CN120473079B - System and method for analyzing artificial intelligent blood data - Google Patents
System and method for analyzing artificial intelligent blood dataInfo
- Publication number
- CN120473079B CN120473079B CN202510981497.8A CN202510981497A CN120473079B CN 120473079 B CN120473079 B CN 120473079B CN 202510981497 A CN202510981497 A CN 202510981497A CN 120473079 B CN120473079 B CN 120473079B
- Authority
- CN
- China
- Prior art keywords
- data
- blood
- feature
- compression
- quantum
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The invention relates to the technical field of computer data analysis, in particular to a system and a method for artificial intelligence blood data analysis, which comprises a data acquisition module, a data synchronization module, a blood data network modeling module, an entropy compression and optimization feature module and a digital twin modeling module. The method not only can efficiently integrate multi-modal blood data to construct an accurate DBN model so as to mine dynamic causal relationship among the data, but also ensures the high efficiency and accuracy of data processing through the entropy compression and optimization feature module, predicts the state of quantum feature vector values by utilizing the digital twin and A3C algorithm model, and finally accurately recommends the administration time to be sent to the medical data terminal, thereby effectively improving the medical data processing efficiency.
Description
Technical Field
The invention relates to the technical field of computer data analysis, in particular to a system and a method for analyzing artificial intelligence blood data.
Background
Blood data contains a variety of modal information, while these data also have spatiotemporal characteristics. The existing analysis method often processes the data modes in isolation, so that multi-mode information cannot be effectively integrated and the time-space synchronism of the data is considered. This results in incomplete understanding of blood data, and difficulty in mining the deep physiological and pathological information behind the data.
The Chinese patent application with publication number of CN114662623B discloses a method and a system for classifying blood samples in blood coagulation detection based on XGBoost, which belong to the technical field of intelligent medical treatment, blood coagulation index data of the blood samples are input into a pre-trained classification model of the blood samples in the blood coagulation detection based on XGBoost, the blood coagulation index data of the blood samples are extracted according to preset weights of various blood coagulation indexes, the extracted low-level features and high-level features are fused to obtain detection item features, and a XGBoost-based classifier is utilized to obtain blood coagulation detection classification results of the blood samples to be detected according to the detection item features. However, the solution still has the problems that the analysis and the prediction of the trend of the blood in the future time are difficult, so that the administration time cannot be recommended according to the trend of the blood, and the analysis efficiency of the administration time is low.
Disclosure of Invention
Therefore, the invention provides an artificial intelligence blood data analysis system and method, which are used for solving the problems that in the prior art, the analysis and the prediction of the trend of blood change in future time are difficult, so that the time recommendation of the drug administration can not be carried out according to the trend of blood change, and the analysis efficiency of the drug administration time is low.
To achieve the above object, in one aspect, the present invention provides a system for artificial intelligence blood data analysis, comprising:
the data acquisition module is used for acquiring blood data;
The data synchronization module is used for carrying out data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set;
The blood data network modeling module is used for inputting the multi-mode biophysical data set into a pre-constructed DBN model to obtain a recommended dynamic causal graph output by the DBN model;
the entropy compression and optimization feature module is used for compressing and optimizing the recommended dynamic causal graph through a compression optimization method to obtain a quantum feature vector value, judging the compression effect condition according to the data compression ratio, correcting the data synchronization method according to the judging result, and adjusting the judging process of the compression effect condition according to the quantum feature similarity;
The digital twin body modeling module is used for inputting the quantum characteristic vector value into a pre-built twin body model to obtain a predicted quantum characteristic vector value state, inputting the predicted quantum characteristic vector value state into a pre-built A3C algorithm model to obtain recommended administration time, and sending the recommended administration time to the medical data terminal.
Further, the data synchronization module performs data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set, and the data synchronization method comprises:
A1, carrying out interpolation processing on blood data by using a linear interpolation algorithm to obtain interpolated blood data, wherein the interpolated blood data comprises an interpolated blood molecular vibration spectrum, an interpolated blood pressure, an interpolated blood flow velocity data, an interpolated cell image, an interpolated blood pH value and an interpolated oxygen partial pressure;
a2, carrying out coordinate space synchronization on the interpolated blood molecule vibration spectrum and the interpolated cell image by using a SIFT algorithm to obtain a synchronized blood molecule vibration spectrum and a synchronized cell image;
A3, carrying out coordinate space synchronization on the interpolated blood pH value and the interpolated oxygen partial pressure and the synchronized cell image by using a coordinate synchronization method to obtain the synchronized blood pH value and the synchronized oxygen partial pressure;
And A4, outputting the synchronized blood molecular vibration spectrum, the interpolated blood pressure, the interpolated blood flow velocity data, the synchronized cell image, the synchronized blood pH value and the synchronized oxygen partial pressure as a multi-mode biophysical data set to obtain the multi-mode biophysical data set.
Further, the SIFT algorithm of the data synchronization module in step A2 includes:
Step A21, detecting the frequency spectrum of the blood molecule vibration after interpolation and the extreme point of the cell image after interpolation through a Gaussian differential pyramid to obtain a stable characteristic point;
Step A22, constructing a feature descriptor according to the stable feature points to obtain a vibration spectrum feature descriptor and a cell image feature descriptor;
step A23, calculating the distance proportion according to the vibration spectrum feature descriptors and the cell image feature descriptors by using a Euclidean distance nearest neighbor algorithm Obtaining the distance proportionProportional distanceComparing with a preset distance ratio db0, judging the distance ratio according to the comparison result, and outputting the characteristic point attribute according to the judgment result, wherein:
When (when) When db0 is not more than, the data synchronization module judges that the distance proportion condition is that the distance proportion is small, and outputs the vibration spectrum feature descriptors and the cell image feature descriptors as matching feature points serving as feature point attributes;
When (when) When db0 is detected, the data synchronization module judges that the distance proportion condition is that the distance proportion is large, and outputs the vibration spectrum feature descriptors and the cell image feature descriptors which are unmatched feature points as feature point attributes;
And step A24, outputting the attribute of the characteristic point as the characteristic point matched with the characteristic point to perform coordinate synchronization, and obtaining the synchronized blood molecule vibration spectrum and the synchronized cell image.
Further, the blood data network modeling module builds a DBN model through a DBN model building method, and inputs the multi-mode biophysical data set into the DBN model to obtain a recommended dynamic causal graph output by the DBN model.
Further, the entropy compression and optimization feature module compresses and optimizes the recommended dynamic causal graph through a compression optimization method to obtain a quantum feature vector value, and the compression optimization method comprises the following steps:
step B1, calculating single feature entropy of the recommended dynamic causal graph to obtain a single feature entropy value;
Step B2, constructing a maximum entropy feature subset according to the single feature entropy value by utilizing a greedy algorithm to obtain a maximum entropy feature subset S;
step B3, performing DNA sequence mapping on the maximum entropy feature subset S by using a DNA sequence mapping method to obtain a DNA sequence;
Step B4, performing quantum optimization on the DNA sequence by using a quantum optimization method to obtain a quantum eigenvector value 。
Further, the entropy compression and optimization feature module calculates a data compression ratio Ys according to a data size Bs of the recommended dynamic causal graph and a data size By of the DNA sequence, sets ys=100% x Bs/By, compares the data compression ratio Ys with a preset compression ratio Ys0, judges the compression effect condition according to the comparison result, and corrects the data synchronization method according to the judgment result, wherein:
When Ys is more than or equal to Ys0, the entropy compression and optimization feature module judges that the compression effect condition is good, and the data synchronization method is not corrected;
When Ys < Ys0, the entropy compression and optimization feature module judges that the compression effect is poor, corrects the data synchronization method, corrects a preset distance proportion db0 according to a correction coefficient Jz, sets jz=0.87-0.11×e -0.7×(Ys0-Ys), wherein e is the base of natural logarithm, obtains a corrected preset distance proportion db0J, sets db0 j=db0×jz, replaces the preset distance proportion db0 with the corrected preset distance proportion db0J, and compares the distance proportion db with the corrected preset distance proportion db0J again.
Further, the entropy compression and optimization feature module is used for compressing and optimizing the entropy compression and optimization feature module according to quantum feature vector valuesAnd the target quantum eigenvector valueSimilarity to quantum characteristicsPerforming calculation and settingSimilarity of quantum characteristicsSimilarity to a presetComparing, judging the similarity condition according to the comparison result, and adjusting the judging process of the compression effect condition according to the judging result, wherein:
When (when) ≥When the entropy compression and optimization feature module judges that the similarity conditions are similar, the judging process of the compression effect conditions is not adjusted;
When (when) <When the entropy compression and optimization feature module judges that the similarity conditions are dissimilar, the judgment process of the compression effect conditions is adjusted, and coefficients are adjustedThe preset compression ratio Ys0 is adjusted and setObtaining the adjusted preset compression ratio Ys0t, setting Ys0 t=ys 0×And replacing the preset compression ratio Ys0 with the adjusted preset compression ratio Ys0t, and re-comparing the data compression ratio Ys with the adjusted preset compression ratio.
Further, the digital twin body modeling module constructs a twin body model through a twin body model constructing method to obtain a twin body model, and inputs the quantum characteristic vector value into the twin body model to obtain a predicted quantum characteristic vector value state S (t) output by the twin body model.
Further, the digital twin body modeling module constructs an A3C algorithm through an A3C algorithm model construction method to obtain an A3C algorithm model, inputs a predicted quantum eigenvector value state S (t) into the A3C algorithm model, acquires recommended administration time output by the A3C algorithm model, and sends the recommended administration time to a medical data terminal.
In another aspect, the present invention also provides a method of artificial intelligence blood data analysis, the method comprising:
Step S1, collecting blood data;
step S2, carrying out data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set;
S3, inputting the multi-mode biophysical data set into a pre-constructed DBN model to obtain a recommended dynamic causal graph output by the DBN model;
s4, compressing and optimizing the recommended dynamic causal graph through a compression optimization method to obtain a quantum characteristic vector value, judging the compression effect condition according to the data compression ratio, correcting the data synchronization method according to the judgment result, and adjusting the judgment process of the compression effect condition according to the quantum characteristic similarity;
And S5, inputting the quantum characteristic vector value into a pre-constructed twin body model to obtain a predicted quantum characteristic vector value state, inputting the predicted quantum characteristic vector value state into a pre-constructed A3C algorithm model to obtain recommended administration time, and transmitting the recommended administration time to a medical data terminal.
Compared with the prior art, the invention has the beneficial effects that the system collects the blood data through the data collection module so as to analyze and predict the future change trend of the blood, the system synchronizes the collected blood data through the data synchronization module so as to make the blood data highly uniform and facilitate the analysis and prediction of the future change trend of the blood, the system also obtains the recommended dynamic causal graph through the blood data network modeling module, knows various feature vectors in the blood data according to the recommended dynamic causal graph so as to compress the blood data, the system also compresses and optimizes the recommended dynamic causal graph through the entropy compression and optimization feature module so as to reduce the calculation amount required for analyzing the blood data and predicting the future data of the blood, improve the analysis and prediction efficiency of the blood data and further improve the analysis efficiency of the medication time, the system also builds a twin model and an A3C algorithm model through a digital twin modeling module, analyzes and predicts the future 24-hour blood change trend, and recommends the administration time according to the predicted blood pressure change trend, so that the analysis efficiency of a medical data analysis terminal on the administration time is improved, the whole-flow intelligent processing from blood data acquisition to recommended administration time is realized through a series of modules such as data acquisition, synchronization, modeling, compression optimization and the like, the accurate DBN model can be built by efficiently integrating multi-mode blood data to mine the dynamic causal relationship among the data, the high efficiency and the accuracy of the data processing are ensured through an entropy compression and optimization feature module, the state of quantum feature vector values is predicted through the digital twin and the A3C algorithm model, and finally the recommended administration time is accurately sent to the medical data terminal, and the medical data processing efficiency is effectively improved.
Drawings
FIG. 1 is a schematic diagram of a system for artificial intelligence blood data analysis according to the present embodiment;
Fig. 2 is a flow chart of a method for analyzing artificial intelligence blood data according to the embodiment.
Detailed Description
The invention will be further described with reference to examples for the purpose of making the objects and advantages of the invention more apparent, it being understood that the specific examples described herein are given by way of illustration only and are not intended to be limiting.
Preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are merely for explaining the technical principles of the present invention, and are not intended to limit the scope of the present invention.
It should be noted that, in the description of the present invention, terms such as "upper," "lower," "left," "right," "inner," "outer," and the like indicate directions or positional relationships based on the directions or positional relationships shown in the drawings, which are merely for convenience of description, and do not indicate or imply that the apparatus or elements must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention.
In addition, it should be noted that, in the description of the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, integrally connected, mechanically connected, electrically connected, directly connected, indirectly connected through an intermediate medium, or in communication between two elements. The specific meaning of the above terms in the present invention can be understood by those skilled in the art according to the specific circumstances.
Referring to fig. 1, a schematic structural diagram of a system for analyzing artificial intelligence blood data according to the present embodiment is shown, where the system includes:
the data acquisition module is used for acquiring blood data;
The data synchronization module is used for carrying out data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set, and is connected with the data acquisition module;
The blood data network modeling module is used for inputting the multi-mode biophysical data set into a pre-constructed DBN model to obtain a recommended dynamic causal graph output by the DBN model, and is connected with the data synchronization module;
The entropy compression and optimization feature module is used for compressing and optimizing the recommended dynamic causal graph through a compression optimization method to obtain a quantum feature vector value, judging the compression effect condition according to the data compression ratio, correcting the data synchronization method according to the judging result, and adjusting the judging process of the compression effect condition according to the quantum feature similarity, and is connected with the blood data network modeling module;
The digital twin modeling module is used for inputting the quantum characteristic vector value into a pre-constructed twin model to obtain a predicted quantum characteristic vector value state, inputting the predicted quantum characteristic vector value state into a pre-constructed A3C algorithm model to obtain recommended administration time, and sending the recommended administration time to the medical data terminal, and is connected with the entropy compression and optimization feature module.
Specifically, the system for analyzing the artificial intelligent blood data is used for a medical data analysis terminal, predicts blood and drug reactions through predicting future change trend of blood, gives recommended administration time, so that the follow-up medical data terminal analyzes drug administration data, wherein the system collects blood data through a data collection module, so that the future change trend of blood is analyzed and predicted later, the system further synchronizes the collected blood data through a data synchronization module, enables the blood data to be highly uniform, facilitates the follow-up analysis and prediction of the future change trend of blood, the system further obtains a recommended dynamic causal graph through a blood data network modeling module, knows various feature vectors in the blood data according to the recommended dynamic causal graph, so that the follow-up blood data is compressed, the system further compresses and optimizes the recommended dynamic causal graph through an entropy compression and optimization feature module, reduces calculation amount required for analyzing the blood data and predicting the future blood data, improves analysis and prediction efficiency of the blood data, accordingly improves analysis efficiency of drug administration time, the system further builds a model of a twin body and A3 model for the blood, the blood data is not integrated with a model for the data is integrated through the data analysis of the blood analysis, the blood analysis results of the blood analysis and the blood analysis data is completely in a three-dimensional model, the data analysis model is completely compressed through the data analysis of the optimal time, and the blood analysis is realized, and the data is not integrated through the data analysis of the data is completely-planned in the medical data, and the data is completely-compressed through the data analysis model, and the data is completely-optimized, and the time-matched with the data analysis model is optimized, and the data is optimized, the entropy compression and optimization feature module is used for guaranteeing the high efficiency and accuracy of data processing, the digital twin and A3C algorithm model is used for predicting the state of the quantum feature vector value, and finally, the recommended administration time is accurately sent to the medical data terminal, so that the medical data processing efficiency is effectively improved.
Specifically, the Chinese language of the DBN model is called a dynamic Bayesian network model, and the English name of the DBN model is Dynamic Bayesian Network.
The data acquisition module is used for acquiring blood data, the blood data comprises a blood molecular vibration spectrum, blood pressure, blood flow rate data, a cell image, a blood pH value and an oxygen partial pressure, the data acquisition module acquires the blood molecular vibration spectrum through a terahertz biological field imager, the terahertz biological field imager is equipment for imaging electromagnetic characteristics of biological tissues or cells by utilizing terahertz waves, such as a blue source peak QT-TO1000, the blood molecular vibration spectrum is a spectrum signal reflecting the vibration energy characteristics of chemical bonds in molecules, the data acquisition module acquires the blood pressure, the blood flow rate data and the cell image through a microfluidic organ chip, the microfluidic organ chip is a bionic micro-device, a micro-scale microfluidic channel and a porous membrane structure are constructed on the chip through a micro-fluidic technology, the blood pressure is a value applied TO a channel wall in the micro-channel, the blood flow rate is a fluid volume passing through the micro-channel, the blood flow sensor is a channel, the blood flow rate is a fluid volume passing through the channel, the blood is a pH value of the cell image is a high-based on the oxygen partial pressure sensor, the oxygen partial pressure is generated in the channel, the oxygen partial pressure sensor is high in the channel is based on the pH value, and the oxygen partial pressure sensor is high in the channel is based on the pressure of the physical state of the oxygen partial pressure, and the oxygen sensor is high in the channel, and the oxygen is sensitive TO the oxygen is in the channel.
Specifically, the data acquisition module acquires blood data so as to carry out intelligent analysis on the blood data subsequently, and the effect of intelligent analysis on the blood data is realized.
Specifically, the data synchronization module performs data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set, and the data synchronization method comprises the following steps:
A1, carrying out interpolation processing on blood data by using a linear interpolation algorithm to obtain interpolated blood data, wherein the interpolated blood data comprises an interpolated blood molecular vibration spectrum, an interpolated blood pressure, an interpolated blood flow velocity data, an interpolated cell image, an interpolated blood pH value and an interpolated oxygen partial pressure;
a2, carrying out coordinate space synchronization on the interpolated blood molecule vibration spectrum and the interpolated cell image by using a SIFT algorithm to obtain a synchronized blood molecule vibration spectrum and a synchronized cell image;
A3, carrying out coordinate space synchronization on the interpolated blood pH value and the interpolated oxygen partial pressure and the synchronized cell image by using a coordinate synchronization method to obtain the synchronized blood pH value and the synchronized oxygen partial pressure;
And A4, outputting the synchronized blood molecular vibration spectrum, the interpolated blood pressure, the interpolated blood flow velocity data, the synchronized cell image, the synchronized blood pH value and the synchronized oxygen partial pressure as a multi-mode biophysical data set to obtain the multi-mode biophysical data set.
Specifically, in this embodiment, since the blood pressure, the blood flow velocity data and the cell image are collected by the microfluidic organ chip, the interpolated blood pressure, the interpolated blood flow velocity data and the interpolated cell image are not subjected to coordinate space synchronization, the linear interpolation algorithm refers to an algorithm for uniformly interpolating the blood data collected at different frequencies onto a common time sequence, the embodiment does not limit the specific implementation of the linear interpolation algorithm, and a person skilled in the art can set the time sequence of the microfluidic organ chip cell image as a reference according to the actual situation, obtain a reference time point, and calculate values at the reference time point for the blood molecular vibration spectrum, the blood pressure, the blood flow velocity data, the blood pH value and the oxygen partial pressure by linear interpolation, and set the values at the reference time point, wherein,As a point in time of the reference,For a first raw acquisition time point adjacent to the reference time point,For the acquisition value of the first original acquisition time point,For a second raw acquisition time point adjacent to the reference time point,For the acquisition value of the second original acquisition time point,As a result of interpolation, the two original collection time points adjacent to the reference time point refer to collection time points which are similar to the reference time point in the collection process of blood molecular vibration spectrum, blood pressure, blood flow velocity data, blood pH value and oxygen partial pressure, the SIFT algorithm refers to a computer vision algorithm, the feature descriptors refer to statistical description of local areas around the image feature points and are used for quantifying visual properties of the feature points so as to facilitate feature matching, the visual properties refer to features or characteristics which can be perceived and understood by eyes in an object, scene, image or visual element through a visual system, the chinese full name of the SIFT algorithm refers to a feature transformation algorithm which is unchanged, the english full name of the SIFT algorithm refers to Scale-InvariantFeatureTransform, the coordinate synchronization method refers to a method for performing spatial synchronization on the interpolated blood pH value and the interpolated oxygen partial pressure and the synchronized cell image by taking the synchronized cell image as a reference, the spatial coordinate synchronization refers to a method for performing the measurement of the same dimensional coordinate system, the motion coordinate data under different directions or the same reference system is converted to the specific coordinate system, and the differential coordinate system is applied to the differential measurement method such as the differential measurement of the current, the differential measurement method is implemented by a mathematical method,),As the abscissa of the quantum sensor,Aligning the coordinates of the quantum sensor with the coordinates of the cell image to obtain aligned coordinates of the quantum sensor,),To align the abscissa of the post-alignment quantum sensor,For the ordinate of the aligned quantum sensor, setWhereinFor a width in the physical resolution of the cell image,Is the length in the physical resolution of the cell image.
Specifically, the data synchronization module performs data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set with uniform time stamps, a coordinate system and aligned spaces, so that the multi-mode biophysical data set with uniform time stamps, the coordinate system and aligned spaces can be analyzed later.
Specifically, the SIFT algorithm of the data synchronization module in step A2 includes:
Step A21, detecting the frequency spectrum of the blood molecule vibration after interpolation and the extreme point of the cell image after interpolation through a Gaussian differential pyramid to obtain a stable characteristic point;
Step A22, constructing a feature descriptor according to the stable feature points to obtain a vibration spectrum feature descriptor and a cell image feature descriptor;
step A23, calculating the distance proportion according to the vibration spectrum feature descriptors and the cell image feature descriptors by using a Euclidean distance nearest neighbor algorithm Obtaining the distance proportionProportional distanceComparing with a preset distance ratio db0, judging the distance ratio according to the comparison result, and outputting the characteristic point attribute according to the judgment result, wherein:
When (when) When db0 is not more than, the data synchronization module judges that the distance proportion condition is that the distance proportion is small, and outputs the vibration spectrum feature descriptors and the cell image feature descriptors as matching feature points serving as feature point attributes;
When (when) When db0 is detected, the data synchronization module judges that the distance proportion condition is that the distance proportion is large, and outputs the vibration spectrum feature descriptors and the cell image feature descriptors which are unmatched feature points as feature point attributes;
And step A24, outputting the attribute of the characteristic point as the characteristic point matched with the characteristic point to perform coordinate synchronization, and obtaining the synchronized blood molecule vibration spectrum and the synchronized cell image.
Specifically, the embodiment is not limited to the specific embodiment of detecting the interpolated blood molecular vibration spectrum and the interpolated extreme point of the cell image by the gaussian differential pyramid, for example, D (xp, σ) = (G (k×σ) -G (σ)) ×i (h) may be set by calculating the extreme point D (xp, σ), xp represents the image pixel coordinates, σ represents the gaussian kernel scale parameter, I (h) is the image gray value, I (h) includes the interpolated image gray value of the cell image and the interpolated image gray value of the blood molecular vibration spectrum, I (h) may be acquired by the image processing software, σ=1.6 may be set, k is the adjacent scale multiple, and k=k is setG (sigma) represents a Gaussian kernel function, the embodiment does not limit the specific implementation mode of constructing the feature descriptors according to stable feature points, for example, the feature descriptors are constructed through a descriptor construction method in a SIFT algorithm, wherein the descriptor construction method in the SIFT algorithm refers to a method part for constructing the descriptors in the SIFT algorithm, and the embodiment does not calculate the distance proportion according to the vibration spectrum feature descriptors and the cell image feature descriptors through a Euclidean distance nearest neighbor algorithmDefining the specific calculation mode of (1) such as each vibration spectrum characteristic descriptor dTHz in the interpolated blood molecule vibration spectrum and each cell image characteristic descriptor in the interpolated cell imageWhen calculating the Euclidean distance OD between the two, settingAccording to the blood molecule vibration spectrum characteristic descriptor after interpolationBlood molecule vibration spectrum characteristic descriptor in cell image after interpolation and after interpolationRecent feature descriptorsAnd blood molecular vibration spectrum characteristic descriptors in cell images after interpolation and after interpolationSecond closest feature descriptorComparative distance ratioPerforming calculation and settingThe saidRefers to a blood molecule vibration spectrum characteristic descriptor after interpolationBlood molecular vibration spectrum characteristic descriptors in cell images after interpolation and after interpolationRecent feature descriptorsEuclidean distance between, saidIs the blood molecular vibration spectrum characteristic descriptor after valueBlood molecular vibration spectrum characteristic descriptors in cell images after interpolation and after interpolationSecond closest feature descriptorThe preset distance ratio refers to a preset value for judging the condition of the distance ratio, in this embodiment, specific values of the preset distance ratio are not limited, for example, specific values of the preset distance ratio can be set according to the experience of experts on image data synchronization, db0 is not less than 0.6 and not more than 0.7, in this embodiment, specific embodiments for outputting the attribute of the feature point as the feature point of matching are not limited, for example, transformation matrix H can be solved through the matching point, and h=is setWherein, it is characterized by,) The coordinate Z1 of the characteristic point of the vibration spectrum of the blood molecules after interpolation is calculated,) For the coordinates Z2 of the matching feature points in the cell image,、、、、AndFor affine change parameters, the embodiment does not limit specific values of affine change parameters, for example, it can be assumed that n pairs of matching points exist,,,),For the abscissa of the interpolated cell image matching points,For the ordinate of the interpolated cell image matching points,For the abscissa of the interpolated blood molecule vibration spectrum,For the ordinate of the vibration spectrum of the interpolated blood molecules, the coordinate error of each pair of matching points is: , wherein, 、Representing the abscissa error and the ordinate error, setting a total error function E, wherein the total error function E refers to the square sum of the abscissa error and the ordinate error of all the matching points, and settingSolving parameters by minimizing the total error function E、、、、、I= {1,2,..n }, i is the matching point order.
Specifically, the data synchronization module realizes high-precision cross-modal registration of blood molecular vibration spectrum and cell image through robust feature matching and space coordinate synchronization, and provides a technical foundation for multi-scale biophysical information fusion.
Specifically, the blood data network modeling module builds a DBN model through a DBN model building method, inputs the multi-mode biophysical data set into the DBN model, and obtains a recommended dynamic causal graph output by the DBN model.
The DBN model is a probability graph model which takes a multi-mode biophysical data set as input and takes a recommended dynamic causal graph as output and is used for modeling time sequence data, dynamic causal relation among variables is described through a directed acyclic graph, the embodiment does not limit specific implementation mode of the DBN model construction through a DBN model construction method, nodes of the DBN can be determined according to biological priori knowledge, and the nodes comprise biological entities of blood molecules, blood pressure, blood flow rate, cells, blood pH value and oxygen partial pressure, causal relation among the nodes is established through priori knowledge, the physical causal relation comprises a physical causal chain, a molecular-cell causal chain and a biochemical causal chain, the physical causal chain comprises blood pressure, flow rate, molecular transmission efficiency and cell microenvironment, the biochemical chain comprises oxygen partial pressure, oxygen carrying capacity of red blood cells, anaerobic metabolism of cells, lactic acid accumulation and pH value, and the molecular-cell causal chain comprises inflammatory factors, white blood cell activation and vascular endothelial injury.
Specifically, the blood data network modeling module builds a DBN model through a DBN model building method, inputs the multi-modal biophysical data set into the DBN model building method to obtain a recommended dynamic causal graph, and can deeply mine complex dynamic association and causal logic among multi-modal information in blood data so as to predict the blood change trend.
Specifically, the entropy compression and optimization feature module compresses and optimizes the recommended dynamic causal graph through a compression optimization method to obtain a quantum feature vector value, and the compression optimization method comprises the following steps:
step B1, calculating single feature entropy of the recommended dynamic causal graph to obtain a single feature entropy value;
Step B2, constructing a maximum entropy feature subset according to the single feature entropy value by utilizing a greedy algorithm to obtain a maximum entropy feature subset S;
step B3, performing DNA sequence mapping on the maximum entropy feature subset S by using a DNA sequence mapping method to obtain a DNA sequence;
Step B4, performing quantum optimization on the DNA sequence by using a quantum optimization method to obtain a quantum eigenvector value 。
In particular, the present embodiment is not limited to a specific implementation in which the single feature entropy of the recommendation dynamic causal graph is calculated, e.g., feature vectors according to the recommendation dynamic causal graph∈{Xh、Xw、Xg、Xs、Xq、Xy}、For causal weights of each feature in the dynamic causal graph, recommending dynamic causal graph feature vectorsState value inRecommending dynamic causal graph feature vectorsState value inProbability density of (2)For single feature entropyPerforming calculation and settingThe embodiment does not limit the specific value of the causal weight of each feature in the dynamic causal graph, for example, the clinical data can be set after analysis according to the expert, the clinical data is a multi-dimensional information set which is systematically collected in medical activities and related to the health condition and diagnosis and treatment process of the patient, the Xh is a cell microenvironment feature in the recommended dynamic causal graph, the Xw is an anaerobic metabolism marker feature in the recommended dynamic causal graph, the Xg is an intravascular loss feature in the recommended dynamic causal graph, the Xs is a tissue oxygenation index feature in the recommended dynamic causal graph, the Xq is an inflammation-ischemia scoring feature in the recommended dynamic causal graph, the Xy is a stress time feature in the recommended dynamic causal graph, and the feature vector is thatState value inRefers to the order of the states of the blood data in the recommended dynamic causal graph, which comprises Xh, xw, xg, xs, xq, xy, the feature vectorState value inProbability density of (2)The method refers to recommending the probability of occurrence of the order of the states of the blood data in the dynamic causal graph, the embodiment is not limited to the concrete implementation of constructing the maximum entropy feature subset according to the single feature entropy value by using the greedy algorithm, and a person skilled in the art can set the maximum entropy feature subset according to the actual requirement, for example, initialize the maximum entropy feature subset S, set s=empty set, and calculate all single feature entropySelecting single feature entropyIs greater than a preset characteristic entropy0, And setting 0.8 to less than or equal to 00.Ltoreq.0.9, the embodiment does not limit the specific implementation of mapping the DNA sequence to the maximum entropy feature subset S by using the DNA sequence mapping method, for example, the features in the maximum entropy feature subset S can be normalized to a [0-1] interval to obtain the normalized maximum entropy feature subset S, the [0,1] interval is divided into 4 equal probability intervals corresponding to 4 bases, the 4 bases comprise A, T, C and G, wherein [0,0.25 ]. Fwdarw.A, [0.25,0.5 ]. Fwdarw.T, [0.5, 0.75) -. Fwdarw.C, [0.75,1 ]. Fwdarw.G, and the normalized maximum entropy feature subset S corresponds to 4 bases and is classified to obtain the DNA sequence according to A/T →.Mapping A/T in DNA sequence to ground stateAccording to C/G →Mapping C/G in DNA sequence to ground StateThe Chinese full name of the DNA is deoxyribonucleic acid, and the English full name of the DNA is DeoxyribonucleicAcid.
Specifically, the entropy compression and optimization feature module reserves uncertainty of a biological mechanism through an information entropy theory, strengthens key pathological association by combining causal weights, and finally realizes light weight and high information density storage of a feature library by utilizing DNA high storage density and quantum computation parallelism.
Specifically, the entropy compression and optimization feature module calculates a data compression ratio Ys according to a data size Bs of the recommended dynamic causal graph and a data size By of the DNA sequence, sets ys=100% x Bs/By, compares the data compression ratio Ys with a preset compression ratio Ys0, judges the compression effect condition according to the comparison result, and corrects the data synchronization method according to the judgment result, wherein:
When Ys is more than or equal to Ys0, the entropy compression and optimization feature module judges that the compression effect condition is good, and the data synchronization method is not corrected;
when Ys < Ys0, the entropy compression and optimization feature module determines that the compression effect is poor, corrects the data synchronization method, corrects the preset distance proportion db0 according to the correction coefficient Jz, sets jz=0.87+0.1xe -0.7×(Ys0-Ys), wherein e is the base of natural logarithm, obtains the corrected preset distance proportion db0J, sets db0 j=db0 xjz, replaces the preset distance proportion db0 with the corrected preset distance proportion db0J, and compares the distance proportion db with the corrected preset distance proportion db0J again.
Specifically, the data size of the recommended dynamic causal graph refers to the information data size of the recommended dynamic causal graph, the data size of the DNA sequence refers to the information data size of the DNA sequence, the embodiment does not limit the data size of the recommended dynamic causal graph and the data size of the DNA sequence, for example, the data size of the recommended dynamic causal graph and the data size of the DNA sequence can be obtained through computer background data, the preset compression ratio refers to a preset value for judging the compression effect condition, the embodiment does not limit the specific value of the preset compression ratio, the embodiment can be set by a person skilled in the art according to the actual situation, for example, according to the analysis efficiency requirement on blood data, when the analysis efficiency requirement is high, 90% is less than or equal to 95%, the compression effect condition refers to the condition of the compression effect of the recommended causal graph judged according to the data compression ratio and the preset compression ratio, and the compression effect condition includes that the compression effect condition is good and the compression effect condition is bad.
Specifically, the entropy compression and optimization feature module corrects the data synchronization method by judging the compression effect condition, sets a correction coefficient which is reduced from 0.98 to 0.87 along with the reduction of the data compression ratio, adjusts the preset distance proportion, and reduces the value of the preset distance proportion so as to increase the matching feature point, so that the synchronization accuracy is higher.
In particular, the entropy compression and optimization feature module is based on quantum feature vector valuesAnd the target quantum eigenvector valueSimilarity to quantum characteristicsPerforming calculation and settingSimilarity of quantum characteristicsSimilarity to a presetComparing, judging the similarity condition according to the comparison result, and adjusting the judging process of the compression effect condition according to the judging result, wherein:
When (when) ≥When the entropy compression and optimization feature module judges that the similarity conditions are similar, the judging process of the compression effect conditions is not adjusted;
When (when) <When the entropy compression and optimization feature module judges that the similarity conditions are dissimilar, the judgment process of the compression effect conditions is adjusted, and coefficients are adjustedThe preset compression ratio Ys0 is adjusted and setE is the base of natural logarithm, the preset compression ratio Ys0t after adjustment is obtained, and Ys0 t=ys 0×issetAnd replacing the preset compression ratio Ys0 with the adjusted preset compression ratio Ys0t, and re-comparing the data compression ratio Ys with the adjusted preset compression ratio.
Specifically, the target quantum characteristic vector value refers to a quantum characteristic vector value to be achieved by the optimized DNA sequence, the specific value of the target quantum characteristic vector value is not limited in this embodiment, a person skilled in the art can set the specific value of the target quantum characteristic vector value according to his own requirement, for example, the specific value of the target quantum characteristic vector value is set according to the requirement of calculation efficiency, the preset similarity refers to a preset value for judging the similarity condition, the specific value of the preset similarity is not limited in this embodiment, for example, 0.9 is less than or equal to Ftx0 is less than or equal to 1 according to the requirement of high information density, the similarity condition refers to the similarity condition of the quantum characteristic vector value judged according to the quantum characteristic similarity and the preset similarity, and the similarity condition includes that the similarity condition is similar and the similarity condition is dissimilar.
Specifically, the entropy compression and optimization feature module judges the similarity condition, adjusts the judging process of the compression effect condition according to the judging result so as to improve the subsequent analysis efficiency of blood data, and reduces the value of a preset compression ratio by setting an adjusting coefficient which is reduced from 0.99 to 0.85 along with the reduction of the quantum feature similarity, so that the compression ratio is reduced, the similarity of the data is improved, the situation that the analysis of the blood data is inaccurate in the subsequent analysis is avoided, and the accuracy of the blood data analysis is improved.
Specifically, the digital twin body modeling module constructs a twin body model through a twin body model constructing method to obtain a twin body model, and inputs quantum characteristic vector values into the twin body model to obtain a predicted quantum characteristic vector value state S (t) output by the twin body model.
Specifically, the twin body model refers to a machine learning model with quantum characteristic vector values as input and predicted quantum characteristic vector value states as output, the predicted quantum characteristic vector value state S (t) refers to the quantum characteristic vector value state at the time t, the specific implementation method of the twin body model construction method is not limited in this embodiment, and the twin body model can be set by a person skilled in the art according to actual situations, for example, according to a state transition matrix Aj, a dosing vector u (t), a drug action matrix Bj, and random noise(T) setting the state transition equation S (t+1)S (t) = [ S1 (t), S2 (t),. The term sd (t) ], S (t) refers to the total collection of quantum eigenvector states at the time t, S1 (t) refers to the quantum eigenvector state at the first time, S2 (t) refers to the quantum eigenvector state at the second time, sd (t) refers to the quantum eigenvector state at the d-th time, d is the predicted node order of the quantum eigenvector states at the time t, S (t+1) refers to the total collection of quantum eigenvector states at the time t+1, the embodiment does not limit the acquisition modes of the state transition matrix and the drug action matrix, for example, the state transition matrix and the drug action matrix can be acquired through statistical fitting of clinical data, the embodiment does not limit the acquisition modes of the drug administration vector, for example, the drug administration vector can be acquired through the conversion of the drug dose in the doctor advice into the vector form, the embodiment does not limit the acquisition modes of random noise, for example, the random noise is determined by the doctor, the clinical data refers to the medical doctor advice system, the medical advice collected in the medical activity system, the medical advice information about the medical advice of the patient, and the medical advice of the patient is the medical advice of the medical condition according to the medical advice.
Specifically, the digital twin body modeling module inputs the quantum characteristic vector value into a twin body model to obtain a predicted quantum characteristic vector value state S (t) output by the twin body model, and predicts the future change trend of blood data so as to recommend the administration time later.
Specifically, the digital twin body modeling module constructs an A3C algorithm through an A3C algorithm model construction method to obtain an A3C algorithm model, inputs a predicted quantum eigenvector value state S (t) into the A3C algorithm model, acquires recommended administration time output by the A3C algorithm model, and sends the recommended administration time to a medical data terminal.
Specifically, the A3C algorithm model refers to a machine learning model with a predicted quantum feature vector value state S (t) as input and a recommended administration time as output, and the embodiment does not limit a specific implementation method of an A3C algorithm model building method, for example, an A3C algorithm model may be built by using an A3C algorithm as a basic frame of the A3C algorithm model, the A3C algorithm refers to a distributed reinforcement learning frame based on a strategy gradient, efficient exploration and strategy optimization are realized through asynchronous parallel training, and the embodiment does not limit a specific implementation mode of sending the recommended administration time to a medical data terminal, for example, the recommended administration time may be sent to the medical data terminal through a wireless signal, and the medical data terminal refers to a terminal for analyzing and processing medical data.
Specifically, the digital twin body modeling module outputs the recommended administration time and sends the recommended administration time to the medical data terminal, so that the calculation load of the medical data terminal is reduced, and the medical data processing efficiency is effectively improved.
Referring to fig. 2, a flow chart of a method for analyzing artificial intelligence blood data according to the present embodiment is shown, the method includes:
Step S1, collecting blood data;
step S2, carrying out data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set;
S3, inputting the multi-mode biophysical data set into a pre-constructed DBN model to obtain a recommended dynamic causal graph output by the DBN model;
s4, compressing and optimizing the recommended dynamic causal graph through a compression optimization method to obtain a quantum characteristic vector value, judging the compression effect condition according to the data compression ratio, correcting the data synchronization method according to the judgment result, and adjusting the judgment process of the compression effect condition according to the quantum characteristic similarity;
And S5, inputting the quantum characteristic vector value into a pre-constructed twin body model to obtain a predicted quantum characteristic vector value state, inputting the predicted quantum characteristic vector value state into a pre-constructed A3C algorithm model to obtain recommended administration time, and transmitting the recommended administration time to a medical data terminal.
Thus far, the technical solution of the present invention has been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of protection of the present invention is not limited to these specific embodiments. Equivalent modifications and substitutions for related technical features may be made by those skilled in the art without departing from the principles of the present invention, and such modifications and substitutions will be within the scope of the present invention.
Claims (4)
1. A system for artificial intelligence blood data analysis, comprising:
the data acquisition module is used for acquiring blood data;
The data synchronization module is used for carrying out data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set;
The blood data network modeling module is used for inputting the multi-mode biophysical data set into a pre-constructed DBN model to obtain a recommended dynamic causal graph output by the DBN model;
the entropy compression and optimization feature module is used for compressing and optimizing the recommended dynamic causal graph through a compression optimization method to obtain a quantum feature vector value, judging the compression effect condition according to the data compression ratio, correcting the data synchronization method according to the judging result, and adjusting the judging process of the compression effect condition according to the quantum feature similarity;
the digital twin body modeling module is used for inputting the quantum characteristic vector value into a pre-built twin body model to obtain a predicted quantum characteristic vector value state, inputting the predicted quantum characteristic vector value state into a pre-built A3C algorithm model to obtain recommended administration time, and sending the recommended administration time to the medical data terminal;
the data synchronization module performs data synchronization on blood data through a data synchronization method to obtain a multi-mode biophysical data set, and the data synchronization method comprises the following steps:
A1, carrying out interpolation processing on blood data by using a linear interpolation algorithm to obtain interpolated blood data, wherein the interpolated blood data comprises an interpolated blood molecular vibration spectrum, an interpolated blood pressure, an interpolated blood flow velocity data, an interpolated cell image, an interpolated blood pH value and an interpolated oxygen partial pressure;
a2, carrying out coordinate space synchronization on the interpolated blood molecule vibration spectrum and the interpolated cell image by using a SIFT algorithm to obtain a synchronized blood molecule vibration spectrum and a synchronized cell image;
A3, carrying out coordinate space synchronization on the interpolated blood pH value and the interpolated oxygen partial pressure and the synchronized cell image by using a coordinate synchronization method to obtain the synchronized blood pH value and the synchronized oxygen partial pressure;
step A4, outputting the synchronized blood molecular vibration spectrum, the interpolated blood pressure, the interpolated blood flow velocity data, the synchronized cell image, the synchronized blood pH value and the synchronized oxygen partial pressure as a multi-mode biophysical data set to obtain a multi-mode biophysical data set;
The SIFT algorithm of the data synchronization module in step A2 includes:
Step A21, detecting the frequency spectrum of the blood molecule vibration after interpolation and the extreme point of the cell image after interpolation through a Gaussian differential pyramid to obtain a stable characteristic point;
Step A22, constructing a feature descriptor according to the stable feature points to obtain a vibration spectrum feature descriptor and a cell image feature descriptor;
step A23, calculating the distance proportion according to the vibration spectrum feature descriptors and the cell image feature descriptors by using a Euclidean distance nearest neighbor algorithm Obtaining the distance proportionComparing the distance ratio db with a preset distance ratio db0, judging the distance ratio according to the comparison result, and outputting the characteristic point attribute according to the judgment result, wherein:
When (when) When db0 is not more than, the data synchronization module judges that the distance proportion condition is that the distance proportion is small, and outputs the vibration spectrum feature descriptors and the cell image feature descriptors as matching feature points serving as feature point attributes;
When (when) When db0 is detected, the data synchronization module judges that the distance proportion condition is that the distance proportion is large, and outputs the vibration spectrum feature descriptors and the cell image feature descriptors which are unmatched feature points as feature point attributes;
Step A24, outputting the attribute of the characteristic point as the characteristic point matched with the characteristic point to perform coordinate synchronization, so as to obtain a synchronized blood molecule vibration spectrum and a synchronized cell image;
the blood data network modeling module builds a DBN model through a DBN model building method, and inputs the multi-mode biophysical data set into the DBN model to obtain a recommended dynamic causal graph output by the DBN model;
The entropy compression and optimization feature module compresses and optimizes the recommended dynamic causal graph through a compression optimization method to obtain a quantum feature vector value, and the compression optimization method comprises the following steps:
step B1, calculating single feature entropy of the recommended dynamic causal graph to obtain a single feature entropy value;
Step B2, constructing a maximum entropy feature subset according to the single feature entropy value by utilizing a greedy algorithm to obtain a maximum entropy feature subset S;
step B3, performing DNA sequence mapping on the maximum entropy feature subset S by using a DNA sequence mapping method to obtain a DNA sequence;
Step B4, performing quantum optimization on the DNA sequence by using a quantum optimization method to obtain a quantum eigenvector value ;
The entropy compression and optimization feature module calculates a data compression ratio Ys according to a data size Bs of the recommended dynamic causal graph and a data size By of the DNA sequence, sets ys=100% x Bs/By, compares the data compression ratio Ys with a preset compression ratio Ys0, judges the compression effect condition according to a comparison result, and corrects the data synchronization method according to a judgment result, wherein:
When Ys is more than or equal to Ys0, the entropy compression and optimization feature module judges that the compression effect condition is good, and the data synchronization method is not corrected;
when Ys < Ys0, the entropy compression and optimization feature module determines that the compression effect is poor, corrects the data synchronization method, corrects a preset distance proportion db0 according to a correction coefficient Jz, sets jz=0.87-0.11×e -0.7×(Ys0-Ys), e is the base of natural logarithm, obtains a corrected preset distance proportion db0J, sets db0 j=db0×jz, replaces the preset distance proportion db0 with the corrected preset distance proportion db0J, and replaces the distance proportion db0 with the corrected preset distance proportion db0J And re-comparing with the corrected preset distance proportion db 0J.
2. The system for artificial intelligence blood data analysis according to claim 1, wherein the entropy compression and optimization feature module is based on quantum feature vector valuesAnd the target quantum eigenvector valueSimilarity to quantum characteristicsPerforming calculation and settingSimilarity of quantum characteristicsSimilarity to a presetComparing, judging the similarity condition according to the comparison result, and adjusting the judging process of the compression effect condition according to the judging result, wherein:
When (when) ≥When the entropy compression and optimization feature module judges that the similarity conditions are similar, the judging process of the compression effect conditions is not adjusted;
When (when) <When the entropy compression and optimization feature module judges that the similarity conditions are dissimilar, the judgment process of the compression effect conditions is adjusted, and coefficients are adjustedThe preset compression ratio Ys0 is adjusted and setObtaining the adjusted preset compression ratio Ys0t, setting Ys0 t=ys 0×And replacing the preset compression ratio Ys0 with the adjusted preset compression ratio Ys0t, and re-comparing the data compression ratio Ys with the adjusted preset compression ratio.
3. The system for analyzing artificial intelligence blood data according to claim 2, wherein the digital twin modeling module constructs a twin model by a twin model constructing method to obtain a twin model, and inputs the quantum characteristic vector value into the twin model to obtain a predicted quantum characteristic vector value state S (t) output by the twin model.
4. The system for analyzing artificial intelligence blood data according to claim 3, wherein the digital twin modeling module constructs an A3C algorithm by an A3C algorithm model constructing method to obtain an A3C algorithm model, inputs a predicted quantum eigenvector value state S (t) into the A3C algorithm model, obtains a recommended administration time output by the A3C algorithm model, and transmits the recommended administration time to the medical data terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510981497.8A CN120473079B (en) | 2025-07-16 | 2025-07-16 | System and method for analyzing artificial intelligent blood data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510981497.8A CN120473079B (en) | 2025-07-16 | 2025-07-16 | System and method for analyzing artificial intelligent blood data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN120473079A CN120473079A (en) | 2025-08-12 |
| CN120473079B true CN120473079B (en) | 2025-09-30 |
Family
ID=96637349
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510981497.8A Active CN120473079B (en) | 2025-07-16 | 2025-07-16 | System and method for analyzing artificial intelligent blood data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120473079B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120258602A (en) * | 2025-03-20 | 2025-07-04 | 华能应城热电有限责任公司 | A coal-fired power plant safety monitoring system and method |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP4208812A4 (en) * | 2020-09-02 | 2024-12-25 | The General Hospital Corporation | Methods for identifying cross-modal features from spatially resolved data sets |
| DE102023108748A1 (en) * | 2023-04-05 | 2024-10-10 | RobotDreams® GmbH | Computer-implemented method for determining in vivo and in vitro conditions by analyzing blood parameters measured in a hematological analyzer |
| WO2025076537A1 (en) * | 2023-10-05 | 2025-04-10 | Landscan Llc | Digital twins for precise agricultural management |
| US12423283B2 (en) * | 2023-12-12 | 2025-09-23 | AtomBeam Technologies Inc. | Unified system for multi-modal data compression with relationship preservation and neural reconstruction |
-
2025
- 2025-07-16 CN CN202510981497.8A patent/CN120473079B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120258602A (en) * | 2025-03-20 | 2025-07-04 | 华能应城热电有限责任公司 | A coal-fired power plant safety monitoring system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN120473079A (en) | 2025-08-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110444263B (en) | Disease data processing method, device, equipment and medium based on federal learning | |
| CN118471542B (en) | Medical health management system based on big data | |
| US12260623B2 (en) | Training method and apparatus for image region segmentation model, and image region segmentation method and apparatus | |
| Deng et al. | On-line pattern analysis by evolving self-organizing maps | |
| WO2023143628A1 (en) | Diabetic retinopathy detection method based on genetic fuzzy tree and deep network | |
| CN116188435A (en) | A Method of Depth Segmentation of Medical Image Based on Fuzzy Logic | |
| CN106021524B (en) | Working method of second-order dependency tree augmented Bayes classifier for big data mining | |
| CN113781385A (en) | A joint attention map convolution method for automatic classification of medical images of the brain | |
| CN119337313B (en) | A multimodal data fusion intelligent decision-making system | |
| Ye et al. | Self-supervised cross-modal visual retrieval from brain activities | |
| CN111402278A (en) | Segmentation model training method, image labeling method and related device | |
| CN120198671A (en) | A multi-scale cardiac image segmentation method and system based on graph neural network | |
| CN120473079B (en) | System and method for analyzing artificial intelligent blood data | |
| CN118629596A (en) | System and method for analyzing psychological state of patients with multiple myeloma and cardiac amyloidosis | |
| Chatterjee et al. | A survey on techniques used in medical imaging processing | |
| CN120636522A (en) | Deep learning-driven spatial multi-omics data analysis and integration methods | |
| CN117764972B (en) | Noninvasive armpit metastasis load prediction method and system for breast cancer early-stage patients | |
| CN117009839B (en) | Patient clustering method and device based on heterogeneous hypergraph neural network | |
| CN116721065B (en) | DDH ultrasonic image analysis method based on contrast learning pre-training | |
| CN119418402A (en) | Human behavior recognition method based on prototype neural network with fusion attention mechanism | |
| CN111860618B (en) | Bidirectional GAN model for pathological data conversion and construction and application methods thereof | |
| CN116630777A (en) | A Completion Method for Missing Data of Medical Indexes Based on Information Bottleneck Graph Compression | |
| Wang et al. | Color tongue image segmentation using fuzzy Kohonen networks and genetic algorithm | |
| CN115797675A (en) | Artificial intelligence image processing method and system | |
| CN120163828B (en) | Gastric cancer pathological image processing method and gastric cancer pathological image processing system based on multitask learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |