Parathyroid blood flow status analysis system and method based on photoacoustic imaging
By real-time image preprocessing and feature extraction of parathyroid samples, combined with BP neural network model, the parathyroid blood circulation status is automatically analyzed, which solves the problem that the analysis results in the prior art are susceptible to human factors, and improves the analysis Accuracy and consistency.
Patent Information
- Application Number
- CN202411052536.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-02
AI Technical Summary
The existing parathyroid blood vessel status analysis system requires professional equipment and technicians to operate, and the results are susceptible to human factors, resulting in inaccurate analysis results.
By collecting real-time images of parathyroid samples for preprocessing and feature extraction, a BP neural network model is constructed for feature data classification and prediction, and the blood vessel status level is automatically analyzed to reduce human interference.
It improves the accuracy and consistency of parathyroid blood circulation status analysis, reduces the influence of human factors, and realizes an automated analysis process.
Smart Images

Figure CN118587209B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of image analysis, and in particular, relates to a parathyroid blood flow status analysis system and method based on photoacoustic imaging. Background Art
[0002] Photoacoustic imaging is a non-invasive and non-ionizing biomedical imaging method. Its principle is to use pulsed laser to irradiate biological tissue. When the laser is absorbed by the light absorption domain in the tissue, part of the light energy is converted into heat energy, causing thermoelastic expansion of the surrounding tissue, thereby generating broadband (megahertz level) ultrasonic emission; these ultrasonic waves are then detected by the ultrasonic transducer, and by detecting these photoacoustic signals, the light absorption distribution image in the tissue can be reconstructed. Photoacoustic imaging combines the high selectivity of pure optical imaging and the deep penetration of pure ultrasonic imaging. It can provide high-resolution and high-contrast tissue images, breaking through the depth limitation of high-resolution optical imaging and achieving high-quality imaging of deep in vivo tissues.
[0003] Multispectral photoacoustic tomography combined with ultrasound techniques has been used to noninvasively characterize a range of thyroid diseases;
[0004] Currently, the systems for analyzing the blood supply status of the parathyroid glands include a parathyroid function imaging system and endoscope proposed in the patent document with application number 202121407683.4; this system uses near-infrared fluorescence technology in the endoscope to locate the parathyroid glands, and laser speckle technology to display the speckle contrast image of the parathyroid glands, thereby understanding whether the blood supply is lost and whether transplantation is needed, to guide the diagnosis of parathyroid function during thyroid surgery.
[0005] Chinese patent application CN117710325A discloses a method for automatic identification and tracking of parathyroid glands during endoscopic thyroid surgery. The method obtains the original image during endoscopic thyroid surgery, extracts its contour and adjusts its pixel value to obtain a labeled enhanced image. The image and the original image are then used to train a pre-built recognition model. Finally, the trained model is used to identify and track the parathyroid glands in real-time endoscopic video. The invention can realize automatic identification and real-time tracking of parathyroid glands, which helps reduce the risk of parathyroid gland damage.
[0006] At present, the analysis of the parathyroid blood circulation status requires professional equipment and technicians to operate, and has high requirements for equipment and operators. The experience and technical level of different operators may affect the results, making the analysis results different due to various human factors. Such analysis results cannot determine whether the parathyroid blood circulation status is normal. Summary of the invention
[0007] In view of the problems in the related art, the present invention proposes a parathyroid blood circulation status analysis system and method to overcome the above-mentioned technical problems existing in the existing related art.
[0008] In order to solve the above technical problems, the present invention is achieved through the following technical solutions:
[0009] The present invention is a method for analyzing the blood supply status of the parathyroid gland, comprising the following steps:
[0010] S1, collecting real-time images of multiple parathyroid samples and preprocessing them to obtain a final parathyroid imaging set; and testing the blood supply status levels of the multiple parathyroid samples to obtain a first blood supply status level data set;
[0011] S2, extracting characteristic data of the final parathyroid imaging set to obtain a parathyroid characteristic data matrix;
[0012] S3, performing iterative optimization classification on each feature data in the parathyroid feature data matrix to obtain an optimal classification center data matrix;
[0013] S4, collecting historical images of the plurality of parathyroid samples and performing preprocessing, and then performing feature extraction on the historical images to obtain a historical parathyroid feature data matrix set;
[0014] Using the historical parathyroid gland feature data matrix to train and test the initial BP neural network model to obtain a final BP neural network model;
[0015] S5, collecting the real-time imaging feature data and the historical imaging feature data of the parathyroid gland to be analyzed, and obtaining a real-time parathyroid feature data set and a historical parathyroid feature data matrix to be analyzed;
[0016] Inputting the historical parathyroid gland feature data matrix to be analyzed into the real-time final BP neural network model for prediction to obtain the future parathyroid gland feature data matrix to be analyzed;
[0017] S6. Calculate the Euclidean distance between each feature data in the future feature data matrix of the parathyroid gland to be analyzed and the real-time parathyroid feature data set and each center data in the optimal classification center data matrix, and combine with the first blood circulation status level data set to obtain the blood circulation status level data of the parathyroid gland to be analyzed.
[0018] First, images of multiple parathyroid samples are collected, and the corresponding blood circulation status levels are measured, thereby forming training samples and training sample data labels; data support is provided for the subsequent acquisition of classification center data of characteristic data of parathyroid imaging; the parathyroid imaging is preprocessed to make the subsequent extraction of characteristic data of parathyroid imaging more convenient; by extracting characteristic data related to the blood circulation status in the parathyroid imaging, it is convenient to find the correlation between the characteristic data and the blood circulation status level data through classification in the subsequent process, so the characteristic data of parathyroid imaging is collected for analysis, and more influencing factors are considered in the analysis process, so that the classification More accurate; since the blood circulation status of the parathyroid gland in the future is also an important factor showing the current blood circulation status of the parathyroid gland, a BP neural network model is constructed to perform prediction operations on the characteristic data of the parathyroid gland to be analyzed at the future moment, thereby obtaining the characteristic data of the parathyroid gland to be analyzed at the future moment, and then classifying the characteristic data at the future moment, thereby determining whether there are hidden risks in the parathyroid gland to be analyzed; taking corresponding measures for prevention; finally, comparing and analyzing the real-time characteristic data of the parathyroid gland to be analyzed with the future characteristic data of the parathyroid gland to be analyzed, obtaining the blood circulation status level data of the parathyroid gland to be analyzed.
[0019] Preferably, the S1 comprises the following steps:
[0020] S11. Collect multiple parathyroid samples to obtain a parathyroid sample set ; Indicates the collected parathyroid samples, represents the total number of parathyroid samples collected;
[0021] Reset the parathyroid blood flow status level set , Indicates the setting The parathyroid blood supply status level, Indicates the total number of set parathyroid blood supply status levels;
[0022] Parathyroid blood flow status level set Analysis and parathyroid sample set The blood flow status level of each parathyroid sample in the first blood flow status level data set is obtained. , Indicates The blood status grade data of each parathyroid sample;
[0023] Preferably, in S11, the parathyroid blood flow status level set Analysis and parathyroid sample set The vascularity level of each parathyroid gland sample was analyzed using laser speckle technique;
[0024] S12, collecting the parathyroid sample set The imaging images of each parathyroid sample in the sample are obtained to obtain the initial parathyroid imaging set , Indicates Image of parathyroid gland samples;
[0025] Preferably, the method of collecting the imaging images of each parathyroid sample in the parathyroid sample set includes autofluorescence imaging technology, photoacoustic imaging technology, or laser speckle technology.
[0026] Of course, other existing technologies can also be used to obtain the imaging images of each parathyroid sample.
[0027] S13, the initial parathyroid imaging set The parathyroid images in the image collection are preprocessed to obtain the final parathyroid image collection. ;
[0028] Laser speckle technology determines whether the blood supply status of the parathyroid gland is normal by measuring the size of the speckle contrast in the regional speckle contrast image; therefore, each level in the parathyroid blood supply status level set is divided according to the size of the speckle contrast in the regional speckle contrast image, so the normality of the parathyroid blood supply status can be determined by the parathyroid blood supply status level.
[0029] Preferably, the S13 comprises the following steps:
[0030] S131, the initial parathyroid imaging set The denoising operation is performed on each parathyroid image in the image to obtain the denoised parathyroid image set , Express The parathyroid image obtained after denoising is specifically comprised of the following steps:
[0031] S1311, the initial parathyroid imaging set Traverse each parathyroid image in, perform morphological opening operation and morphological closing operation on each parathyroid image in turn, and obtain a parathyroid image set after opening operation and a parathyroid image set after closing operation;
[0032] S1312, setting noise threshold , according to the noise threshold , the parathyroid imaging set after opening operation and the parathyroid imaging set after closing operation set noise detection formula ;as follows,
[0033] ;
[0034] In the formula, and They respectively represent the pixel abscissa and pixel ordinate of each parathyroid image in the parathyroid image set after the closing operation; represents the number of parathyroid images on each parathyroid image set after the opening operation. The pixel point corresponds to the initial parathyroid image set on the parathyroid image The difference between pixels; represents the number of parathyroid images on each parathyroid image set after the closing operation. The pixel point corresponds to the initial parathyroid image set on the parathyroid image The difference between pixels;
[0035] when When the value of is 1 or -1, the parathyroid imaging set after the closed operation contains the parathyroid imaging The pixel point is a noise point, and a noise point set is obtained; the parathyroid imaging in the parathyroid imaging set after the closing operation is Pixels are not noise points;
[0036] S1313, denoising is performed on each noise point in the noise point set, and the denoising formula is as follows:
[0037] ;
[0038] In the formula, represents a noise point of the noise point set; and Respectively express Perform one-step morphological calculation and two-step morphological calculation; and All represent square structural elements; represents the closing operation; represents the black point operator, which is used to highlight the dark areas in the image or the parts that do not match the structural elements;
[0039] right and Find the mean ; So far, the denoising process of each noise point in the noise point set is completed, and the denoised parathyroid imaging set is obtained. ;
[0040] S132, the parathyroid imaging set after denoising The contrast of each parathyroid image in the image is enhanced to obtain the final parathyroid image set. , Express Parathyroid imaging obtained after contrast-enhanced procedures;
[0041] Preferably, the contrast enhancement operation in S132 adopts an ACE algorithm;
[0042] By denoising and enhancing the contrast of the parathyroid images, the features in the parathyroid images are made clearer, thus facilitating subsequent extraction.
[0043] Preferably, S2 comprises the following steps:
[0044] S21. Set parathyroid imaging feature type set , Indicates the setting Parathyroid imaging features, Indicates the total number of parathyroid imaging feature types that are set;
[0045] According to the parathyroid imaging feature type set Final parathyroid imaging set Each parathyroid image in is subjected to feature extraction to obtain the parathyroid feature data matrix ;as follows,
[0046] ;
[0047] in, Final parathyroid imaging set Middle Parathyroid scintigraphy feature data.
[0048] Preferably, S3 comprises the following steps:
[0049] S31. Establishing the first male mayfly population and the first female mayfly population , and Respectively represent the first male mayfly population The first male mayfly and the first female mayfly population A female mayfly, and Respectively represent the size of the first male mayfly population and the first female mayfly population; then set the maximum number of iterations of the first male mayfly population and the first female mayfly population to , the current number of iterations is And the dimension of the search space is ;
[0050] S32, according to the parathyroid characteristic data matrix The initial position of each male mayfly in the first male mayfly population and the initial position of each female mayfly in the first female mayfly population are set; specifically, the following steps are included:
[0051] S321, from the parathyroid gland characteristic data matrix Random selection The row data is used as the initial position matrix of the first male mayfly in the first male mayfly population; and then the parathyroid characteristic data matrix Random selection The row data is used as the initial position matrix of the first female mayfly in the first female mayfly population;
[0052] S322, repeat S321 until the initial positions of all male mayflies in the first male mayfly population and the initial positions of all female mayflies in the first female mayfly population are initialized, and the first male mayfly initial position matrix set is obtained. and the initial position matrix set of the first female mayfly , and Respectively represent the first male mayfly population The initial position matrix of the male mayflies and the first female mayfly population The initial position matrix of the female mayflies is as follows:
[0053] ;
[0054] ;
[0055] in, and Respectively Representative The first The position component of the feature data and Representative The first The position component of the feature data;
[0056] S33, calculating the parathyroid characteristic data matrix The Euclidean distance matrix is obtained by calculating the Euclidean distance between each row of data in the initial position of each mayfly. ;as follows,
[0057] ;
[0058] in, Represents the parathyroid feature data matrix Middle Row data to the initial position of each mayfly Euclidean distance of row data;
[0059] Select the Euclidean distance matrix The parathyroid feature data matrix corresponding to the smallest Euclidean distance data in each row The data in is classified into the data of the corresponding row in the initial position of each mayfly, and the initial parathyroid feature data classification matrix set is obtained. , Indicates The initial parathyroid feature data classification matrix is as follows:
[0060] ;
[0061] in, Indicates The first The parathyroid gland feature data, Indicates The total number of parathyroid feature data in the classification of initial parathyroid feature data; ;
[0062] Set the fitness function of the first male mayfly population and the first female mayfly population ;as follows,
[0063] ;
[0064] In the formula, Indicates The total number of parathyroid feature data in the initial parathyroid feature data classification, Indicates The first of the initial parathyroid feature data classification The Euclidean matrix between the characteristic data set of each parathyroid gland and the corresponding row data of the mayfly;
[0065] S34, start iterating, and in each round of iteration, according to the fitness function Calculating the fitness values of each mayfly in the first male mayfly population and the first female mayfly population; selecting the position of the mayfly individual with the highest fitness in the first male mayfly population and the first female mayfly population, and recording them as the first best fitness, the second best fitness, the first best position, and the second best position, respectively; updating the position and speed of each male mayfly and the female mayfly according to the first best fitness, the second best fitness, the first best position, and the second best position; and mating the male mayfly individual with the highest fitness in the first male mayfly population with a random female mayfly individual in the first female mayfly population;
[0066] S35, when When , stop the iteration and get the first optimal position matrix And the first optimal classification matrix set , Represents the first optimal classification matrix set parathyroid feature data classification matrix; and They are as follows:
[0067] ; ;
[0068] in, Indicates the optimal position matrix corresponding to The first in the classification center dataset of parathyroid feature data classification feature data, Indicates The optimal classification matrix The parathyroid gland feature data, Indicates The total number of parathyroid feature data in the optimal classification; ;
[0069] S36. Based on the first blood supply status level data set Determine the first optimal classification matrix set Whether the blood flow status level data corresponding to the parathyroid feature data sets in each optimal classification matrix are the same;
[0070] If they are the same, the first optimal position matrix As the optimal classification center data matrix Otherwise, continue the iterative operation in S34 until the first optimal classification matrix set The parathyroid gland feature data sets in each optimal classification matrix correspond to the same blood flow status level data, and the optimal classification center data matrix is obtained. ;
[0071] The mayfly optimization algorithm has the characteristics of fast convergence speed and strong global search ability, and has a wide range of applications in many fields. By constructing a mayfly population, the classification centers of the parathyroid feature data matrix are continuously iterated and optimized. Since the classification center data of the initial feature data are randomly selected, the data in the optimal solution space of the mayfly population is richer, making the optimization efficiency higher. Ultimately, the blood circulation status data corresponding to each feature data assigned to each classification are the same, thereby obtaining a set of optimal classification center data, which provides a basis for the subsequent classification of the parathyroid feature data to be analyzed.
[0072] Preferably, S4 comprises the following steps:
[0073] S41. Setting a historical time point set ; Indicates A historical point in time, Indicates the total number of set historical time points;
[0074] S42: Collecting the historical time point set Final parathyroid imaging set at each time point The characteristic data of each parathyroid imaging in the parathyroid gland is obtained to obtain the historical parathyroid gland comprehensive data matrix set , Indicates A comprehensive parathyroid data matrix at each historical time point;
[0075] S43, constructing an initial BP neural network model; using the historical parathyroid feature data matrix set Train and test the initial BP neural network model to obtain the final BP neural network model;
[0076] By collecting historical characteristic data of the parathyroid glands to be analyzed, the initial BP neural network model is trained and tested, so that the BP neural network model has a better prediction effect on the characteristic data of the parathyroid glands, and provides model support for the subsequent prediction of the characteristic data of the parathyroid glands to be analyzed at future moments.
[0077] Preferably, the S43 comprises the following steps:
[0078] S431, constructing an initial BP neural network model and setting a training data ratio, a training error threshold, and a test accuracy threshold;
[0079] S432: According to the training data ratio, the historical parathyroid feature data matrix set is Perform data division to obtain a historical parathyroid comprehensive training data matrix set and a historical parathyroid comprehensive test data matrix set;
[0080] S433, inputting the historical parathyroid comprehensive training data matrix set into the initial BP neural network model for training; when the training error during the training process is less than the training error threshold, stopping the training to obtain a trained BP neural network model;
[0081] S434, input the historical parathyroid comprehensive test data matrix set into the trained BP neural network model for testing to obtain a test accuracy rate; when the test accuracy rate is greater than or equal to a test accuracy rate threshold, use the trained BP neural network model as the final BP neural network model; otherwise, return to S433 to continue training the trained BP neural network model until the test accuracy rate is greater than or equal to the test accuracy rate threshold, to obtain a final BP neural network model;
[0082] By dividing the historical parathyroid feature data matrix set, setting the training error threshold and the test accuracy threshold, the effect of the subsequent training and testing process is determined.
[0083] Preferably, S5 comprises the following steps:
[0084] S51, collect the parathyroid image to be analyzed in real time, and then use S13 to denoise and extract features to obtain a real-time parathyroid feature data set , Indicates Real-time characteristic data of parathyroid glands to be analyzed;
[0085] S52: Based on the historical time point set Photoacoustic imaging equipment is used to collect images of the parathyroid glands to be analyzed, and then S13 method is used to denoise and extract features to obtain the historical parathyroid feature data matrix to be analyzed. ;as follows,
[0086] ;
[0087] in, Indicates The parathyroid gland to be analyzed at the historical time point Characteristic data;
[0088] S53. Setting future time point set , Indicates A future time point, Indicates the total number of set future time points;
[0089] Based on the future time point The historical parathyroid characteristic data matrix to be analyzed Input into the final BP neural network model for prediction, and obtain the future characteristic data matrix of the parathyroid gland to be analyzed ;as follows,
[0090] ;
[0091] in, Indicates The first parathyroid gland to be analyzed at a future time point Characteristic data;
[0092] By collecting the real-time data of the parathyroid gland to be analyzed, the real-time blood circulation of the parathyroid gland to be analyzed is analyzed, and once an abnormal situation occurs, it can be discovered in time; by using the final BP neural network model to predict the future characteristic data of the parathyroid gland to be analyzed, and then classifying the predicted future characteristic data, once an abnormal situation occurs, relevant measures can be taken in time to prevent it.
[0093] Preferably, the S6 comprises the following steps:
[0094] S61. Calculate real-time parathyroid feature dataset To the optimal classification center data matrix The Euclidean distance of each row of data in the first Euclidean distance set is obtained; the optimal classification center data matrix corresponding to the smallest Euclidean distance in the first Euclidean distance set is obtained. The blood supply status level data of the row data is used as the blood supply status level data of the real-time parathyroid feature data set, and is recorded as the first blood supply status level data;
[0095] S62. Calculate the future characteristic data matrix of the parathyroid gland to be analyzed Each row of data in the optimal classification center data matrix The Euclidean distance of each row of data in the second Euclidean distance matrix is obtained; the optimal classification center data matrix corresponding to the smallest Euclidean distance in each row of the second Euclidean distance matrix is obtained. The blood flow status level data of the row data is used as the future characteristic data matrix of the parathyroid gland to be analyzed The blood circulation status level data of the feature data set corresponding to the row in the image is obtained to obtain a second blood circulation status level data set;
[0096] S63, merging the second blood supply status level data set with the first blood supply status level data set to obtain a third blood supply status level data set; using the data with the lowest blood supply status level in the third blood supply status level data set as the blood supply status level data of the parathyroid gland to be analyzed;
[0097] Finally, the real-time characteristic data of the parathyroid gland to be analyzed and the future characteristic data of the parathyroid gland to be analyzed are compared and analyzed to make the analysis more comprehensive and accurate.
[0098] A parathyroid blood circulation status analysis system includes a parathyroid real-time imaging data acquisition module, a preprocessing module, a feature extraction module, an optimization classification module, a parathyroid data acquisition module to be analyzed, a prediction module and a final blood circulation status determination module.
[0099] Preferably, the parathyroid real-time imaging data acquisition module adopts a photoacoustic imaging device of the prior art, specifically a parathyroid blood circulation status analysis system based on photoacoustic imaging.
[0100] Of course, the parathyroid real-time imaging data acquisition module may also be other prior art autofluorescence imaging devices or laser speckle imaging devices. Other prior art methods for obtaining imaging images of each parathyroid sample are also applicable to the present invention.
[0101] The present invention has the following beneficial effects:
[0102] 1. The present invention collects characteristic data of parathyroid imaging for analysis, so that classification is more accurate; by classifying characteristic data at future moments, it is determined whether the parathyroid to be analyzed has hidden risks; corresponding measures are taken for prevention; finally, the real-time characteristic data of the parathyroid to be analyzed is compared and analyzed with the future characteristic data of the parathyroid to be analyzed, so as to obtain the blood supply status level data of the parathyroid to be analyzed; by adopting the technical solution of the present invention, automatic image feature recognition and fixed standards are used to classify and analyze the characteristic data of the parathyroid, so that the analysis process is more fixed, and thus the analysis result is more accurate.
[0103] 2. In the present invention, the classification centers of the parathyroid characteristic data matrix are continuously iterated and optimized by constructing the mayfly population. Since the classification center data of the initial characteristic data are randomly selected, the optimization efficiency is higher, and a set of optimal classification center data is obtained, thereby providing a judgment basis for the subsequent classification of the parathyroid characteristic data to be analyzed.
[0104] 3. In the present invention, the initial BP neural network model is trained and tested by collecting historical characteristic data of the parathyroid gland to be analyzed, so that the BP neural network model has a better prediction effect on the characteristic data of the parathyroid gland, and provides model support for the subsequent prediction of the characteristic data of the parathyroid gland to be analyzed at a future time.
[0105] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS
[0106] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.
[0107] Figure 1 It is a schematic diagram of the process of analyzing the parathyroid blood supply status by the parathyroid blood supply status analysis system of the present invention;
[0108] Figure 2 It is a schematic diagram of the prediction process of future characteristic data of the parathyroid gland to be analyzed in the parathyroid gland blood supply status analysis system of the present invention. DETAILED DESCRIPTION
[0109] The following will be combined with the drawings in the embodiments of the invention to clearly and completely describe the technical solutions in the embodiments of the invention. Obviously, the described embodiments are only part of the embodiments of the invention, not all of the embodiments. Based on the embodiments in the invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the invention.
[0110] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.
[0111] Example 1
[0112] This embodiment is a method for analyzing the blood supply status of the parathyroid gland, comprising the following steps:
[0113] S1, collecting real-time images of multiple parathyroid samples and preprocessing them to obtain a final parathyroid imaging set; and testing the blood supply status levels of the multiple parathyroid samples to obtain a first blood supply status level data set;
[0114] The S1 comprises the following steps:
[0115] S11. Collect multiple parathyroid samples to obtain a parathyroid sample set ; Indicates the collected parathyroid samples, represents the total number of parathyroid samples collected;
[0116] Reset the parathyroid blood flow status level set , Indicates the setting The parathyroid blood supply status level, Indicates the total number of set parathyroid blood supply status levels;
[0117] Parathyroid blood flow status level set Analysis and parathyroid sample set The blood flow status level of each parathyroid sample in the first blood flow status level data set is obtained. , Indicates The blood status grade data of each parathyroid sample;
[0118] Preferably, in S11, the parathyroid blood flow status level set Analysis and parathyroid sample set The vascularity level of each parathyroid gland sample was analyzed using laser speckle technique;
[0119] S12, collecting the parathyroid sample set The imaging images of each parathyroid sample in the sample are obtained to obtain the initial parathyroid imaging set , Indicates Image of parathyroid gland samples;
[0120] S13, the initial parathyroid imaging set The parathyroid images in the image collection are preprocessed to obtain the final parathyroid image collection. ;
[0121] The S13 comprises the following steps:
[0122] S131, the initial parathyroid imaging set The denoising operation is performed on each parathyroid image in the image to obtain the denoised parathyroid image set , Express The parathyroid image obtained after denoising is specifically comprised of the following steps:
[0123] S1311, the initial parathyroid imaging set Traverse each parathyroid image in, perform morphological opening operation and morphological closing operation on each parathyroid image in turn, and obtain a parathyroid image set after opening operation and a parathyroid image set after closing operation;
[0124] S1312, setting noise threshold , according to the noise threshold , the parathyroid imaging set after opening operation and the parathyroid imaging set after closing operation set noise detection formula ;as follows,
[0125] ;
[0126] In the formula, and They respectively represent the pixel abscissa and pixel ordinate of each parathyroid image in the parathyroid image set after the closing operation; represents the number of parathyroid images on each parathyroid image set after the opening operation. The pixel point corresponds to the initial parathyroid image set on the parathyroid image The difference between pixels;
[0127] represents the number of parathyroid images on each parathyroid image set after the closing operation. The pixel point corresponds to the initial parathyroid image set on the parathyroid image The difference between pixels;
[0128] when When the value of is 1 or -1, the parathyroid imaging set after the closed operation contains the parathyroid imaging The pixel point is a noise point, and a noise point set is obtained; the parathyroid imaging in the parathyroid imaging set after the closing operation is Pixels are not noise points;
[0129] S1313, denoising is performed on each noise point in the noise point set, and the denoising formula is as follows:
[0130] ; ;
[0131] In the formula, represents a noise point of the noise point set; and Respectively express Perform one-step morphological calculation and two-step morphological calculation; and All represent square structural elements; represents the closing operation; represents the black point operator, which is used to highlight the dark areas in the image or the parts that do not match the structural elements;
[0132] right and Find the mean ; So far, the denoising process of each noise point in the noise point set is completed, and the denoised parathyroid imaging set is obtained. ;
[0133] S132, the parathyroid imaging set after denoising The contrast of each parathyroid image in the image is enhanced to obtain the final parathyroid image set. , Express Parathyroid imaging obtained after contrast-enhanced procedures;
[0134] S2, extracting characteristic data of the final parathyroid imaging set to obtain a parathyroid characteristic data matrix;
[0135] The S2 comprises the following steps:
[0136] S21. Set parathyroid imaging feature type set , Indicates the setting Parathyroid imaging features, Indicates the total number of parathyroid imaging feature types that are set;
[0137] According to the parathyroid imaging feature type set Final parathyroid imaging set Each parathyroid image in is subjected to feature extraction to obtain the parathyroid feature data matrix ;as follows,
[0138] ;
[0139] in, Final parathyroid imaging set Middle Parathyroid scintigraphy Characteristic data;
[0140] S3, performing iterative optimization classification on each feature data in the parathyroid feature data matrix to obtain an optimal classification center data matrix;
[0141] The S3 comprises the following steps:
[0142] S31. Establishing the first male mayfly population and the first female mayfly population , and Respectively represent the first male mayfly population The first male mayfly and the first female mayfly population A female mayfly, and Respectively represent the size of the first male mayfly population and the first female mayfly population; then set the maximum number of iterations of the first male mayfly population and the first female mayfly population to , the current number of iterations is And the dimension of the search space is ;
[0143] S32, according to the parathyroid characteristic data matrix The initial position of each male mayfly in the first male mayfly population and the initial position of each female mayfly in the first female mayfly population are set; specifically, the following steps are included:
[0144] S321, from the parathyroid gland characteristic data matrix Random selection The row data is used as the initial position matrix of the first male mayfly in the first male mayfly population; and then the parathyroid characteristic data matrix Random selection The row data is used as the initial position matrix of the first female mayfly in the first female mayfly population;
[0145] S322, repeat S321 until the initial positions of all male mayflies in the first male mayfly population and the initial positions of all female mayflies in the first female mayfly population are initialized, and the first male mayfly initial position matrix set is obtained. and the initial position matrix set of the first female mayfly , and Respectively represent the first male mayfly population The initial position matrix of the male mayflies and the first female mayfly population The initial position matrix of the female mayflies is as follows:
[0146] ;
[0147] ;
[0148] in, and Respectively Representative The first The position component of the feature data and Representative The first The position component of the feature data;
[0149] S33, calculating the parathyroid characteristic data matrix The Euclidean distance matrix is obtained by calculating the Euclidean distance between each row of data in the initial position of each mayfly. ;as follows,
[0150] ;
[0151] in, Represents the parathyroid feature data matrix Middle Row data to the initial position of each mayfly Euclidean distance of row data;
[0152] Select the Euclidean distance matrix The parathyroid feature data matrix corresponding to the smallest Euclidean distance data in each row The data in is classified into the data of the corresponding row in the initial position of each mayfly, and the initial parathyroid feature data classification matrix set is obtained. , Indicates The initial parathyroid feature data classification matrix is as follows:
[0153] ;
[0154] in, Indicates The first The parathyroid gland feature data, Indicates The total number of parathyroid feature data in the classification of initial parathyroid feature data; ;
[0155] Set the fitness function of the first male mayfly population and the first female mayfly population ;as follows,
[0156] ;
[0157] In the formula, Indicates The total number of parathyroid feature data in the initial parathyroid feature data classification, Indicates The first of the initial parathyroid feature data classification The Euclidean matrix between the characteristic data set of each parathyroid gland and the corresponding row data of the mayfly;
[0158] S34, start iterating, and in each round of iteration, according to the fitness function Calculating the fitness values of each mayfly in the first male mayfly population and the first female mayfly population; selecting the position of the mayfly individual with the highest fitness in the first male mayfly population and the first female mayfly population, and recording them as the first best fitness, the second best fitness, the first best position, and the second best position, respectively; updating the position and speed of each male mayfly and the female mayfly according to the first best fitness, the second best fitness, the first best position, and the second best position; and mating the male mayfly individual with the highest fitness in the first male mayfly population with a random female mayfly individual in the first female mayfly population;
[0159] S35, when When , stop the iteration and get the first optimal position matrix And the first optimal classification matrix set , Represents the first optimal classification matrix set parathyroid feature data classification matrix; and They are as follows:
[0160] ; ;
[0161] in, Indicates the optimal position matrix corresponding to The first in the classification center dataset of parathyroid feature data classification feature data, Indicates The optimal classification matrix The parathyroid gland feature data, Indicates The total number of parathyroid feature data in the optimal classification; ;
[0162] S36. Based on the first blood supply status level data set Determine the first optimal classification matrix set Whether the blood flow status level data corresponding to the parathyroid feature data sets in each optimal classification matrix are the same;
[0163] If they are the same, the first optimal position matrix As the optimal classification center data matrix Otherwise, continue the iterative operation in S34 until the first optimal classification matrix set The parathyroid gland feature data sets in each optimal classification matrix correspond to the same blood flow status level data, and the optimal classification center data matrix is obtained. ;
[0164] S4, collecting historical images of the plurality of parathyroid samples and performing preprocessing, and then performing feature extraction on the historical images to obtain a historical parathyroid feature data matrix set;
[0165] Using the historical parathyroid gland feature data matrix to train and test the initial BP neural network model to obtain a final BP neural network model;
[0166] The S4 comprises the following steps:
[0167] S41. Setting a historical time point set ; Indicates A historical point in time, Indicates the total number of set historical time points;
[0168] S42: Collecting the historical time point set Final parathyroid imaging set at each time point The characteristic data of each parathyroid imaging in the parathyroid gland is obtained to obtain the historical parathyroid gland comprehensive data matrix set , Indicates A comprehensive parathyroid data matrix at each historical time point;
[0169] S43, constructing an initial BP neural network model; using the historical parathyroid feature data matrix set Train and test the initial BP neural network model to obtain the final BP neural network model;
[0170] The S43 comprises the following steps:
[0171] S431, constructing an initial BP neural network model and setting a training data ratio, a training error threshold, and a test accuracy threshold;
[0172] S432: According to the training data ratio, the historical parathyroid feature data matrix set is Perform data division to obtain a historical parathyroid comprehensive training data matrix set and a historical parathyroid comprehensive test data matrix set;
[0173] S433, inputting the historical parathyroid comprehensive training data matrix set into the initial BP neural network model for training; when the training error during the training process is less than the training error threshold, stopping the training to obtain a trained BP neural network model;
[0174] S434, input the historical parathyroid comprehensive test data matrix set into the trained BP neural network model for testing to obtain a test accuracy rate; when the test accuracy rate is greater than or equal to a test accuracy rate threshold, use the trained BP neural network model as the final BP neural network model; otherwise, return to S433 to continue training the trained BP neural network model until the test accuracy rate is greater than or equal to the test accuracy rate threshold, to obtain a final BP neural network model;
[0175] S5, collecting the real-time imaging feature data and the historical imaging feature data of the parathyroid gland to be analyzed, and obtaining a real-time parathyroid feature data set and a historical parathyroid feature data matrix to be analyzed;
[0176] Inputting the historical parathyroid gland feature data matrix to be analyzed into the real-time final BP neural network model for prediction to obtain the future parathyroid gland feature data matrix to be analyzed;
[0177] The S5 comprises the following steps:
[0178] S51, collect the parathyroid image to be analyzed in real time, and then use S13 to denoise and extract features to obtain a real-time parathyroid feature data set , Indicates Real-time characteristic data of parathyroid glands to be analyzed;
[0179] S52: Based on the historical time point set Photoacoustic imaging equipment is used to collect images of the parathyroid glands to be analyzed, and then S13 method is used to denoise and extract features to obtain the historical parathyroid feature data matrix to be analyzed. ;as follows,
[0180] ;
[0181] in, Indicates The parathyroid gland to be analyzed at the historical time point Characteristic data;
[0182] S53. Setting future time point set , Indicates A future time point, Indicates the total number of set future time points;
[0183] Based on the future time point The historical parathyroid characteristic data matrix to be analyzed Input into the final BP neural network model for prediction, and obtain the future characteristic data matrix of the parathyroid gland to be analyzed ;as follows,
[0184] ;
[0185] in, Indicates The first parathyroid gland to be analyzed at a future time point Characteristic data;
[0186] S6, calculating the Euclidean distance between each feature data in the future feature data matrix of the parathyroid gland to be analyzed and the real-time parathyroid feature data set and each center data in the optimal classification center data matrix, and combining with the first blood circulation status level data set to obtain the blood circulation status level data of the parathyroid gland to be analyzed;
[0187] The S6 comprises the following steps:
[0188] S61. Calculate real-time parathyroid feature dataset To the optimal classification center data matrix The Euclidean distance of each row of data in the first Euclidean distance set is obtained; the optimal classification center data matrix corresponding to the smallest Euclidean distance in the first Euclidean distance set is obtained. The blood supply status level data of the row data is used as the blood supply status level data of the real-time parathyroid feature data set, and is recorded as the first blood supply status level data;
[0189] S62. Calculate the future characteristic data matrix of the parathyroid gland to be analyzed Each row of data in the optimal classification center data matrix The Euclidean distance of each row of data in the second Euclidean distance matrix is obtained; the optimal classification center data matrix corresponding to the smallest Euclidean distance in each row of the second Euclidean distance matrix is obtained. The blood flow status level data of the row data is used as the future characteristic data matrix of the parathyroid gland to be analyzed The blood circulation status level data of the feature data set corresponding to the row in the image is obtained to obtain a second blood circulation status level data set;
[0190] S63. Merge the second blood supply status level data set with the first blood supply status level data to obtain a third blood supply status level data set; and use the data with the lowest blood supply status level in the third blood supply status level data set as the blood supply status level data of the parathyroid gland to be analyzed.
[0191] Example 2
[0192] This embodiment discloses a parathyroid blood supply status analysis system, which can implement the method of the above embodiment, including a parathyroid real-time imaging data acquisition module, a preprocessing module, a feature extraction module, an optimization classification module, a parathyroid data acquisition module, a prediction module and a final blood supply status determination module;
[0193] The parathyroid real-time imaging data acquisition module is used to acquire real-time images of the parathyroid glands of a plurality of parathyroid gland samples;
[0194] Preferably, the parathyroid real-time imaging data acquisition module uses a prior art photoacoustic imaging device or an autofluorescence imaging device or a laser speckle imaging device. In addition, other prior art methods for obtaining imaging images of each parathyroid sample are also applicable to the present invention.
[0195] The preprocessing module preprocesses the real-time images of the parathyroid glands of the plurality of parathyroid gland samples to obtain a final parathyroid gland imaging set;
[0196] The feature extraction module is used to extract feature data of the final parathyroid imaging set to obtain a parathyroid feature data matrix;
[0197] The optimization classification module performs iterative optimization classification on each feature data in the parathyroid feature data matrix to obtain an optimal classification center data matrix;
[0198] The parathyroid data collection module is used to collect real-time parathyroid characteristic data of the parathyroid gland to be analyzed and historical parathyroid characteristic data to be analyzed;
[0199] The prediction module is used to use the final BP neural network model to predict the real-time parathyroid feature data set of the parathyroid to be analyzed and the historical parathyroid feature data matrix to be analyzed to obtain the future feature data matrix of the parathyroid to be analyzed;
[0200] The final blood supply status determination module is used to obtain the blood supply status level data of the parathyroid gland to be analyzed based on the future characteristic data matrix of the parathyroid gland to be analyzed, the real-time parathyroid characteristic data set and the optimal classification center data matrix.
[0201] Preferably, it also includes a prediction model construction module, which is used to construct a final BP neural network model through an optimal classification center data matrix.
[0202] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0203] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.
Claims
1. The parathyroid blood flow status analysis system based on photoacoustic imaging is characterized by: include: A real-time imaging data acquisition module, used to acquire real-time images of multiple parathyroid samples; A preprocessing module, used for preprocessing the real-time images of multiple parathyroid samples to obtain a final parathyroid imaging set; A feature extraction module, used for extracting feature data of the final parathyroid imaging set to obtain a parathyroid feature data matrix; An optimization classification module is used to iteratively optimize and classify the parathyroid feature data matrix to obtain the optimal classification center data matrix; The method for iterative optimization classification of parathyroid feature data matrix includes: Construct a first male mayfly population and a first female mayfly population; then set the maximum number of iterations of the first male mayfly population and the first female mayfly population to , the current number of iterations is ; Setting the initial position of each male mayfly in the first male mayfly population and the initial position of each female mayfly in the first female mayfly population according to the parathyroid gland characteristic data matrix; Calculating the Euclidean distance from each piece of data in the parathyroid characteristic data matrix to each row of data in the initial position of each mayfly to obtain a Euclidean distance matrix; selecting the data in the parathyroid characteristic data matrix corresponding to the minimum Euclidean distance data in each row of the Euclidean distance matrix and classifying them into the data in the corresponding row in the initial position of each mayfly to obtain an initial parathyroid characteristic data classification matrix set; setting the fitness function of the first male mayfly population and the first female mayfly population; An iterative operation is started, and in each round of iteration, the fitness value of each mayfly in the first male mayfly population and the first female mayfly population is calculated according to the fitness function; the position of the mayfly individual with the highest fitness in the first male mayfly population and the first female mayfly population is selected, and recorded as the first best fitness, the second best fitness, the first best position and the second best position respectively; the position and speed of each male mayfly and the female mayfly are updated according to the first best fitness, the second best fitness, the first best position and the second best position; at the same time, the male mayfly individual with the highest fitness in the first male mayfly population is mated with a random female mayfly individual in the first female mayfly population; when When , the iteration is stopped, and the first optimal position matrix and the first optimal classification matrix set are obtained; Determine whether the blood supply status level data corresponding to the characteristic data sets of the parathyroid glands in each optimal classification matrix in the first optimal classification matrix set are the same according to the first blood supply status level data set; if they are the same, use the first optimal position matrix as the optimal classification center data matrix; otherwise, continue the iterative operation in S34 until the blood supply status level data corresponding to the characteristic data sets of the parathyroid glands in each optimal classification matrix in the first optimal classification matrix set are the same, and obtain the optimal classification center data matrix; A parathyroid data collection module, used for collecting real-time parathyroid characteristic data of the parathyroid gland to be analyzed and historical parathyroid characteristic data to be analyzed; The prediction module uses the final BP neural network model constructed by the optimal classification center data matrix to perform prediction operations on the real-time parathyroid characteristic data and the historical parathyroid characteristic data to be analyzed, and obtains the future characteristic data matrix of the parathyroid to be analyzed; The final blood supply status determination module is used to obtain the blood supply status grade data of the parathyroid gland to be analyzed based on the future characteristic data matrix of the parathyroid gland to be analyzed, the real-time parathyroid characteristic data set and the optimal classification center data matrix.
2. The parathyroid blood supply status analysis system based on photoacoustic imaging according to claim 1, characterized in that: The method for collecting real-time images of the plurality of parathyroid samples comprises: Collecting a plurality of parathyroid samples to obtain a parathyroid sample set; then setting a parathyroid blood flow status level set; and analyzing the blood flow status level of each parathyroid sample in the parathyroid sample set in conjunction with the parathyroid blood flow status level set to obtain a first blood flow status level data set; The imaging images of each parathyroid sample in the parathyroid sample set are collected to obtain an initial parathyroid imaging set.
3. The parathyroid blood supply status analysis system based on photoacoustic imaging according to claim 2, characterized in that: The preprocessing method for real-time images of multiple parathyroid samples includes: performing a denoising operation on each parathyroid image in the initial parathyroid image set to obtain a denoised parathyroid image set; A contrast enhancement operation is performed on each parathyroid image in the denoised parathyroid image set to obtain a final parathyroid image set.
4. The parathyroid blood supply status analysis system based on photoacoustic imaging according to claim 3, characterized in that: The method of constructing the final BP neural network model through the optimal classification center data matrix includes: Setting a historical time point set; collecting characteristic data of each parathyroid image in the final parathyroid image set at each time point in the historical time point set to obtain a historical parathyroid characteristic data matrix set; An initial BP neural network model is constructed; and the historical parathyroid gland feature data matrix set is used to train and test the initial BP neural network model to obtain a final BP neural network model.
5. The parathyroid blood flow status analysis system based on photoacoustic imaging according to claim 4, characterized in that: The method of using the historical parathyroid feature data matrix set to train and test the initial BP neural network model includes: Construct the initial BP neural network model and set the training data ratio, training error threshold and test accuracy threshold; Performing data division on the historical parathyroid characteristic data matrix set according to the training data ratio to obtain a historical parathyroid comprehensive training data matrix set and a historical parathyroid comprehensive test data matrix set; Inputting the historical parathyroid comprehensive training data matrix set into the initial BP neural network model for training; when the training error during the training process is less than the training error threshold, stopping the training to obtain a trained BP neural network model; The historical parathyroid comprehensive test data matrix set is input into the trained BP neural network model for testing to obtain the test accuracy; when the test accuracy is greater than or equal to the test accuracy threshold, the trained BP neural network model is used as the final BP neural network model; otherwise, return and continue to train the trained BP neural network model until the test accuracy is greater than or equal to the test accuracy threshold to obtain the final BP neural network model.
6. The parathyroid blood supply status analysis system based on photoacoustic imaging according to claim 5, characterized in that: The method for collecting real-time parathyroid characteristic data of the parathyroid gland to be analyzed and historical parathyroid characteristic data to be analyzed includes: The images of the parathyroid glands to be analyzed are collected in real time, and then the images of the parathyroid glands to be analyzed collected in real time are denoised and features are extracted to obtain a real-time parathyroid imaging feature data set to be analyzed; The parathyroid glands to be analyzed are imaged according to the historical time point set, and then the acquired historical images of the parathyroid glands to be analyzed are denoised and features are extracted to obtain a historical parathyroid gland feature data matrix to be analyzed.
7. The parathyroid blood supply status analysis system based on photoacoustic imaging according to claim 6, characterized in that: The method of predicting operation includes: A future time point set is set; according to the future time point set, the historical parathyroid characteristic data matrix to be analyzed is input into the final BP neural network model for prediction to obtain the future parathyroid characteristic data matrix to be analyzed.
8. The parathyroid blood supply status analysis system based on photoacoustic imaging according to claim 7, characterized in that: The method for obtaining the blood supply status grade data of the parathyroid gland to be analyzed according to the future characteristic data matrix of the parathyroid gland to be analyzed, the real-time parathyroid characteristic data set and the optimal classification center data matrix comprises: Calculate the Euclidean distance from the real-time parathyroid feature data set to each row of data in the optimal classification center data matrix to obtain a first Euclidean distance set; use the blood circulation status level data of the row data of the optimal classification center data matrix corresponding to the smallest Euclidean distance in the first Euclidean distance set as the blood circulation status level data of the real-time parathyroid feature data set, recorded as the first blood circulation status level data; Calculate the Euclidean distance from each row of data in the parathyroid future characteristic data matrix to be analyzed to each row of data in the optimal classification center data matrix to obtain a second Euclidean distance matrix; use the blood circulation status level data of the row data of the optimal classification center data matrix corresponding to the minimum Euclidean distance in each row of the second Euclidean distance matrix as the blood circulation status level data of the characteristic data set of the corresponding row in the parathyroid future characteristic data matrix to be analyzed to obtain a second blood circulation status level data set; The second blood supply status level data set is combined with the first blood supply status level data set to obtain a third blood supply status level data set; and the data with the lowest blood supply status level in the third blood supply status level data set is used as the blood supply status level data of the parathyroid gland to be analyzed.
9. A method for analyzing the parathyroid blood flow status system, characterized in that: The following steps are included: S1, collecting real-time images of multiple parathyroid samples and preprocessing them to obtain a final parathyroid imaging set; and testing the blood supply status levels of the multiple parathyroid samples to obtain a first blood supply status level data set; S2, extracting characteristic data of the final parathyroid imaging set to obtain a parathyroid characteristic data matrix; S3, performing iterative optimization classification on each feature data in the parathyroid feature data matrix to obtain an optimal classification center data matrix; The S3 comprises the following steps: S31, constructing a first male mayfly population and a first female mayfly population; and then setting the maximum number of iterations of the first male mayfly population and the first female mayfly population to , the current number of iterations is ; S32, setting the initial position of each male mayfly in the first male mayfly population and the initial position of each female mayfly in the first female mayfly population according to the parathyroid gland characteristic data matrix; S33, calculating the Euclidean distance from each data in the parathyroid characteristic data matrix to each row of data in the initial position of each mayfly, to obtain a Euclidean distance matrix; selecting the data in the parathyroid characteristic data matrix corresponding to the minimum Euclidean distance data in each row in the Euclidean distance matrix and classifying them into the data in the corresponding row in the initial position of each mayfly, to obtain an initial parathyroid characteristic data classification matrix set; setting the fitness function of the first male mayfly population and the first female mayfly population; S34, start an iterative operation, and in each round of iteration, calculate the fitness value of each mayfly in the first male mayfly population and the first female mayfly population according to the fitness function; select the position of the mayfly individual with the highest fitness in the first male mayfly population and the first female mayfly population, and record them as the first best fitness, the second best fitness, the first best position and the second best position respectively; update the position and speed of each male mayfly and the female mayfly according to the first best fitness, the second best fitness, the first best position and the second best position; and mate the male mayfly individual with the highest fitness in the first male mayfly population with a random female mayfly individual in the first female mayfly population; S35, when When , the iteration is stopped, and the first optimal position matrix and the first optimal classification matrix set are obtained; Determine whether the blood supply status level data corresponding to the characteristic data sets of the parathyroid glands in each optimal classification matrix in the first optimal classification matrix set are the same according to the first blood supply status level data set; if they are the same, use the first optimal position matrix as the optimal classification center data matrix; otherwise, continue the iterative operation in S34 until the blood supply status level data corresponding to the characteristic data sets of the parathyroid glands in each optimal classification matrix in the first optimal classification matrix set are the same, and obtain the optimal classification center data matrix; S4, collecting historical images of the plurality of parathyroid samples and performing preprocessing, and then performing feature extraction on the historical images to obtain a historical parathyroid feature data matrix set; Using the historical parathyroid gland feature data matrix to train and test the initial BP neural network model to obtain a final BP neural network model; S5, collecting the real-time imaging feature data and the historical imaging feature data of the parathyroid gland to be analyzed, and obtaining a real-time parathyroid feature data set and a historical parathyroid feature data matrix to be analyzed; Inputting the historical parathyroid gland feature data matrix to be analyzed into the real-time final BP neural network model for prediction to obtain the future parathyroid gland feature data matrix to be analyzed; S6. Calculate the Euclidean distance between each feature data in the future feature data matrix of the parathyroid gland to be analyzed and the real-time parathyroid feature data set and each center data in the optimal classification center data matrix, and combine with the first blood circulation status level data set to obtain the blood circulation status level data of the parathyroid gland to be analyzed.
Citation Information
Patent Citations
Method and system for automatically identifying and tracking parathyroid gland in endoscopic thyroid surgery
CN117710325A
Parathyroid gland function imaging system and endoscope
CN219895706U
Parallel spectral clustering method based on KD tree and chaos mayfly naiad optimization algorithm
CN113128618A
Parathyroid gland identification and blood supply judgment method and system in endoscopic thyroid surgery
CN117710323A
Population classification using point cloud features
DE102022206025A1
