An abnormal auricle morphology data model construction method and system

Abstract

The application relates to the technical field of intelligent medical treatment, in particular to an abnormal auricle morphology data model construction method and system. The method comprises the following steps: acquiring auricle morphology data, and preprocessing the auricle morphology data to obtain auricle morphology pretreatment data; a normal auricle morphology threshold is constructed, the auricle morphology pretreatment data is subjected to abnormal feature identification according to the normal auricle morphology threshold to obtain abnormal auricle morphology data; abnormal auricle morphology fusion data is obtained according to the abnormal auricle morphology data to construct an abnormal auricle morphology joint feature tensor; abnormal auricle morphology space features are obtained by performing abnormal space mapping according to the abnormal auricle morphology joint feature tensor; and an abnormal auricle morphology data model is constructed based on the abnormal auricle morphology space features. The abnormal auricle morphology data model is constructed, a quantitative basis is provided for individualized treatment of abnormal auricles, and the accuracy of abnormal auricle diagnosis is significantly improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent medical technology, and in particular to a method and system for constructing a data model of abnormal auricle morphology. Background Technology

[0002] Traditional auricular morphology analysis primarily relies on two-dimensional imaging or manual measurement, assessing morphological abnormalities through auricular geometric parameters. Some methods combine mechanical properties such as elastic modulus to analyze tissue characteristics. In terms of data processing, the standard workflow includes registration, noise reduction, feature extraction, and threshold comparison. Anomaly identification is based on a single static threshold or expert experience. For example, the diagnosis of microtia typically involves comparing the geometric template of a normal auricle with differences in elastic modulus to determine tissue developmental abnormalities; cup ear is assessed by measuring the deviation of the helix curl angle from the normal range.

[0003] The shortcomings of traditional methods include: First, limited multimodal data integration capabilities; traditional methods struggle to effectively fuse multimodal data, leading to insufficient information utilization. Second, the noise reduction and feature extraction mechanisms lack adaptability; fixed-parameter filtering methods cannot adapt to individual differences or dynamic physiological changes. Third, anomaly identification relies on a single threshold or linear model, neglecting multi-scale and multi-level abnormal features, easily leading to missed or misdiagnosed cases. Furthermore, feature fusion strategies are simplistic, failing to consider nonlinear interactions and spatiotemporal evolution between features, resulting in insufficient model dynamism and difficulty in capturing the spatial distribution and temporal development trends of anomalies. These shortcomings make traditional methods less sensitive and specific in the diagnosis of complex auricular deformities, failing to meet the needs of precision medicine. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for constructing a data model of abnormal auricle morphology.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a method for constructing an abnormal auricle morphology data model, the method comprising the following steps: acquiring auricle morphology data and preprocessing the auricle morphology data to obtain auricle morphology preprocessed data; constructing a normal auricle morphology threshold, and performing abnormal feature identification on the auricle morphology preprocessed data based on the normal auricle morphology threshold to obtain abnormal auricle morphology data; obtaining abnormal auricle morphology fusion data based on the abnormal auricle morphology data to construct an abnormal auricle morphology joint feature tensor; performing abnormal space mapping based on the abnormal auricle morphology joint feature tensor to obtain abnormal auricle morphology spatial features; and constructing an abnormal auricle morphology data model based on the abnormal auricle morphology spatial features. This invention achieves high-precision modeling and diagnosis of auricular deformities through multimodal data fusion and dynamic analysis technology; data preprocessing improves the signal-to-noise ratio and registration accuracy of the original data; a progressive anomaly recognition mechanism enhances the sensitivity of complex deformity identification; a fusion strategy effectively integrates sensitive features of auricular morphology to construct a joint feature tensor to quantify the spatial distribution and evolution trend of abnormalities; finally, an auricular morphology data model is constructed based on a dynamic evolution tensor network, embedding a biomechanical constraint mechanism to provide a quantitative basis for personalized treatment; it overcomes the limitations of traditional static threshold methods and significantly improves the accuracy of auricular abnormality diagnosis.

[0006] Optionally, the step of acquiring auricle morphology data and preprocessing the auricle morphology data to obtain auricle morphology preprocessed data includes: acquiring geometric point cloud data, elastic modulus matrix, and blood flow heatmap of the auricle to construct multimodal data, and using the multimodal data as the auricle morphology data; constructing biomechanical constraints, registering the auricle morphology data based on the biomechanical constraints to obtain auricle morphology registration data; establishing an adaptive manifold denoising mechanism, combining the auricle morphology registration data to obtain auricle morphology denoising data, and using the auricle morphology denoising data as the auricle morphology preprocessed data. This invention significantly improves the accuracy of auricle morphology analysis through multimodal data and biomechanical constraint registration; the data preprocessing process significantly improves the data signal-to-noise ratio and structural consistency, providing high-quality input for subsequent abnormal feature identification and model construction, and contributing to the refined analysis of complex auricle deformities.

[0007] Optionally, establishing the adaptive manifold denoising mechanism includes: obtaining the blood flow oscillation index of the auricle based on the blood flow heatmap, and obtaining the dynamic denoising intensity based on the blood flow oscillation index; constructing a multimodal curvature manifold of the auricle by performing multimodal joint curvature analysis on the auricle morphology data; obtaining the biomechanical edge of the auricle based on the biomechanical constraints and the multimodal curvature manifold; and constructing the adaptive manifold denoising mechanism by combining the dynamic denoising intensity, the multimodal curvature manifold, and the biomechanical edge. This invention significantly improves the signal-to-noise ratio and structural fidelity of auricle morphology data by dynamically adjusting the denoising intensity; and by dynamically adjusting the denoising intensity, multimodal manifold modeling, and fusion of biomechanical priors, it retains anomaly-sensitive features while suppressing random noise, providing a high-quality data foundation for subsequent anomaly identification.

[0008] Optionally, the step of constructing a normal auricle morphology threshold and obtaining abnormal auricle morphology data by identifying abnormal features in the preprocessed auricle morphology data based on the normal auricle morphology threshold includes: constructing a developmental trajectory function of a normal auricle; obtaining the normal auricle morphology threshold based on the developmental trajectory function; the normal auricle morphology threshold including a curvature threshold, an elastic modulus threshold, and a developmental state calibration threshold of the normal auricle; establishing a progressive abnormality identification mechanism based on the normal auricle morphology threshold; and combining the progressive abnormality identification mechanism and the preprocessed auricle morphology data to obtain the abnormal auricle morphology data. This invention constructs a dynamic developmental trajectory function and a progressive abnormality identification mechanism, significantly improving the accuracy and adaptability of auricle abnormality detection. It effectively integrates morphological and biomechanical information, accurately identifies auricle deformities, and provides a quantitative basis for early intervention.

[0009] Optionally, the step of establishing a progressive anomaly recognition mechanism based on the normal auricle morphology threshold includes: obtaining the auricle confidence level; constructing a multi-level confidence interval for the auricle by combining the normal auricle morphology threshold; performing coupled analysis on the preprocessed auricle morphology data by combining the developmental trajectory function and the multi-level confidence interval for the auricle to obtain the auricle developmental deviation-lesion correlation index; and constructing the progressive anomaly recognition mechanism based on the multi-level confidence interval for the auricle and the auricle developmental deviation-lesion correlation index. This invention uses multi-level confidence intervals to hierarchically screen abnormal regions, combines the developmental trajectory function to quantify the correlation between morphological deviation and lesions, and dynamically adapts to individual developmental stage differences; effectively distinguishes between normal variations and pathological abnormalities, improves the sensitivity of auricle deformity recognition, reduces the missed diagnosis rate, and provides a quantitative basis for accurate diagnosis.

[0010] Optionally, the step of obtaining abnormal auricle morphology fusion data based on the abnormal auricle morphology data to construct an abnormal auricle morphology joint feature tensor includes: extracting abnormal auricle morphology sensitive features from the abnormal auricle morphology data at multiple scales to obtain abnormal auricle morphology sensitive features; establishing a feature fusion strategy, combining the abnormal auricle morphology sensitive features and the feature fusion strategy to obtain the abnormal auricle morphology fusion data; and performing spatiotemporal evolution based on the abnormal auricle morphology fusion data to obtain the abnormal auricle morphology joint feature tensor. This invention significantly improves the representational ability of abnormal auricle morphology through multi-scale feature extraction and spatiotemporal evolution analysis; provides multi-dimensional quantitative evidence for the accurate diagnosis and pathological mechanism analysis of abnormal auricles, and enhances the model's ability to characterize dynamic abnormal processes.

[0011] Optionally, establishing the feature fusion strategy includes: obtaining a sensitive feature interaction matrix and a sensitive feature distribution entropy based on the abnormal sensitive features; combining the sensitive feature interaction matrix and the sensitive feature distribution entropy to obtain an entropy-weighted feature interaction matrix; and establishing the feature fusion strategy based on the entropy-weighted feature interaction matrix. This invention quantifies feature associations through interaction matrices, assesses feature uncertainty through distribution entropy, and optimizes feature weights through entropy weighting, effectively integrating multimodal abnormal information and improving the robustness of fused data in representing complex auricular deformities.

[0012] Optionally, the step of obtaining the spatial features of the abnormal auricle morphology by performing abnormal spatial mapping based on the joint feature tensor of the abnormal auricle morphology includes: constructing a biomechanical potential field of the abnormal auricle based on the joint feature tensor of the abnormal auricle morphology; performing abnormal spatial mapping on the biomechanical potential field of the abnormal auricle to obtain an abnormal spatial feature vector; and obtaining the spatial features of the abnormal auricle morphology based on the abnormal spatial feature vector. This invention dynamically quantifies the mechanical distribution and spatial correlation of abnormal regions by constructing a biomechanical potential field of the abnormal auricle; the abnormal spatial mapping transforms the potential field into a feature vector, accurately characterizing the spatial location and range of abnormalities, improving the spatial positioning accuracy of auricle deformities, and providing a quantitative spatial basis for personalized correction plans.

[0013] Optionally, the step of constructing an abnormal auricle morphology data model based on the spatial features of the abnormal auricle morphology includes: constructing a dynamic evolutionary tensor network as the model architecture; establishing a model training mechanism based on the biomechanical constraints; and constructing the abnormal auricle morphology data model by combining the spatial features of the abnormal auricle morphology, the model architecture, and the model training mechanism. This invention quantifies the spatiotemporal evolution of abnormalities through a dynamic evolutionary tensor network architecture, combines biomechanical constraints to ensure the model conforms to the mechanical properties of auricle tissue, effectively integrates multimodal spatial features, and constructs an abnormal auricle morphology data model that combines morphological, mechanical, and physiological correlations; providing quantitative dynamic prediction for personalized treatment plans and improving the clinical intervention effect for complex auricle deformities.

[0014] Secondly, this invention provides a system for constructing a data model of abnormal auricle morphology. The system executes the method for constructing a data model of abnormal auricle morphology provided by this invention. The system includes an input device, an output device, a processor, and a memory, all interconnected. The memory stores a computer program, which includes program instructions, and the processor is configured to call these instructions. This invention enables real-time input of multimodal data through the input device, and, combined with the powerful parallel computing capabilities of the processor, rapidly executes the construction of a dynamic evolutionary tensor network and biomechanical constraint training. The memory provides large-capacity, high-speed data read / write support, ensuring the continuity of data processing. The output device presents abnormal spatial features and model evolution results, providing stable and efficient hardware support for the accurate diagnosis and treatment of complex auricle deformities. Attached Figure Description

[0015] Figure 1 This is a flowchart of a method for constructing an abnormal auricle morphology data model according to an embodiment of the present invention;

[0016] Figure 2 This is a system framework diagram for constructing an abnormal auricle morphology data model according to an embodiment of the present invention. Detailed Implementation

[0017] Specific embodiments of the present invention will now be described in detail. It should be noted that the embodiments described herein are for illustrative purposes only and are not intended to limit the invention. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be apparent to those skilled in the art that these specific details are not necessary to practice the invention. In other instances, well-known circuits, software, or methods have not been specifically described to avoid obscuring the invention.

[0018] Throughout this specification, references to "an embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the invention. Therefore, the phrases "in an embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. Furthermore, specific features, structures, or characteristics can be combined in one or more embodiments or examples in any suitable combination and / or sub-combination. Moreover, those skilled in the art will understand that the illustrations provided herein are for illustrative purposes and are not necessarily drawn to scale.

[0019] Please see Figure 1 An embodiment of the present invention provides a method for constructing an abnormal auricle morphology data model, the method comprising the following steps:

[0020] S1. Obtain auricular morphology data and preprocess the auricular morphology data to obtain auricular morphology preprocessed data.

[0021] Specifically, S1 includes the following steps:

[0022] S11. Obtain geometric point cloud data, elastic modulus matrix and blood flow heat map of the auricle to construct multimodal data, and use the multimodal data as the auricle morphology data.

[0023] In this embodiment, a high-precision handheld 3D scanner is used to scan the surface of the auricle using structured light projection technology to obtain raw point cloud data. The scanning resolution is set to 0.1 mm and the sampling frequency to 10 Hz to ensure coverage of all anatomical structures, including the helix, antihelix, and earlobe. During the scanning process, the scanning head must be kept perpendicular to the auricle surface, and multi-angle scanning is employed to avoid occlusion. Noise reduction is performed on the raw point cloud data, outliers are removed, and multi-view point cloud alignment is performed. Holes are filled using Poisson surface reconstruction, ultimately generating geometrically uniform point cloud data.

[0024] In this embodiment, a biomechanical testing platform was used to measure the mechanical properties of auricular tissue. Twenty to thirty feature points were selected on the auricular surface, and a compressive force was applied at a speed of 0.5 mm / s using a cylindrical probe with a diameter of 2 mm. The compressive force-auricular deformation curve was recorded. The elastic modulus was obtained by analyzing this curve, and considering tissue anisotropy, the test was repeated in the X, Y, and Z directions. The elastic modulus values ​​were mapped to the corresponding nodes in the geometric point cloud data to construct an elastic modulus matrix.

[0025] In this embodiment, a high-resolution infrared thermal imager was used to capture the surface temperature distribution of the auricle in a constant-temperature environment. During imaging, the lens was kept 30 cm away from the auricle, the exposure time was 1 / 60 s, and data was continuously acquired for 30 seconds to obtain stable data. Simultaneously, physiological parameters such as the subject's heart rate and blood pressure were recorded, and the relationship between blood flow and temperature was calibrated using a linear regression model. An adaptive threshold algorithm was used to segment the auricle region, and interpolation was used to improve the resolution of the thermal map to match the geometric point cloud data, ultimately obtaining a blood flow thermal map.

[0026] Furthermore, using geometric point cloud data as the reference coordinate system, a coordinate transformation method based on feature points is employed: 10-15 anatomical landmarks are extracted from the surface of the auricle to establish a spatial mapping relationship between the elastic modulus matrix and the blood flow heatmap. An interpolation algorithm is then used to map the elastic modulus matrix and the blood flow heatmap to each node of the geometric point cloud, ultimately generating a five-dimensional feature vector containing three-dimensional coordinates, elastic modulus, and blood flow velocity to construct a multimodal dataset.

[0027] S12. Construct biomechanical constraints, and register the auricle morphology data based on the biomechanical constraints to obtain auricle morphology registration data.

[0028] In this embodiment, biomechanical constraints are constructed by combining the mechanical properties and anatomical features of auricular tissue. First, a database of mechanical parameters for normal auricles is established: the elastic modulus of normal auricular tissue is tested, the anisotropic properties of regions such as the helix, antihelix, and earlobe are distinguished, and Poisson's ratio is measured. Second, anatomical landmarks on the surface of the auricle are extracted as spatial constraint references. Finally, biomechanical constraints are constructed based on the theory of continuum mechanics, including constraints on the range of mechanical parameters, spatial constraints on anatomical landmarks, and constraints on minimizing deformation energy.

[0029] Furthermore, the registration process uses geometric point clouds as a reference and combines biomechanical constraints to achieve multimodal data registration. First, preliminary spatial alignment of the auricle morphology data is performed based on coordinate transformation and interpolation. Then, biomechanical constraints are introduced to optimize the registration accuracy: spatial constraints of anatomical landmarks are used as hard constraints to obtain the positional deviation of the landmarks. At the same time, the deformation process of the auricle is simulated by combining finite element analysis, and the constraint of minimizing deformation energy is used as a soft constraint, requiring the elastic modulus distribution after registration to satisfy the force balance equation. Finally, auricle morphology registration data is obtained through iterative optimization to ensure the spatial consistency between geometric shape and mechanical properties.

[0030] S13. Establish an adaptive manifold noise reduction mechanism, combine the auricle morphology registration data to obtain auricle morphology noise reduction data, and use the auricle morphology noise reduction data as the auricle morphology preprocessing data.

[0031] In this embodiment, the blood flow oscillation index of the auricle is obtained based on the blood flow heatmap, and the dynamic noise reduction intensity is obtained based on the blood flow oscillation index; a multimodal joint curvature analysis is performed on the auricle morphology data to construct a multimodal curvature manifold of the auricle; the biomechanical edge of the auricle is obtained based on biomechanical constraints and the multimodal curvature manifold; and an adaptive manifold noise reduction mechanism is constructed by combining the dynamic noise reduction intensity, the multimodal curvature manifold, and the biomechanical edge.

[0032] The temporal temperature change curves of various points on the auricle surface were extracted based on blood flow thermograms. The temperature signal was segmented using a sliding window method, and the spectral characteristics of each segment were analyzed using short-time Fourier transform to calculate the blood flow oscillation index. The blood flow oscillation index is defined as the ratio of energy in the low-frequency (0.1Hz-0.5Hz) to the high-frequency (0.5Hz-2Hz) bands, reflecting the stability of local blood flow: a higher blood flow oscillation index indicates smoother blood flow fluctuations and stronger noise interference. Subsequently, the noise reduction intensity was dynamically adjusted according to the blood flow oscillation index value: in the region where the blood flow oscillation index > 1.5 (the stable blood flow region), high-intensity noise reduction was set (e.g., a filter kernel radius of 3mm); in the region where the blood flow oscillation index < 1.0 (the fluctuating blood flow region), the noise reduction intensity was reduced (e.g., a filter kernel radius of 1mm) to preserve potential abnormal signals.

[0033] Based on auricular morphology registration data, a curvature manifold is constructed by integrating geometric, mechanical, and blood flow modal information. First, geometric curvature is calculated from geometric point cloud data to reflect surface concavity and convexity; mechanical curvature is calculated from the elastic modulus matrix, and the change in elastic modulus gradient is analyzed using the finite difference method; temperature curvature is obtained based on blood flow thermograms, and blood flow dynamics are evaluated based on temperature gradients; then, a weighted fusion strategy is adopted: in mechanically sensitive areas such as the helix and antihelix, the weight of mechanical curvature is increased; in blood flow sensitive areas such as the earlobe and triangular fossa, the weight of temperature curvature is increased; geometric curvature is used as the basic weight; finally, a multimodal curvature manifold is obtained to quantify the comprehensive morphological characteristics of various points on the auricular surface.

[0034] By combining biomechanical constraints with a multimodal curvature manifold, the edges of key auricular structures are identified. First, the auricle is simulated under stress and deformation to extract mechanically active regions; simultaneously, a 2mm radius around anatomical landmarks is designated as a hard constraint region. Then, in the multimodal curvature manifold, curvature abrupt change points are selected as candidate edges, and the intersection of the mechanically active and hard constraint regions is used for verification: if a candidate edge is located within either the mechanically active or hard constraint region, it is retained as a biomechanical edge; otherwise, it is discarded. Finally, the biomechanical edges of the auricle are obtained, clearly defining the boundaries of key structures that need to be retained.

[0035] An adaptive manifold denoising mechanism is constructed by combining dynamic denoising intensity, multimodal curvature manifold, and biomechanical edge parameters. In non-biomechanical edge regions, anisotropic diffusion filtering is employed, with the diffusion coefficient jointly determined by dynamic denoising intensity and multimodal curvature, ensuring stronger diffusion in low-curvature, high-denoising-intensity regions and effectively smoothing noise. In biomechanical edge regions, edge-preserving filtering is switched to, with spatial and value domain weights set to retain edge details. During filtering, the consistency between blood flow signals and mechanical parameters is monitored in real time. If the blood flow oscillation index changes by more than 10% after filtering, the process reverts to the previous step and readjusts the parameters to avoid over-denoising.

[0036] In this embodiment, the auricle morphology registration data is denoised using an adaptive manifold denoising mechanism to obtain denoised auricle morphology data. After filtering, the denoising quality is evaluated in three dimensions: geometric dimension checks the point cloud density (target > 50 points / mm). 2 The following dimensions were verified: mechanical dimension for modulus gradient continuity (gradient variance < 0.05 kPa / mm), and blood flow thermogram and geometry registration error (error < 0.2 mm). If any dimension fails to meet the standard, the dynamic noise reduction intensity or filtering parameters are adjusted and reprocessed to finally obtain the auricle morphology noise reduction data.

[0037] S2. Construct a normal auricle shape threshold, and perform abnormal feature identification on the auricle shape preprocessing data based on the normal auricle shape threshold to obtain abnormal auricle shape data.

[0038] In this embodiment, a developmental trajectory function of a normal auricle is constructed, and a normal auricle morphology threshold is obtained based on the developmental trajectory function. The normal auricle morphology threshold includes the curvature threshold, elastic modulus threshold, and developmental state calibration threshold of a normal auricle. A progressive anomaly recognition mechanism is established based on the normal auricle morphology threshold. Abnormal auricle morphology data is obtained by combining the progressive anomaly recognition mechanism and auricle morphology preprocessing data.

[0039] A developmental trajectory function for normal auricles was constructed based on large-scale healthy population data. First, geometric, mechanical, and blood flow parameters of normal auricles at different age groups were collected. Then, age-related analysis was performed on each indicator to establish the developmental trajectory function. Based on the trajectory function, thresholds were determined by statistically analyzing normal auricle data: the curvature threshold was the upper confidence limit of geometric curvature; the elastic modulus threshold differentiated regions (helix E < 1.2 kPa, earlobe E < 0.6 kPa); and the developmental state calibration threshold was determined through residual analysis of geometric dimensions and age. If the developmental data deviated from the trajectory function evaluation value by more than two standard deviations, it was marked as developmental abnormality. Finally, the curvature threshold, elastic modulus threshold, and developmental state calibration threshold were integrated to form a multi-dimensional normal auricle morphology threshold system.

[0040] The above developmental trajectory function satisfies the following relationship:

[0041]

[0042] in, The mean of the features after age calibration. The baseline characteristic mean, This represents the peak development coefficient. Let be the growth function. For growth rate factor, This refers to the individual's actual age. The age of the growth inflection point This is a gender correction factor. To label an individual's gender.

[0043] In this embodiment, the confidence level of the auricle is obtained, and a multi-level confidence interval of the auricle is constructed by combining the normal auricle morphology threshold; the auricle morphology preprocessing data is coupled and analyzed by combining the developmental trajectory function and the multi-level confidence interval of the auricle to obtain the auricle developmental deviation-lesion correlation index; a progressive abnormality identification mechanism is constructed based on the multi-level confidence interval of the auricle and the auricle developmental deviation-lesion correlation index.

[0044] The confidence level of the auricle was obtained by combining anatomical structural features. First, based on auricular anatomy, three levels of confidence were defined: high-confidence regions (such as the helix and antihelix) were assigned level 1 due to their regular shape and high data acquisition stability; medium-confidence regions (such as the triangular fossa and concha) were assigned level 2 due to their complex structure but controllable data quality; and low-confidence regions (such as the earlobe edge) were assigned level 3 due to their susceptibility to motion artifacts. Then, based on normal auricular morphology thresholds, a threshold range was defined for each confidence level: a strict threshold (e.g., curvature <0.7mm⁻) was used for level 1 regions. 1 Elastic modulus <1.0kPa); Level 2 region relaxed by 10% (curvature <0.8mm⁻) 1 Elastic modulus <1.1kPa); Level 3 is further relaxed by 15% (curvature <0.9mm⁻) 1 (Elastic modulus < 1.2 kPa). Finally, multi-level confidence intervals for the auricle were obtained, clarifying the sensitivity of abnormality detection in different regions.

[0045] Furthermore, firstly, the expected mean value of features for each age group is calculated based on the developmental trajectory function; simultaneously, the actual values ​​of corresponding points in the preprocessed auricular morphology data are extracted. Subsequently, within each confidence interval, the deviation between the actual value and the expected value is calculated; finally, an auricular developmental deviation-lesion association index is constructed to quantify the association strength between developmental deviation and potential lesions.

[0046] The above-mentioned auricular developmental deviation-lesion correlation index satisfies the following relationship:

[0047]

[0048] in, The auricular development deviation-lesion association index, The degree of deviation from the developmental trajectory. This represents the initial anomaly intensity. For developmental deviations from the population standard deviation, The standard deviation of the abnormal intensity group. The nearest boundary distance of the anatomical functional area. The anatomical attenuation coefficient is... The second gradient of the elastic modulus, This is a reference value for the elastic modulus.

[0049] A progressive anomaly identification mechanism is implemented: Level 1 screening targets the high-confidence region; if the actual value exceeds the Level 1 threshold and the auricular development deviation-lesion correlation index is >1.5, it is marked as a mild abnormality. Level 2 screening targets the medium-confidence region; if the actual value exceeds the Level 2 threshold and the auricular development deviation-lesion correlation index is >2.0, it is marked as a moderate abnormality. Level 3 screening targets the low-confidence region; if the actual value exceeds the Level 3 threshold and the auricular development deviation-lesion correlation index is >2.5, it is marked as a severe abnormality. Simultaneously, dynamic feedback is introduced: if a region is marked in a lower-level screening, its confidence level is automatically increased, and the threshold and correlation index are reassessed to avoid missed detections.

[0050] In this embodiment, a progressive anomaly recognition mechanism is used to identify anomalies in the preprocessed auricle morphology data to obtain abnormal auricle morphology data. First, the preprocessed auricle morphology data is input into the progressive anomaly recognition mechanism, which automatically loads the multi-level confidence interval and developmental trajectory function of the auricle. Based on the auricle developmental deviation-lesion correlation index combined with the progressive anomaly recognition mechanism, a three-level screening is performed. Finally, abnormal auricle morphology data containing the abnormality type, severity, and spatial location are output, providing accurate input for subsequent feature fusion.

[0051] S3. Obtain abnormal auricle morphology fusion data based on the abnormal auricle morphology data to construct an abnormal auricle morphology joint feature tensor.

[0052] In this embodiment, abnormal auricle morphology sensitive features are extracted from the abnormal auricle morphology data at multiple scales; a feature fusion strategy is established, and abnormal auricle morphology fusion data is obtained by combining the abnormal auricle morphology sensitive features and the feature fusion strategy; and a joint feature tensor of abnormal auricle morphology is obtained by spatiotemporal evolution based on the abnormal auricle morphology fusion data.

[0053] A multi-resolution analysis framework was constructed based on preprocessed geometric point cloud data. A Gaussian pyramid model was used to perform three-level downsampling on the original data, forming auricular surface representations at different scales (original scale, 1 / 2 scale, and 1 / 4 scale). At each scale level, local feature analysis based on geometric measures was performed: at the original scale level, the distribution density of principal curvature extrema was calculated to identify micro-protrusions or depressions on the surface; at the 1 / 2 scale level, anisotropic diffusion equations were used for surface smoothing to suppress noise while preserving edge information, followed by calculation of the normal direction consistency index to capture morphological variations over a larger range; at the 1 / 4 scale level, wavelet transform was applied to extract frequency domain features, focusing on analyzing anomalous fluctuation patterns in low-frequency components. A cross-scale correlation mechanism was simultaneously established, mapping and aligning anomalous indicators at different levels by constructing a feature pyramid, and using a dynamic time warping algorithm to quantify the correlation of morphological variations between scales. For the screening of abnormal sensitive features, confidence evaluation is performed on features at each scale: the coefficient of variation of each feature point between adjacent scales is calculated, and when the coefficient of variation exceeds a preset threshold (such as 0.8), it is determined to be a high-sensitivity region. At the same time, the candidate features are classified and verified by support vector machine, and features with a classification accuracy of more than 90% are retained as sensitive features of abnormal auricle morphology, providing sensitive feature input with spatiotemporal correlation for subsequent feature fusion.

[0054] The above coefficients of variation satisfy the following relationship:

[0055]

[0056] in, The coefficient of variation of the feature points. For the number of scale levels, For the scale level index variable, For feature points In scale eigenvectors, For feature points The mean of multi-scale features For the Euclidean norm, It is the stability constant. Weights for anatomical locations.

[0057] In this embodiment, a sensitive feature interaction matrix is ​​constructed to quantify the correlation between features at different scales. The mutual information values ​​between each pair of features are calculated to form an N-order symmetric matrix. The element values ​​of this matrix range [0,1], representing the proportion of shared information between feature pairs. Simultaneously, the sensitive feature distribution entropy is calculated. The Shannon entropy formula is applied to each feature dimension, and the feature distribution curve is fitted using the kernel density estimation method to obtain the entropy vector of each feature as the sensitive feature distribution entropy. An entropy weighting strategy is designed to weight and fuse the entropy-weighted feature interaction matrix with the sensitive feature distribution entropy to obtain the entropy-weighted feature interaction matrix. This results in a nonlinear enhancement of the interaction strength between high-entropy features (rich in information) and low-entropy features (high determinism), while suppressing noisy interactions of low-information features.

[0058] The above entropy weighting strategy satisfies the following relationship:

[0059]

[0060] in, The elements of the entropy-weighted feature interaction matrix, The elements of the entropy-weighted feature interaction matrix, Features The entropy value, Features The entropy value.

[0061] Establish a feature fusion strategy based on the entropy-weighted feature interaction matrix: set a threshold. (e.g., 0.6), when Time-determined features and A strong correlation exists, so a weighted sum fusion strategy is adopted, with the fusion weights determined by entropy normalization; when A maximum value retention strategy is adopted to prevent the loss of information from weakly correlated features.

[0062] In an optional embodiment, abnormal auricle morphology sensitive features are weighted and fused according to the entropy-weighted feature interaction matrix to obtain abnormal auricle morphology fusion data. First, the feature space is reconstructed by matrix multiplication to enhance the interaction strength between high-entropy and low-entropy features. Then, the weighted feature sequence is slidably cut according to the time window, and the dynamic evolution law of morphological features is captured by the temporal transmission of hidden states. Finally, the effective aggregation of multi-scale abnormal features is achieved to obtain fusion data.

[0063] Furthermore, a spatiotemporal evolution constraint is introduced, and the fused data of abnormal auricle morphology is input into a long short-term memory network. The dynamic weight adjustment of cross-scale features is achieved through the hidden state propagation of time steps, and finally a third-order feature tensor is output as the joint feature tensor of abnormal auricle morphology.

[0064] S4. Based on the abnormal auricle morphology joint feature tensor, perform abnormal space mapping to obtain the abnormal auricle morphology spatial features.

[0065] Specifically, S4 includes the following steps:

[0066] S41. Construct the biomechanical potential field of the abnormal auricle based on the joint feature tensor of the abnormal auricle morphology.

[0067] In this embodiment, firstly, the multi-scale spatiotemporal feature tensor is mapped to the physical space domain. Using a finite element analysis framework, the surface of the auricle is discretized into triangular mesh elements, and each node is assigned attribute parameters based on the feature tensor. By applying boundary conditions simulating acoustic wave excitation, the displacement field distribution of the mesh nodes under steady-state vibration is calculated, generating principal stress direction maps and strain energy density cloud maps as the biomechanical potential field of the abnormal auricle. Energy accumulation regions are located in the potential field; these regions correspond to the mechanical imbalance points caused by auricle morphological abnormalities. For example, biomechanical simulations of patients with congenital microtia show that there are 3-5 high strain energy concentration areas in the sound wave propagation path of the abnormal auricle, and their distribution pattern is significantly different from that of the normal auricle.

[0068] S42. Obtain the abnormal space feature vector by performing abnormal space mapping on the abnormal auricle biomechanical potential field.

[0069] In this embodiment, wavelet thresholding denoising algorithm is used to preprocess the potential field data. A soft threshold function is selected according to the noise level to suppress Gaussian white noise while retaining micro-deformation features at the 0.1mm-0.5mm level. A field line tracking algorithm is introduced, using the potential energy gradient direction as a guide, and extracting the energy flow path in the triangular mesh cell through gradient descent method. The discontinuity of continuous field lines is defined as anomaly spatial feature points. Through non-negative matrix decomposition, the topological features are modally fused with geometric parameters such as the curvature and perimeter of the auricle surface to generate anomaly spatial feature vectors that combine spatial structural information and biomechanical properties.

[0070] S43. Obtain the spatial features of the abnormal auricle shape based on the abnormal spatial feature vector.

[0071] In this embodiment, nonnegative matrix factorization is used to perform modal fusion of topological and geometric features. The abnormal spatial feature vector, along with geometric parameters such as the curvature and perimeter of the auricle surface, is input into a multilayer perceptron. The feature weights are optimized through cross-validation to establish a mapping relationship between abnormal spatial features and biological attributes, ultimately obtaining the abnormal auricle morphological spatial features, including a 128-dimensional feature vector. The first 32 dimensions represent topological anomalies, the middle 64 dimensions reflect geometric deformations, and the last 32 dimensions correspond to biomechanical inaccuracies.

[0072] S5. Construct an abnormal auricle morphology data model based on the spatial features of the abnormal auricle morphology.

[0073] In this embodiment, a dynamic evolution tensor network is constructed as the model architecture; a model training mechanism is established based on biomechanical constraints; and an abnormal auricle morphology data model is constructed by combining the spatial features of abnormal auricle morphology, the model architecture, and the model training mechanism.

[0074] Specifically, a dynamic evolutionary tensor network is used as the model architecture, with the spatial features of abnormal auricle morphology as input. The first layer of the network constructs a spatiotemporal graph structure, discretizing the auricle surface into a node graph structure, with each node associated with five nearest neighbor nodes to form an edge. The graph weights are determined by calculating the Euclidean distance between nodes using a Gaussian kernel function. The second layer introduces a dynamic evolution mechanism, using a temporal convolutional network to capture the temporal dependence of features, setting the receptive field to eight time steps, and using dilated convolution to achieve multi-scale temporal pattern extraction. The third layer integrates biomechanical potential field information, using the strain energy density field obtained from finite element analysis as prior knowledge, and dynamically adjusting the node weights through a graph attention network to enhance the feature representation of abnormal regions. Finally, the output layer adopts a fully connected structure to generate anomaly representation vectors.

[0075] Furthermore, the training process introduces a dual-constraint loss function: a data-driven term and a physics-driven term. The data-driven term employs cross-entropy loss, calculating the classification error based on clinically labeled deformity levels. The physics-driven term constructs a biomechanical regularization term, comprising two aspects: first, by comparing the model-predicted auricular deformation field with the ground truth field from finite element simulation, the mean square error is calculated to constrain the physical rationality of the deformation pattern; second, a strain energy density threshold constraint is introduced, triggering a penalty term when the local strain energy output by the model exceeds the 95% confidence interval of a normal auricle. Training employs a phased optimization strategy, with initial iterations only updating the data-driven term parameters, and subsequent iterations jointly optimizing the dual-constraint term. To enhance model generalization, a transfer learning mechanism is introduced, initializing parameters on a pre-training set containing 1000 normal auricular data points, and then fine-tuning on 200 abnormal data points.

[0076] The spatial features of abnormal auricle morphology are encoded; processed through three layers in a dynamic evolution tensor network; mapped to the biomechanical parameter space through a fully connected layer; the total loss is calculated; gradient updates are performed using an optimizer with dynamic decay of the learning rate; after training with 200 cases of abnormal auricle data, morphological features are preserved, and a quantitative evaluation report is obtained based on the abnormal auricle morphology data model, including quantitative indicators such as the degree of anatomical variation, mechanical stability index, and surgical correction priority.

[0077] Please see Figure 2In one optional embodiment, the present invention provides a system for constructing an abnormal auricle morphology data model. The system includes an input device, an output device, a processor, and a memory, all interconnected. The memory stores a computer program, which includes program instructions. The processor is configured to invoke the program instructions to execute specific steps as described in the relevant embodiments of the abnormal auricle morphology data model construction method provided by the present invention. The abnormal auricle morphology data model construction system provided by the present invention is structurally complete, objective, and stable, enhancing the overall applicability and practical application capabilities of the present invention.

[0078] In summary, the present invention provides a method and system for constructing an abnormal auricle morphology data model. This method utilizes multimodal data acquisition and adaptive noise reduction preprocessing; accurately extracts abnormal features based on developmental trajectory functions and progressive recognition mechanisms; combines multi-scale sensitive feature extraction, entropy-weighted fusion strategies, and spatiotemporal evolution to generate joint feature tensors; obtains spatial features through abnormal spatial mapping; and then constructs an abnormal auricle morphology data model based on dynamic tensor networks and biomechanical constraints. This achieves multi-dimensional quantitative representation of complex auricle deformities, improves the sensitivity of abnormality detection and spatial positioning accuracy, and provides support for personalized treatment plans. The method is easy to understand, computationally simple, requires minimal workload, and is convenient for practical application, providing a theoretical foundation and technical support for the further development of intelligent medical technology.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for constructing a data model of abnormal auricle morphology, characterized in that, Includes the following steps: Acquire auricular morphology data and preprocess the auricular morphology data to obtain auricular morphology preprocessed data; A normal auricle shape threshold is constructed, and abnormal auricle shape data is obtained by performing abnormal feature identification on the auricle shape preprocessing data based on the normal auricle shape threshold. Abnormal auricle morphology fusion data is obtained based on the abnormal auricle morphology data to construct the abnormal auricle morphology joint feature tensor. The spatial features of the abnormal auricle morphology are obtained by performing anomaly space mapping based on the joint feature tensor of the abnormal auricle morphology. A data model of abnormal auricle morphology is constructed based on the spatial features of the abnormal auricle morphology. The process of acquiring auricle morphology data and preprocessing the auricle morphology data to obtain auricle morphology preprocessed data includes: Geometric point cloud data, elastic modulus matrix and blood flow heat map of the auricle are acquired to construct multimodal data, and the multimodal data is used as the auricle morphology data; Biomechanical constraints are constructed, and the auricular morphology data is registered based on the biomechanical constraints to obtain auricular morphology registration data. An adaptive manifold noise reduction mechanism is established, and auricular morphology noise reduction data is obtained by combining the auricular morphology registration data. The auricular morphology noise reduction data is used as the auricular morphology preprocessing data. The step of obtaining abnormal auricle morphology fusion data based on the abnormal auricle morphology data to construct an abnormal auricle morphology joint feature tensor includes: Multi-scale abnormality sensitive feature extraction is performed on the abnormal auricle morphology data to obtain abnormal auricle morphology sensitive features; A feature fusion strategy is established, and the abnormal auricle morphology fusion data is obtained by combining the abnormal auricle morphology sensitive features and the feature fusion strategy. The joint feature tensor of the abnormal auricle morphology is obtained by performing spatiotemporal evolution based on the fusion data of the abnormal auricle morphology. The step of obtaining spatial features of abnormal auricle morphology by performing abnormal space mapping based on the joint feature tensor of the abnormal auricle morphology includes: A biomechanical potential field of the abnormal auricle is constructed based on the joint feature tensor of the abnormal auricle morphology. Anomaly space feature vectors are obtained by performing anomaly space mapping on the biomechanical potential field of the abnormal auricle. The spatial features of the abnormal auricle shape are obtained based on the abnormal spatial feature vector.

2. The method for constructing an abnormal auricle morphology data model according to claim 1, characterized in that, The establishment of the adaptive manifold noise reduction mechanism includes: The blood flow oscillation index of the auricle is obtained based on the blood flow heat map, and the dynamic noise reduction intensity is obtained based on the blood flow oscillation index. Multimodal joint curvature analysis was performed on the auricle morphology data to construct a multimodal curvature manifold of the auricle; The biomechanical edge of the auricle is obtained based on the biomechanical constraints and the multimodal curvature manifold. The adaptive manifold noise reduction mechanism is constructed by combining the dynamic noise reduction intensity, the multimodal curvature manifold, and the biomechanical edge.

3. The method for constructing an abnormal auricle morphology data model according to claim 1, characterized in that, The process of constructing a normal auricle morphology threshold and then performing abnormal feature identification on the preprocessed auricle morphology data based on the normal auricle morphology threshold to obtain abnormal auricle morphology data includes: A developmental trajectory function for a normal auricle is constructed, and a normal auricle morphology threshold is obtained based on the developmental trajectory function. The normal auricle morphology threshold includes the curvature threshold, elastic modulus threshold, and developmental state calibration threshold of the normal auricle. A progressive anomaly recognition mechanism is established based on the normal auricle morphology threshold. The abnormal auricle morphology data is obtained by combining the progressive anomaly recognition mechanism and the auricle morphology preprocessing data.

4. The method for constructing an abnormal auricle morphology data model according to claim 3, characterized in that, The progressive anomaly recognition mechanism based on the normal auricle morphology threshold includes: Obtain the confidence level of the auricle and construct a multi-level confidence interval for the auricle by combining it with the normal auricle morphology threshold. By combining the developmental trajectory function and the multi-level confidence interval of the auricle, the auricle morphology preprocessing data is coupled and analyzed to obtain the auricle developmental deviation-lesion correlation index. The progressive abnormality identification mechanism is constructed based on the multi-level confidence interval of the auricle and the auricle developmental deviation-lesion correlation index.

5. The method for constructing an abnormal auricle morphology data model according to claim 1, characterized in that, The established feature fusion strategy includes: Based on the aforementioned abnormal sensitive features, obtain the sensitive feature interaction matrix and the sensitive feature distribution entropy; The entropy-weighted feature interaction matrix is ​​obtained by combining the sensitive feature interaction matrix and the sensitive feature distribution entropy; The feature fusion strategy is established based on the entropy-weighted feature interaction matrix.

6. The method for constructing an abnormal auricle morphology data model according to claim 1, characterized in that, The construction of the abnormal auricle morphology data model based on the spatial features of the abnormal auricle morphology includes: A dynamically evolving tensor network is constructed as the model architecture; A model training mechanism is established based on the aforementioned biomechanical constraints; The abnormal auricle morphology data model is constructed by combining the spatial features of the abnormal auricle morphology, the model architecture, and the model training mechanism.

7. A system for constructing a data model of abnormal auricle morphology, characterized in that, The system includes an input device, an output device, a processor, and a memory, which are interconnected. The memory stores a computer program, which includes program instructions. The processor is configured to invoke the program instructions to execute the abnormal auricle morphology data model construction method as described in any one of claims 1-6.

Images