Breast health data anomaly detection method and system

Abstract

The invention relates to the technical field of health monitoring, in particular to a mammary gland health data anomaly detection method and system, and the method comprises the following steps: extracting an anomaly attribute set based on temperature and conductive characteristics, analyzing gradient and vector directions, screening cross-regional anomalies, combining a conductivity matrix and thermal field characteristics, evaluating a deviation degree, and calibrating an anomaly level. And matching the response list with the priority sequence, screening the partition numbers needing to be adjusted, and outputting a mammary gland health repair adjustment partition number list. According to the method, synchronous analysis and trend capture of multi-source data are realized by establishing a temperature track and conductive response partition joint analysis system, steady-state confirmation of an abnormal linkage region is realized, the abnormal grade division capability is enhanced through deviation trend evaluation, and dynamic adjustment of a repair range is promoted through coverage range comparison; therefore, the health state judgment reliability and the anomaly positioning accuracy are improved, the risk assessment sensitivity and consistency are optimized, misjudgment and omission are reduced, and it is ensured that a repair strategy has pertinence and coordination.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of health monitoring, in particular to a breast health data anomaly detection method and system. BACKGROUND

[0002] The technical field of health monitoring involves the continuous collection, analysis and evaluation of various physiological parameters and health data of the human body, aiming to realize real-time monitoring and trend judgment of individual health status through multi-dimensional data sensing. The core matters of this field include sensor data collection technology, physiological signal feature extraction and analysis, health index modeling and abnormal data determination mechanism. Health monitoring technology relies on the collaborative operation of wearable devices, mobile terminals and cloud data platforms to obtain multi-source physiological information such as heart rate, blood pressure, body temperature, blood oxygen and gait, and to realize individual health management, early warning and intervention support through data fusion and statistical analysis means. This field has important research significance in terms of data accuracy, detection sensitivity and optimization of abnormal recognition algorithms.

[0003] Among them, the traditional breast health data anomaly detection method refers to a means of collecting related physiological data on the surface or shallow layer of breast tissue, analyzing its change rule to identify potential health risks. This method is based on the detection of physical quantities such as temperature distribution, tissue conductivity characteristics or acoustic parameters, and uses thermal imaging instruments, ultrasonic probes or electrical impedance measurement devices to obtain data. Subsequently, by using experience threshold comparison, statistical feature extraction or pattern matching, the spatial distribution characteristics, time change trend and difference index of the collected data are analyzed to determine whether the data is abnormal. Traditional methods rely on fixed measurement points and manually set abnormal standards, and reflect the state difference of breast tissue through single parameter changes, lacking comprehensive analysis and dynamic evaluation ability of the correlation between multi-dimensional data.

[0004] Traditional breast health detection relies on single parameter collection and fixed measurement point mode, which cannot simultaneously depict the dynamic correlation of temperature distribution and conductivity characteristics, lacks linkage analysis means between spatial structure information and time sequence changes, leading to the inability to identify cross-regional abnormalities in a timely manner. The fixed threshold judgment method cannot adapt to the data drift caused by individual differences, resulting in significant fluctuations in abnormal determination results. The single feature comparison method cannot present the complex abnormal trend caused by deep tissue changes, the correlation between data cannot be effectively utilized, making the abnormal region boundary fuzzy and the trend inference insufficient. In the face of the demand for multi-source information fusion, there is a risk of recognition lag and misjudgment, which affects the stability and accuracy of breast health evaluation. SUMMARY

[0005] The purpose of the present application is to solve the shortcomings in the prior art and to provide a breast health data anomaly detection method and system.

[0006] In order to achieve the above object, the present application adopts the following technical scheme, a breast health data anomaly detection method, comprising the following steps: S1: based on the breast tissue surface temperature distribution and the conductivity characteristic, the detection partition is divided, the temperature change trajectory and the conductivity response curve of the partition are extracted, the synchronous abnormal area of the temperature fluctuation interval and the conductivity response mutation point on the time axis is identified, the corresponding partition number and the spatial coordinates are extracted, and the breast partition abnormal attribute set is generated; S2: according to the breast partition abnormal attribute set, the temperature gradient direction and the conductivity response vector of the partition are identified, the direction consistency of the two at the partition boundary is analyzed, the partition with connectivity and temperature gradient anomaly is screened out, and the cross-zone abnormal connectivity data group is formed; S3: based on the cross-zone abnormal connectivity data group, the tissue conductivity distribution matrix and the local thermal field characteristics in the partition are extracted, the conductivity uniformity and the thermal field consistency are analyzed, the matching abnormal area is screened out, and the abnormal defect linkage area group is obtained; S4: according to the abnormal defect linkage area group, the conductivity distribution trend of the corresponding partition is analyzed, the deviation degree from the original health state curve is evaluated, the abnormal grade of the partition is calibrated according to the deviation amplitude, and the breast health partition detection risk grade evaluation list is output.

[0007] As a further scheme of the present application, the breast partition abnormal attribute set includes temperature mutation partition number, conductivity response abnormal point, partition coordinate mark, time sequence mutation mark, the cross-zone abnormal connectivity data group includes temperature gradient abnormal partition mark, partition boundary connectivity unit, partition junction consistency block, the abnormal defect linkage area group includes conductivity abnormal continuous partition, thermal field irregular area, conductivity performance mutation coincidence area, linkage abnormal partition number, and the breast health partition detection risk grade evaluation list includes risk grade label, partition response deviation value, local conductivity abnormality index and health state deviation grade.

[0008] As a further scheme of the present application, the acquisition step of the breast partition abnormal attribute set is specifically: S111: based on the breast tissue surface temperature distribution and the conductivity characteristic, the detection partition is divided, the temperature change trajectory and the conductivity response curve of the partition are extracted, the difference value calculation is carried out on the two kinds of data in the same partition, and the temperature and conductivity response difference trend value is obtained; S112: according to the temperature and conductivity response difference trend value, the fluctuation interval in the temperature change curve and the mutation point of the conductivity response trajectory are identified, the two kinds of values are time superimposed, the time interval of the fluctuation exceeding the preset reference value and the mutation point being in the set range is extracted, and the high-frequency abnormal interval time period set is generated; S113: For the high-frequency abnormal interval time period set, match the corresponding partition number and spatial coordinate information, extract the partition position where the signal occurs, and generate a breast partition abnormal attribute set.

[0009] As a further scheme of the present application, the acquisition step of the cross-zone abnormal connectivity data group is specifically: S211: Identify the temperature gradient direction and the conductive response vector of the partition in the breast partition abnormal attribute set, extract the projection trajectory of the two at the partition boundary, identify the distribution number and aggregation degree of the intersection point in the partition, and obtain the partition boundary consistency atlas; S212: According to the partition boundary consistency atlas, filter the boundary area with an aggregation degree higher than the average level, compare the spatial boundary of the whole breast structure atlas, identify the continuous and belonging to the same partition in the boundary area, and obtain the temperature gradient abnormal zoning in the breast partition. S213: Based on the temperature gradient abnormal zoning in the breast partition, integrate the partition boundary consistency, temperature gradient dispersion, conductivity uniformity and conductive response delay, and perform partition matching according to the response block in the partition to form a cross-zone abnormal connectivity data group.

[0010] As a further scheme of the present application, the acquisition step of the abnormal defect linkage area group is specifically: S311: Based on the cross-zone abnormal connectivity data group, extract the conductivity distribution matrix and local thermal field characteristics of the numbered partition in the partition, align the time stamp of the data in the partition, identify the conductivity fluctuation value and thermal field consistency offset, and obtain the breast local defect response feature set; S312: According to the breast local defect response feature set, jointly analyze the conductivity uniformity and thermal field consistency in the partition, and use the formula: ; Calculate the conductivity thermal field coupling characteristic value, filter the partition unit of the conductivity thermal field coupling degree in the partition, and establish the conductivity thermal field cooperative response spatial distribution map; Among them, represents the conductivity thermal field coupling characteristic value, represents the conductivity value of the i-th unit, represents the temperature value of the i-th unit, represents the spatial coordinate change of the i-th unit in the x-axis direction, represents the spatial coordinate change of the i-th unit in the y-axis direction, represents the thermal diffusion coefficient of unit volume of material, represents the thermal coefficient of unit volume of conductivity; S313: calling the conductivity thermal field synergistic response spatial distribution map, clustering the partitions exceeding the synergistic identification benchmark in the coupling characteristic value partition, marking the partition code and coordinates corresponding to the continuous abnormal region, and obtaining the abnormal defect linkage region group.

[0011] As a further scheme of the present application, the obtaining of the breast health partition detection risk level evaluation list specifically comprises: S411: extracting the partition conductivity performance distribution curve under the specified number according to the abnormal defect linkage region group, performing time unification processing, identifying the unit time conductivity performance change amount, and obtaining the partition conductivity performance abnormal change rate set; S412: identifying the conductivity performance distribution curve of the original health state stage in the original data or the benchmark data according to the partition conductivity performance abnormal change rate set, comparing the current conductivity performance change sequence with the reference curve, identifying the partition conductivity performance deviation level, extracting and marking the partitions with the deviation level exceeding the upper warning limit, and obtaining the deviation sudden increase partition set; S413: binding the deviation level value of each partition with the position number in the breast structure spatial map according to the deviation sudden increase partition set, and using the formula: ; calculating the health risk index of the partition, sorting according to the risk level, and outputting the breast health partition detection risk level evaluation list; wherein, represents the health risk index of the partition, represents the deviation level value, represents the mean value of all partition deviation level values, represents the standard deviation of all partition deviation level values, represents the spatial distance of the th partition position in the breast structure spatial map, represents the risk weight coefficient of the th partition position, represents the total number of involved partition positions.

[0012] As a further scheme of the present application, the method further comprises the S5 step: S5: calling the breast health partition detection risk level evaluation list, identifying the corresponding number of the partition in the breast function map, calling the repair response unit list, comparing the response level with the breast protection priority sequence, screening the partition numbers requiring adjustment of the response coverage range, and outputting the breast health repair adjustment partition number list; The breast protection priority sequence refers to the sequence according to the breast function criticality and protection demand. The breast health repair adjustment subregion number list comprises an adjustment target subregion number, a response level adjustment parameter, a protection priority comparison item, and a linkage response trigger type.

[0013] As a further scheme of the present application, the breast health repair adjustment subregion number list is obtained by: S511: The breast health subregion detection risk level evaluation list is called, the subregion number in the breast function map is extracted, the subregion risk level value is mapped with the region coordinate boundary, the subregion information corresponding to the breast protection level is identified, and a breast subregion risk distribution map is generated; S512: Based on the breast subregion risk distribution map, the repair response unit number and the response level are extracted, the subregion risk level is matched with the repair response level, the unit number with insufficient response coverage is identified, and a breast subregion response risk disconnection list is obtained; S513: According to the breast subregion response risk disconnection list, according to the level number in the breast protection priority sequence, the key subregion number that needs to improve the response coverage range is extracted, the adjustment control parameter linked with the original repair unit is output in sequence, and the breast health repair adjustment subregion number list is output.

[0014] The breast health data anomaly detection system is used to execute the above-mentioned breast health data anomaly detection method, and the system comprises: The temperature monitoring module divides the detection subregion based on the breast tissue surface temperature distribution and the conductivity characteristic, compares the temperature fluctuation interval and the conductivity response mutation point in the same time period, screens the synchronous abnormal area of both, extracts the subregion number and the spatial coordinate, summarizes the abnormal time period and the subregion number, and generates a breast subregion anomaly attribute set; The subregion positioning module identifies the consistency of the temperature gradient direction and the conductivity response vector based on the breast subregion anomaly attribute set, calibrates the subregion boundary number, matches the breast overall structure map, extracts the subregion number range of the temperature gradient abnormal area, and establishes a cross-region abnormal connectivity data group; The defect linkage module calls the conductivity distribution matrix and the continuous data sequence of the local thermal field characteristic in the region based on the cross-region abnormal connectivity data group, judges the conductivity abnormal boundary connectivity and the thermal field consistency, marks the subregion number meeting the linkage threshold of both, and outputs an abnormal defect linkage area group; The conductivity early warning module analyzes the conductivity performance distribution trend and the original health state curve deviation degree of the corresponding subregion based on the subregion number of the abnormal defect linkage area group, extracts the subregion number of the deviation trend, completes the level identification according to the risk grading standard, and generates a breast health subregion detection risk level evaluation list; The repair optimization module detects the risk level evaluation list based on the breast health partition, finds the corresponding position number of the risk level partition in the breast function map, calls the current repair response unit configuration list, compares the breast protection priority with the current response level matching condition, filters the partition that needs to be updated, and outputs the breast health repair adjustment partition number list.

[0015] Compared with the prior art, the advantages and positive effects of the present application are that: In the present application, the synchronous analysis and trend capturing of multi-source data are realized by establishing a partition joint analysis system of temperature trajectory and conductive response, the identification ability of abnormal areas is strengthened from the dual angles of time fluctuation and spatial correlation, the fineness of cross-zone abnormal association judgment is improved through connectivity screening mechanism and gradient direction consistency analysis, the multi-dimensional stable confirmation of abnormal linkage areas is realized under the matching filtering of conductivity matrix and local thermal field characteristics, the quantitative ability and risk presentation ability of abnormal level division are enhanced through deviation trend evaluation method, the dynamic adjustment of key area repair range is promoted through the comparison and screening of strategy coverage range, so as to realize the comprehensiveness and continuity of abnormal identification, improve the reliability and stability of partition-level health state judgment, strengthen the accuracy and trend expression ability of abnormal positioning, optimize the sensitivity and consistency of risk assessment, reduce the misjudgment and omission, ensure that the health repair strategy has pertinence and coordination, and promote the effective implementation of breast health management. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is a workflow schematic diagram of the present application; Figure 2 It is a flowchart of acquisition of abnormal attribute set of breast partition in the present application; Figure 3 It is a flowchart of acquisition of cross-zone abnormal connectivity data group in the present application; Figure 4 It is a flowchart of acquisition of abnormal defect linkage area group in the present application; Figure 5 It is a flowchart of acquisition of breast health partition detection risk level evaluation list in the present application; Figure 6 It is a flowchart of acquisition of breast health repair adjustment partition number list in the present application. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application is further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application, and are not used to limit the present application.

[0018] In the description of the present application, it should be understood that the terms "length", "width", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, in the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0019] Embodiment 1: please refer to Figure 1 The present application provides a technical solution, a breast health data anomaly detection method, comprising the following steps: S1: based on the breast tissue surface temperature distribution and the conductive characteristic, the detection partition is divided, the temperature change trajectory and the conductive response curve of the partition are extracted, the synchronous abnormal area of the temperature fluctuation interval and the conductive response mutation point on the time axis is identified, the corresponding partition number and spatial coordinates are extracted, and the breast partition abnormal attribute set is generated; S2: according to the breast partition abnormal attribute set, the temperature gradient direction and the conductive response vector of the partition are identified, the direction consistency of the two at the partition boundary is analyzed, the partition with connectivity and temperature gradient anomaly is screened out combined with the connectivity of the partition boundary, and the cross-zone abnormal connectivity data group is formed; S3: based on the cross-zone abnormal connectivity data group, the tissue conductivity distribution matrix and the local thermal field characteristics in the partition are extracted, the conductivity uniformity and the thermal field consistency are analyzed, the matching abnormal area is screened out, and the abnormal defect linkage area group is obtained; S4: according to the abnormal defect linkage area group, the conductive performance distribution trend corresponding to the partition is analyzed, the deviation degree from the original health state curve is evaluated, the abnormal grade of the partition is calibrated according to the deviation amplitude, and the breast health partition detection risk grade evaluation list is output; S5: calling the breast health partition detection risk grade evaluation list, identifying the corresponding number of the partition in the breast function graph, calling the repair response unit list, comparing the response grade and the breast protection priority sequence, screening the partition number which needs to adjust the response coverage range, and outputting the breast health repair adjustment partition number list; The breast protection priority sequence refers to the sequence according to the breast function criticality and protection demand.

[0020] The breast partition abnormality attribute set includes temperature mutation partition number, conductive response abnormal point, partition coordinate marker, time sequence mutation mark, cross-zone abnormal connectivity data set includes temperature gradient abnormal partition mark, partition boundary connectivity unit, partition junction consistency block, abnormal defect linkage area group includes conductivity abnormal continuous partition, thermal field irregular area, conductivity performance mutation overlap area, linkage abnormal partition number, breast health partition detection risk level evaluation list includes risk level label, partition response deviation value, local conductivity abnormality index, health state deviation level, breast health repair adjustment partition number list includes adjustment target partition number, response level adjustment parameter, protection priority comparison item, linkage response trigger type.

[0021] Please refer to Figure 2 The acquisition step of the breast partition abnormality attribute set is specifically: S111: Based on the breast tissue surface temperature distribution and the conductivity characteristic, the detection partition is divided, the temperature change trajectory and the conductivity response curve of the partition are extracted, and the difference value calculation is performed on the two types of data in the same partition to obtain the temperature and conductivity response difference trend value; Based on the breast tissue surface temperature distribution and the conductivity characteristic, the detection partition is divided, the breast surface temperature distribution data is acquired by high-precision infrared thermal imager scanning, the precision is 0.1 Celsius degree, the matrix specification is 256x256, the breast conductivity characteristic data is acquired by bioimpedance analyzer, the precision is 0.01 S / m, and the matrix specification is 256x256. The temperature distribution of a certain area is 36.5 Celsius degree to 37.8 Celsius degree, the conductivity distribution is 0.5 S / m to 0.7 S / m, the preset physiological partition standard is divided into 8 detection partitions, for example, the left upper outer partition contains 64x64 pixel points, the temperature and conductivity data in the partition are averaged, the temperature change trajectory and the conductivity response curve of the partition are extracted, the average temperature value and the average conductivity value are recorded in the continuous monitoring period (every 10 minutes), and the time sequence is formed. For example, the temperature sequence data and the conductivity sequence data of partition A are calculated, the difference value calculation is performed on the two types of data in the same partition, the difference between the temperature value at each monitoring time point and the temperature value at the previous time point is calculated, and the difference between the conductivity value and the conductivity value at the previous time point is calculated. For example, the temperature change amount of partition A is 0.1 Celsius degree, and the conductivity change amount is 0.02 S / m, the temperature and conductivity response difference trend value is obtained, and the weighted or Euclidean distance of the temperature change amount and the conductivity change amount is calculated.

[0022] S112: According to the temperature and conductivity response difference trend value, the fluctuation interval in the temperature change curve and the mutation point of the conductivity response trajectory are identified, the time overlap of the two types of values is performed, the time interval in which the fluctuation exceeds the preset reference value and the mutation point is located in the set range is extracted, and the high-frequency abnormal interval time period set is generated; According to the temperature and conductivity response difference trend value, the fluctuation interval is identified by calculating the local standard deviation of the temperature change sequence, and the local standard deviation (3 time points) exceeds the temperature fluctuation threshold of 0.05 degrees Celsius, and the mutation point is identified by calculating the second derivative of the conductivity change sequence, and the second derivative or CUSUM value exceeds the conductivity mutation threshold of 0.03 S / m. For example, a certain segment of the difference trend value sequence in partition A between the 6th and 7th time points, the local standard deviation of the temperature change increases from 0.03 degrees Celsius to 0.07 degrees Celsius, and the second derivative of the conductivity change increases from 0.01 S / m to 0.04 S / m. Temperature fluctuations and conductivity mutations are identified. The two types of values are time overlapped, and the temperature fluctuation interval and the conductivity mutation point are aligned on the time axis. For example, the temperature fluctuation interval lasts from time point 5 to time point 8, the conductivity mutation point appears at time points 6 and 7, and the overlapping time period is time points 6 and 7. Extract the time interval where the fluctuation exceeds the preset reference value and the mutation point is within the set range. The preset reference value is defined as the average difference trend value of the partition health state plus two standard deviations. For example, the average difference trend value of partition A in the healthy state is 0.05, the standard deviation is 0.01, the reference value is 0.07, the set mutation amplitude threshold is 0.03 S / m, the difference trend value at time point 6 is 0.15, which exceeds the reference value 0.07, and there is a conductivity mutation. Time point 6 is extracted, a high-frequency abnormal interval time period set is generated, and time intervals that meet the conditions are collected.

[0023] S113: For the high-frequency abnormal interval time period set, match the corresponding partition number and spatial coordinate information, extract the partition position where the signal occurs, and generate a breast partition abnormal attribute set; For the high-frequency abnormal interval time period set, take out the abnormal interval, for example, "time point 6-time point 8" is the abnormal interval, match the partition number and spatial coordinate information, for each abnormal time interval, find the abnormal partition number (P01) and its spatial coordinate range (pixel points (10, 10) to (74, 74)), match the partition division and spatial mapping based on the pre-defined partition division. For example, P01 is identified as a high-frequency abnormality during "time point 6-time point 8", the partition number P01, the spatial coordinate range ((10, 10), (74, 74)), the partition position where the signal occurs is extracted, and the matched partition number and spatial coordinate information are extracted and stored. For example, P01 and the coordinate range ((10, 10), (74, 74)) are stored, a breast partition abnormal attribute set is generated, the set is composed of records, and each record contains a partition number, an abnormal time period, and partition spatial coordinate information.

[0024] Please refer to Figure 3 The acquisition steps of the cross-zone abnormal connectivity data group are as follows: S211: Identify the temperature gradient direction and the conductive response vector in the partition of the breast partition abnormal attribute set, extract the projection trajectory of the two at the partition boundary, identify the distribution number and aggregation degree of the intersection point in the partition, and obtain the partition boundary consistency atlas; Identify the breast partition abnormal attribute set, obtain the number and spatial coordinate information of each abnormal partition, for example, the spatial coordinate range of partition P01 is ((10, 10), (74, 74)), retrieve the original temperature distribution data and conductive characteristic data of the partition and its adjacent partitions, identify the temperature gradient direction, and calculate the temperature change rate vector within the partition and between adjacent partitions. For example, at the boundary between P01 and P02, point (x1, y1) has a temperature of 37.0 degrees Celsius, point (x1+1, y1) has a temperature of 37.1 degrees Celsius, the temperature gradient is 0.1 degrees Celsius / mm, the conductive response vector is identified, the conductivity change rate vector within the partition and between adjacent partitions is calculated, for example, point (x1, y1) has a conductivity of 0.60 S / m, point (x1+1, y1) has a conductivity of 0.61 S / m, the conductive response is 0.01 S / m / mm, and the projection trajectory of the two at the partition boundary is extracted. At each intersection point, the temperature gradient vector and the conductive response vector are projected onto the normal and tangent directions of the intersection line to obtain a set of components that constitute the projection trajectory. For example, intersection point A, temperature gradient vector GT, conductive response vector GC, projection trajectory includes the components of GT in the normal and tangent directions and the components of GC in the normal and tangent directions, identify the distribution number and aggregation degree of the intersection point in the partition, count the number of intersection points on the intersection line whose projection trajectory similarity reaches the threshold (angle less than 10 degrees and modulus ratio 0.8 to 1.2), and evaluate the spatial aggregation degree. For example, 80 out of 100 sampling points on the P01 and P02 intersection line have a similarity that meets the threshold and are continuously distributed, the aggregation degree is high, the partition boundary consistency atlas is obtained, the atlas is represented in matrix form, and the elements represent the consistency score of a pair of adjacent partition boundaries. The score is calculated based on the distribution number and aggregation degree of the intersection points.

[0025] S212: According to the partition boundary consistency atlas, filter the boundary regions with aggregation degree higher than the average level, compare the spatial boundary of the whole breast structure atlas, identify the continuous and belonging to the same partition of the intersection concentrated area, and obtain the temperature gradient abnormal area in the breast partition; According to the consistency map of the partition boundary, the boundary consistency scores of each adjacent partition pair are obtained, for example, P01 and P02 score 0.85, P02 and P03 score 0.70, the boundary regions with aggregation degree higher than the average level are screened, and the average value of the boundary consistency scores of all adjacent partitions is calculated. For example, the average value is 0.75, the regions with scores higher than 0.75 are screened, such as the boundary region of P01 and P02 (0.85 is higher than 0.75), the spatial boundary of the overall breast structure is compared, the high-aggregation-degree boundary region screened is superimposed on the breast anatomical structure map, the continuity is checked and aligned with the structure boundary. For example, the high-aggregation-degree boundary region of P01 and P02 is consistent with the trend of a main duct of the breast, matches the breast structure boundary, the continuous area in the boundary concentration area which has continuity and belongs to the same partition is identified, the high-aggregation-degree boundary region which matches the structure boundary is tracked, and the continuous area which forms a whole in function is determined, all the boundary points in the area belong to the same functional partition, for example, the three partitions P01, P02 and P03 have boundary consistency scores higher than the average level, forming a continuous area which is determined as a certain lobe of the breast, and then the boundary concentration area is identified, and the temperature gradient abnormal area in the breast partition is obtained. The temperature gradient and the conductive response in the area show consistency and match the physiological structure, and are marked as the temperature gradient abnormal area in the breast partition.

[0026] S213: Based on the temperature gradient abnormal area in the breast partition, the integrated analysis of the partition boundary consistency, temperature gradient dispersion, conductivity balance and conductive response delay is performed, the response blocks in the partition are matched according to the partition, and the cross-zone abnormal connectivity data set is formed; Based on the temperature gradient abnormal area in the breast partition, the temperature gradient abnormal area information is obtained, for example, the "lobe A abnormal area" covers the P01, P02 and P03 partitions, and the integrated analysis of the partition boundary consistency, temperature gradient dispersion, conductivity balance and conductive response delay is performed; Partition boundary consistency: the boundary consistency scores between two partitions are extracted, the average value is calculated, for example, P01-P02 is 0.85, P02-P03 is 0.70, and P01-P03 is 0.65, and the average value is 0.73; Temperature gradient dispersion: the standard deviations of the modulus and direction of all temperature gradient vectors in the partition are calculated, for example, the temperature gradient modulus standard deviation of P01 partition is 0.02 ℃ / mm, and the direction standard deviation is 5 degrees; Conductivity balance: the coefficient of variation of all conductivity values in the partition is calculated, for example, the conductivity variation coefficient of P01 partition is 0.05; Conductive response delay: the time difference between temperature change and conductive response is recorded, the average value and standard deviation are calculated, for example, the average conductive response delay of P01 partition is 5 seconds; The four indicators are normalized (0 to 1 range), weighted summation, and the integrated abnormal score of the abnormal area division is obtained. The weight: boundary consistency 0.3, temperature gradient dispersion 0.25, conductivity uniformity 0.25, conductivity response delay 0.2, the integrated abnormal score of the "adenoid leaf A abnormal area division" is 0.8845, according to the response block in the partition, the integrated abnormal score is compared with the preset "abnormal response threshold" (0.8), if the score exceeds the threshold, the division is a strong response block, which is accurately matched with the breast partition number, for example, 0.8845 is higher than 0.8, then the "adenoid leaf A abnormal area division" is confirmed as a response block, which is bound with P01, P02, P03 partition number, forming a cross-zone abnormal connectivity data group, including strong response block abnormal area division, covering breast partition number and connectivity relationship.

[0027] Please refer to Figure 4 The acquisition steps of the abnormal defect linkage area group are as follows: S311: Based on the cross-zone abnormal connectivity data group, the conductivity distribution matrix and local thermal field characteristics of the partition number partition are extracted, the data in the partition is time stamped, the conductivity fluctuation value and the thermal field consistency offset are identified, and the local defect response feature set of the breast is obtained; Based on the cross-zone abnormal connectivity data group, the abnormal division corresponding to the connectivity ID and the partition number contained therein are extracted. For example, C01 contains P01, P02, P03, the conductivity distribution matrix and local thermal field characteristics of the partition number partition are extracted, for each partition (P01, P02, P03), the conductivity distribution matrix and local temperature distribution data in the abnormal period are called, and the local thermal field characteristics include the average temperature, temperature standard deviation, and maximum local temperature gradient of the partition. For example, the conductivity distribution matrix of P01 is a 64x64 numerical matrix, and the local thermal field characteristics include the average temperature of 37.5 degrees Celsius and the temperature standard deviation of 0.3 degrees Celsius, the data in the partition is time stamped, and the conductivity data and temperature data are collected at the same time point or time period. For example, the conductivity data is collected at 10:00, 10:10 and 10:20, and the temperature data is sampled at the same time point, the conductivity fluctuation value and the thermal field consistency offset are identified, the conductivity fluctuation value is obtained by calculating the conductivity standard deviation or the difference between the maximum and minimum values of the conductivity matrix of all pixel points in the partition in the abnormal period. For example, the conductivity standard deviation of P01 partition is 0.08 S / m, and the thermal field consistency offset is obtained by comparing the local thermal field characteristics in the partition with the average temperature difference under healthy state, for example, the current average temperature of P01 partition is 37.5 degrees Celsius, and the average temperature under healthy state is 37.0 degrees Celsius, the offset is 0.5 degrees Celsius, the local defect response feature set of the breast is obtained, and the feature set contains the conductivity fluctuation value and the thermal field consistency offset of each abnormal partition in the abnormal period.

[0028] S312: According to the breast local defect response feature set, the conductivity uniformity and thermal field consistency in the partition are jointly analyzed, and the formula is adopted: ; The conductivity thermal field coupling characteristic value is calculated, the partition unit of the conductivity thermal field coupling degree in the partition is screened, and the conductivity thermal field coordination response space distribution map is established; Among them, represents the conductivity thermal field coupling characteristic value, represents the conductivity value of the i th unit, represents the temperature value of the i th unit, represents the spatial coordinate change of the i th unit in the x axis direction, represents the spatial coordinate change of the i th unit in the y axis direction, represents the thermal diffusion coefficient of unit volume of material, represents the thermal coefficient of unit volume of conductivity; According to the breast local defect response feature set, the conductivity fluctuation value and the thermal field consistency offset of each partition are obtained, for example, the conductivity fluctuation value of the partition P01 is 0.08 S / m, and the thermal field consistency offset is 0.5℃, the conductivity uniformity and thermal field consistency in the partition are jointly analyzed, the analysis here is the comprehensive evaluation of the conductivity fluctuation value and the thermal field consistency offset, not simply looking at a single value, for example, the conductivity fluctuation value is high but the thermal field consistency offset is low, which indicates different types of tissue abnormalities, while both indicate more serious defects, and the formula is adopted to calculate the conductivity thermal field coupling characteristic value, wherein, represents the conductivity thermal field coupling characteristic value, represents the conductivity value of the i th unit, For example, for the partition P01, the average conductivity value 0.65 S / m in the abnormal time period can be taken, or the maximum conductivity value in the partition is taken, here the average conductivity value is taken, which is obtained by averaging the conductivity data of all monitoring points in the partition P01 in the abnormal time period, represents the temperature value of the i th unit, For example, for the partition P01, the average temperature value 37.5℃ in the abnormal time period is taken, which is obtained by averaging the temperature data of all monitoring points in the partition P01 in the abnormal time period, represents the spatial coordinate change of the i th unit in the x axis direction, For example, for the partition P01, the span in the X axis direction is , represents the spatial coordinate change of the i th unit in the y axis direction, The spatial coordinate variation of a unit in the y-axis direction, similarly, , represents the thermal diffusivity of unit volume of material, which characterizes the ability of the material to transfer heat, and its value can be obtained by experimental measurement of breast tissue or by consulting the database of biotissue thermophysical properties. For healthy breast tissue, a typical value is For abnormal tissue, this value changes, and here it is set to , represents the thermal coefficient of electrical conductivity per unit volume, which characterizes the sensitivity of electrical conductivity to temperature changes. Its value can be calculated by measuring the electrical conductivity of breast tissue samples at different temperatures and then performing linear fitting, etc. For breast tissue, a typical value is and here it is set to ; The operation logic and purpose of the formula: The formula aims to quantify the coupling degree between the electrical conductivity and the thermal field in a local region of the breast tissue, i.e. when the temperature and electrical conductivity change in a certain region, the molecular term represents the product of electrical conductivity and temperature, reflecting the strength of the thermoelectric effect; the denominator in the denominator represents the spatial scale of the unit, which is used to normalize the coupling strength to eliminate the influence of unit size and ensure comparability between different sizes of sub-regions; the rightmost is the normalization factor, where and are material intrinsic properties, which are used to convert the coupling characteristic value into a dimensionless or uniformly dimensioned index, making comparisons between different tissue types have physical significance; The benefit of this formula is that by introducing spatial scale normalization and material intrinsic property correction, the calculated coupling characteristic value can more accurately reflect the inherent physical properties of local defects, rather than being affected by sub-region size or general thermoelectric properties of the tissue, thereby improving the accuracy of abnormality detection; Now, substitute the values of P01 sub-region for calculation: , , , , , ; ; Screen the sub-region units of the electrical conductivity thermal field coupling degree in the sub-region, compare the calculated value with the preset coupling characteristic value threshold , for example, by analyzing a large number of healthy and diseased breast tissue samples, set to If partitioned Value higher than If so, the partition is considered to have high conductivity and thermal field coupling, for example, Higher than Therefore, P01 was selected, and a spatial distribution map of the conductivity thermal field cooperative response was established. All the selected high coupling degree partition units were visualized according to their actual positions in the mammary gland structure map to form a spatial distribution map, in which each high coupling degree partition was marked.

[0029] S313: Call the conductivity thermal field co-response spatial distribution map, cluster the partitions in the coupled feature value partitions that exceed the co-identification benchmark, label the partition codes and coordinates corresponding to the continuous abnormal regions, and obtain the abnormal defect linkage region group. The conductive thermal field co-response spatial distribution map is used to obtain all selected high-coupling partition units and their coupling characteristic values. For example, the coupling characteristic value of partition P01 is 1.5869 x 10^11. Partitions with coupling characteristic values ​​exceeding the co-identification benchmark are clustered. The co-identification benchmark is 1.2 x 10^11. Any partition with a coupling characteristic value higher than this benchmark is considered to exceed the benchmark. Since P01's coupling characteristic value of 1.5869 x 10^11 is higher than the benchmark of 1.2 x 10^11, P01 is identified as exceeding the benchmark. Adjacent partitions exceeding the benchmark are spatially clustered into a continuous anomalous region. For example, P01, P02, and P03 are adjacent and all have coupling characteristic values ​​exceeding the benchmark, so they are clustered into a continuous anomalous region. The partition codes and coordinates corresponding to the continuous anomalous region are labeled. Each clustered continuous anomalous region is assigned a unique identifier (A01), and the overall spatial coordinate range of all partition codes (P01, P02, P03) on the breast structure map is recorded. For example, region A01 covers pixels (10, 10) to (138, 74) to obtain an abnormal defect linkage region group. The group contains all the identified continuous abnormal regions and their associated partition codes and spatial coordinate information.

[0030] Please see Figure 5 The specific steps to obtain the breast health zoning risk level assessment list are as follows: S411: Based on the abnormal defect linkage area group, extract the partition conductivity distribution curve under the specified number, perform time uniform processing, identify the conductivity change per unit time, and obtain the partition conductivity abnormal change rate set. According to the abnormal defect linkage area group, extract each abnormal area (A01) and the subarea number (P01, P02, P03) contained therein, extract the conductivity distribution curve of the subarea under the specified number, extract the conductivity time series data of each subarea (P01) during the entire monitoring period, and construct the conductivity distribution curve. For example, the P01 conductivity curve is a continuous conductivity value E1, E2,..., Ek corresponding to the time point t1, t2,..., tk. Perform time uniform processing, synchronize the conductivity distribution curves of different subareas in time, and synchronize all curve data points to the same time stamp. For example, all curves are interpolated to one data point every 5 minutes, the conductivity performance change amount per unit time is identified, and the conductivity performance change amount between adjacent time points on the conductivity distribution curve is calculated for each subarea. Divide by the corresponding time interval to obtain the conductivity performance change rate per unit time. For example, the P01 subarea has a conductivity of 0.60 S / m at time t1 and a conductivity of 0.62 S / m at time t2 (t2-t1=10 minutes). The conductivity performance change amount per unit time is 0.002 S / (m·min). Obtain the subarea conductivity performance abnormal change rate set, and the unit time conductivity performance change amount set of all abnormal subareas at all monitoring time points.

[0031] S412: According to the subarea conductivity performance abnormal change rate set, identify the conductivity performance distribution curve of the original health state stage in the original data or the reference data, compare the current conductivity performance change sequence with the reference curve, identify the subarea conductivity performance deviation level, extract and mark the subarea whose deviation level exceeds the upper limit of the warning boundary, and obtain the deviation surge subarea set; According to the abnormal change rate set of the partition conductive performance, the conductive performance change rate of each abnormal partition at each time point is obtained. For example, the conductive performance change rate sequence of the P01 partition, which is obtained by analyzing the conductive performance data of the patient in the healthy period, or using a healthy population statistical model to construct a reference curve, for example, the average conductive performance change rate distribution curve of the P01 partition in the healthy state fluctuates between (-0.0005, 0.0005) S / (m·min), compare the current conductive performance change sequence with the reference curve, compare the current conductive performance change rate sequence with the healthy reference curve for each partition. For example, calculate the DTW distance between the current sequence of the P01 partition and the healthy reference curve, identify the deviation level of the partition conductive performance, and divide the deviation degree into different levels according to the comparison result. For example, 5 deviation levels: level 1 to level 5, the level division is based on the quantitative difference threshold. For example, the DTW distance is 0-0.5 for level 1, and greater than 2.0 for level 5, the DTW distance of the P01 partition is 1.7, and the deviation level is level 4. Extract and mark the partitions whose deviation level exceeds the upper warning limit, and the upper warning limit is the deviation level threshold 3, any partition whose deviation level is higher than 3 is extracted and marked as a "deviation sudden increase partition", for example, the deviation level of P01 is level 4, which is higher than level 3, and P01 is marked as a deviation sudden increase partition. Obtain the set of deviation sudden increase partitions, which contains all the partition numbers marked as deviation levels exceeding the upper warning limit.

[0032] S413: According to the set of deviation sudden increase partitions, bind the deviation level value of each partition with the position number in the breast structure space graph, using the formula: ; Calculate the health risk index of the partition, sort according to the risk level, and output the breast health partition detection risk level evaluation list; Wherein, represents the health risk index of the partition, represents the deviation level value, represents the mean of all partition deviation level values, represents the standard deviation of all partition deviation level values, represents the spatial distance of the partition position in the breast structure space graph, represents the risk weight coefficient of the partition position, represents the total number of involved partition positions; According to the set of deviant burst partitions, the deviant level value of each partition (e.g. P01, P04, P07) is obtained, for example, the deviant level value of P01 is 4, the deviant level value of P04 is 3.5, and the deviant level value of P07 is 4.2 (for the convenience of calculation, the level value can be refined here), and the deviant level value of each partition is bound with the position number in the breast structure space graph, and the level value of each deviant burst partition is associated with its specific position (represented by partition number and spatial coordinates) in the breast structure space graph, for example, the deviant level value 4 of P01 is bound with the position of partition P01 in the graph, and the formula is: ; the health risk index of the partition is calculated, represents the health risk index of the partition, represents the deviant level value, for example, the deviant level value of P01 is , represents the mean value of the deviant level values of all partitions, for example, in the set of deviant burst partitions {P01, P04, P07}, the deviant level values are , then , represents the standard deviation of the deviant level values of all partitions, which is calculated according to the above deviant level values, ; represents the spatial distance of the position of the th partition in the breast structure space graph, where the spatial distance refers to the distance from the deviant burst partition to its nearest known high-risk area (e.g. the position of a tumor in a historical case or a dense breast area), if there is no known high-risk area, it can be defined as the distance to the center point of the breast, here it is defined as the distance to the nearest dense breast area, in millimeters, for example, the distance of P01 partition to the nearest dense breast area is , the distance of P04 partition is , and the distance of P07 partition is , represents the risk weight coefficient of the position of the th partition, which reflects the importance or susceptibility of the position in the breast structure, for example, a specific quadrant of the breast (such as the outer upper quadrant) has a higher risk due to rich glandular tissue, which can be set according to anatomical knowledge, for example, the weight of the outer upper quadrant is , and the weight of the inner upper quadrant is , here it is assumed that P01, P04, and P07 are located in the outer upper quadrant, so the risk weight coefficient is , represents the total number of involved partition positions, which is the number of partitions in the set of deviant burst partitions, i.e. ; Operational logic and purpose of the formula: The formula aims to comprehensively assess the health risk of each partition, taking into account not only the abnormality degree of its conductivity performance deviation from the grade, but also its spatial location risk in the breast structure. The left side of the formula is the Z-score standardization of the deviation grade value, which measures the "abnormality" degree of the current partition's deviation from the average level of all deviation partitions, avoiding the absolute comparison of the original deviation grade value and better reflecting the relative risk. The right side of the formula introduces the spatial risk factor, where represents the product of each deviation partition and its spatial location risk, and then the sum and square root are taken, aiming to reflect the potential risk impact of all deviation partitions as a whole in space. The square root operation can reduce the excessive influence of extreme distance or weight on the overall risk; The benefit of this formula is that it combines the degree of physiological abnormality (deviation grade) with the inherent risk of spatial location, providing a more comprehensive and quantitative health risk assessment index, making the evaluation results more biologically meaningful and clinically valuable; Now, calculate the health risk index of P01 partition: , , , , ; For the sum term on the right side, we need to calculate the sum of all deviation surge partitions' spatial risk: ; ; ; ; Similarly, the health risk indices of P04 and P07 can be calculated: ; ; According to the risk level, sort all deviation surge partitions in descending order according to the calculated health risk index value, for example, P04( )>P07( )>P01( ). This sorting result, together with the partition number, deviation grade value and spatial coordinates, constitutes the breast health partition detection risk level evaluation list, as shown in Table 1; Table 1: Breast health partition detection risk level evaluation list

[0033] The results show that P07 partition has the highest health risk, which further indicates that its tissue abnormalities are more serious or its location has higher clinical risk, and the deviation level value of P04 is 3.5, although its health risk index R value is higher, but the result shows that its health risk is not as high as P07, and the list directly constitutes the result content of this step.

[0034] Please refer to Figure 6 The acquisition step of the breast health repair adjustment partition number list is specifically: S511: Call the breast health partition detection risk level evaluation list, extract the number of the partition in the breast function map, map the partition risk level value and the regional coordinate boundary, identify the partition information corresponding to the breast protection level, and generate a breast partition risk distribution map; Call the breast health partition detection risk level evaluation list, extract the number, health risk index R value and deviation level value of each partition. For example, the R of P04 partition is 11.546, and the deviation level value is 3.5. The breast partition number is mapped with the breast function map to determine the function attribution of the partition. For example, P04 partition belongs to a specific lobe of the breast, and the risk level value of the partition is mapped with the regional coordinate boundary. For example, the coordinate range of P04 partition is ((xa, ya), (xb, yb)), and the risk index 11.546 is bound to this region. Identify the partition information corresponding to the breast protection level, according to the clinical guidelines or preset strategy, divide the breast region into different protection levels (1st level: high risk, 2nd level: medium risk, 3rd level: low risk), and identify the partition information corresponding to each risk level. For example, set R greater than or equal to 10 as the 1st protection zone, 5 less than or equal to R less than 10 as the 2nd protection zone, and R less than 5 as the 3rd protection zone. P04 (R=11.546) is the 1st protection zone, P07 (R=8.641) is the 2nd protection zone, and P01 (R=2.885) is the 3rd protection zone. Generate a breast partition risk distribution map, which visualizes the protection level of each partition of the breast. Different protection level partitions are marked on the breast structure map with different colors or markers.

[0035] S512: Based on the breast partition risk distribution map, extract the repair response unit number and response level, match the partition risk level with the repair response level, identify the unit number with insufficient response coverage, and obtain a list of breast partition response risk disconnection; Based on the breast partition risk distribution map, the protection level information of each partition and its map marker are obtained. For example, P04 is a first protection zone (red highlight), P07 is a second protection zone (yellow), and P01 is a third protection zone (green). The repair response unit is an intervention measure, and each unit has a response level indicating the applicable risk level and treatment intensity. For example, the physical therapy (unit F01) response level is "mild (applicable to a third protection zone)", the drug therapy (unit F02) response level is "moderate (applicable to a second protection zone)", and the surgical treatment (unit F03) response level is "severe (applicable to a first protection zone)". The risk level of each partition is matched with the response level of the corresponding repair response unit. For example, P04 (first protection zone) is matched with F03 (severe), P07 (second protection zone) is matched with F02 (moderate), and P01 (third protection zone) is matched with F01 (mild). Each partition protection level is matched with the response level of the corresponding repair response unit. For example, P04 (first protection zone) is matched with F03 (severe), P07 (second protection zone) is matched with F02 (moderate), and P01 (third protection zone) is matched with F01 (mild). The unit number of insufficient response coverage is identified, and it is checked whether each partition obtains repair response coverage. For example, the first protection zone P04 currently only matches F02 (moderate) instead of F03 (severe), which is insufficient response coverage. The unit number (F02) and its service area (P04) are identified, and a breast partition response risk disconnection list is obtained. The list lists the partition numbers with insufficient response coverage and the current matched (or unmatched) repair response unit information.

[0036] S513: According to the breast partition response risk disconnection list, according to the level number in the breast protection priority sequence, extract the key partition number that needs to improve the response coverage range, output the adjustment control parameter linked with the original repair unit in order, and output the breast health repair adjustment partition number list; According to the breast partition response risk disconnection list, the partition number with insufficient response coverage and its current response unit and expected response unit information are obtained, for example, P04 partition current response unit F02 (moderate), expected response unit F03 (severe), the key partition number that needs to improve the response coverage range is extracted, and the breast protection priority sequence guides the intervention priority. From the disconnection list, according to the priority sequence, the key partition that needs to immediately improve the response coverage range is extracted. For example, P04 (1st protection area) has high priority and is confirmed as a key partition that needs to improve the response coverage range. The adjustment control parameters linked with the original repair unit are output in order, and for each key partition, the existing repair unit adjustment parameters are determined according to the expected repair response level. The adjustment parameters include increasing the treatment scheme intensity, adjusting the drug dose, increasing the treatment frequency, or introducing advanced treatment methods, for example, P04 partition current F02 (moderate), expected F03 (severe), adjustment control parameters include: "F02 treatment intensity is increased to 80%, F03 intervention frequency is increased to 2 times per week", the parameters are linked with the original repair unit and output, and the breast health repair adjustment partition number list is output, which contains all the key partition numbers that need to be adjusted and their detailed adjustment control parameters.

[0037] The breast health data anomaly detection system is used to perform the breast health data anomaly detection method described above, and the system comprises: The temperature monitoring module divides the detection partition based on the breast tissue surface temperature distribution and the electrical conductivity characteristics, compares the temperature fluctuation range and the electrical conductivity response mutation point in the same time period, screens the synchronous abnormal area of the two, extracts the partition number and spatial coordinates, and summarizes the abnormal time period and the partition number to generate a breast partition abnormal attribute set; The partition positioning module identifies the consistency of the temperature gradient direction and the electrical conductivity response vector based on the breast partition abnormal attribute set, labels the partition boundary number, matches the overall breast structure diagram, extracts the partition number range of the temperature gradient abnormal area, and establishes a cross-zone abnormal connectivity data group; The defect linkage module retrieves the electrical conductivity distribution matrix and the continuous data sequence of the local thermal field characteristics in the region based on the cross-zone abnormal connectivity data group, judges the electrical conductivity abnormal boundary connectivity and the thermal field consistency, marks the partition number that meets the linkage threshold of the two, and outputs the abnormal defect linkage area group; The electrical conductivity warning module analyzes the electrical conductivity performance distribution trend and the degree of deviation from the original health state curve of the corresponding partition based on the partition number of the abnormal defect linkage area group, extracts the partition number of the deviation trend, completes the level identification according to the risk grading standard, and generates a breast health partition detection risk level evaluation list; The repair optimization module detects the risk level evaluation list based on the breast health partition, finds the corresponding position number of the risk level partition in the breast function map, calls the current repair response unit configuration list, compares the breast protection priority with the current response level matching condition, filters the partition that needs to be updated, and outputs the breast health repair adjustment partition number list.

[0038] The above is only a preferred embodiment of the present application, and does not limit the present application in other forms. Any person skilled in the art can use the disclosed technical content to make changes or modifications as equivalent embodiments applied to other fields, but any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present application without departing from the technical solution content of the present application still belongs to the protection scope of the technical solution of the present application.

Claims

1. A method for detecting abnormal breast health data, characterized in that, Includes the following steps: S1: Based on the surface temperature distribution and conductivity characteristics of breast tissue, detection zones are divided, the temperature change trajectory and conductivity response curve of each zone are extracted, and synchronous abnormal regions of temperature fluctuation range and conductivity response mutation point are identified on the time axis. The corresponding zone number and spatial coordinates are extracted to generate a set of abnormal attributes of breast zones. S2: Based on the set of abnormal attributes of the mammary gland partitions, identify the temperature gradient direction and conductivity response vector of the partitions, analyze the consistency of their directions at the partition boundaries, and combine the partition boundary connectivity to screen partitions with connectivity and abnormal temperature gradients to form a cross-regional abnormal connectivity data group. S3: Based on the cross-regional abnormal connectivity data group, extract the tissue conductivity distribution matrix and local thermal field characteristics within the partition, analyze the conductivity uniformity and thermal field consistency, screen and match abnormal regions, and obtain abnormal defect linkage region groups. S4: Based on the abnormal defect linkage area group, analyze the conductivity distribution trend of the corresponding partition, assess the degree of deviation from the original healthy state curve, mark the abnormality level of the partition according to the deviation range, and output the breast health partition detection risk level assessment list.

2. The method for detecting abnormal breast health data according to claim 1, characterized in that, The set of abnormal attributes for breast regions includes temperature mutation region number, abnormal conductivity response points, region coordinate markers, and time series mutation identifiers. The cross-regional abnormal connectivity data group includes temperature gradient abnormal region identifiers, region boundary connectivity units, and region boundary consistency blocks. The abnormal defect linkage region group includes continuous conductivity abnormal regions, irregular thermal field regions, overlapping conductivity performance mutation regions, and linkage abnormal region numbers. The breast health region detection risk level assessment list includes risk level labels, region response deviation values, local conductivity abnormality indicators, and health status deviation levels.

3. The method for detecting abnormal breast health data according to claim 1, characterized in that, The specific steps for obtaining the set of abnormal attributes of the breast region are as follows: S111: Based on the surface temperature distribution and conductivity characteristics of breast tissue, detection zones are divided, the temperature change trajectory and conductivity response curve of each zone are extracted, and the difference between the two types of data in the same zone is calculated to obtain the trend value of the difference between temperature and conductivity response. S112: Based on the trend value of the difference between temperature and conductivity response, identify the fluctuation range in the temperature change curve and the abrupt change point of the conductivity response trajectory, perform time superposition on the two types of values, extract the time interval where the fluctuation exceeds the preset benchmark value and the abrupt change point is within the set range, and generate a set of high-frequency abnormal interval time periods. S113: For the set of high-frequency abnormal interval time periods, match the corresponding partition number and spatial coordinate information, extract the partition location where the signal occurred, and generate a set of abnormal attributes of the breast partition.

4. The method for detecting abnormal breast health data according to claim 3, characterized in that, The specific steps for obtaining the cross-regional abnormal connectivity data group are as follows: S211: Identify the temperature gradient direction and conductivity response vector of the abnormal attribute set of the breast region, extract the projection trajectory of the two at the boundary of the region, identify the distribution number and aggregation degree of the boundary point in the region, and obtain the boundary consistency map of the region. S212: Based on the partition boundary consistency map, filter the boundary areas with a higher degree of aggregation than the average level, compare with the spatial boundary of the overall breast structure map, identify continuous and concentrated boundary areas belonging to the same partition, and obtain the temperature gradient abnormality zone within the breast partition. S213: Based on the abnormal temperature gradient zoning within the breast region, an integrated analysis is performed on the consistency of the region boundary, the dispersion of the temperature gradient, the uniformity of conductivity, and the delay of conductivity response. Based on the response blocks in the region, the region is matched to form a cross-regional abnormal connectivity data group.

5. The method for detecting abnormal breast health data according to claim 4, characterized in that, The specific steps for obtaining the abnormal defect linkage region group are as follows: S311: Based on the cross-regional abnormal connectivity data group, extract the conductivity distribution matrix and local thermal field features of the numbered partitions within the partition, align the data within the partition with timestamps, identify the conductivity fluctuation value and thermal field consistency offset, and obtain the breast local defect response feature set. S312: Based on the aforementioned local defect response feature set of the breast, a joint analysis of the conductivity uniformity and thermal field consistency within the partition is performed, using the following formula: ； Calculate the characteristic value of conductivity-thermal field coupling, filter the partitioning units of conductivity-thermal field coupling degree in the partition, and establish a spatial distribution map of conductivity-thermal field cooperative response; in, Represents the characteristic value of electrical conductivity coupled with thermal field. This represents the conductivity value of the i-th unit. This represents the temperature value of the i-th unit. This represents the change in the spatial coordinates of the i-th unit along the x-axis. This represents the change in the spatial coordinates of the i-th unit along the y-axis. The thermal diffusivity of a material per unit volume. Thermal coefficient, representing electrical conductivity per unit volume; S313: Call the conductivity thermal field cooperative response spatial distribution map, cluster the partitions in the coupled feature value partitions that exceed the cooperative identification benchmark, label the partition codes and coordinates corresponding to the continuous abnormal regions, and obtain the abnormal defect linkage region group.

6. The method for detecting abnormal breast health data according to claim 5, characterized in that, The specific steps for obtaining the breast health zone detection risk level assessment list are as follows: S411: Based on the abnormal defect linkage area group, extract the partition conductivity distribution curve under the specified number, perform time uniform processing, identify the conductivity change per unit time, and obtain the partition conductivity abnormal change rate set. S412: Based on the set of abnormal change rates of conductivity performance of the partitions, identify the conductivity distribution curve of the original healthy state stage in the original data or reference data, compare the current conductivity change sequence with the reference curve, identify the deviation level of conductivity performance of the partitions, extract and mark the partitions whose deviation level exceeds the upper limit of the warning, and obtain the set of partitions with sudden increase in deviation. S413: Based on the aforementioned set of deviation abrupt increase partitions, the deviation level value of each partition is bound to its position number in the breast structure spatial map, using the formula: ； Calculate the health risk index of each region, sort them by risk level, and output a list of breast health region detection risk level assessments. in, The health risk index represents the region. This represents a deviation from the grade value. This represents the mean deviation of all partitions from their rank values. This represents the standard deviation of all partitions from their respective rank values. The first one in the spatial diagram representing the breast structure Spatial distance between partition locations Representing the Risk weight coefficient for each partition location This represents the total number of partition locations involved.

7. The method for detecting abnormal breast health data according to claim 1, characterized in that, The method also includes step S5: S5: Call the breast health zoning detection risk level assessment list, identify the corresponding number of the zoning in the breast function map, retrieve the repair response unit list, compare the response level with the breast protection priority sequence, filter the zoning numbers that need to adjust the response coverage, and output the breast health repair adjustment zoning number list. The breast protection priority sequence refers to the ranking based on the criticality of breast function and the need for protection. The list of adjustment zones for breast health repair includes the target zone number, response level adjustment parameters, protection priority comparison items, and linkage response trigger types.

8. The method for detecting abnormal breast health data according to claim 7, characterized in that, The specific steps for obtaining the list of breast health repair and adjustment zone numbers are as follows: S511: Call the breast health zone detection risk level assessment list, extract the zone number in the breast function map, map the zone risk level value to the regional coordinate boundary, identify the zone information corresponding to the breast protection level, and generate a breast zone risk distribution map. S512: Based on the breast region risk distribution map, extract the repair response unit number and response level, match the region risk level with the repair response level, identify the unit number with insufficient response coverage, and obtain the breast region response risk disconnect list. S513: Based on the breast region response risk disconnect list, extract the key region numbers that need to improve the response coverage according to the level number in the breast protection priority sequence, output the adjustment control parameters linked with the original repair unit in sequence, and output the breast health repair adjustment region number list.

9. A breast health data abnormality detection system, characterized in that, The system is used to implement the breast health data abnormality detection method according to any one of claims 1-8, the system comprising: The temperature monitoring module divides the detection zones based on the surface temperature distribution and conductivity characteristics of breast tissue. It compares the temperature fluctuation range and conductivity response abrupt change points within the same time period, filters out synchronous abnormal areas, extracts the zone number and spatial coordinates, summarizes the abnormal time period and zone number, and generates a set of abnormal attributes of breast zones. Based on the set of abnormal attributes of the breast region, the partition localization module identifies the consistency between the temperature gradient direction and the conductivity response vector, marks the partition boundary number, matches the overall breast structure map, extracts the partition number range of the abnormal temperature gradient region, and establishes a cross-regional abnormal connectivity data group. Based on the cross-regional abnormal connectivity data group, the defect linkage module retrieves the conductivity distribution matrix and the local thermal field characteristics in the region as a continuous data sequence, determines the consistency between the conductivity abnormal boundary connectivity and the thermal field, marks the partition number that meets the linkage threshold of both, and outputs the abnormal defect linkage region group. Based on the partition number of the abnormal defect linkage area group, the conductivity early warning module analyzes the distribution trend of conductivity performance of the corresponding partition and the degree of deviation from the original health state curve, extracts the partition number of the deviation trend, completes the level identification according to the risk classification standard, and generates a breast health partition detection risk level assessment list. The repair and optimization module, based on the breast health partition detection risk level assessment list, finds the corresponding position number of the risk level partition in the breast function diagram, retrieves the current repair response unit configuration list, compares the matching situation of breast protection priority with the current response level, filters the partitions that need to be updated, and outputs a list of breast health repair adjustment partition numbers.

Images