EAR CANAL BIOMETRIC VERIFICATION SYSTEM BASED ON VARIABLE ACOUSTIC SCANNING SIGNAL
Patent Information
- Application Number
- TR202612640
- Authority / Receiving Office
- TR · TR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2026-07-28
- Publication Date
- 2026-09-21
Smart Images
Figure 00000016_0000
Abstract
Description
1 TARIFF EAR CANAL SCAN BASED ON VARIABLE ACOUSTIC SCANNING SIGNAL BIOMETRIC IDENTITY VERIFICATION SYSTEM TECHNICAL AREA 5 The invention relates to wearable acoustic devices, acoustic biometrics, and electronic identification. The invention relates to the technical field of verification systems. Specifically, it concerns each identity verification system. an acoustic scan signal whose frequency composition is modified during the verification session applied to the user's ear canal, the echo signal returning from the ear canal perception, user-specific acoustic transmission from the perceived echo signal 10 extracting the characteristic of the acoustic transmission feature converting the resulting feature data and matching it with the registered user profile. a biometric authentication system based on comparison It is related. STATE OF THE ART Today, passwords and personal information are used to verify user identity. In addition to identity number and similar information-based verification methods, facial recognition is also used. utilizing user biometric data such as fingerprints, iris, and voice systems are used. In biometric authentication systems, 20 are obtained from the user. Comparison of the obtained measurement data with a previously recorded reference data. and generally, a decision regarding user identity is made based on the comparison result. It is known as. Such as facial image, fingerprint, and previously recorded voice sample. Verification using static or reproducible biometric data 25 systems that prevent the copying, replaying, or imitation of the data in question security vulnerability to attacks carried out through this method It can demonstrate this. In voice-based verification systems, ambient noise, however, microphone position, changes in the user's speech pattern, and recorded data. Factors such as audio playback negatively affect verification accuracy. 30 It can have an impact. Hearing aids are wearable devices that are placed in or around the ear. In acoustic devices, sound is detected from the user or the user's environment. processing of signals and extracting user-related features from the processed signals 2 It is known that this can be determined. In this context, patent number EP3709115A1 in the document, configured for use by a specific user The hearing aid and user identification unit are explained. The document describes the voice characteristics of the person using the device and how they hear from their mouth. Evaluation of the characteristics of the acoustic channel extending from the device to the microphone 5 in this way, it can be determined whether the person using the device is a specific user or not. An identification signal is generated. User identification is mentioned in the document. The process involves the user's own voice utilizing the acoustic channel between the mouth and the microphone. It is based on its features. However, in each validation session A frequency-modified acoustic scanning signal is inserted into the ear canal at 10°C. The procedure involves the return signal from the ear canal to the scanning signal in question. the perception of echo and the acoustics specific to the ear canal derived from the perceived echo. The method of deriving the transfer characteristic is not explained. Another document regarding the known state of the art is EP4075423A1, A test ear used to evaluate the performance of headphone devices. It relates to the model. In this document, a layered artificial eardrum is described. plate-shaped elements arranged and containing holes by means of physically simulating an individual's external ear canal This is explained. The solution in question involves the acoustics of a headset device. It is intended to evaluate its performance under laboratory conditions and user 20 a regulation regarding the biometric verification of identity It does not include. In the current state of the art, a user with a hearing aid is also involved. Authentication is used to establish a reliable connection between the application and the system. Systems that use cryptographic data such as the key and key identifier 25 It is known that such a reliable link is mentioned in document EP3334188A1. The procedures for its creation are explained. However, the aforementioned The approach involves a biometric measurement physically taken from the user's ear canal. not acoustic data, but verification data used between devices It is based on. 30 Document number WO2021136962A1 shows a device that takes images of the surroundings. a camera, a microphone that detects ambient sounds, and the captured image or sound data a processor that changes the way the hearing aid system works depending on the connection This document explains how the audio processing mode changes depending on environmental conditions. 3 It aims to modify the individual acoustic characteristics of the ear canal. a technical framework for its use in user authentication It does not include. The geometry of the ear canal and the anatomical structures within the ear, 5. It is known that this is the case. Accordingly, the acoustic responses received from the ear canal... There are ongoing studies regarding its evaluation as biometric data. However, a fixed frequency or a predetermined uniform stimulation signal in the structures where it is used, measurement data is recorded and reused. The verification system uses previously obtained responses to the same physical stimulus. 10 The risk of being misled may arise. In addition, the raw echo signal received from the ear canal is reflected in the environment. the effects, along with the device location and the characteristics of the signal generation unit. assessment directly of user-specific ear canal characteristics This can make it difficult to determine. The raw echo signal may only have amplitude or 15 Comparison based on specific frequency components, small differences between users inadequate for reliably distinguishing anatomical differences It can remain. In current systems, the user experiences changes over time. Ignoring anatomical or physical variations is also a separate technique. 20 This creates a limitation. Small differences in headphone placement affect the ear. the moisture, temperature or similar physical conditions inside and in the ear canal Changes that occur over time are compared with a constant reference profile. This can lead to an increase in the difference between the measurement data and the reference profile. Uncontrolled updating of data obtained from an unauthorized user constitutes a violation of Article 25. This can create the risk of being included in the system. In conclusion, under the current state of the art; Copying of static or reusable biometric data and It is vulnerable to replay attacks, A physical acoustic stimulus that changes in each authentication session for the user 30 non-implementation the scanning signal sent via the echo received from the ear canal not being separated by deconvolution, 4 Reliable acoustic transmission characteristics specific to the ear canal inability to remove, The verification profile is derived only from successful verifications due to the fact that the data is not updated in a controlled and limited manner 5 in terms of safety, measurement stability and user-specific distinguishability. There are technical aspects that are open to improvement. THE TECHNICAL PROBLEM THAT IS INTENDED TO BE SOLVED In the known state of the art, facial recognition, fingerprinting, voice recording, and Similar static biometric data can be obtained and copied beforehand. 10 or can be reused. The same or a predetermined stimulus. In validation systems where data is used repeatedly, previously re-applying a recorded response received from the actual user It can be considered a current measurement. The acoustic response obtained from the ear canal is used as biometric data. 15 In structures where it is used, only the raw echo signal or a specific frequency is available. Comparison of components, effect of transmitted acoustic signal on the user the unique ear canal characteristics differing sufficiently from each other This can lead to indistinguishability. This situation requires different validation methods. comparability of measurements obtained in sessions and users 20 This reduces the discernibility of anatomical differences between them. Also, small differences in the placement of the headset or in-ear acoustic device changes, along with variations in moisture and temperature in the ear canal, affect the user anatomical differences that have occurred over time, current measurements This leads to the data deviating from the fixed reference profile initially created. 25 It can open. Using the reference profile without modification allows the real user to... incorrect rejection of related measurements; uncontrolled updating This leads to unverified data being included in the user profile. It is possible. In the known state of the technique, 30 mimic the physical shape of the ear canal. The echo signal obtained from an inanimate object is transmitted to the actual user's ear. There is also a risk that it may be interpreted as an echo signal received from the channel. Therefore, not only the morphological features extracted from the echo signal, but also the live signal itself... also the evaluation of acoustic damping behavior associated with soft tissue It is needed. The fundamental technical problem that the invention aims to solve is; each authentication During the session, an acoustic scanning signal with a modified frequency composition is transmitted to the ear. applied to the ear canal, the echo returning from the ear canal is combined with the applied scanning signal 5 By working together, they isolate and obtain the user-specific acoustic transmission characteristics. A verification process that compares the acquired current acoustic data with the registered user profile. It is the creation of its structure. This fundamental technical problem also includes: Different validation of a previously recorded ear canal response 10 reducing reusability in the session, From the raw echo signal detected in the ear canal, the applied scan deconvolution of the signal's effect a comparable user-specific ear canal transmission characteristic conversion to feature data, 15 Angular similarity between current feature data and registered user profile and determined through differences in form, live by evaluating the attenuation time of the echo signal acoustic responses originating from soft tissue and inanimate material separation, 20 Identity verification processes can be performed when there is no external network connection. Continuation of the embedded processor on the device and adaptive profile Synchronizing updates with the remote server, Current attribute associated only with a successful validation result Adding the data to the registered user profile with a limited weight and 25 in the ear canal or in the placement of the acoustic device over time the stability of the reference profile against minor changes that occur The technical problems related to its protection are being addressed. The technical problem that is intended to be solved in this context is not just an identity. It is not about making a decision about the information; it is about a variable physical acoustic signal 30 the generation of that signal, its transmission to the ear canal, and the return from the ear canal The detection of a physical echo involves subjecting the echo signal to signal processing operations. and technically specific acoustic transmission characteristics of the user 6 a reliable and repeatable verification chain of determination processes It is carried out within. A BRIEF DESCRIPTION OF THE INVENTION The invention relates the acoustic characteristics specific to the user's ear canal to biometric data. 5 an identity verification process performed by being evaluated as such It is related to the system. In the system described in the invention, the frequency structure is present in each validation session. A modified acoustic scanning signal is transmitted to the ear canal, and the ear An echo returning from the channel is detected. Deconvolution from the detected echo is 10. This process provides the acoustic transmission characteristics of the ear canal. a feature data containing characteristic mel-frequency cepstral coefficients is being converted and based on cosine similarity with the registered user profile. They are being compared. The acoustic scan signal covers the audible frequency band, ultrasonic frequency 15 the band or encompassing both audible and ultrasonic frequency bands The dynamic chirp is generated as an acoustic scanning signal. signal start frequency, end frequency and phase verification sessions between It is being changed. The invention also includes the current acoustic 20 obtained as a result of successful verification. by including feature data in the registered user profile with a ten percent weighting The profile is updated in a controlled manner. This ensures that the saved information is properly stored. The resilience to echo data reuse is being increased, ear Limited changes in channel and device placement are being adapted to and This is applicable in wearable acoustic devices containing in-ear microphones and speakers. A verification structure is being created. In one application of the invention, the echo signal detected by the microphone The damping time is determined, and that damping time is observed in the living organism. This is compared with the recorded damping time metric for soft tissue. The acoustic liveness data obtained was analyzed using user profiles over 30 years. This comparison is considered in the authentication decision. In another application of the invention, deconvolution, feature extraction, and Decision-making processes are embedded in the wearable acoustic device. 7 The update data regarding the adaptive profile database is being processed by the processor. This is synchronized with the remote server via a communication interface. LIST OF FIGURES Figure 1. Showing the flow of the ear canal biometric authentication system. 5 schematic view The Definitions of the Numbers in the Figures 1. Dynamic signal generator 2. Acoustic sensor and hardware interface 10 3. Deconvolution engine 4. Feature extraction module 5. Pure cosine decision engine 6. Adaptive profile database DETAILED DESCRIPTION OF THE INVENTION The invention includes a dynamic signal generator (1), an acoustic sensor and hardware interface (2), deconvolution engine (3), feature extraction module (4), pure cosine decision engine (5) and includes the adaptive profile database (6). Dynamic signal generator (1), 20 to be used in each authentication session audible, with a randomly determined starting and ending frequency frequency band, ultrasonic frequency band or audible and ultrasonic frequency a dynamic chirp acoustic scan signal encompassing both bands It produces the initial stage of the dynamic chirp acoustic scan signal. The frequency, end frequency, and phase are changed between verification sessions. 25 Thus, each validation session uses the same methods as in previous validation sessions. with an acoustic scan signal different from acoustic scan signals is being carried out. Acoustic scanning signal generated by dynamic signal generator (1), The acoustic sensor and hardware interface (2) are transmitted. Acoustic sensor and 30 The hardware interface (2) transmits the acoustic scan signal to the user's ear canal. a speaker and a microphone that detects the echo signal returning from the ear canal It includes. 8 The speaker and microphone are positioned facing the ear canal. It is positioned. The acoustic scan transmitted into the ear canal by the speaker. the signal depends on the individual anatomical variations of the ear canal is reflected and the resulting echo signal is transmitted to the microphone. It is perceived by. 5 Acoustic sensor and hardware interface (2) speaker facing the ear canal and because it includes a microphone, the acoustic response obtained directly from the ear canal is captured. This structure enables perception. This design contains a microphone that is actually positioned inside the ear. wireless stereo headphones or those with active noise cancellation It can be applied to headphones. 10 Because the microphone is outward-facing in standard wired headphones. Potential electromagnetic cable leakage and crosstalk can occur, especially at low levels. This can affect the measurement data in the frequency ranges. Therefore, acoustics The sensor and hardware interface (2) contains a microphone facing the ear canal. It is preferred to be applied in the structure of wearable acoustic devices. 15 The echo signal detected by the microphone is sent to the deconvolution motor (3) is transmitted. The deconvolution motor (3) transmits acoustic signals to the ear canal. It processes the scan signal together with the echo signal returning from the ear canal. In the deconvolution motor (3), acoustic scan sent into the ear canal The frequency domain representation of the signal is the input signal, detected by the microphone, 20 The frequency domain representation of the echo signal is shown as the output signal. The acoustic transmission characteristics specific to the ear canal are being evaluated. H(f) = Y(f) / X(f) It is determined using the relationship. Here, X(f) is the transmitted acoustic scan. Y(f) is the perceived echo signal and H(f) is the acoustic transmission signal of the ear canal. It expresses its characteristic. The acoustic scan signal sent as a result of the deconvolution process The effect on the echo signal is being isolated, and the acoustics specific to the ear canal are being analyzed. The transfer characteristic is obtained. Thus, in the authentication process, the raw data is used. Instead of directly using the echo signal, 30 Acoustic transmission characteristics are being evaluated. In one application of the invention, the deconvolution motor (3) uses a microphone time-dependent attenuation behavior of the echo signal detected by This determines the attenuation time of the echo signal. 9 and the determined damping time is the recorded damping time of live soft tissue. It is compared using the duration criterion. The determined damping time is recorded for living soft tissue. if it matches the damping time criterion, it represents the live acoustic response. an acoustic viability data, or 5 from non-viable material if there is no match. acoustic vibrancy data representing the resulting acoustic response is being created. The generated acoustic liveness data is fed into a pure cosine decision engine (5) is transmitted. The pure cosine decision engine (5) uses current acoustic feature data with the main the result of the comparison between the acoustic profiles and the acoustic vibrancy in question 10 By evaluating the data together, it determines whether to grant or deny access. Acoustic transmission obtained by the deconvolution engine (3) The characteristic is transferred to the feature extraction module (4). Feature extraction module (4) describes the acoustic transmission characteristic of mel-frequency cepstral It processes using coefficients and converts them to a three-dimensional feature matrix. It transforms. The three-dimensional feature matrix is based on the audio output obtained during the validation session. current acoustic property data representing the acoustic transmission characteristics of the channel This creates acoustic data obtained from different verification sessions. Transfer characteristics are made comparable through a common feature representation. is being brought. Current acoustic properties generated by the feature extraction module (4) The data is transmitted to the pure cosine decision engine (5). The pure cosine decision engine (5), current acoustic property data found in the adaptive profile database (6) and It compares to the main acoustic profile associated with the user. 25 The comparison process involves comparing current acoustic property data with the main acoustic profile. This is done based on the angular and shape similarities between them. Pure In the cosine decision engine (5), only the absolute amplitude values of the feature data No, it is the relative trend and structural similarity of the data in question. is being evaluated. 30 A pure cosine decision engine (5) made a decision as a result of the comparison. It generates output. The current acoustic properties data forms the main acoustic profile. If the match is correct, access is granted; if it does not match, access is denied. is being brought. Adaptive profile database (6), the user’s main acoustic profile It stores the main acoustic profile, which is based on the user's previous successful performance. It includes acoustic property data obtained from verifications. 5 If the identity verification process is successful, the relevant verification will be sent. Current acoustic property data obtained in the session were added to the adaptive profile database (6) is being transmitted. Current acoustic characteristic data is used as the main acoustic data, with a ten percent weighting. It is included in the profile, and the main acoustic profile is updated accordingly. In one application of the invention, the deconvolution engine (3) feature extraction 10 The operations performed by the module (4) and the pure cosine decision engine (5), by an embedded processor located on the wearable acoustic device This is being carried out. Thus, the acoustic transmission characteristics of the ear canal are determined. identification, creation of a three-dimensional feature matrix, and authentication. The decision is made without the need for an external network connection. (wearable 15) This is performed on an acoustic device. The system also provides bidirectional communication between the wearable acoustic device and the remote server. It includes a communication interface that enables data transfer to the adaptive profile database. (6) Update data related to the communication interface is sent to the remote server. The current main acoustic profile is being transmitted and the adaptive profile 20 is received from the remote server. is recorded in the database (6). If there is no connection to the remote server, the authentication process will fail. Master acoustic profile and embedded processor stored on the wearable acoustic device. It is maintained using adaptive technology. If the connection is re-established, adaptive technology will be used. Update data relating to the profile database (6) is synchronized with the remote server 25 is being done. Profile updates can only be processed after successful authentication. This is carried out using the obtained acoustic property data. Access Feature data obtained in a validation session that resulted in rejection Not included in the acoustic profile. 30 Current acoustic properties data is included in the main acoustic profile with a ten percent weighting. the main acoustic profile of the data obtained from a single validation session 11 This prevents it from completely changing. However, in the cartilage structure... micro-changes, weight gain and loss, and microphone holding habits Successful validations of limited differences arising from changes This makes it possible to reflect it onto the main acoustic profile. In the system in question, the authentication process is performed by a dynamic signal generator (1) 5 by generating an acoustic scanning signal specific to the verification session It begins. The generated acoustic scanning signal, acoustic sensor and equipment. is transmitted to the user's ear canal via the interface (2) and from the ear canal A returning echo signal is detected. The acoustic scan signal sent along with the detected echo signal, 10 processed together in the deconvolution motor (3) to create acoustics specific to the ear canal. The transmission characteristics are determined. The determined acoustic transmission characteristics, three features containing mel-frequency cepstral coefficients in the feature extraction module (4) It is converted into a three-dimensional feature matrix. The current acoustic characteristic data obtained was used in the pure cosine decision engine (5), 15 The adaptive profile is compared with the main acoustic profile recorded in the database (6). If a match is identified as a result of the comparison, access will be granted. is being created and current acoustic property data is weighted by ten percent as the main acoustic data. It is included in the profile. Access will be denied if no match is found. It is being created and the main acoustic profile is not being changed. 20 Between verification sessions by dynamic signal generator (1) The use of varying acoustic scanning signals, a previously recorded echo the signal can be used directly in a different validation session This makes it difficult. The deconvolution motor (3) sends a signal to the ear canal. The relationship between perceived echo and specific acoustic transmission of the ear canal 25 It determines its characteristics. The feature extraction module (4) analyzes the acoustic transmission characteristic in question. converting into comparable feature data; pure cosine decision engine (5), evaluating the similarity between current characteristic data and the main acoustic profile and adaptive profile database (6), data from successful validations limited to 30 It incorporates a significant amount of weight into the main acoustic profile. Thus, in the system in question, the acoustics specific to the verification session are as follows: the generation of a scanning signal, its transmission to the ear canal, and its return from the ear canal. The perception of echo is a characteristic of acoustic transmission in the ear canal. 12 extraction, conversion to feature matrix, with the recorded master acoustic profile comparison and successful verification of the main acoustic profile The update processes are carried out within the same technical infrastructure. 10 20 30
Claims
13 REQUESTS 1. An ear canal biometric authentication system, characterized by: - start frequency, end frequency and phase for each validation session randomly changing the audible frequency band, ultrasonic frequency 5 the band or both audible and ultrasonic frequency bands dynamic signal that includes dynamic chirp acoustic scanning signal producing dynamic signal producer (1), - the dynamic chirp acoustic scan signal in question is transmitted to the user's ear. a 10 positioned facing the ear canal, transmitting signals to the ear canal The ear canal detects the echo signal returning from the speaker. acoustic channel containing a microphone positioned facing the channel sensor and hardware interface (2), - with a dynamic chirp acoustic scanning signal sent into the ear canal Deconvolution of the echo signal detected by the microphone 15 by subjecting it to a process that provides acoustic transmission specific to the ear canal. deconvolution engine (3) which determines its characteristic - the acoustic transmission characteristic in question is mel-frequency cepstral an attribute that converts its coefficients into a three-dimensional feature matrix inference module (4), 20 - current acoustic property data created with a three-dimensional feature matrix Angular and shape-related main acoustic profile associated with the user compares based on similarity and reaches an agreement as a result of the comparison. Pure cosine decision engine (5) which generates approval or access denial and - Stores the main acoustic profile associated with the user, access permission 25 Current acoustic characteristics obtained in the created verification session including the data with a ten percent weight in the main acoustic profile and denying access Current acoustic characteristics obtained in the created verification session adaptive profile database (6) which does not include data in the main acoustic profile It includes. 30 14 2. According to Claim 1, the ear canal is a biometric authentication system, and its features are: dynamic chirp acoustic scanning signal is generated in the ultrasonic frequency band. It includes a dynamic signal generator (1).
3. According to Claim 1, the ear canal is a biometric authentication system, and its features are: 5 where the speaker and microphone are positioned facing the ear canal true wireless stereo headphones or with active noise cancellation It includes an acoustic sensor and hardware interface (2) suitable for the headphone structure.
4. According to Claim 1, the ear canal is a biometric authentication system, and its features are: frequency of dynamic chirp acoustic scan signal sent into ear canal The frequency of the echo signal detected by the microphone, as shown in the field, is 10 processing its representation in the field together, the dynamic chirp acoustics sent to isolate the effect of the scanning signal on the echo signal and the ear to obtain the acoustic transmission characteristics specific to the channel It includes a structured deconvolution engine (3).
5. According to Claim 1, the ear canal is a biometric authentication system, and its feature is; 15 the attenuation time of the echo signal detected by the microphone to determine and the determined damping time of living soft tissue Acoustic vibrancy data compared with recorded damping time metric. deconvolution engine structured to generate (3) and the said acoustic vibrancy data combined with current acoustic characteristics data and main acoustic profile 20 to be evaluated together with the comparison result It includes a structured pure cosine decision engine (5).
6. According to Claim 1, the ear canal is a biometric authentication system, and its features are: deconvolution motor (3) located on the wearable acoustic device 25 by feature extraction module (4) and pure cosine decision engine (5) an embedded processor configured to execute the operations performed It includes.
7. According to claim 6, the ear canal is a biometric authentication system, and its feature is; Bidirectional data transfer between wearable acoustic device and remote server. a communication interface structured to provide and update data 30 transferring data to the remote server via the communication interface in question, remote configured to record the current master acoustic profile received from the server. It includes an adaptive profile database (6). 10 20 30