A model construction and detection method for ear impression scanning distortion detection

Abstract

The embodiment of the present invention discloses a model construction and detection method for detecting distortion in ear impression scanning, which relates to the technical field of hearing aid digital model processing. The model construction method includes: obtaining multiple ear impressions and the annotation results of the distorted patches in each ear impression; solving the patch mean and standard deviation of each ear impression; continuously adjusting the variable parameters #imgabs4# and #imgabs5# in the ear impression scanning distortion detection model #imgabs3# based on the patch mean #imgabs1# and standard deviation #imgabs2# of each ear impression #imgabs0#, and determining the model effect based on the judgment result of the detection model and the annotation result after each adjustment; determining the final detection model based on the values of #imgabs6# and #imgabs7# with the best model effect; wherein the detection model indicates: the patch in the ear impression #imgabs8# with an area #imgabs9# greater than #imgabs10# is judged as a distorted patch. This embodiment can achieve rapid detection of ear impression scanning distortion, and the detection effect is excellent.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of hearing aid digital model processing, and in particular to a model construction and detection method for ear impression scanning distortion detection. Background Art

[0002] In order to make a customized hearing aid for a user, it is necessary to first obtain the user's ear impression. The conventional way to obtain an ear impression is: use a syringe to inject the ear sample paste into the user's external auditory canal, and gradually withdraw it along the reverse thrust of the ear sample paste; wait for the ear sample paste to solidify, and then remove the solidified ear sample from the user's ear; then, use a 3D scanner to perform a 3D scan of the user's ear sample to obtain a digital 3D model, i.e., the ear impression. However, when scanning the 3D ear sample, due to factors such as different brands of scanners and the placement of the ear sample in the scanner, the scanned ear impression may be distorted. For example, some parts may not be scanned sufficiently, or they may not be scanned due to placement reasons, resulting in differences between the ear impression and the user's actual ear sample in these parts.

[0003] A qualified ear impression is the key to determining the quality of a custom-made hearing aid. A distorted ear impression will result in poor fit between the custom-made hearing aid and the ear canal, and may even mislead the selection of custom-made hearing aid models, leading to problems such as poor wearing comfort, howling, and reduced gain.

[0004] Currently, model makers usually judge whether the key parts of hearing aids are distorted through careful observation. This not only requires judging each part of the ear canal one by one, which is time-consuming and labor-intensive, but also involves a certain degree of subjectivity. For inexperienced model makers, the accuracy of the judgment cannot be guaranteed. Summary of the Invention

[0005] The embodiment of the present invention provides an ear impression scanning distortion detection model construction and detection method to solve the above technical problems.

[0006] In a first aspect, an embodiment of the present invention provides a method for constructing an ear impression scanning distortion detection model, comprising:

[0007] Obtain multiple ear impressions and the annotation results of the distorted patches in each ear impression;

[0008] Calculate the mean and standard deviation of each ear impression patch;

[0009] According to the ear impression The patch mean and standard deviation , continuously adjust the ear impression scanning distortion detection model Variable parameters in and , and after each adjustment, determine the model effect according to the judgment result of the detection model and the annotation result;

[0010] According to the best model and Take the value and determine the final detection model;

[0011] Wherein, the detection model represents: each ear impression Medium area Greater than The patch is judged as a distorted patch.

[0012] In a second aspect, an embodiment of the present invention provides a method for detecting distortion in ear impression scanning, comprising:

[0013] obtaining an ear impression to be tested and different ear-like regions in the ear impression;

[0014] Calculating the mean and standard deviation of the ear impressions;

[0015] Substituting the patch mean and standard deviation into the final detection model to obtain the distorted patch in the ear impression;

[0016] Calculate the total area of the distorted patch in each ear sample area;

[0017] Whether scanning distortion exists in each ear sample region is determined based on the ratio of the total area of the distorted patches in each ear sample region to the total area of the region.

[0018] In a third aspect, an embodiment of the present invention further provides an electronic device, comprising:

[0019] one or more processors;

[0020] a memory for storing one or more programs;

[0021] When the one or more programs are executed by the one or more processors, the one or more processors implement the ear impression scanning distortion detection model construction method or ear impression scanning distortion detection method described in any embodiment.

[0022] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the ear impression scanning distortion detection model construction method or ear impression scanning distortion detection method described in any embodiment.

[0023] In summary, the embodiment of the present invention provides a method for constructing an ear impression scanning distortion detection model. The basic structure of the detection model is determined theoretically. Combined with the specific characteristics of scanning distortion and the influence of scanning point density on distortion detection, the fixed parameters in the model are optimized and variable parameters are introduced. and dynamically adjust parameters Finally, the grid search method was used to determine the parameter combination with the best model performance, and a distortion detection model suitable for all ear impression scanning was obtained.

[0024] This embodiment also provides a method for detecting distortion in ear impression scanning, which can quickly determine whether a 3D ear impression scanned by a scanner is distorted, without the need for a technician. This method decouples distortion detection from the production process, enabling detection as soon as the 3D ear sample is generated. This method is extremely fast, with the detection time for each ear impression being controlled within 100ms, saving significant manpower and material resources. Furthermore, this method offers excellent detection results and can clearly identify the distorted area, providing an important reference for rescanning and fully ensuring the quality of the hearing aid product. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0026] Figure 1 This is a flow chart of a method for constructing an ear impression scanning distortion detection model provided by an embodiment of the present invention;

[0027] Figure 2 This is a flow chart of a method for detecting ear impression scanning distortion provided by an embodiment of the present invention;

[0028] Figure 3 1 is a schematic diagram of ear impression region segmentation provided by an embodiment of the present invention;

[0029] Figure 4 is a schematic diagram of a distortion area provided by an embodiment of the present invention;

[0030] Figure 5 A schematic structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0031] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention are described clearly and completely below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are also within the scope of protection of the present invention.

[0032] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0033] In the description of the present invention, it should also be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood broadly. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on the specific circumstances.

[0034] Figure 1 This is a flow chart of a method for constructing a distortion detection model for ear impression scanning provided by an embodiment of the present invention. This method is applicable to the case where distortion patch detection is performed on an ear impression scanned by a scanning device, and is executed by an electronic device. Figure 1 As shown, the method specifically includes:

[0035] S110: Acquire multiple ear impressions and the labeling results of the distorted patches in each ear impression.

[0036] This embodiment first acquires a large amount of ear impression distortion data as a dataset for constructing an ear impression scanning distortion detection model. Each piece of ear impression distortion data includes the model data for the ear impression, as well as the distorted patches within the ear impression, annotated by experienced ear shapers. This data can be automatically uploaded by individual ear shapers or stores. For example, approximately 20,000 pieces of distortion data can be acquired to construct the dataset.

[0037] Optionally, the model data of the ear impression can be expressed as: grid data consisting of m points and n triangular facets, wherein the coordinate matrix P of each point can be expressed as:

[0038] (1)

[0039] in, Represent the three coordinate values of the i-th point in three-dimensional space.

[0040] The triangle patch matrix T can be expressed as:

[0041] (2)

[0042] in, 、 and They represent the three points in the Qth (Q=1,2,…,n) triangle in the ear impression.

[0043] S120. Calculate the mean and standard deviation of each ear impression.

[0044] Specifically, taking any ear impression in the above data set as an example, first calculate the area of each facet in the ear impression. 、 and Taking the triangle patch as an example, the coordinates of the three points are:

[0045] (3)

[0046] (4)

[0047] (5)

[0048] The area S of the patch can be calculated by the following formula:

[0049] (6)

[0050] (7)

[0051] (8)

[0052] (9)

[0053] Among them, AB and AC are the vectors of the two edges in the triangle. means multiplying the corresponding elements of two vectors. They are the three elements in the multiplication result.

[0054] The above calculation is performed on each patch to obtain the area of each patch; thus, the average area of all patches is obtained. and standard deviation :

[0055] (10)

[0056] (11)

[0057] in, It means to sum the area S of each of the n patches.

[0058] As mentioned above, the same operation is performed on each ear impression in the data set to obtain each ear impression The mean and standard deviation .

[0059] S130, according to each ear impression The patch mean and standard deviation , continuously adjusting the variable parameters in the ear impression scanning distortion detection model, and determining the model effect according to the judgment result of the detection model and the annotation result after each adjustment.

[0060] Optionally, the ear impression scanning distortion detection model used in this embodiment is:

[0061] ,or (12)

[0062] This model represents the area of the ear impression r Greater than patches, and the area Less than The patches are all judged as distorted patches.

[0063] The model comes from Specifically, for a certain ear impression, although there are distorted patches with too large or too small areas, the areas of most patches are relatively uniform, so the area can be Patches outside the interval are considered as outliers, i.e. distorted patches, where and Represent the mean and standard deviation of the ear impression area respectively. The range of the distortion patch is too arbitrary. In actual application, the effect of a patch that is too small (i.e., too dense scanning points) on the entire ear impression is much smaller than the effect of a patch that is too large (e.g., sparse scanning points, resulting in incomplete ear sample structure). Moreover, how far the distance from the mean is too large or too small also needs to be determined based on the specific characteristics of the ear impression. Therefore, in this embodiment, the above Fixed parameter 3 in is optimized to an adjustable parameter and , the judgment model shown in formula (12) was constructed. In this model, the parameters and Mutually independent, ensuring that oversized areas and undersized areas can have different criteria for judgment; and variable parameters and The mean value of each ear impression patch The impact on distortion detection is to achieve dynamic adjustment of the distortion area threshold with the density of scanning points. For example, in some ear impressions with a high overall density of scanning points, the average value of the patch is The area of the distorted patch is also smaller (compared to the ear impressions with lower overall point density). The judgment threshold of the distorted patch can be lowered while still maintaining effective recognition of the distorted patch.

[0064] After the model structure is determined, the variable parameters in the model are also 、 、 and Optimize and determine a set of optimal parameter values. In one embodiment, the process may include the following steps:

[0065] Step 1: Take the ear impressions The patch mean and standard deviation Substitute them into formula (12) to obtain the detection model of each ear impression. That is, in the detection model of formula (12), the mean and standard deviation are not necessarily the same for each ear impression, but 、 、 and This is the same for every ear impression.

[0066] Step 2: Determine variable parameters 、 、 and For example, for and , we can determine a value range centered on the fixed parameter 3; for and , a value range can be determined based on the average patch area of most scanned models; this embodiment does not impose specific limitations.

[0067] Step 3: Use grid search method to determine 、 、 and Multiple value combinations; and for each value combination, perform operations S0 to S2 respectively:

[0068] S0, traverse each ear impression, and perform the following operations for each ear impression: 、 、 and The current value combination of is substituted into the detection model of the current ear impression, and the ear impression with an area greater than patches, and areas smaller than The patches are all judged to be distorted patches; then, the model effect of the current ear impression is determined according to the judgment result and the labeling result of the current ear impression.

[0069] Optionally, the model effect can be expressed by the F1 value. Specifically, the following four indicators can be calculated based on the judgment results and annotation results:

[0070] True Positive (TP), that is, the number of positive samples correctly predicted by the model as positive;

[0071] False Positive (FP), that is, the number of negative samples that the model incorrectly predicts as positive;

[0072] True Negative (TN), that is, the number of negative samples correctly predicted by the model as negative;

[0073] False Negative (FN): The number of positive samples that the model incorrectly predicts as negative;

[0074] Among them, the positive sample index is the sample of the distorted patch, and the negative sample index is the sample of the non-distorted patch.

[0075] Based on the above indicators, the model’s:

[0076] Precision: , which measures the proportion of samples predicted to be positive that are actually positive, reflecting the accuracy of the model in predicting the positive class; and:

[0077] Recall: , also known as sensitivity, is used to indicate the proportion of actual positive samples that are correctly predicted as positive, reflecting the model's ability to capture positive samples.

[0078] Thus we get the model F1 value: The F1 value is the harmonic mean of precision and recall, which takes into account the balance between the two and can more comprehensively evaluate the model performance.

[0079] S2, average the model effects of each ear impression as the final model effect under the current value combination. Optionally, after the operation of S0, each ear impression corresponds to an F1 value, and all F1 values are averaged to obtain the final model effect of the current value combination. The average can reflect the 、 、 and The generalization ability of the current value combination of for different ear impressions.

[0080] Step 4: Take the model with the best effect 、 、 and The value combination of is taken as the optimal parameter combination. That is, the value combination with the largest F1 mean is taken as the optimal parameter combination.

[0081] S140: Determine a final earprint scanning distortion detection model based on the value combination with the best model effect.

[0082] Specifically, 、 、 and The value combination of will be substituted into formula (12) to obtain the final detection model, which is applicable to all ear impressions and can maintain high detection performance.

[0083] Furthermore, in another specific embodiment, when executing step 4 of S130 above, the F1 value under each value combination can be calculated separately. 、 、 and The correlation between the four parameters and the F1 value. If the correlation of a parameter is less than the set threshold, the parameter can be discarded. and The F1 value does not show any correlation, so this embodiment removes these two parameters and adjusts the ear impression scanning distortion detection model to:

[0084] (13)

[0085] That is, only when the patch is too large will it cause non-negligible model distortion. For the sake of distinction and description, the detection model in formula (12) will be referred to as the initial ear impression scanning distortion detection model.

[0086] For the model shown in formula (13), the operations of S110-S140 can be re-executed to re-determine and The optimal value combination of 、 、 and From the optimal value combination of and The value of is directly substituted into formula (13). This embodiment is not specifically limited.

[0087] In summary, this embodiment provides a method for constructing an ear impression scanning distortion detection model based on The basic structure of the detection model is determined theoretically. Combined with the specific characteristics of scanning distortion and the influence of scanning point density on distortion detection, the fixed parameter 3 is optimized and variable parameters are introduced into the detection model. and , and dynamically adjust parameters and ; Then, a grid search method was used to determine the parameter combination with the best model performance, and a distortion detection model suitable for all ear impression scans was obtained; 、 、 and The influence of the four parameters on the model performance is further removed 、 These two parameters further simplify the model size and improve detection efficiency while ensuring detection performance.

[0088] Based on the above detection model, Figure 2 FIG. 1 is a flow chart of a method for detecting ear impression scanning distortion provided by an embodiment of the present invention. Figure 2 As shown, the method specifically includes:

[0089] S210: Acquire an ear impression to be detected and different ear-sample areas in the ear impression.

[0090] Optionally, the ear-like area includes a bend, two bends, tragus, concha cavity, antitragus, and helix, such as Figure 3 As shown, different colors represent different areas.

[0091] S220: Calculate the surface mean and standard deviation of the ear impression.

[0092] Optionally, the method shown in formulas (6)-(11) can be used to calculate the mean and standard deviation of the current ear impression.

[0093] S230: Substitute the patch mean and standard deviation into the final detection model to obtain the distorted patch in the ear impression.

[0094] Substituting the patch mean and standard deviation calculated by S220 into the model corresponding to the optimal parameter combination formula (12) or (13), the detection model of the current ear impression can be obtained. According to this model, the distorted patch in the ear impression can be determined, such as Figure 4 The gray area in the red frame is shown.

[0095] S240: Calculate the total area of the distorted patches in each ear sample region.

[0096] For each ear sample region t, calculate the total area of the distorted patch in the region , and the total area of all patches in the region Optional, t=1,2,…6, corresponding to the first bend, second bend, tragus, cavum concha, antitragus, and helix respectively.

[0097] S250: Determine whether scanning distortion exists in each ear sample region according to the ratio of the total area of the distorted patches in each ear sample region to the total area of the region.

[0098] For each ear sample area t, if If the value is greater than a set threshold (called the distortion ratio threshold), the area is judged to be distorted. For the entire ear impression, as long as there is distortion in one area, the ear impression is judged to have scan description distortion.

[0099] Furthermore, different distortion ratio thresholds can be determined for different ear sample areas of different types of customized machines according to the type of customized machine.

[0100] For ITE (ITE-full shell) custom-made devices, full-area distortion detection is used. This means that all areas must be distortion-free. Therefore, a low distortion ratio threshold is set for all areas.

[0101] For HS (ITE-half shell) and ITC (In the Canal) custom-made earphones, the distortion in the helix area can be ignored, so a higher distortion ratio threshold is set for this area.

[0102] For CIC (Completely in the Canal) type custom machines, the distortion of the antitragus, tragus, lower half of the concha cavity, and the helix can be ignored, so a higher distortion ratio threshold is set for these areas;

[0103] For MINI CIC (ultra-small complete in-canal) type custom machines, only the distortion of the first and second bends is considered, so a lower distortion ratio threshold is set for these areas, and a higher distortion ratio threshold is set for other areas.

[0104] On this basis, the distortion ratio threshold of each area under each customized machine type can be set manually or automatically determined using the data set and grid search method. Optionally, 20,000 distorted data and 40,000 undistorted data can be collected (distortion and undistortion here are the annotation results); and based on the basic requirements of the thresholds under the above different customized machine types, the value range of the distortion ratio threshold of each area under each customized machine type is determined; then a grid search of the distortion ratio thresholds of different areas is performed, and the distortion area under different distortion ratio thresholds and the judgment result of whether it is distorted are calculated, and compared with the above distorted and undistorted annotation results, the threshold when the F1 value is the largest is taken as the optimal threshold. More details are similar to the above embodiments and will not be repeated here.

[0105] Finally, the distortion ratio thresholds of the first bend, second bend, tragus, concha cavity, antitragus, and helix regions under hearing aid type M are recorded as , then the distortion detection function of each region can be expressed as:

[0106] (14)

[0107] (15)

[0108] Among them, when hour, is 1, the current area is distorted; when When , the detection result F is 1, and the ear impression is distorted.

[0109] In summary, this embodiment provides a method for detecting distortion in ear impression scanning, which can quickly determine whether a three-dimensional ear impression scanned by a scanner is distorted, without the need for a modeler to participate in the entire process. This method decouples distortion detection from the production process, allowing detection to be performed when the three-dimensional ear sample is generated. The detection speed is extremely fast. For example, under the final detection model shown in formula (13), the detection time for each ear impression can be controlled within 100ms, saving a lot of manpower and material resources. At the same time, this method sets different distortion ratio thresholds for different areas according to different customized machine types, further improving the detection effect and clearly identifying the distorted area, providing an important reference for rescanning and fully ensuring the quality of hearing aid products.

[0110] Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention is shown in FIG. Figure 5 As shown, the device includes a processor 60, a memory 61, an input device 62 and an output device 63; the number of processors 60 in the device can be one or more. Figure 5In the embodiment, a processor 60 is used as an example; the processor 60, the memory 61, the input device 62 and the output device 63 in the device can be connected by a bus or other means. Figure 5 The bus connection is taken as an example.

[0111] Memory 61, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the ear impression scanning distortion detection method in the embodiments of the present invention. Processor 60 executes the software programs, instructions, and modules stored in memory 61 to perform various functional applications and data processing of the device, thereby implementing the aforementioned ear impression scanning distortion detection model construction method or ear impression scanning distortion detection method.

[0112] The memory 61 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function; the data storage area may store data generated based on the use of the terminal. Furthermore, the memory 61 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state memory device. In some instances, the memory 61 may further include memory remotely located relative to the processor 60, and these remote memories may be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0113] The input device 62 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 63 may include a display device such as a display screen.

[0114] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method for constructing an ear impression scanning distortion detection model or the ear impression scanning distortion detection method described in any embodiment is implemented.

[0115] The computer storage medium of the embodiments of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device.

[0116] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0117] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0118] Computer program code for performing the operations of the present invention can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as C or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for constructing a distortion detection model for ear impression scanning, characterized in that: include: Obtain multiple ear impressions and the annotation results of the distorted patches in each ear impression; Calculate the mean and standard deviation of each ear impression patch; According to the ear impression The patch mean and standard deviation , continuously adjust the ear impression scanning distortion detection model Variable parameters in and , and after each adjustment, determine the model effect according to the judgment result of the detection model and the annotation result, wherein, Indicates the area of the ear impression patch; According to the best model and Take the value and determine the final detection model; Wherein, the detection model represents: each ear impression Medium area Greater than The patch is judged as a distorted patch.

2. The ear impression scanning distortion detection model construction method according to claim 1, characterized in that: According to the ear impression The patch mean and standard deviation , continuously adjust the ear impression scanning distortion detection model Variable parameters in and ,include: Make each ear impression The patch mean and standard deviation , respectively substituted into the ear impression scanning distortion detection model , obtain the detection model of each ear impression; Determine variable parameters and The value range of In the range of values, the grid search method is used to determine and Multiple value combinations of .

3. The ear impression scanning distortion detection model construction method according to claim 1, characterized in that: Determining the model effect according to the judgment result of the detection model and the labeling result after each adjustment includes: against and For each value combination of , perform the following operations respectively: S1. Traverse each ear impression and perform the following operations on each ear impression: and The current value combination of is substituted into the detection model of the current ear impression, and the area greater than The face patch is judged as a distorted face patch; and a model effect of the current ear impression is determined according to the judgment result and the labeling result of the current ear impression; S2. The model effects of each ear impression are averaged and the final model effect is obtained.

4. The ear impression scanning distortion detection model construction method according to claim 1, characterized in that: Determining the model effect according to the judgment result of the detection model and the labeling result includes: Calculate the true positives, false positives, true negatives, and false negatives of any ear print in the model detection according to the detection model judgment result and the labeling result of the ear print; The F1 value of any ear impression detection model is calculated based on the true positive examples, false positive examples, true negative examples, and false negative examples, wherein the F1 value is the harmonic mean of precision and recall.

5. The ear impression scanning distortion detection model construction method according to claim 1, characterized in that: According to the ear impression The patch mean and standard deviation , continuously adjust the ear impression scanning distortion detection model Variable parameters in and Previously, it also included: According to the ear impression The patch mean and standard deviation , continuously adjusting the variable parameters in the distortion detection model of the initial ear impression scan 、 、 and , wherein the initial ear impression scanning distortion detection model is: ,or , The initial ear impression scanning distortion detection model represents: Medium area Greater than patches, and the area Less than The patches are all judged as distorted patches; After each adjustment, the model effect is determined based on the judgment result of the initial ear impression scanning distortion detection model and the annotation result, and the model effect is determined. 、 、 and The influence of each parameter on the model effect; Remove parameters with low impact and , the initial ear impression scanning distortion detection model is adjusted to .

6. A method for detecting distortion in ear impression scanning, characterized in that: include: obtaining an ear impression to be tested and different ear-like regions in the ear impression; Calculating the mean and standard deviation of the ear impressions; Substituting the patch mean and standard deviation into the final detection model according to any one of claims 1 to 5 to obtain the distorted patch in the ear impression; Calculate the total area of the distorted patch in each ear sample area; According to the ratio of the total area of the distorted patches in each ear sample region to the total area of the region, it is determined whether each ear sample region has scanning distortion.

7. The ear impression scanning distortion detection method according to claim 6, characterized in that: The determining whether scanning distortion exists in each ear sample region according to the ratio of the total area of the distorted patches to the total area of the region includes: According to the type of the customization machine, a threshold value of the ratio of the total area of the distorted patches to the total area of each ear sample region of the ear impression is called; If the ratio of any ear sample area is greater than a corresponding ratio threshold, it is determined that scanning distortion exists in any ear sample area.

8. The ear impression scanning distortion detection method according to claim 7, characterized in that: Before calling the threshold value of the ratio of the total area of the distorted patches to the total area of the region in each ear sample region of the ear impression according to the type of the customization machine, the method includes: Obtain multiple ear impressions and label results of whether each ear sample area in each ear impression is distorted; For any ear sample area of any type of customized machine, determine the value range of the threshold value of the ratio of the total area of the distorted surface patches to the total area of the region; Within the value range, a grid search method is used to determine multiple values of the ratio threshold; The final value of the ratio threshold is determined according to the judgment result and the marking result of whether any ear sample area is distorted under each value.

9. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the ear impression scanning distortion detection model construction method described in any one of claims 1-5, or the ear impression scanning distortion detection method described in any one of claims 6-8.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, which, when executed by a processor, implements the ear impression scanning distortion detection model construction method described in any one of claims 1-5, or the ear impression scanning distortion detection method described in any one of claims 6-8.

Images