw but also complex rotational movements. For example, the opening and closing of the jaw during chewing is primarily a linear movement, while the lateral grinding, forward and backward movement, and rotation of the temporomandibular joint are rotational movements. These movements generate minute vibrations in the ear with different directions and amplitudes, and the six-axis IMU can simultaneously capture these multi-dimensional motion information, providing more comprehensive sensory data.

[0040] In some embodiments, step S201 involves obtaining the target sampling sequence; It should be noted that, in the embodiments of this application, the dietary behavior monitoring headphones employ a six-axis inertial measurement unit as a vibration sensor to perceive the vibration of the target object, thereby acquiring a sensing data stream. This process begins with acquiring the target sampling sequence. The determination of the target sampling sequence is based on the analysis of the characteristics of the chewing action signal, particularly the consideration of the signal frequency range. The vibration signal generated by the chewing action is mainly concentrated in the low-frequency range, such as 1 to 3 Hz, and may also contain high-frequency impact components, such as the 30 to 50 Hz signal generated by tooth collision. According to the Nyquist sampling theorem, the sampling frequency should be at least twice the highest frequency of the signal. Therefore, in some embodiments, the target sampling sequence needs to be set above 100 Hz to ensure that these signals can be completely captured. In practical applications, in order to improve signal fidelity and provide sufficient data redundancy for subsequent processing, a higher sampling frequency, such as 200 Hz, is usually selected. This setting of the sampling sequence lays the foundation for the accurate acquisition of subsequent vibration signals.

[0041] Reference Figure 3 According to some embodiments of this application, step S201, obtaining the target sampling sequence, may include: Step S301: Obtain computing resource constraints and power consumption limits; Step S302: Perform chewing vibration spectrum measurement on the target object to obtain vibration test spectrum sample; Step S303: Perform spectral analysis on the vibration test spectrum sample to obtain the test spectrum analysis results; Step S304: Calculate the sampling rate based on the test spectrum analysis results to obtain the theoretical sampling rate; Step S305: Adjust the theoretical sampling rate to the target vibration sampling rate based on computational resource constraints and power consumption limits. Step S306: Based on the target vibration sampling rate, set the target sampling sequence.

[0042] In some embodiments, step S301 involves obtaining computational resource constraints and power consumption constraints; It's important to note that, firstly, computational resource constraints and power consumption limits need to be identified. This is because monitoring headphones typically rely on limited battery capacity and the computing power of microcontrollers, and these constraints directly impact the efficiency of data acquisition and processing. For example, if the headphones use a low-power microcontroller, its processing speed and memory capacity are limited. Therefore, a balance needs to be found between the sampling rate and data processing complexity to ensure the device can operate stably for extended periods.

[0043] In step S302 of some embodiments, a chewing vibration spectrum measurement is performed on the target object to obtain a vibration test spectrum sample; It should be noted that the chewing vibration spectrum is measured on the target object to obtain vibration test spectrum samples. This process involves collecting vibration signals generated by actual chewing actions to obtain their spectral characteristics. By deploying high-precision data acquisition equipment at the ear location, chewing samples from different users in different eating scenarios are collected. These samples will include the fundamental frequency of chewing actions (e.g., 1 to 3 Hz) and high-frequency impact components (e.g., 30 to 50 Hz). These vibration test spectrum samples provide the basic data for subsequent spectral analysis.

[0044] In some embodiments, step S303 involves performing spectral analysis on the vibration test spectrum sample to obtain the test spectrum analysis results. It should be noted that spectral analysis is performed on the vibration test spectrum samples to obtain the test spectrum analysis results. The purpose of spectral analysis is to determine the main frequency components and their distribution of the chewing vibration signal. Using methods such as Fast Fourier Transform, the time-domain signal can be converted into a frequency-domain signal, thus clearly showing the energy distribution of different frequency components. For example, the analysis results may show that the dominant frequencies of chewing motion are concentrated between 1 and 3 Hz, while high-frequency impacts mainly occur between 30 and 50 Hz. These analysis results provide a scientific basis for determining the appropriate sampling rate.

[0045] In some embodiments, step S304 involves calculating the sampling rate based on the test spectrum analysis results to obtain the theoretical sampling rate. It should be noted that the theoretical sampling rate is calculated based on the test spectrum analysis results. According to the Nyquist sampling theorem, the sampling rate should be at least twice the highest frequency of the signal to avoid aliasing. For example, if the test spectrum analysis results show that the highest frequency of the chewing signal is 50 Hz, then the theoretical sampling rate should be at least 100 Hz. However, to ensure signal integrity and the accuracy of subsequent processing, a higher sampling rate, such as 200 Hz, is usually chosen to provide richer signal details.

[0046] In step S305 of some embodiments, the theoretical sampling rate is adjusted to the target vibration sampling rate based on computational resource constraints and power consumption constraints. It's important to note that determining the theoretical sampling rate also requires consideration of computational resource and power consumption constraints. Therefore, based on these constraints, the theoretical sampling rate is adjusted to the target vibration sampling rate. For example, if the microcontroller's processing power and battery capacity are limited, the theoretical sampling rate may need to be adjusted from 200 Hz to 160 Hz or lower to ensure the device can operate for extended periods without frequent charging. This adjustment process requires finding the optimal balance between signal fidelity and device performance.

[0047] In some embodiments, step S306 involves setting a target sampling sequence based on the target vibration sampling rate.

[0048] It should be noted that a target sampling sequence is set based on the target vibration sampling rate. Setting the target sampling sequence ensures the synchronization and stability of data acquisition. For example, if the target vibration sampling rate is 160 Hz, and the sampling sequence (time interval) is expressed as 1 / sampling frequency, then the sampling sequence can be set to acquire data once every 6.25 milliseconds. This setting not only ensures the accuracy of data acquisition but also provides a stable input for subsequent data processing and analysis. Through this series of steps, the dietary behavior monitoring headset can achieve efficient and accurate acquisition of chewing vibration signals while meeting practical application requirements.

[0049] In step S202 of some embodiments, based on the target sampling sequence, the linear vibration generated by the jaw movement of the target object is captured by a six-axis inertial measurement unit to obtain the linear vibration sensing component. It should be noted that, based on a defined target sampling sequence, the six-axis inertial measurement unit (IMU) begins to capture the linear vibrations generated by the mandibular movements of the target object. The triaxial accelerometer within the IMU measures the linear acceleration changes of the mandible in three orthogonal directions (typically the X, Y, and Z axes). These acceleration changes reflect the linear motion characteristics of the mandible during chewing. For example, when a user chews food, the up-and-down movement of the mandible generates an acceleration signal in the vertical direction (Z-axis), while the left-and-right movement generates corresponding acceleration signals in the horizontal directions (X and Y axes). By sampling these acceleration signals according to the target sampling sequence, linear vibration sensing components can be obtained. These components contain important information such as the amplitude, frequency, and direction of the chewing motion, providing fundamental data for subsequent analysis of the force, rhythm, and pattern of chewing.

[0050] In step S203 of some embodiments, based on the target sampling sequence, the rotation component of the target object's jaw opening and closing is captured by a six-axis inertial measurement unit to obtain the opening and closing rotation sensing component. It should be noted that, based on the target sampling sequence, the six-axis inertial measurement unit (IMU) also captures the rotational components of the target object's mandible during opening and closing. The three-axis gyroscope within the six-axis IMU measures the angular velocity changes of the mandible around three orthogonal axes during opening and closing. These angular velocity signals reflect the rotational motion characteristics of the mandible; for example, the opening and closing motion of the mandible primarily generates rotation around the vertical axis (Z-axis), while lateral chewing motion generates rotational components around the horizontal axes (X-axis and Y-axis). By sampling these angular velocity signals according to the target sampling sequence, the opening and closing rotational sensing components can be obtained. These rotational components not only provide information on the rotational angle and velocity of chewing motions but also help distinguish different chewing patterns, such as unilateral and bilateral chewing, and assess the uniformity and coordination of chewing.

[0051] In step S204 of some embodiments, the linear vibration sensing component and the opening and closing rotation sensing component are integrated to obtain the vibration sensing data corresponding to the target sampling sequence of the sensing data stream.

[0052] It should be noted that the linear vibration sensing component and the opening / closing rotation sensing component are integrated to obtain the vibration sensing data corresponding to the target sampling step sequence in the sensing data stream. This integration process needs to ensure the consistency of the two components in time and space. Since linear vibration and rotational motion occur simultaneously, some embodiments require aligning the accelerometer and gyroscope data at the same timestamp to accurately reflect the complete characteristics of chewing actions. The integrated vibration sensing data contains comprehensive information about chewing actions, including the amplitude, frequency, direction, and velocity of linear and rotational motions. This data provides a rich foundation for subsequent signal processing and behavior analysis, enabling the system to more accurately identify and analyze eating behavior patterns, thereby providing users with detailed feedback and suggestions regarding their eating habits.

[0053] Reference Figure 4 According to some embodiments of this application, step S204 integrates the linear vibration sensing component and the opening / closing rotation sensing component to obtain the vibration sensing data corresponding to the target sampling step sequence of the sensing data stream, which may include: Step S401: Obtain the characterization conditions for linear vibration of chewing and the characterization conditions for opening and closing rotation of chewing. Step S402: Calculate the amplitude based on the linear vibration sensing component of each target sampling step to obtain the corresponding linear vibration amplitude. Step S403: If the linear vibration amplitude satisfies the chewing linear vibration characterization condition, the target sampling step sequence corresponding to the linear vibration amplitude is determined as the key verification step sequence. Step S404: Calculate the rotation angle based on the opening and closing rotation sensing components of each key verification step to obtain the corresponding opening and closing rotation angle. Step S405: If the opening and closing rotation angle of the key verification steps meets the chewing opening and closing rotation characterization condition, integrate the linear vibration sensing component and the opening and closing rotation sensing component corresponding to the key verification steps to obtain the vibration sensing data of the sensing data stream corresponding to each key verification step.

[0054] In the embodiments of this application, the process of integrating linear vibration sensing components and opening / closing rotation sensing components to obtain a sensing data stream is a sophisticated multi-step operation designed to ensure that only signals highly correlated with chewing actions are integrated and analyzed.

[0055] In some embodiments, step S401 involves obtaining the characterization conditions for linear vibration of chewing and the characterization conditions for opening and closing rotation of chewing. It should be noted that this process begins with obtaining characterization conditions for linear chewing vibrations and chewing opening-closing rotation. These characterization conditions are thresholds or characteristic patterns set based on in-depth research into the physical properties of chewing movements, used to distinguish chewing movements from vibration signals generated by other non-chewing movements (such as speaking, nodding, etc.). For example, characterization conditions for linear chewing vibrations may include the range of vibration amplitude, frequency characteristics, and periodicity patterns of vibration; while characterization conditions for chewing opening-closing rotation may involve specific patterns of rotation angle, rotation speed, and rotation direction.

[0056] In step S402 of some embodiments, the amplitude is calculated based on the linear vibration sensing component of each target sampling step sequence to obtain the corresponding linear vibration amplitude. It should be noted that the amplitude is calculated based on the linear vibration sensing component of each target sampling step to obtain the corresponding linear vibration amplitude. This step utilizes accelerometer data from the six-axis inertial measurement unit to calculate the linear vibration amplitude at each sampling point. By analyzing the accelerometer outputs in the three orthogonal directions (X, Y, and Z axes), a comprehensive vibration amplitude value can be obtained. This amplitude value reflects the intensity of the linear movement generated by the mandible during chewing, providing a quantitative basis for subsequent signal screening.

[0057] In step S403 of some embodiments, if the linear vibration amplitude satisfies the chewing linear vibration characterization condition, the target sampling step sequence corresponding to the linear vibration amplitude is determined as the key verification step sequence. It should be noted that the embodiments of this application determine whether the linear vibration amplitude meets the chewing linear vibration characterization conditions. If the linear vibration amplitude of a target sampling step meets the pre-set chewing characteristic conditions, such as the vibration amplitude being within the typical range of chewing action and exhibiting the periodic characteristics of chewing action, then this step will be identified as a key verification step. This screening process ensures that only those signal steps that may be related to chewing action enter the next verification stage, thereby reducing the amount of data processed subsequently and improving the efficiency and accuracy of the system.

[0058] In step S404 of some embodiments, the rotation angle is calculated based on the opening and closing rotation sensing components of each key verification step to obtain the corresponding opening and closing rotation angle. It should be noted that, for each key verification step, this embodiment further calculates the rotation angle based on its opening and closing rotation sensing component to obtain the corresponding opening and closing rotation angle. This step utilizes gyroscope data from the six-axis inertial measurement unit to calculate the rotation angle of each key verification step. By analyzing the angular velocity output of the gyroscope on the three orthogonal axes, the rotation angle information during mandibular opening and closing can be obtained. This angle information reflects the rotational motion characteristics of the mandible during chewing, providing another important dimension of data for the final signal integration.

[0059] In step S405 of some embodiments, if the opening and closing rotation angle of the key verification step sequence satisfies the chewing opening and closing rotation characterization condition, the linear vibration sensing component and the opening and closing rotation sensing component corresponding to the key verification step sequence are integrated to obtain the vibration sensing data of the sensing data stream corresponding to each key verification step sequence.

[0060] It should be noted that the embodiments of this application determine whether the opening and closing rotation angle features of the key verification steps meet the chewing opening and closing rotation characterization conditions. If the rotation angle features of a certain key verification step conform to the typical pattern of chewing action, for example, the rotation angle is within a reasonable range of chewing action, and the rotation speed and direction conform to the characteristics of chewing action, then the system will integrate the linear vibration sensing component and the opening and closing rotation sensing component corresponding to that step. By integrating these two components, the vibration sensing data corresponding to each key verification step is obtained. This integration process ensures that the final sensing data not only includes the linear motion features of chewing action but also the rotational motion features, thereby providing a more comprehensive and accurate signal foundation for subsequent behavior pattern recognition and dietary behavior analysis.

[0061] Through this series of steps, the diet behavior monitoring headphones can effectively filter signals highly correlated with chewing actions from a large amount of sensory data, and integrate these signals to form high-quality vibration sensing data. This data provides a solid foundation for subsequent feature extraction and behavior pattern recognition, enabling the system to more accurately identify and analyze the user's diet behavior patterns, thereby achieving precise monitoring of diet behavior.

[0062] Reference Figure 5 According to some embodiments of this application, the diet behavior monitoring headphones further include an electromyography (EMG) sensor. Step S101 involves sensing vibrations in the target object to obtain a sensing data stream, and may further include: Step S501: Vibration sensing of the target object is performed using a vibration sensor to obtain a vibration sensing data stream; wherein, the sampling rate of the vibration sensing data stream is the target vibration sampling rate; Step S502: Electromyography (EMG) is performed on the target object using an EMG sensor to obtain an EMG sensing data stream; wherein the sampling rate of the EMG sensing data stream is the target EMG sampling rate. Step S503: Determine the sampling ratio relationship based on the target vibration sampling rate and the target electromyography sampling rate; Step S504: Align the vibration sensing data stream and the electromyography sensing data stream based on the sampling ratio relationship; Step S505: The aligned vibration sensing data stream and electromyography sensing data stream are identified as sensing data streams.

[0063] In some embodiments of this application, the eating behavior monitoring headphones are equipped not only with a six-axis inertial measurement unit as a vibration sensor, but also with an electromyography (EMG) sensor to enhance the monitoring capability of chewing behavior. This multi-sensor fusion design aims to improve the accuracy and reliability of monitoring through the synergistic effect of vibration sensing and EMG sensing.

[0064] In step S501 of some embodiments, vibration sensing is performed on the target object using a vibration sensor to obtain a vibration sensing data stream; wherein, the sampling rate of the vibration sensing data stream is the target vibration sampling rate; It should be noted that, firstly, the vibration sensor senses the vibration of the target object, obtaining a vibration sensing data stream. The sampling rate of the vibration sensing data stream is set based on the target vibration sampling rate determined in the previous steps. This sampling rate is the optimal value obtained after comprehensive analysis of computational resource constraints, power consumption limits, and signal frequency characteristics. The vibration sensing data stream can capture the linear vibrations and rotational motions generated by chewing actions, providing basic mechanical vibration information for subsequent signal processing.

[0065] In step S502 of some embodiments, electromyography (EMG) sensors are used to sense the target object to obtain an EMG sensing data stream; wherein the sampling rate of the EMG sensing data stream is the target EMG sampling rate. It should be noted that, simultaneously, the electromyography (EMG) sensor performs EMG sensing on the target object, obtaining an EMG sensing data stream. The sampling rate of the EMG sensing data stream is the target EMG sampling rate, which is usually set individually based on the characteristics of the EMG signal and monitoring requirements. EMG signals reflect the electrophysiological activity of the masticatory muscles, directly indicating the contraction and relaxation states of the muscles, thus providing an independent and important signal source for monitoring masticatory behavior. Because the frequency range and dynamic characteristics of EMG signals differ from those of vibration signals, their sampling rate may also differ from the vibration sampling rate.

[0066] In some embodiments, step S503 determines the sampling ratio relationship based on the target vibration sampling rate and the target electromyography sampling rate; It's important to note that to effectively integrate two data streams with different sampling rates, a sampling ratio needs to be determined based on the target vibration sampling rate and the target electromyography (EMG) sampling rate. The sampling ratio refers to the temporal correspondence between the vibration-sensing data stream and the EMG-sensing data stream. For example, if the vibration sampling rate is 160 Hz and the EMG sampling rate is 500 Hz, then every three EMG sampling points correspond to one vibration sampling point. Determining this ratio is a crucial step in achieving data alignment, ensuring accurate temporal correspondence between the two signals, thus providing a foundation for subsequent fusion analysis.

[0067] In some embodiments, step S504 aligns the vibration sensing data stream and the electromyography sensing data stream based on the sampling ratio relationship; It should be noted that the vibration sensing data stream and the electromyography (EMG) sensing data stream are aligned based on the sampling ratio. The alignment process involves synchronizing the two data streams on the time axis, ensuring that each vibration sampling point has a corresponding EMG sampling point. This process may require data interpolation or downsampling to ensure temporal consistency between the two signals. For example, given the aforementioned sampling ratio, the EMG data can be downsampled to match the sampling rate of the vibration data, or the vibration data can be interpolated to match the sampling rate of the EMG data. The aligned data stream reflects the synchronous changes in the mechanical vibration and electromyographic activity of chewing movements, providing richer information for subsequent signal processing and behavioral analysis.

[0068] In some embodiments, step S505 determines the aligned vibration sensing data stream and electromyography sensing data stream as a sensing data stream.

[0069] It should be noted that the aligned vibration sensing data stream and electromyography (EMG) sensing data stream are determined as the final sensing data stream. This sensing data stream integrates information from both mechanical vibration and muscle electrical activity, providing a more comprehensive reflection of the characteristics of chewing behavior. For example, vibration signals can provide information on the intensity and rhythm of chewing movements, while EMG signals can provide information on the activity intensity and coordination of the chewing muscles. Through this multi-dimensional data fusion, the embodiments of this application can more accurately identify and analyze eating behavior patterns, thereby providing users with more precise eating behavior monitoring results. This fusion method not only improves the accuracy of monitoring but also enhances robustness, enabling it to operate stably in complex usage environments.

[0070] In step S102 of some embodiments, noise bands are filtered out for the sensing data stream to obtain the target sensing signal; It should be noted that after obtaining the sensing data stream, this embodiment of the application proceeds to the stage of "filtering out noise frequency bands from the sensing data stream to obtain the target sensing signal." This step uses digital signal processing techniques, such as Fast Fourier Transform and digital filters, to analyze and process the original sensing data stream. Since the sensing data stream may contain interference from environmental noise, head movements, and other non-eating-related actions, filtering out these noise frequency bands can effectively extract the pure signal related to chewing actions. For example, by setting a low-pass filter to remove high-frequency noise and using a notch filter to remove interference at specific frequencies (such as 50 Hz power frequency interference), a more accurate and purer target sensing signal is obtained. This process not only improves the signal quality but also provides a more reliable data foundation for subsequent feature extraction and behavioral pattern recognition.

[0071] It should be noted that the sensor data stream is directly acquired from the vibration sensor in the headphones, containing valid signals generated by chewing movements as well as various noise components. This noise may originate from environmental vibrations, non-chewing head movements (such as nodding or shaking), electromagnetic interference from electronic devices, etc. Sending the raw sensor data stream directly to the target terminal without noise filtering will lead to numerous problems.

[0072] Because the sensing data stream contains a large amount of invalid information, this invalid information not only consumes valuable transmission bandwidth but also increases the processing burden on the target terminal. The amount of data that the target terminal needs to process will increase significantly, which may lead to processing delays and affect the real-time performance of monitoring results. For example, if the sensing data stream contains a large amount of low-frequency noise, this noise may overlap with the frequency range of the chewing signal, requiring the target terminal to use additional algorithms to distinguish between valid signals and noise during the feature extraction stage, thus increasing computational complexity.

[0073] Secondly, the presence of noise degrades signal quality, making subsequent feature extraction and behavioral pattern recognition difficult. Noise may mask subtle features of the chewing signal, leading to inaccurate feature extraction and consequently affecting the accuracy of behavioral pattern recognition. For example, the amplitude and frequency characteristics of the chewing signal are crucial for determining chewing intensity and rhythm; noise interference may cause these features to be misjudged, resulting in incorrect dietary behavior patterns.

[0074] The aforementioned problems can be effectively solved by filtering out noise bands in the sensor data stream within the eating behavior monitoring headset. The process of filtering out noise bands involves digital signal processing techniques. For example, low-pass filters are used to remove high-frequency noise, and band-stop filters are used to remove interference within specific frequency ranges. The resulting target sensor signal is cleaner, retaining valuable information related to chewing actions while eliminating most noise components. This processing method reduces the burden of transmitting the target sensor signal to the target terminal.

[0075] On the one hand, the amount of target perception signal data is significantly reduced after noise removal, and the time and bandwidth required for transmission are also reduced accordingly, improving data transmission efficiency. On the other hand, the target perception signal received by the target terminal is of higher quality and has more obvious features, which makes the target terminal more efficient in feature extraction and behavior pattern recognition, reducing the consumption of computing resources and improving the overall system performance.

[0076] Furthermore, noise band filtering can improve the robustness of the system. In complex usage environments, such as noisy restaurants or sports scenes, environmental noise can significantly affect the sensing data stream. By performing noise filtering at the earpiece end, it can be ensured that even in these complex environments, the embodiments of this application can stably extract effective chewing signals, thereby guaranteeing the accuracy and reliability of dietary behavior monitoring.

[0077] Based on this, filtering out noise bands in the sensing data stream in the diet behavior monitoring headset can not only reduce the burden of sending the target sensing signal to the target terminal, but also improve signal quality, enhance the robustness of the system, and ultimately achieve more efficient and accurate diet behavior monitoring.

[0078] Reference Figure 6 According to some embodiments of this application, step S102, which involves filtering out noise bands in the sensing data stream to obtain the target sensing signal, may include: Step S601: Generate the target vibration sensing spectrum based on the vibration sensing data corresponding to each target sampling step sequence in the sensing data stream; Step S602: Perform spectral analysis on the target vibration sensing spectrum to obtain the target spectral analysis results; Step S603: Perform filtering and denoising operations on the sensing data stream based on the target spectrum analysis results to obtain the target sensing signal.

[0079] In some embodiments of this application, the process of filtering out noise bands in the sensing data stream to obtain the target sensing signal is a systematic signal processing step aimed at improving the quality and reliability of the monitoring data.

[0080] In some embodiments, step S601 generates a target vibration sensing spectrum based on the vibration sensing data corresponding to each target sampling step sequence in the sensing data stream. It should be noted that the target vibration sensing spectrum is generated based on the vibration sensing data corresponding to each target sampling step in the sensing data stream. This process involves converting the vibration sensing data in the time domain into frequency domain data to more clearly identify the distribution of different frequency components. Using mathematical tools such as Fourier transform, the vibration sensing data can be decomposed into sinusoidal wave components of different frequencies, thereby generating the target vibration sensing spectrum. This spectrum can visually demonstrate which frequency components dominate the vibration signal and which frequency components may be introduced by noise.

[0081] In some embodiments, step S602 involves performing spectral analysis on the target vibration sensing spectrum to obtain the target spectral analysis result. It should be noted that spectral analysis is performed on the target vibration sensing spectrum to obtain the target spectral analysis results. The purpose of spectral analysis is to identify which frequency bands in the sensing spectrum belong to the effective signal and which belong to noise. For example, the vibration signal generated by chewing action is usually concentrated in the low-frequency band of 1 to 3 Hz, while the high-frequency band (such as above 50 Hz) may mainly contain environmental noise or interference from electronic devices. By analyzing the target vibration sensing spectrum, the range of noise frequency bands can be determined, providing a basis for subsequent filtering and denoising operations. This analysis process may involve calculating the energy of each frequency component, identifying peaks within a specific frequency range, and evaluating the overall distribution of the spectrum.

[0082] In some embodiments, step S603 involves performing a filtering and denoising operation on the sensing data stream based on the target spectrum analysis results to obtain the target sensing signal.

[0083] It should be noted that filtering and denoising operations are performed on the sensing data stream based on the target spectral analysis results to obtain the target sensing signal. This step utilizes the results of spectral analysis to remove signal components in noisy frequency bands by designing appropriate filters. For example, if the spectral analysis shows that the high-frequency band mainly contains noise, a low-pass filter can be designed to remove these high-frequency components; if there is interference at specific frequencies, a band-stop filter can be used to eliminate this interference. The filter design needs to retain as many characteristics of the effective signal as possible while removing noise to ensure that the final target sensing signal can accurately reflect the vibration characteristics of chewing movements. Through filtering and denoising operations, noise in the sensing data stream is effectively screened out, resulting in a purer and more reliable target sensing signal, providing a high-quality data foundation for subsequent behavior pattern recognition and dietary behavior analysis.

[0084] This series of steps ensures the extraction of high-quality target sensing signals from the raw sensing data stream. Through frequency domain analysis and filtering denoising, the signal-to-noise ratio of the signal is effectively improved, enabling the monitoring system to more accurately capture and analyze the vibration characteristics of chewing behavior.

[0085] Reference Figure 7 According to some embodiments of this application, step S603, which performs filtering and denoising operations on the sensing data stream based on the target spectrum analysis results to obtain the target sensing signal, may include: Step S701: Obtain the target vibration sampling rate corresponding to the target sampling step sequence; Step S702: Set the low-pass filter cutoff frequency based on the target vibration sampling rate, and perform low-pass filtering on the sensing data stream based on the low-pass filter cutoff frequency to obtain the first intermediate data stream; Step S703: Based on the target spectrum analysis results, determine the target interference frequency; Step S704: Based on the target interference frequency, preset notch filter parameters are set, and the first intermediate data stream is processed point by point through the notch filter with the set parameters to obtain the second intermediate data stream. Step S705: Perform random noise removal processing on the second intermediate data stream to obtain the target sensing signal.

[0086] In some embodiments of this application, the process of performing filtering and denoising operations on the sensing data stream based on the target spectrum analysis results to obtain the target sensing signal is a fine signal processing flow performed in steps.

[0087] In some embodiments, step S701 involves obtaining the target vibration sampling rate corresponding to the target sampling step sequence. It should be noted that the target vibration sampling rate corresponding to the target sampling step sequence is obtained. This sampling rate was determined after comprehensively considering computational resource constraints, power consumption limitations, and signal frequency characteristics analysis. It determines the data acquisition frequency and data volume. Accurate acquisition of the sampling rate is crucial for the subsequent filter design, because the filter parameter settings need to match the sampling rate to ensure the accuracy and effectiveness of the filtering effect.

[0088] In some embodiments, step S702 involves setting a low-pass filter cutoff frequency based on the target vibration sampling rate, and performing low-pass filtering on the sensed data stream based on the low-pass filter cutoff frequency to obtain a first intermediate data stream. It should be noted that the low-pass filter cutoff frequency is set based on the target vibration sampling rate, and the sensed data stream is then subjected to low-pass filtering to obtain the first intermediate data stream. The purpose of low-pass filtering is to remove high-frequency noise, which may originate from interference from electronic devices or environmental vibrations. The cutoff frequency is typically set based on the results of spectrum analysis, selecting a frequency value that can effectively remove high-frequency noise while preserving the main characteristics of the signal. For example, if spectrum analysis shows that the chewing signal is mainly concentrated between 1 and 3 Hz, while high-frequency noise increases significantly from 50 Hz, then the low-pass filter cutoff frequency can be set at around 40 Hz. Through low-pass filtering, high-frequency noise in the sensed data stream is effectively suppressed, resulting in a relatively clean first intermediate data stream.

[0089] In some embodiments, step S703 involves determining the target interference frequency based on the target spectrum analysis results. It should be noted that the target interference frequencies are determined based on the target spectrum analysis results. These target interference frequencies may be caused by specific environmental factors or equipment characteristics, such as 50 Hz power frequency interference. Through spectrum analysis, the interference components at these specific frequencies can be accurately identified, providing target frequencies for subsequent notch filtering.

[0090] In some embodiments, step S704 involves setting parameters for a preset notch filter based on the target interference frequency, and then performing point-by-point calculations on the first intermediate data stream using the notch filter with the set parameters to obtain a second intermediate data stream. It should be noted that, based on the target interference frequency, parameters are set for a preset notch filter, and the first intermediate data stream is processed point-by-point using the notch filter with the set parameters to obtain the second intermediate data stream. A notch filter is a filter that can effectively remove specific frequency components, and its parameter settings need to be precisely matched to the target interference frequency. Processing the first intermediate data stream with a notch filter can further remove interference at specific frequencies, resulting in a cleaner second intermediate data stream. This process ensures that the data stream no longer contains known specific frequency interference, improving signal quality and reliability.

[0091] In some embodiments, step S705 involves performing random noise removal processing on the second intermediate data stream to obtain the target sensing signal.

[0092] It should be noted that random noise removal processing is performed on the second intermediate data stream to obtain the target perception signal. Random noise is usually caused by sensor thermal noise, minor environmental vibrations, or other unpredictable interference. Although these noises are small in amplitude, they can affect the accuracy and stability of the signal. By performing random noise removal processing, such as using moving average filtering, median filtering, or other adaptive filtering techniques, the data stream can be further smoothed, random fluctuations can be removed, and a high-quality target perception signal can be obtained. This signal can more accurately reflect the vibration characteristics of chewing actions, providing a reliable data foundation for subsequent behavioral pattern recognition and dietary behavior analysis.

[0093] Through this series of meticulous filtering and noise reduction steps, the diet behavior monitoring headphones can effectively extract high-quality target perception signals from the raw perception data stream, thereby improving the accuracy and reliability of monitoring.

[0094] In step S103 of some embodiments, the target perception signal is sent to the target terminal so that the target terminal can extract features from the target perception signal to obtain the target behavior features of the target object. It should be noted that the target perception signal is sent to the target terminal, which then performs feature extraction on the signal to obtain the target object's behavioral characteristics. This step leverages the powerful computing capabilities of modern smart devices (such as smartphones or tablets) to transmit the pre-processed target perception signal to these devices for further analysis. On the target terminal, complex algorithms extract features from the target perception signal, which may include chewing frequency, force, rhythm, and duration. These features reflect the user's eating habits, such as chewing speed and food texture (judged by chewing force). By distributing the data processing tasks to the terminal device, not only is the computational burden on the headset reduced, but the high-performance processor and storage capabilities of the terminal device can also be utilized to achieve more accurate feature extraction.

[0095] In step S104 of some embodiments, behavioral pattern recognition is performed on the target behavioral characteristics in the target terminal to determine the dietary behavior pattern of the target object, and dietary behavior analysis is performed based on the dietary behavior pattern and target behavioral characteristics to obtain dietary behavior monitoring results.

[0096] It should be noted that behavioral pattern recognition is performed on the target terminal based on target behavioral characteristics to determine the target object's dietary behavior pattern. Based on the dietary behavior pattern and target behavioral characteristics, dietary behavior analysis is performed to obtain dietary behavior monitoring results. This step is the core of the entire monitoring method. Machine learning or deep learning algorithms are used to analyze the extracted target behavioral characteristics to identify the user's dietary behavior pattern. For example, the algorithm can determine whether the user is eating solid or liquid food based on chewing frequency and force, or whether the user is chewing slowly and thoroughly based on chewing rhythm. Ultimately, based on these behavioral patterns and characteristics, the embodiments of this application can generate detailed dietary behavior monitoring results, such as the user's eating speed, estimated food intake, and dietary habit assessment. These results not only provide users with feedback on their dietary behavior but can also be used in health monitoring, dietary management, and other application scenarios to help users improve their eating habits and promote the formation of a healthy lifestyle.

[0097] It should be understood that the dietary behavior monitoring method of this application embodiment not only solves the privacy and environmental interference problems existing in traditional monitoring methods, but also achieves unobtrusive monitoring through headphones, an everyday wearable device, and utilizes the computing power of the terminal device to achieve accurate monitoring and analysis of dietary behavior.

[0098] In some embodiments, step S105 involves receiving a target perception signal from the diet behavior monitoring headphones and extracting features from the target perception signal to obtain the target behavior features of the target object. It should be noted that in this embodiment, the target perception signal is received from the eating behavior monitoring headphones. This signal, after undergoing preliminary processing steps such as vibration sensing, noise removal, and multi-sensor data fusion, is already a relatively pure signal stream containing rich information. The target perception signal not only contains linear vibration and rotational motion information of chewing movements, but may also incorporate electromyographic signals, thus providing a multi-dimensional data foundation for subsequent feature extraction.

[0099] Next, this embodiment of the application performs feature extraction on the target perception signal to obtain the target behavior features of the target object. The feature extraction process transforms key information in the original signal into quantitative indicators that can be used for analysis and identification. For example, for vibration signals, features such as chewing frequency, chewing force (calculated through vibration amplitude), and the regularity of chewing rhythm (e.g., periodic changes) can be extracted. If electromyography (EMG) signals are fused, features such as the intensity of EMG activity, the frequency of EMG bursts, and the temporal relationship between EMG and vibration signals can also be extracted. These features can reflect the physiological and mechanical characteristics of chewing actions from different perspectives, providing rich data support for subsequent behavior pattern recognition.

[0100] In some embodiments, step S106 involves performing behavioral pattern recognition on the target behavioral characteristics to determine the dietary behavior pattern of the target object, and performing dietary behavior analysis based on the dietary behavior pattern and the target behavioral characteristics to obtain dietary behavior monitoring results.

[0101] It should be noted that after obtaining the target behavioral characteristics, this embodiment of the application enters the behavioral pattern recognition stage. The goal of this stage is to determine the target object's eating behavior pattern by analyzing the target behavioral characteristics. Behavioral pattern recognition typically relies on machine learning or deep learning algorithms, which learn the feature representations of different eating behavior patterns through training data. For example, this embodiment of the application can identify whether a user is eating solid or liquid food, whether they are chewing quickly or slowly, and whether there are intermittent eating patterns. The recognition of these patterns is based on the comparison and matching of the target behavioral characteristics with known behavioral pattern characteristics.

[0102] Ultimately, this application embodiment analyzes dietary behavior based on determined dietary behavior patterns and target behavioral characteristics to obtain dietary behavior monitoring results. The process of dietary behavior analysis combines identified behavioral patterns with specific dietary behavior characteristics to generate detailed monitoring reports. For example, this application embodiment can estimate the texture and intake of food based on chewing frequency and force; assess the healthiness of a user's eating habits based on the regularity of chewing rhythm; and even provide feedback on the degree of fatigue in chewing muscles based on the intensity and duration of electromyographic activity. These monitoring results not only provide users with detailed information about their dietary behavior but can also be used in applications such as health monitoring, diet management, or medical research.

[0103] Through this series of steps, the dietary behavior monitoring system can extract valuable information from the raw sensory signals received by the headphones and transform it into specific dietary behavior monitoring results. This process not only demonstrates the advantages of multi-sensor data fusion and advanced signal processing technologies but also showcases the powerful capabilities of machine learning algorithms in behavioral pattern recognition. Ultimately, the embodiments of this application provide users with a comprehensive, accurate, and personalized dietary behavior monitoring solution.

[0104] Reference Figure 8 , Figure 8 This illustration shows an ear-hook structure of a headphone for monitoring eating behavior according to an embodiment of this application, designed to monitor and analyze the user's chewing behavior. Figure 8 The earphone shell is visible, providing not only physical protection but also housing all the internal components. Inside the earphones are several key electronic modules, each with its specific function, working together to achieve the earphones' intelligent monitoring capabilities.

[0105] Figure 8The location of the vibration sensor is marked on the image. This sensor is the core component for monitoring chewing movements; it captures the minute vibrations generated when the user's jaw moves. These vibration signals are then sent to a processor for further analysis and processing. The precise placement and high sensitivity of the sensor are crucial for capturing the details of chewing movements, thus ensuring the accuracy of the monitoring data.

[0106] Secondly, the power module provides the necessary power to the headphones. This typically includes a small battery and possibly charging circuitry, ensuring the headphones can operate continuously for extended periods. The design of the power module needs to consider the size limitations of the headphones and the user's wearing comfort, while also ensuring a stable power supply to support all headphone functions.

[0107] The processor is the brain of the headphones; it's responsible for processing the data collected by the vibration sensors. The processor executes complex algorithms to analyze characteristics such as chewing rhythm, force, and timing, and transforms this information into useful monitoring results. The processor's performance directly impacts the speed and accuracy of data processing, therefore a balance between computing power and energy consumption must be struck during the design phase.

[0108] The memory is used to store data analyzed by the processor, as well as the headset's firmware or software programs. This can include temporary storage to buffer data streams and long-term storage to store the user's historical chewing data. The size and read / write speed of the memory affect the headset's data storage capacity and response speed.

[0109] The communication module is responsible for transmitting the data collected and processed by the earphones to the target terminal, such as a smartphone or tablet. Data transmission is achieved wirelessly, such as via Bluetooth. The design of the communication module needs to ensure the stability and security of data transmission, while also considering power consumption and compatibility issues.

[0110] Reference Figure 9 This illustration shows another in-ear structure of the dietary behavior monitoring earphone according to an embodiment of this application, including a shell, a power module, a processor, a memory, a vibration sensor, and a communication module. The power module provides power to the earphone, the processor processes chewing motion data captured by the vibration sensor, the memory stores data and programs, the vibration sensor is specifically used to monitor vibrations during chewing, and the communication module is responsible for wirelessly transmitting the processed data to an external device. These components work together to enable the earphone to monitor and analyze the user's dietary behavior in real time.

[0111] In conclusion, Figure 8 and Figure 9 The presentation showcases the key components and functions within the dietary behavior monitoring headset. The sophisticated design and collaborative work of these components enable the headset to accurately monitor the user's eating habits and effectively transmit the data to the target terminal for further analysis.

[0112] This application also provides a computer program product, which includes a computer program. The processor of a computer device reads and executes the computer program, causing the computer device to perform the above-described dietary behavior monitoring method.

[0113] This application also provides a computer program product, which includes a computer program. The processor of a computer device reads and executes the computer program, causing the computer device to perform the above-described dietary behavior monitoring method.

[0114] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in this disclosure and the foregoing drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “including,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatuses.

[0115] It should be understood that in this disclosure, "at least one item" means one or more, and "more than one" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0116] It should be understood that in the description of the embodiments of this application, "multiple" means two or more, "greater than", "less than", "exceeding" etc. are understood to exclude the number itself, and "above", "below", "within" etc. are understood to include the number itself.

[0117] In the several embodiments provided in this disclosure, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0118] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0119] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0120] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium may include: a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and other media capable of storing program code.

[0121] It should also be understood that the various implementation methods provided in this application can be combined arbitrarily to achieve different technical effects.

[0122] The above is a detailed description of the embodiments of this disclosure. However, this disclosure is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this disclosure. All such equivalent modifications or substitutions are included within the scope defined by the claims of this disclosure.

Claims

1. A method for monitoring dietary behavior, characterized in that, Headphones used for monitoring eating behavior include: Vibration sensing is performed on the target object to obtain a sensing data stream; The target sensing signal is obtained by filtering out noise frequency bands from the sensing data stream. The target perception signal is sent to the target terminal so that the target terminal can extract features from the target perception signal to obtain the target behavior features of the target object; In the target terminal, behavioral pattern recognition is performed on the target behavioral characteristics to determine the dietary behavior pattern of the target object. Based on the dietary behavior pattern and the target behavioral characteristics, dietary behavior analysis is performed to obtain dietary behavior monitoring results.

2. The method according to claim 1, characterized in that, The dietary behavior monitoring earphone uses a six-axis inertial measurement unit as a vibration sensor. The vibration sensing of the target object and the resulting sensing data stream include: Obtain the target sampling sequence; Based on the target sampling sequence, the linear vibration generated by the jaw movement of the target object is captured by the six-axis inertial measurement unit to obtain the linear vibration sensing component; Based on the target sampling sequence, the rotation component of the target object's jaw opening and closing is captured by the six-axis inertial measurement unit to obtain the opening and closing rotation sensing component; By integrating the linear vibration sensing component and the opening / closing rotation sensing component, the vibration sensing data corresponding to the target sampling step sequence of the sensing data stream is obtained.

3. The method according to claim 2, characterized in that, The integration of the linear vibration sensing component and the opening / closing rotation sensing component to obtain the vibration sensing data corresponding to the target sampling step sequence in the sensing data stream includes: Obtain the characterization conditions for linear vibration of chewing and the characterization conditions for chewing opening and closing rotation; The amplitude of the linear vibration sensing component is calculated based on each target sampling step sequence to obtain the corresponding linear vibration amplitude. If the linear vibration amplitude satisfies the chewing linear vibration characterization condition, the target sampling step sequence corresponding to the linear vibration amplitude is determined as the key verification step sequence; Based on the opening and closing rotation sensing components of each key verification step, the rotation angle is calculated to obtain the corresponding opening and closing rotation angle. If the opening and closing rotation angle of the key verification step satisfies the chewing opening and closing rotation characterization condition, the linear vibration sensing component and the opening and closing rotation sensing component corresponding to the key verification step are integrated to obtain the vibration sensing data of the sensing data stream corresponding to each key verification step.

4. The method according to claim 2, characterized in that, The steps for obtaining the target sampling sequence include: Obtain computational resource constraints and power consumption limits; The chewing vibration spectrum of the target object was measured to obtain a vibration test spectrum sample. Spectral analysis was performed on the vibration test spectrum sample to obtain the test spectrum analysis results; The sampling rate is calculated based on the test spectrum analysis results to obtain the theoretical sampling rate; Based on the computational resource constraints and the power consumption constraints, the theoretical sampling rate is adjusted to the target vibration sampling rate; Based on the target vibration sampling rate, the target sampling sequence is set.

5. The method according to claim 2, characterized in that, The dietary behavior monitoring headphones also include an electromyography (EMG) sensor; the step of sensing vibrations in the target object to obtain a sensing data stream further includes: The vibration sensor detects vibration in the target object, resulting in a vibration sensing data stream; wherein the sampling rate of the vibration sensing data stream is the target vibration sampling rate. The electromyography (EMG) sensor is used to detect the electromyography of the target object, thereby obtaining an EMG data stream; wherein the sampling rate of the EMG data stream is the target EMG sampling rate; Based on the target vibration sampling rate and the target electromyography sampling rate, the sampling ratio relationship is determined; Based on the sampling ratio relationship, the vibration sensing data stream and the electromyography sensing data stream are aligned; The aligned vibration sensing data stream and electromyography sensing data stream are identified as the sensing data stream.

6. The method according to claim 1, characterized in that, The step of filtering out noise frequency bands in the sensing data stream to obtain the target sensing signal includes: Based on the vibration sensing data corresponding to each target sampling step sequence in the sensing data stream, a target vibration sensing spectrum is generated. Spectral analysis is performed on the vibration sensing spectrum of the target to obtain the target spectral analysis results; Based on the target spectrum analysis results, a filtering and denoising operation is performed on the sensing data stream to obtain the target sensing signal.

7. The method according to claim 6, characterized in that, The step of performing filtering and denoising operations on the sensing data stream based on the target spectrum analysis results to obtain the target sensing signal includes: Obtain the target vibration sampling rate corresponding to the target sampling step sequence; A low-pass filter cutoff frequency is set based on the target vibration sampling rate, and the sensing data stream is subjected to low-pass filtering based on the low-pass filter cutoff frequency to obtain a first intermediate data stream. Based on the target spectrum analysis results, the target interference frequency is determined; Based on the target interference frequency, preset notch filter parameters are set, and the first intermediate data stream is processed point by point through the notch filter with the set parameters to obtain the second intermediate data stream. Random noise removal processing is performed on the second intermediate data stream to obtain the target sensing signal.

8. A method for monitoring dietary behavior, characterized in that, Applied to target terminals, including: Target perception signal is received from the diet behavior monitoring earphone; wherein, the target perception signal is obtained in the diet behavior monitoring earphone through the following steps: vibration sensing is performed on the target object to obtain a perception data stream, and noise frequency band is filtered out from the perception data stream to obtain the target perception signal; Feature extraction is performed on the target perception signal to obtain the target behavior features of the target object; Behavioral pattern recognition is performed on the target behavioral characteristics to determine the dietary behavior pattern of the target object. Based on the dietary behavior pattern and the target behavioral characteristics, dietary behavior analysis is performed to obtain dietary behavior monitoring results.

9. A pair of headphones for monitoring eating behavior, characterized in that, include: The device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the dietary behavior monitoring method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The storage medium stores a program that is executed by a processor to implement the dietary behavior monitoring method as described in any one of claims 1 to 8.

Images