Chromosome abnormality detection method based on image analysis

Abstract

The invention belongs to the technical field of biomedical image processing and analysis, and particularly discloses a chromosome abnormality detection method based on image analysis, which comprises the following steps: obtaining a user chromosome microscopic image set and extracting a multi-modal feature data set for processing to obtain abnormal chromosome feature data of an abnormal tracking area; and performing collection processing to obtain an abnormal risk assessment index, performing correction in combination with abnormal identification coefficients of a plurality of historical similar abnormal samples to obtain an abnormal risk correction index, and performing assessment processing according to an obtained clinical detection data set corresponding to the abnormal tracking area in combination with genetic background feature data to obtain an abnormal tracking area genetic risk judgment index. And comparing the abnormal risk correction index with the abnormal risk correction index to obtain an abnormal evaluation matching degree coefficient, and verifying an abnormal evaluation result through threshold comparison, so that efficient identification and accurate evaluation of various complex chromosome abnormality types are realized, and comprehensive risk evaluation is performed in combination with genetic background information and clinical data to improve the reliability of a detection result.

Description

Technical Field

[0001] The present invention belongs to the technical field of biomedical image processing and analysis, and specifically relates to a method for detecting chromosomal abnormalities based on image analysis. Background Art

[0002] With the rapid development of medical image analysis technology, the method for detecting chromosomal abnormalities based on image analysis has gradually become a research hotspot in the field of molecular diagnostics. By combining computer vision, deep learning, and image processing technologies, this method can significantly improve the efficiency and accuracy of chromosomal abnormality detection, providing important support for clinical diagnosis. However, the existing related technical solutions still have deficiencies in aspects such as image processing accuracy, automation degree, and the ability to identify complex abnormality types, which limit their popularization and application in practical use.

[0003] In the prior art, the patent with the publication number CN115619774B proposed a method for identifying chromosomal abnormalities. This solution mainly targets specific types of chromosomal abnormalities and lacks broad applicability to other complex or rare chromosomal abnormalities. In addition, this method relies on high-quality input images. If the image quality is poor or there is noise, it may lead to a decrease in recognition accuracy. At the same time, its model training process requires a large amount of labeled data, increasing the time and cost of data preparation. Another prior art, the patent with the publication number CN112288706B, proposed an automated method for chromosomal karyotype analysis and abnormality detection. This solution may have certain limitations when dealing with complex chimeras or the coexistence of multiple abnormality types. In addition, due to its model design focusing on the rough localization of the target area and semantic feature extraction, it may be less sensitive to subtle structural abnormalities, increasing the risk of missed or false detections. At the same time, this method has high requirements for the computing power of hardware devices and may be difficult to widely apply in resource-constrained environments.

[0004] The above problems indicate that the existing methods for detecting chromosomal abnormalities based on image analysis still have certain deficiencies in aspects such as the broad applicability of abnormality types, the adaptability to complex scenarios, the sensitivity to subtle structural abnormalities, and the demand for hardware resources. Therefore, it is urgent to provide a technical solution to solve the above problems.

[0005] At the same time, there are obvious deficiencies in the existing methods for detecting chromosomal abnormalities in terms of psychological assessment, case matching, and misdiagnosis control. Summary of the Invention

[0006] In order to solve the above technical problems, the present invention provides a method for detecting chromosomal abnormalities based on image analysis, and the following technical solutions: Collect multiple sets of chromosomal microscopic images of a user at multiple time periods within a preset period, and obtain the user's genetic background information and clinical test reports; Perform multi-modal feature extraction on the chromosome microscopic image sets in each time period to obtain a chromosome microscopic image feature data set; Extract a clinical test data set according to the clinical test report, including a chromosome structural abnormality index, a chromosome number abnormality index, and a genetic disease risk index, and extract genetic background feature data according to the genetic background information; Process and identify the corresponding chromosome microscopic image feature data sets of the user in each time period to obtain abnormal chromosome feature data, and mark the corresponding abnormal chromosome identification regions as abnormal tracking regions; Aggregate the multiple abnormal chromosome feature data corresponding to the abnormal tracking regions to obtain an abnormal tracking region feature data set, and process it to obtain an abnormal risk assessment index; Perform similarity matching on the abnormal chromosome feature data of the abnormal tracking region through a preset chromosome abnormality feature database to obtain multiple historical similar abnormal samples that meet the requirements and their corresponding abnormal identification coefficients, and perform correction processing on the abnormal risk assessment index to obtain an abnormal risk correction index; Perform evaluation processing on the clinical test data set corresponding to the abnormal tracking region in each time period in combination with the genetic background feature data through a preset genetic risk assessment model to obtain an abnormal tracking region genetic risk determination index; Compare and process the abnormal tracking region genetic risk determination index and the abnormal risk correction index to obtain an abnormal evaluation matching degree coefficient for the abnormal tracking region, and compare it with a preset abnormal identification matching degree threshold to verify the abnormal evaluation result.

[0007] In an alternative manner, the performing multi-modal feature extraction on the chromosome microscopic image sets in each time period to obtain a chromosome microscopic image feature data set includes: Pick up image feature information from the chromosome microscopic image sets in each time period through a preset multi-modal feature extraction model to obtain a chromosome microscopic image feature data set for each time period; Extract corresponding chromosome microscopic image feature data from the chromosome microscopic image feature data set, including chromosome morphological feature data, chromosome distribution feature data, and chromosome texture feature data.

[0008] In an alternative manner, the extracting a clinical test data set according to the clinical test report, including a chromosome structural abnormality index, a chromosome number abnormality index, and a genetic disease risk index, and extracting genetic background feature data according to the genetic background information includes: Extract a clinical test data set according to the clinical test report, including the chromosomal structural abnormality index, chromosomal number abnormality index, and genetic disease risk index within each time period; Extract genetic background feature data according to the genetic background information, including family genetic history data, gene mutation frequency data, and environmental exposure risk data.

[0009] In an alternative manner, the processing and recognition of the chromosomal microscopic image feature data set corresponding to the user within each time period to obtain abnormal chromosome feature data, and marking the corresponding abnormal chromosome identification area as an abnormal tracking area includes: Process and recognize the chromosomal microscopic image feature data set corresponding to the user within each time period through a preset chromosomal abnormality detection model to obtain abnormal chromosome feature data; The abnormal chromosome feature data includes abnormal chromosome identification data, abnormal chromosome contour size data, abnormal chromosome breakpoint position data, and abnormal chromosome recombination pattern data; Obtain the abnormal chromosome identification area corresponding to the identification according to the abnormal chromosome identification data, and mark it as an abnormal tracking area.

[0010] In an alternative manner, the aggregation of multiple abnormal chromosome feature data corresponding to the abnormal tracking area to obtain an abnormal tracking area feature data set, and the processing to obtain an abnormal risk assessment index includes: Aggregate multiple abnormal chromosome feature data corresponding to the abnormal tracking area within the preset period to obtain an abnormal tracking area feature data set; Process the abnormal tracking area feature data set through a preset abnormal risk assessment model to obtain the abnormal risk assessment index of the abnormal tracking area; the calculation formula of the abnormal risk assessment index is: ; where, is the abnormal risk assessment index, is the chromosomal structural abnormality index in the th preset time period, is the chromosomal number abnormality index in the th preset time period, is the chromosomal texture abnormality index in the th preset time period, is the preset risk coefficient of the abnormal area, is the number of time periods within the preset period, , and are the first preset weight coefficients.

[0011] In an alternative approach, the abnormal chromosome feature data of the abnormal tracking region is subjected to similarity matching through a preset chromosome abnormality feature database to obtain multiple historical similar abnormal samples that meet the requirements and their corresponding abnormal identification coefficients, and the abnormal risk assessment index is corrected to obtain an abnormal risk correction index, including: Perform similarity matching processing on the abnormal chromosome feature data of the abnormal tracking region through a preset chromosome abnormality feature database to obtain multiple historical similar abnormal samples that meet the similarity requirements; Extract the corresponding abnormal identification coefficients for the final inspection of the samples based on the multiple historical similar abnormal samples; Perform correction processing on the abnormal risk assessment index based on the multiple abnormal identification coefficients to obtain an abnormal risk correction index; the correction calculation formula for the abnormal risk correction index is: ; where is the abnormal risk correction index, is the abnormal identification coefficient of the th historical similar abnormal sample, is the number of historical similar abnormal samples, is the th preset weight coefficient of the historical similar abnormal sample.

[0012] In an alternative approach, the clinical test data set corresponding to the abnormal tracking region in each time period is combined with the genetic background feature data and evaluated through a preset genetic risk assessment model to obtain a genetic risk determination index for the abnormal tracking region, including: Perform evaluation processing on the chromosome structure abnormality index, chromosome number abnormality index, and genetic disease risk index corresponding to the abnormal tracking region in each time period through a preset genetic risk assessment model to obtain genetic risk identification data for the abnormal tracking region; Perform correction processing on the genetic risk identification data of the abnormal tracking region and the family genetic history data, gene mutation frequency data, and environmental exposure risk data to obtain the genetic risk determination index for the abnormal tracking region of the user within the preset period; The calculation formula for the genetic risk identification data of the abnormal tracking region is: ; where is the genetic risk identification data of the abnormal tracking region, is the genetic disease risk index in the th preset time period, , , are the second preset weight coefficients; The correction calculation formula for the genetic risk determination index of the abnormal tracking region is: ; wherein, is the genetic risk determination index of the abnormal tracking area, is the data of family genetic history, is the gene mutation frequency data, is the environmental exposure risk data, , , are the third preset weight coefficients.

[0013] In an optional manner, comparing and processing the genetic risk determination index of the abnormal tracking area with the abnormal risk correction index to obtain the abnormal evaluation matching degree coefficient of the abnormal tracking area, and comparing it with a preset abnormal identification matching degree threshold for threshold comparison to verify the abnormal evaluation result, including: Comparing and processing the genetic risk determination index of the abnormal tracking area of the user with the abnormal risk correction index to obtain the abnormal evaluation matching degree coefficient of the abnormal tracking area; Comparing the abnormal evaluation matching degree coefficient with a preset abnormal identification matching degree threshold, and verifying the abnormal evaluation result according to the threshold comparison result; The calculation formula of the abnormal evaluation matching degree coefficient is: ; wherein, is the abnormal evaluation matching degree coefficient, is the genetic risk determination index of the abnormal tracking area, is the abnormal risk correction index, and are the fourth preset weight coefficients.

[0014] In an optional manner, after obtaining the abnormal risk correction index, perform the following psychological intervention integration steps: a) Start a three-level psychological assessment mechanism: Call the PHQ-9 and GAD-7 standardized scales for electronic evaluation to generate a psychological state score; Monitor the user's physiological behavior index dataset (including sleep duration, activity trajectory, eating frequency), and calculate the deviation degree from the preset baseline value; Load the structured interview protocol, and match the keywords through voice interaction to diagnose the cognitive level; b) Construct a dynamic intervention matrix: When the psychological state score exceeds the first threshold, activate the cognitive behavioral therapy execution path When the behavior deviation degree exceeds the second threshold for 3 consecutive days, trigger the generation of a multidisciplinary consultation instruction Generate a three-dimensional heat map visualization interface, with the horizontal axis associated with the abnormal evaluation matching degree coefficient, the vertical axis mapping the psychological state score, and the color scale representing the intervention priority.

[0015] In an alternative approach, the historical similar abnormal sample matching process includes: Establish a case decision tree engine, associate the historical similar case library according to the diagnostic code, and extract the treatment cycle and prognosis timeline atlas, the home care plan knowledge base, and the genetic risk assessment correction parameters; When it is detected that the user's retrieval behavior contains a preset keyword, automatically push the disposal records of similar cases with a matching degree higher than 85%; Highlight the topological differences in chromosome breakpoints between the current abnormal tracking area and historical samples in the visualization interface; It also includes an uncertainty management mechanism: a) When the confidence level of chromosome image analysis is lower than 90%: Mark a dynamic flashing review identifier in the abnormal tracking area; Associate the clinical manifestation database to generate a pop-up window containing at least 3 differential diagnosis suggestions; b) Establish a closed-loop feedback system: Automatically initiate on the 7th and 30th days after generating the abnormal assessment result: 1) Chromosome microscopic image re-sampling instruction 2) Psychological state score re-test instruction Dynamically adjust the weighting parameters of the preset genetic risk assessment model according to the deviation amplitude between the re-test data and the baseline value.

[0016] Compared with the prior art, the beneficial effects of the present invention are: The chromosome abnormality detection method based on image analysis provided by the present invention obtains the abnormal chromosome feature data of the abnormal tracking area by acquiring the user's chromosome microscopic image set and extracting the multi-modal feature data set, and obtains the abnormal risk assessment index through combined processing. Then, it is corrected by combining the abnormal identification coefficients of multiple historical similar abnormal samples to obtain the abnormal risk correction index.

[0017] According to the clinical test data set corresponding to the obtained abnormal tracking area and the genetic background feature data, the genetic risk determination index of the abnormal tracking area is evaluated and processed. Then, it is compared with the abnormal risk correction index to obtain the abnormal assessment matching degree coefficient, and the abnormal assessment result is verified through threshold comparison; thereby realizing the efficient identification and accurate assessment of various complex chromosome abnormality types, and combining genetic background information and clinical data for comprehensive risk assessment to improve the reliability of the detection result.

[0018] Through a three - level psychological assessment mechanism (scale assessment, behavior monitoring, structured interview), it comprehensively captures the subjective feelings and objective physiological changes of patients, significantly improving the accuracy of psychological state recognition and the timeliness of intervention. The dynamic intervention matrix can quickly locate patients who need to be given priority for consultation, improving the efficiency of clinical decision - making support.

[0019] The case - matching engine outputs historical cases that are highly similar to the abnormal regions of the current patient through a feature - matching algorithm, providing evidence - based medical support and optimizing treatment decisions. The topological difference visualization technology helps doctors quickly interpret the positions of chromosomal abnormalities, significantly enhancing the efficiency of genetic counseling.

[0020] The confidence management system significantly reduces the false - positive rate and karyotype misjudgment rate through a quality assessment sub - module, improving the review rate of suspicious results. The closed - loop feedback mechanism optimizes the genetic risk assessment model through retest data, continuously enhancing the prediction accuracy of the model and the sensitivity of key feature recognition.

[0021] Through regular whole - genome re - examination and structured telephone follow - up, the progression of chromosomal abnormalities and treatment side effects can be detected early, ensuring timely prognosis assessment of patients and improving the overall treatment effect. The push of treatment record videos and real - time case comparison enhance patients' understanding of treatment plans, reduce decision - making anxiety, and improve patients' compliance and satisfaction with treatment.

[0022] In summary, through innovative technical means, the present invention fully demonstrates outstanding substantive features and significant technological progress compared with the prior art, meeting the requirements of the patent law for the height of creativity.

[0023] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on an examination of the following text, or can be taught from the practice of the present invention. Brief Description of the Drawings

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following - described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0025] Figure 1 It is a schematic flow chart of a method for detecting chromosomal abnormalities based on image analysis. Detailed Embodiments

[0026] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0027] Embodiment 1 Figure 1 The flowchart of an embodiment of a chromosome abnormality detection method based on image analysis provided by the present invention is shown. As Figure 1 shown, the method includes the following steps: S1. Collect multiple sets of chromosome microscopic images of a user in multiple time periods within a preset period, and obtain the user's genetic background information and clinical test reports; S2. Extract multi-modal features from the set of chromosome microscopic images in each time period to obtain a chromosome microscopic image feature data set; S3. Extract a clinical test data set according to the clinical test report, including a chromosome structure abnormality index, a chromosome number abnormality index, and a genetic disease risk index, and extract genetic background feature data according to the genetic background information; S4. Process and identify the corresponding chromosome microscopic image feature data set of the user in each time period to obtain abnormal chromosome feature data, and mark the corresponding abnormal chromosome identification area as an abnormal tracking area; S5. Aggregate the multiple abnormal chromosome feature data corresponding to the abnormal tracking area to obtain an abnormal tracking area feature data set, and process it to obtain an abnormal risk assessment index; S6. Perform similarity matching through a preset chromosome abnormality feature database according to the abnormal chromosome feature data of the abnormal tracking area to obtain multiple historical similar abnormal samples that meet the requirements and their corresponding abnormal identification coefficients, and perform correction processing on the abnormal risk assessment index to obtain an abnormal risk correction index; S7. Perform evaluation processing through a preset genetic risk assessment model according to the clinical test data set corresponding to the abnormal tracking area in each time period in combination with the genetic background feature data to obtain an abnormal tracking area genetic risk determination index; S8. Compare the abnormal tracking area genetic risk determination index with the abnormal risk correction index to obtain an abnormal evaluation matching degree coefficient of the abnormal tracking area, and compare it with a preset abnormal identification matching degree threshold to verify the abnormal evaluation result.

[0028] In an alternative manner, S2 includes: Pick up image feature information from the set of chromosome microscopic images in each time period through a preset multi-modal feature extraction model to obtain a chromosome microscopic image feature data set for each time period; Extract the corresponding chromosome microscopic image feature data according to the chromosome microscopic image feature data set, including chromosome morphological feature data, chromosome distribution feature data, and chromosome texture feature data.

[0029] In an alternative manner, S3 includes: Extract the clinical test data set according to the clinical test report, including the chromosome structure abnormality index, chromosome number abnormality index, and genetic disease risk index within each time period; Extract the genetic background feature data according to the genetic background information, including family genetic history data, gene mutation frequency data, and environmental exposure risk data.

[0030] In an alternative manner, S4 includes: Process and identify the corresponding chromosome microscopic image feature data set of the user within each time period through a preset chromosome abnormality detection model to obtain abnormal chromosome feature data; The abnormal chromosome feature data includes abnormal chromosome identification data, abnormal chromosome contour size data, abnormal chromosome breakpoint position data, and abnormal chromosome recombination pattern data; Obtain the abnormal chromosome identification area corresponding to the corresponding identification according to the abnormal chromosome identification data, and mark it as an abnormal tracking area.

[0031] In an alternative manner, S5 includes: Aggregate the multiple abnormal chromosome feature data corresponding to the abnormal tracking area within the preset period to obtain an abnormal tracking area feature data set; Process the abnormal tracking area feature data set through a preset abnormal risk assessment model to obtain the abnormal risk assessment index of the abnormal tracking area; The calculation formula of the abnormal risk assessment index is: ; where is the abnormal risk assessment index, is the chromosome structure abnormality index in the th preset time period, is the chromosome number abnormality index in the th preset time period, is the chromosome texture abnormality index in the th preset time period, is the preset risk coefficient of the abnormal area, is the number of time periods within the preset period, , and are the first preset weight coefficients.

[0032] In an alternative manner, S6 includes: Performing similarity matching processing through a preset chromosomal abnormality feature database based on the abnormal chromosomal feature data of the abnormal tracking region to obtain multiple historical similar abnormal samples that meet the similarity requirements; Extracting the corresponding abnormal identification coefficients for the final inspection of the samples based on the multiple historical similar abnormal samples; Performing correction processing on the abnormal risk assessment index according to the multiple abnormal identification coefficients to obtain an abnormal risk correction index; The correction calculation formula for the abnormal risk correction index is: ; where is the abnormal risk correction index, is the abnormal identification coefficient of the th historical similar abnormal sample, is the number of historical similar abnormal samples, is the th preset weight coefficient of the historical similar abnormal sample.

[0033] In an alternative manner, S7 includes: Performing assessment processing through a preset genetic risk assessment model based on the chromosomal structure abnormality index, chromosomal number abnormality index, and genetic disease risk index corresponding to the abnormal tracking region within each time period to obtain abnormal tracking region genetic risk identification data; Performing correction processing based on the abnormal tracking region genetic risk identification data, the family genetic history data, the gene mutation frequency data, and the environmental exposure risk data to obtain the abnormal tracking region genetic risk determination index of the user within the preset period; The calculation formula for the abnormal tracking region genetic risk identification data is: ; where is the abnormal tracking region genetic risk identification data, is the genetic disease risk index within the th preset time period, , , are the second preset weight coefficients; The correction calculation formula for the abnormal tracking region genetic risk determination index is: ; where is the abnormal tracking region genetic risk determination index, is the family genetic history data, is the gene mutation frequency data, is the environmental exposure risk data, , , are the third preset weight coefficients.

[0034] In an alternative approach, S8 includes: Comparing the genetic risk determination index of the abnormal tracking region of the user with the abnormal risk correction index to obtain the abnormal evaluation matching degree coefficient of the abnormal tracking region; Comparing the abnormal evaluation matching degree coefficient with a preset abnormal identification matching degree threshold, and verifying the abnormal evaluation result according to the threshold comparison result; The calculation formula of the abnormal evaluation matching degree coefficient is: ; where is the abnormal evaluation matching degree coefficient, is the genetic risk determination index of the abnormal tracking region, is the abnormal risk correction index, and is the fourth preset weight coefficient.

[0035] In this embodiment, it should be noted that: A method for chromosome abnormality detection based on image analysis according to the present invention, the core of which is to realize the efficient recognition and accurate evaluation of complex chromosome abnormality types through multi-modal feature extraction, dynamic weight allocation mechanism and multi-level abnormality classification model. Specifically: 1) According to Figure 1 the overall process shown, in this embodiment, multiple groups of chromosome microscopic image sets of the user are collected within a preset period, and the genetic background information and clinical test reports of the user are synchronously obtained. These data form the basis for subsequent analysis. Specifically, the chromosome microscopic image sets are obtained through high-resolution microscope equipment to ensure that the image quality meets the analysis requirements. The genetic background information includes family genetic history data, gene mutation frequency data and environmental exposure risk data, while the clinical test reports cover chromosome structure abnormality index, chromosome number abnormality index and genetic disease risk index. The collection process of these data needs to strictly follow the standardized operating procedures to ensure the accuracy and consistency of the data.

[0036] 2) Entering the multi-modal feature extraction stage, in this embodiment, a preset multi-modal feature extraction model is used to process the chromosome microscopic image sets in each time period, and chromosome morphological feature data, chromosome distribution feature data and chromosome texture feature data are extracted. These feature data together constitute the chromosome microscopic image feature data set. Among them, the chromosome morphological feature data is used to describe the overall shape and contour of the chromosome; the chromosome distribution feature data reflects the spatial distribution law of the chromosomes in the cell nucleus; the chromosome texture feature data captures the microscopic texture changes on the surface of the chromosomes. The key to this stage lies in the design of the multi-modal feature extraction model, which adopts deep learning algorithms internally and can automatically learn and extract the key features in the image, thereby improving the efficiency and accuracy of feature extraction.

[0037] 3) In this embodiment, the chromosomal microscopic image feature dataset of the user in each time period is processed and recognized to obtain abnormal chromosome feature data, and the corresponding abnormal chromosome identification area is marked as an abnormal tracking area. This process relies on a preset chromosomal abnormality detection model, which can accurately identify the identification data, contour size data, breakpoint position data, and recombination pattern data of abnormal chromosomes through learning a large number of historical samples. The generation of the abnormal tracking area is completed by mapping the abnormal chromosome identification data into the original image, and the marking result is presented in a visual form for subsequent analysis. The technical difficulty in this stage lies in how to design an efficient abnormality detection model so that it can accurately distinguish normal and abnormal chromosomes under complex backgrounds.

[0038] 4) After obtaining the abnormal tracking area, the system aggregates the abnormal chromosome feature data within a preset period to form an abnormal tracking area feature dataset, and calculates the abnormal risk assessment index through a preset abnormal risk assessment model. The calculation formula for the abnormal risk assessment index is: ; The physical meaning of this formula is to comprehensively consider the abnormalities in the three aspects of chromosome structure, number, and texture, and highlight the importance of different features through a dynamic weight allocation mechanism. For example, in certain specific diseases, chromosome structure abnormalities may be more diagnostically valuable than number abnormalities, and this feature can be reflected by adjusting the weight coefficient at this time.

[0039] 5) To further improve the accuracy of abnormal risk assessment, this embodiment performs similarity matching between the abnormal chromosome feature data of the abnormal tracking area and a preset chromosomal abnormality feature database to obtain multiple historical similar abnormal samples that meet the requirements and their corresponding abnormal identification coefficients. Subsequently, the system corrects the abnormal risk assessment index based on these abnormal identification coefficients to obtain an abnormal risk correction index. The correction calculation formula is: ; The core idea of this formula is to correct the current assessment result by introducing the abnormal identification coefficients of historical samples, thereby reducing the error that may be brought by a single model.

[0040] 6) At the same time, this embodiment also combines the user's clinical test dataset and genetic background feature data to calculate the genetic risk determination index of the abnormal tracking area through a preset genetic risk assessment model. The calculation of the genetic risk determination index is divided into two steps: First, calculate the genetic risk identification data of the abnormal tracking area according to the chromosome structure abnormality index, chromosome number abnormality index, and genetic disease risk index. The formula is: ; Second, combine the family genetic history data, gene mutation frequency data, and environmental exposure risk data to correct the genetic risk identification data to obtain the final genetic risk determination index. The formula is: This process fully reflects the importance of genetic background information and clinical data in the detection of chromosomal abnormalities.

[0041] 7) After obtaining the abnormal risk correction index and the genetic risk determination index, this embodiment compares the two, calculates the abnormal assessment matching degree coefficient, and verifies the abnormal assessment result through threshold comparison. The calculation formula of the abnormal assessment matching degree coefficient is: ; The design of this formula aims to comprehensively consider the two factors of genetic risk and abnormal risk, and balance their impacts through a dynamic weight allocation mechanism. Finally, the system compares the abnormal assessment matching degree coefficient with the preset abnormal identification matching degree threshold. If the matching degree coefficient is higher than the threshold, it is determined as abnormal; otherwise, it is determined as normal.

[0042] For example, assume that a user has undergone chromosomal tests in three preset time periods (n = 3). In the genetic background information: the family genetic history data (F) = 0.8 (there is a significant family history), the gene mutation frequency data (M) = 0.6 (medium to high frequency), and the environmental exposure risk data (E) = 0.4 (medium exposure). In the clinical test report: the chromosomal structure abnormality index for each preset time period ( = 0.5, = 0.7, = 0.6), the chromosomal number abnormality index ( = 0.3, = 0.4, = 0.5), the genetic disease risk index ( = 0.6, = 0.7, = 0.8). Extract the chromosomal microscopic image feature data for the three time periods: the chromosomal morphological feature data (such as contour size), the chromosomal distribution feature data (spatial arrangement), and the chromosomal texture feature chromosomal distribution feature (surface details). Assume that the chromosomal texture abnormality index ( = 0.4, = 0.5, = 0.6). The abnormal risk assessment index = 0.664. Match two historical similar abnormal samples (m = 2) from the historical database: Sample 1: = 0.8 (high risk), weight = 0.6; Sample 2: = 0.5 (medium risk), weight = 0.4; The abnormal risk correction index = 1.004. The genetic risk identification data for the abnormal tracking area is calculated to be 2.05, and the genetic risk determination index for the abnormal tracking area is calculated to be 2.71. Then the abnormal assessment matching degree coefficient is 5.476, and the preset abnormal identification matching degree threshold is defaulted to 5.0. Since 5.476 > 5.0, it is determined that the user has chromosomal abnormalities.

[0043] The technical solution of this embodiment obtains the abnormal chromosome feature data of the abnormal tracking area by acquiring the user's chromosome microscopic image set and extracting and processing the multi-modal feature data set, and obtains the abnormal risk assessment index through collective processing. Then, it is corrected by combining the abnormal identification coefficients of multiple historical similar abnormal samples to obtain the abnormal risk correction index. According to the clinical test data set corresponding to the obtained abnormal tracking area and combining the genetic background feature data, the genetic risk determination index of the abnormal tracking area is evaluated and processed. Then, it is compared with the abnormal risk correction index to obtain the abnormal evaluation matching degree coefficient, and the abnormal evaluation result is verified through threshold comparison; thus, it realizes the efficient identification and accurate evaluation of various complex chromosomal abnormality types, and combines genetic background information and clinical data for comprehensive risk assessment to improve the reliability of the test results.

[0044] Embodiment 2 On the other hand, based on Embodiment 1, the present invention provides an embodiment of a chromosomal abnormality detection system based on image analysis. The system includes: A first processing module, configured to collect multiple chromosome microscopic image sets of a user in multiple time periods within a preset period, and obtain the user's genetic background information and clinical test report; A second processing module, configured to perform multi-modal feature extraction on the chromosome microscopic image sets within each time period to obtain a chromosome microscopic image feature data set; A third processing module, configured to extract a clinical test data set according to the clinical test report, including a chromosomal structure abnormality index, a chromosomal number abnormality index, and a genetic disease risk index, and extract genetic background feature data according to the genetic background information; A fourth processing module, configured to process and identify the corresponding chromosome microscopic image feature data set of the user within each time period to obtain abnormal chromosome feature data, and mark the corresponding abnormal chromosome identification area as an abnormal tracking area; A fifth processing module, configured to collect the multiple abnormal chromosome feature data corresponding to the abnormal tracking area to obtain an abnormal tracking area feature data set, and perform processing to obtain an abnormal risk assessment index; A sixth processing module, configured to perform similarity matching on the abnormal chromosome feature data of the abnormal tracking area through a preset chromosomal abnormality feature database to obtain multiple historical similar abnormal samples that meet the requirements and their corresponding abnormal identification coefficients, and perform correction processing on the abnormal risk assessment index to obtain an abnormal risk correction index; A seventh processing module, configured to perform an evaluation process on the basis of the clinical test data set corresponding to the abnormal tracking area in each time period and the genetic background feature data through a preset genetic risk assessment model, so as to obtain a genetic risk determination index for the abnormal tracking area; An abnormality inspection module, configured to perform a comparison process on the basis of the genetic risk determination index of the abnormal tracking area and the abnormal risk correction index, so as to obtain an abnormality evaluation matching degree coefficient of the abnormal tracking area, and perform a threshold comparison with a preset abnormality identification matching degree threshold to verify the abnormality evaluation result.

[0045] In an optional manner, the second processing module is specifically configured to: Pick up image feature information through a preset multimodal feature extraction model according to the chromosome microscopic image sets in each time period, so as to obtain a chromosome microscopic image feature data set for each time period; Extract corresponding chromosome microscopic image feature data according to the chromosome microscopic image feature data set, including chromosome morphological feature data, chromosome distribution feature data, and chromosome texture feature data.

[0046] It should be noted that the beneficial effects of the chromosome abnormality inspection system based on image analysis provided in the above embodiments are the same as those of the chromosome abnormality inspection method based on image analysis, and will not be elaborated here.

[0047] Embodiment 3 This embodiment introduces new steps on the basis of Embodiment 1: such as a psychological intervention integration step and an uncertainty management mechanism, and the specific implementation is as follows: After obtaining the abnormal risk correction index, perform the following psychological intervention integration steps: Integrate the PHQ-9 and GAD-7 electronic evaluation systems to generate a dynamic questionnaire interface on the user terminal. The system automatically calculates the depression and anxiety index scores according to the user's real-time answer data, and generates a psychological state change curve with a time stamp.

[0048] Continuously collect the user's physiological behavior data through a smart wearable device (such as a smart watch), including sleep duration, activity trajectory, and eating frequency. Use an acceleration sensor to detect the frequency of limb movement at night to evaluate the sleep duration, combine the GPS positioning data with the geofencing technology to count the daily activity radius of the user, and obtain the eating frequency data through a preset meal record voice input interface. Establish a baseline value calculation model, take the moving average of the user's data for the first 7 days as the dynamic baseline, and calculate the deviation percentage in real time.

[0049] Deploy a voice interaction robot and load an interview protocol library containing 20 standard questions (e.g., Are you aware of the impact of chromosomal abnormalities on offspring?). Extract keywords (such as don't understand, fear of inheritance) through natural language processing technology and match them with the preset awareness levels (low / medium / high).

[0050] When the PHQ-9 score > 15 points and persists for 3 days, activate the cognitive behavioral therapy (CBT) digital course push system and unlock treatment units (such as cognitive restructuring training, exposure therapy simulation) according to the preset path.

[0051] If the activity radius deviates from the baseline value by > 40% and persists for 72 hours, trigger a multidisciplinary consultation order generator. Automatically generate a consultation application file containing chromosome detection data, psychological assessment reports, and behavioral monitoring charts, and send consultation requests to the genetics department, psychiatry department, and social work department simultaneously through the hospital HIS system.

[0052] The visualization engine generates a three-dimensional heat map based on real-time data. The X-axis is associated with the abnormal assessment matching coefficient (0-1 standardization), the Y-axis maps the comprehensive psychological state score (weighted value of PHQ-9 + GAD-7), and the color scale represents the intervention priority (red: intervention required within 24 hours; yellow: intervention required within 72 hours; green: routine monitoring). Highlight the abnormal area of the heat map on the doctor workstation interface and associate it with the whole genome data visualization panel of the corresponding patient.

[0053] In an optional manner, the historical similar abnormal sample matching process includes the following:[[]]END]] Case decision tree engine: Establish a structured case database. Each historical case contains chromosome breakpoint coordinates (based on the GRCh38 genome reference sequence), a treatment cycle map (marking key time nodes for chemotherapy / surgery / follow-up), and a set of prognostic parameters (5-year survival rate, complication incidence rate, etc.).

[0054] Develop a feature matching algorithm. The input layer receives the abnormal tracking area coordinates of the current patient. The hidden layer calculates the overlap degree with the breakpoints of historical cases (using the Jaccard similarity coefficient). The output layer outputs the top 5 similar cases with a matching degree > 85% and extracts their home care plans (such as nutritional supplement formulas, rehabilitation training plans).

[0055] Intelligent push system: Deploy a keyword monitoring module on the patient's mobile terminal. When it detects that the search record contains preset keywords such as genetic risk, fertility advice, etc., automatically call the case matching interface and push the disposal record video of similar cases (including medical animation explanations).

[0056] A chromosome topology comparison tool was developed, using Circos visualization technology to present the current abnormal regions and the breakpoints of historical cases in the form of concentric circles, and flashing polygon boxes were used to identify the microdeletion / microduplication regions unique to the current patient.

[0057] In an optional approach, an uncertainty management mechanism is also included: Confidence management system: A quality assessment submodule is embedded in the chromosome image analysis module to calculate the confidence index of karyotype analysis based on the chromosome band clarity (grayscale gradient value) and the number of mitotic phases (≥20 valid mitotic phases).

[0058] When the confidence level is <90%, the abnormal area is marked with an orange flashing border (frequency 2 Hz) in the electronic report, and a list of differential diagnosis recommendations is displayed in a pop-up window (such as the recommendation to perform additional FISH to verify the inter-arm inversion of chromosome 9).

[0059] Closed-loop feedback system: On the 7th day after the initial diagnosis, chromosome re-sampling instructions are automatically sent to the laboratory LIS system (including specific band type re-test requirements) and psychological re-test instructions are sent to the patient's mobile terminal (shortened version of the PHQ-4 scale). On the 30th day, the whole genome chip re-examination instructions (compared with the initial CNV changes) and structured telephone follow-up (including 10 prognostic assessment questions) are initiated.

[0060] The retest data is input into the genetic risk assessment model, and the back propagation algorithm is used to adjust the feature weights. If the psychological score deteriorates, the behavioral data weight coefficient is increased; if the chromosome abnormality area expands, the microscopic image feature weight is increased.

[0061] The above embodiments are exemplary and should not be construed as limiting the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A method for detecting chromosome abnormalities based on image analysis, characterized in that: include: Collect multiple chromosome microscopic image sets of the user at multiple time periods within a preset period, and obtain the user's genetic background information and clinical test report; Performing multimodal feature extraction on the chromosome microscopic image set in each time period to obtain a chromosome microscopic image feature data set; Extracting a clinical test data set according to the clinical test report, including a chromosome structural abnormality index, a chromosome number abnormality index, and a genetic disease risk index, and extracting genetic background feature data according to the genetic background information; Processing and identifying the chromosome microscopic image feature data set corresponding to the user in each time period to obtain abnormal chromosome feature data, and marking the corresponding abnormal chromosome identification area as an abnormal tracking area; Collecting a plurality of abnormal chromosome feature data corresponding to the abnormal tracking region to obtain an abnormal tracking region feature data set, and processing the data to obtain an abnormal risk assessment index; Performing similarity matching through a preset chromosome abnormality feature database according to the abnormal chromosome feature data of the abnormal tracking area, obtaining a plurality of historical similar abnormal samples that meet the requirements and their corresponding abnormal identification coefficients, and correcting the abnormal risk assessment index to obtain an abnormal risk correction index; According to the clinical test data set corresponding to the abnormal tracking area in each time period, combined with the genetic background feature data, an evaluation is performed through a preset genetic risk assessment model to obtain a genetic risk determination index for the abnormal tracking area; The abnormality tracking region genetic risk determination index is compared with the abnormality risk correction index to obtain the abnormality assessment matching coefficient of the abnormality tracking region, and the abnormality assessment result is verified by threshold comparison with a preset abnormality identification matching threshold.

2. The method for detecting chromosome abnormality based on image analysis according to claim 1, characterized in that: The multimodal feature extraction is performed on the chromosome microscopic image set in each time period to obtain a chromosome microscopic image feature data set, including: According to the chromosome microscopic image set in each time period, image feature information is picked up by a preset multimodal feature extraction model to obtain a chromosome microscopic image feature data set in each time period; Corresponding chromosome microscopic image feature data are extracted according to the chromosome microscopic image feature data set, including chromosome morphology feature data, chromosome distribution feature data and chromosome texture feature data.

3. The method for detecting chromosome abnormality based on image analysis according to claim 1, characterized in that: The extracting of clinical test data sets according to the clinical test report, including a chromosome structural abnormality index, a chromosome number abnormality index, and a genetic disease risk index, and extracting genetic background feature data according to the genetic background information, include: Extracting a clinical test data set according to the clinical test report, including a chromosome structural abnormality index, a chromosome numerical abnormality index, and a genetic disease risk index in each time period; Genetic background characteristic data are extracted based on the genetic background information, including family genetic disease history data, gene mutation frequency data and environmental exposure risk data.

4. The method for detecting chromosome abnormality based on image analysis according to claim 1, characterized in that: The processing and identifying of the chromosome microscopic image feature data set corresponding to the user in each time period to obtain abnormal chromosome feature data, and marking the corresponding abnormal chromosome identification area as an abnormal tracking area, includes: Processing and identifying the chromosome microscopic image feature data set corresponding to the user in each time period through a preset chromosome abnormality detection model to obtain abnormal chromosome feature data; The abnormal chromosome characteristic data includes abnormal chromosome identification data, abnormal chromosome outline size data, abnormal chromosome breakpoint location data and abnormal chromosome recombination pattern data; The abnormal chromosome identification area corresponding to the identification is obtained according to the abnormal chromosome identification data, and marked as an abnormal tracking area.

5. The method for detecting chromosome abnormality based on image analysis according to claim 3, characterized in that: The step of collecting a plurality of abnormal chromosome feature data corresponding to the abnormal tracking region to obtain an abnormal tracking region feature data set, and processing the data to obtain an abnormal risk assessment index includes: Collecting a plurality of abnormal chromosome feature data corresponding to the abnormal tracking area within the preset period to obtain an abnormal tracking area feature data set; The abnormal tracking area feature data set is processed by a preset abnormal risk assessment model to obtain an abnormal risk assessment index of the abnormal tracking area; the calculation formula of the abnormal risk assessment index is: ;in, is the abnormal risk assessment index, For the The chromosome structural abnormality index within a preset time period, For the The chromosome number abnormality index within a preset time period, For the The chromosome texture abnormality index within a preset time period, Preset risk factors for abnormal areas, is the number of time periods within the preset period, , and is the first preset weight coefficient.

6. The method for detecting chromosome abnormality based on image analysis according to claim 5, characterized in that: The abnormal chromosome feature data of the abnormal tracking area is matched similarly through a preset chromosome abnormal feature database to obtain a plurality of historical similar abnormal samples that meet the requirements and their corresponding abnormal identification coefficients, and the abnormal risk assessment index is corrected to obtain an abnormal risk correction index, including: Performing similarity matching processing based on the abnormal chromosome feature data of the abnormal tracking area through a preset chromosome abnormal feature database to obtain multiple historical similar abnormal samples that meet similarity requirements; Extracting corresponding abnormality identification coefficients for final sample inspection according to the plurality of historical similar abnormal samples; The abnormal risk assessment index is corrected according to the plurality of abnormal identification coefficients to obtain an abnormal risk correction index; the correction calculation formula of the abnormal risk correction index is: ;in, is the abnormal risk correction index, For the The anomaly identification coefficient of historically similar anomaly samples, is the number of historical similar abnormal samples, For the The preset weight coefficient of historically similar abnormal samples.

7. The method for detecting chromosome abnormality based on image analysis according to claim 6, characterized in that: The step of evaluating and processing the clinical test data set corresponding to the abnormal tracking area in each time period in combination with the genetic background feature data through a preset genetic risk assessment model to obtain a genetic risk determination index for the abnormal tracking area includes: According to the chromosome structure abnormality index, chromosome number abnormality index and genetic disease risk index corresponding to the abnormal tracking area in each time period, an evaluation is performed through a preset genetic risk evaluation model to obtain genetic risk identification data of the abnormal tracking area; Perform correction processing based on the abnormal tracking area genetic risk identification data and the family genetic disease history data, gene mutation frequency data and environmental exposure risk data to obtain the abnormal tracking area genetic risk determination index of the user within the preset period; The calculation formula for the abnormal tracking area genetic risk identification data is: ;in, To identify genetic risk data for abnormal tracking regions, For the Genetic disease risk index within a preset time period, , , is a second preset weight coefficient; The modified calculation formula for the abnormal tracking region genetic risk determination index is: ;in, To determine the genetic risk index for abnormal tracking regions, For family genetic disease history data, is the gene mutation frequency data, For environmental exposure risk data, , , is the third preset weight coefficient.

8. The method for detecting chromosome abnormality based on image analysis according to claim 7, characterized in that: The step of comparing the abnormality tracking region genetic risk determination index with the abnormality risk correction index to obtain the abnormality assessment matching coefficient of the abnormality tracking region, and performing a threshold comparison with a preset abnormality identification matching threshold to verify the abnormality assessment result includes: Compare the abnormality tracking region genetic risk determination index of the user with the abnormality risk correction index to obtain an abnormality assessment matching coefficient of the abnormality tracking region; Performing a threshold comparison between the anomaly assessment matching coefficient and a preset anomaly identification matching threshold, and verifying the anomaly assessment result according to the threshold comparison result; The calculation formula of the abnormality assessment matching coefficient is: ;in, is the abnormal assessment matching coefficient, To determine the genetic risk index for abnormal tracking regions, is the abnormal risk correction index, and is the fourth preset weight coefficient.

9. The method for detecting chromosome abnormality based on image analysis according to claim 1, characterized in that: After obtaining the abnormal risk modification index, perform the following psychological intervention integration steps: a) Initiate a three-level psychological assessment mechanism: The PHQ-9 and GAD-7 standardized scales were used for electronic assessment to generate psychological status scores; Monitor the user's physiological behavior indicator data set and calculate the deviation from the preset baseline value; The structured interview protocol was loaded, and the key words were recorded through voice interaction to match the diagnostic awareness level; b) Build a dynamic intervention matrix: When the psychological state score exceeds the first threshold, the cognitive behavioral therapy execution path is activated When the behavior deviation exceeds the second threshold for three consecutive days, a multidisciplinary consultation instruction is generated. Generate a three-dimensional heat map visualization interface, with the horizontal axis correlating the abnormal assessment matching coefficient, the vertical axis mapping the psychological state score, and the color scale representing the intervention priority.

10. The method for detecting chromosome abnormality based on image analysis according to claim 1, characterized in that: The historical similar abnormal sample matching process includes: Establish a case decision tree engine to associate historical similar case libraries based on diagnostic codes, extract treatment cycle and prognosis timeline maps, family care program knowledge base, and genetic risk assessment correction parameters; When it is detected that the user's search behavior contains preset keywords, similar case disposal records with a matching degree higher than 85% are automatically pushed; The topological differences of chromosome breakpoints between the current abnormal tracking region and historical samples are highlighted in the visualization interface; It also includes uncertainty management mechanisms: a) When the confidence level of chromosome image analysis is less than 90%: Mark the abnormal tracking area with a dynamic flashing review mark; The clinical manifestation database is linked to generate a pop-up window containing at least 3 differential diagnosis suggestions; b) Establish a closed-loop feedback system: Automatically initiated on the 7th and 30th day after the abnormal assessment result is generated: 1) Chromosome microscopic image re-collection instructions 2) Mental status score retest instructions The weighting parameters of the preset genetic risk assessment model are dynamically adjusted according to the deviation degree between the retest data and the baseline value.

Images