A method and related device for identifying the material composition based on single-energy CT images

Abstract

The present invention provides a method and related device for identifying material components based on single-energy CT images, which are applied to the technical field of data processing. In this application, each slice image of the single-energy CT image is grouped to generate a number of single-energy CT slice image groups; the initial material component recognition model is processed based on the training sample set to generate a target material component recognition model; the CT image of the target object is subjected to semantic segmentation processing to generate a number of image regions and image region types matching the number of image regions; each pixel in the number of image regions is processed separately to generate a characteristic parameter value corresponding to each pixel; the characteristic parameter values of each pixel in the number of image regions and the image region types matching the number of image regions are processed to generate material component information corresponding to each pixel region; the material component information corresponding to each pixel region is processed to generate attribute information of the pixel region.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a method for identifying material components based on a single-energy CT image and related devices. Background Art

[0002] CT technology has been widely used in various fields. However, currently, it mainly focuses on obtaining three-dimensional structural information (i.e., structural imaging) inside an object. Material identification and multi-parameter imaging (such as relative electron density, equivalent atomic number, relative ray tissue ability) are important development directions of CT functional imaging, and there are clear requirements in applications and research such as medical diagnosis, security inspection, and radiotherapy. The pixel values of clinical CT images are usually represented in the form of HU values. The HU value is a physical quantity related to X-ray energy, material density, and atomic number. Therefore, in single-energy CT, different materials may show the same or nearly the same HU value due to their different material densities and atomic numbers. By collecting two or more groups of X-rays with different energies, two or more groups of CT images at different energies can be reconstructed respectively. By virtue of the different numerical values of the HU value in CT images with different energies, the ambiguity that may exist in ordinary CT images can be eliminated to a certain extent, and material identification can be achieved.

[0003] The core scientific assumption of traditional material component modeling methods: It is assumed that material components can be approximately expressed by a certain combination of a pair of basic materials (water and iodine) and a pair of certain basis functions (approximate photoelectric effect, approximate Compton scattering response), that is, bone is considered to be composed of a certain content of water and a certain content of iodine, or a certain degree of photoelectric effect function and a certain degree of Compton effect function. Air is considered to be composed of water with a content of 0 and iodine with a content of 0. This method has great limitations. For example, material component modeling completely depends on prior assumptions, and its objective rationality is questionable. At the same time, the component quantification algorithm highly depends on experimental parameters, the reliability of the material component modeling effect is relatively poor, and the anti-noise performance is relatively poor.

[0004] In recent years, spectral CT based on dual X-ray sources, rapid switching of high and low voltages of X-ray sources, multi-layer detectors, and photon technology detectors has entered clinical use. By finely modeling the X-ray energy spectrum of the X-ray source, spectral CT can obtain a group (at least two) of single-energy CT images corresponding to different energies. However, currently, no material component quantification method based on single-energy CT images has been seen.

[0005] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The purpose of the present application is to provide a method for modeling the material composition based on single - energy CT images and related devices, which can at least overcome the problems existing in the prior art to a certain extent. Through feature parameter extraction, the purpose of improving the material composition discrimination ability and multi - parameter imaging accuracy is achieved.

[0007] Other features and advantages of the present application will become apparent from the following detailed description, or will be learned in part through the practice of the present invention.

[0008] According to one aspect of the present application, a method for modeling the material composition based on single - energy CT images is provided, including: obtaining a CT image of a target object, where the CT image of the target object includes at least one set of single - energy CT images; performing grouping processing on each slice image of the single - energy CT images to generate a number of single - energy CT slice image groups; obtaining a training sample set and an initial material composition recognition model, where the training sample set includes single - energy CT slice image groups corresponding to known reference materials, and the initial material composition recognition model is composed of the change trends of CT image values of different reference materials at different energy levels. The initial material composition recognition model includes reference values for characterizing the characteristic parameters corresponding to the material composition and their corresponding parameter spaces; processing the initial material composition recognition model based on the training sample set to generate a target material composition recognition model; performing semantic segmentation processing on the CT image of the target object to generate a number of image regions and image region types matching the number of image regions; processing each pixel in a number of image regions based on the target material composition recognition model to generate a characteristic parameter value corresponding to each pixel; processing the characteristic parameter value of each pixel in a number of image regions and the image region types matching the number of image regions based on the target material composition recognition model to generate material composition information corresponding to each pixel region; processing the material composition information corresponding to each pixel region to generate attribute information of the pixel region, where the attribute information of the pixel region includes the relative electron density, equivalent atomic number, and relative ray tissue ability of each pixel.

[0009] In an embodiment of the present application, processing each pixel in a number of image regions based on the target material composition recognition model to generate a characteristic parameter value corresponding to each pixel includes:

[0010] The target material composition recognition model includes a calculation formula for obtaining the characteristic parameter value, and the calculation formula is:

[0011] CTHU(keV)=a·e b·keV +c;

[0012] a=ω1·a bone +ω2·a fat ;

[0013] b = ω1·b bone + ω2·b fat ;

[0014] c = ω1·c bone + ω2·c fat ;

[0015] ω1 + ω2 = 1;

[0016] Wherein, a, b, and c represent characteristic parameters, e represents the exponential function, keV represents the energy level corresponding to the single-energy CT image, x1 and x2 are reference substances respectively, and ω1 and ω2 are the content information of the reference substances respectively.

[0017] In an embodiment of the present application, based on the target substance composition recognition model, the characteristic parameter values of each pixel in a number of image regions and the image region types matching the number of image regions are processed to generate the substance composition information corresponding to each pixel region, including:

[0018] The target substance composition recognition model includes a calculation formula for obtaining substance composition information, and the calculation formula is:

[0019]

[0020] Wherein, p x,y,z represents the substance composition information corresponding to the pixel at the coordinate position x, y, z, ω i represents the content of the i-th reference substance, M i represents the i-th reference substance, and n represents the number of reference substances constituting the substance composition of the pixel.

[0021] In an embodiment of the present application, the target substance composition recognition model includes obtaining multi-parameter information such as relative electron density, equivalent atomic number, and relative ray tissue ability, and the calculation formulas are respectively:

[0022] The target substance composition recognition model includes a calculation formula for obtaining relative electron density, and the calculation formula is:

[0023]

[0024] Wherein, RED x,y,z represents the relative electron density corresponding to the pixel at the coordinate position x, y, z, ω i represents the content of the i-th reference substance, i represents the i-th reference substance, ρ i represents the density of the i-th reference substance, ρ water represents the density of water, RED irepresents the electron density of the i-th reference substance;

[0025] The target substance component recognition model includes a calculation formula for obtaining the effective atomic number, and the calculation formula is:

[0026]

[0027] where Zeff x,y,z represents the effective atomic number corresponding to the pixel at the coordinate position x, y, z, ω i represents the content of the i-th reference substance, i represents the i-th reference substance, and Z i represents the effective atomic number of the i-th reference substance;

[0028] The target substance component recognition model includes a calculation formula for obtaining the relative ray tissue ability, and the calculation formula is:

[0029]

[0030] where SPR represents the relative ray tissue ability and RED represents the electron density of the reference substance.

[0031] In an embodiment of the present application, the initial substance component recognition model is processed based on a training sample set to generate a target substance component recognition model, including: preprocessing the training sample set to generate a training sample set with identification information, where the training sample set includes a number of standard substance information; training the initial substance component recognition model based on the training sample set with identification information to generate a target substance component recognition model.

[0032] In an embodiment of the present application, preprocessing the training sample set to generate a training sample set with identification information includes: extracting features from the training sample set to determine an original feature library; dividing each feature data set according to the original feature library to generate a training data set and a test data set; using a classifier to predict each test data set divided from the original feature library to determine a prediction result; using a preset algorithm to train each training data set divided from the original feature library to obtain a test set class prediction result; generating a training sample with identification information according to the prediction result and the test set class prediction result.

[0033] In an embodiment of the present application, extracting features from the training sample set to determine an original feature library includes: processing the standard substance information based on a preset processing rule to generate a standardized feature, where the standardized feature is a feature of a sample or phantom with a clear and known substance composition; processing the standardized feature based on a preset feature screening and dimensionality reduction rule to generate an original feature; generating an original feature library from a number of original features.

[0034] Another aspect of the present application is a substance composition recognition device based on a single - energy - level CT image, which is characterized by including: an acquisition module, configured to acquire a CT image of a target object, wherein the CT image of the target object includes at least one set of single - energy - level CT images; acquire a training sample set and an initial substance composition recognition model, wherein the training sample set includes a set of single - energy - level CT slice images corresponding to known reference substances, and the initial substance composition recognition model is constituted based on the change trends of CT image values of different reference substances at different energy levels. The initial substance composition recognition model includes reference values for characterizing the characteristic parameters corresponding to the substance composition and their corresponding parameter spaces; a processing module, configured to perform grouping processing on each slice image of the single - energy - level CT image to generate several sets of single - energy - level CT slice images; process the initial substance composition recognition model based on the training sample set to generate a target substance composition recognition model; perform semantic segmentation processing on the CT image of the target object to generate several image regions and image region types matching the several image regions; process each pixel within the several image regions based on the target substance composition recognition model to generate a characteristic parameter value corresponding to each pixel; process the characteristic parameter value of each pixel within the several image regions and the image region types matching the several image regions based on the target substance composition recognition model to generate substance composition information corresponding to each pixel region; process the substance composition information corresponding to each pixel region to generate attribute information of the pixel region, wherein the attribute information of the pixel region includes the relative electron density, equivalent atomic number, and relative ray tissue ability of each pixel.

[0035] According to yet another aspect of the present application, an electronic device is provided, which is characterized by including: a first processor; and a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the above - mentioned substance composition recognition method based on a single - energy - level CT image by executing the executable instructions.

[0036] According to yet another aspect of the present application, a computer - readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a second processor, the above - mentioned substance composition recognition method based on a single - energy - level CT image is implemented.

[0037] According to yet another aspect of the present application, a computer program product is provided, including a computer program, which is characterized in that when the computer program is executed by a third processor, the above - mentioned substance composition recognition method based on a single - energy - level CT image is implemented.

[0038] A method and related device for identifying material components based on a single-energy CT image provided by this application. The server obtains the single-energy CT image of the target object and groups it. Then, using the CT slice image group of known materials as training samples, combined with the initial material component recognition model, a target material component recognition model is obtained through machine learning training. Next, semantic segmentation is performed on the CT image to identify different regions and their types. The trained model is used to analyze the characteristic parameter values of each pixel, and then the material component information is determined. Finally, the attribute information of each pixel region, such as relative electron density, equivalent atomic number, and relative ray tissue ability, is calculated to comprehensively describe the material attributes. Based on the new material component modeling method of the single-energy CT image formed by spectral CT, through feature parameter extraction, the ability to distinguish material components and the accuracy of multi-parameter imaging are improved.

[0039] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and do not limit the present disclosure. Brief Description of the Drawings

[0040] Figure 1 A flowchart showing a method for identifying material components based on a single-energy CT image provided by an embodiment of this application;

[0041] Figure 2 A schematic structural diagram showing a device for identifying material components based on a single-energy CT image provided by an embodiment of this application;

[0042] Figure 3 A schematic structural diagram showing an electronic device provided by an embodiment of this application;

[0043] Figure 4 A schematic diagram showing a storage medium provided by an embodiment of this application;

[0044] Figure 5 A schematic diagram showing the fitting of a part of CTHU>0 material provided by an embodiment of this application;

[0045] Figure 6 A schematic diagram showing a part of CTHU < 0 material fitting diagram. Detailed Embodiments

[0046] The following describes the preferred embodiments of the present invention with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0047] The following is combined with Figure 1A method for identifying the material composition based on a single - energy CT image according to an exemplary embodiment of the present application is described. It should be noted that the following application scenarios are only shown for the convenience of understanding the spirit and principle of the present application, and the embodiments of the present application are not limited in this regard. On the contrary, the embodiments of the present application can be applied to any applicable scenario.

[0048] In one embodiment, the present application also proposes a method for identifying the material composition based on a single - energy CT image and related devices. Figure 1 Schematically shown is a flowchart of a method for identifying the material composition based on a single - energy CT image according to an embodiment of the present application. As Figure 1 shown, this method is applied to a server and includes:

[0049] S101, obtaining a CT image of a target object.

[0050] In one embodiment, the CT image of the target object includes at least one set of single - energy CT images. For example, by collecting twice successively, using Siemens dual - source, GE high - and low - energy rapid alternating switching, or Philips double - layer detectors, high - and low - energy CT data can be collected to form a series of single - energy CT images. Among them, the single - energy CT images include, but are not limited to, high - single - energy CT images and low - single - energy CT images, which respectively simulate CT images at different energies. The energy span generally ranges from 50 keV to 140 keV.

[0051] S102, performing grouping processing on each slice image of the single - energy CT image to generate a number of single - energy CT slice image groups.

[0052] In one embodiment, a preset size value is obtained to process the single - energy CT image to generate a single - energy CT image of a preset size. The single - energy CT image is subjected to multi - layer slicing processing to generate multiple slice image groups of different depth levels. The slice image groups are processed based on time sequence or attribute information to generate a number of slice images. The grouping processing is performed on each slice image of the single - energy CT image in the order of energy priority, and corresponding slice images can be generated in sequence according to the order or type of the slices, that is, tomographic images corresponding to the same spatial position. Different - depth - level slice images are obtained based on different scale information, thereby generating multiple slice image groups of different depth levels, and then obtaining the same image at different scales. For example, from a macroscopic scale or a microscopic scale, the same image is seen. The scale from small to large is the process of seeing fewer details of the objective object.

[0053] S103, obtaining a training sample set and an initial material composition recognition model.

[0054] In one implementation, the training sample set includes a group of single-energy CT slice images corresponding to known reference substances. The initial substance composition recognition model is constructed based on the variation trends of CT image values of different reference substances at different energy levels. The initial substance composition recognition model includes reference values for characterizing the characteristic parameters corresponding to the substance composition and their corresponding parameter spaces. Obtain the group of single-energy CT slice images of several reference substances with known compositions, and establish an initial substance composition recognition model that matches the single-energy CT images according to the variation trends of CT image values (CTHU values) of different reference substances at different energy levels.

[0055] Obtain an initial substance composition recognition model and a training sample set for training the initial substance composition recognition model. Among them, the training sample set includes several standard substance information; preprocess the training sample set to generate a training sample set with identification information, where the identification information is used to characterize the attributes of various types of substance information. Train the initial substance composition recognition model based on the training sample set with identification information to generate a target substance composition recognition model. Specifically, obtain the data classification results corresponding to the training data set. Since the classification results of each training data set can be determined in advance and can be directly obtained from the outside. Compare the prediction results with the data classification results to determine the first comparison result; compare the prediction results of the test set class with the data classification results to determine the second comparison result; when judging whether the second comparison result and the first comparison result meet the preset requirements, when they meet the preset requirements, it indicates that the detection result of the current substance composition recognition model is relatively accurate. At this time, the current substance composition recognition model can be used as the target substance composition recognition model. The substance composition recognition model includes a reference parameter space for characterizing the distribution position of the characteristic parameters corresponding to the substance composition.

[0056] In addition, the standard substance information is a sample or phantom with a clear and known substance composition, including pure water (H2O), CaCO3 with different densities, fat, muscle, bones with different compactness, sodium iodide aqueous solutions with different densities, and common medical implant metals (such as stainless steel, titanium).

[0057] S104, process the initial substance composition recognition model based on the training sample set to generate a target substance composition recognition model.

[0058] In one implementation, preprocess the training sample set to generate a training sample set with identification information, where the training sample set includes several standard substance information, and train the initial substance composition recognition model based on the training sample set with identification information to generate a target substance composition recognition model.

[0059] Specifically, information on reference materials is collected. This information may include the chemical composition, physical properties, source, etc. of the materials, and these data can be obtained from the national reference material resource sharing platform. Data preprocessing is a crucial step in generating the training sample set, including data cleaning (removing incorrect and inconsistent data), data transformation (converting the data into a format suitable for model input), and feature engineering (selecting and constructing features required for model training). To improve the generalization ability of the model, data augmentation techniques are used to expand the training set. For example, the diversity of samples can be increased by means of rotation, scaling, color transformation, etc. In the research on the identification of organic and inorganic substances, the dataset is augmented by adding gray borders to the shorter side to make the pictures square and rotating the angles to ensure that each sample has correct identification information, and these identifications will be used as labels during model training. The annotation should be accurate to avoid the model learning incorrect information.

[0060] The dataset is divided into a training set, a validation set, and a test set. The training set is used to train the model, the validation set is used for model selection and hyperparameter tuning, and the test set is used for the final evaluation of the model performance. Through training and validation, a target substance composition identification model is obtained, which can accurately identify and classify different substance compositions. Remember to monitor and record the performance metrics of the model, such as accuracy, recall, and F1 score, during the model training process to ensure the effectiveness and reliability of the model.

[0061] In another implementation, feature extraction is performed on the training sample set to determine the original feature library; each feature dataset is divided according to the original feature library to generate a training dataset and a test dataset; a classifier is used to predict each test dataset divided from the original feature library to determine the prediction results; a preset algorithm is used to train each training dataset divided from the original feature library to obtain the prediction results of the test set class; and training samples with identification information are generated according to the prediction results and the prediction results of the test set class.

[0062] Feature extraction is performed on the training data set to determine the original feature library. Feature extraction includes four features: original features, statistical features, frequency domain features, and time domain features. The commonly used features extracted are specifically introduced as follows: Original features: Original features are the collected data, which are preprocessed and use all the information in each sample data matrix for model training. In order not to lose sample information, each curve of sample Xi is directly unfolded and stretched into a row vector. Statistical features: Statistical features consider the change trend of all data points on the curve. By extracting statistical features, the dimension of data samples can be reduced, which is convenient for analyzing genetic data and accelerating the convergence speed during model training. The extracted statistical features include maximum value, minimum value, mean, variance, and standard deviation. Frequency domain features: Wavelet transform is performed on the data set, and the coefficients obtained by the second-order wavelet transform are used as new features to form frequency domain features. Time domain features: Time domain features are mainly based on the time perspective to invent the changing laws of signals and systems. Time domain features can reflect the information of curve data in time changes. This process mainly extracts the first-order forward difference of the data set to form the first-order difference time domain features, and uses the exponential moving average feature processing method.

[0063] In another embodiment, the standard material information is processed based on preset processing rules to generate standardized features, wherein the standardized features are features of samples or models whose material composition is clear and known, and the standardized features are processed based on preset feature screening and dimensionality reduction rules to generate original features, and a number of original features are used to generate an original feature library.

[0064] The process of processing the standard material information based on the preset processing rules and generating standardized features involves data cleaning, conversion and normalization to ensure the consistency and comparability of the data. Specifically, the data is cleaned, including removing or correcting outliers, missing values ​​and erroneous data to ensure that the information of all substances is up-to-date and accurate. Key features are extracted from the detailed information of the standard material, such as chemical composition, concentration, physical state, spectral characteristics, etc. The features are standardized to have a uniform dimension and range, such as using Z-score standardization or minimum-maximum normalization. Standardized features refer to the features of samples or models with clear and known material composition. According to the preset feature screening rules, the most informative features for material composition identification are selected. Statistical tests, model selection or knowledge-based rules can be used to determine which features should be retained. Dimensionality reduction techniques such as principal component analysis (PCA), linear discriminant analysis (LDA) or autoencoders are applied to reduce the number of features while retaining the key information in the original data as much as possible.

[0065] Collect the screened and dimension-reduced features to form an original feature library, which will serve as the basis for training a machine learning model. Maintain the feature library, including regularly updating and optimizing the feature set to adapt to new data and discoveries, ensuring the quality and relevance of the feature library to support effective identification of material components.

[0066] S105, perform semantic segmentation on the CT image of the target object to generate a number of image regions and image region types matching the number of image regions.

[0067] In one implementation, perform HU value threshold segmentation based on the 70keV monoenergetic CT image to segment the bone, lung, and soft tissue regions. When performing HU value threshold segmentation on the 70keV monoenergetic CT image, it is usually necessary to set appropriate thresholds according to the specific HU value ranges exhibited by different human tissues in the CT image. The following are the thresholds for bone, lung, and soft tissue region segmentation:

[0068] The CT value of bone is usually higher than 400HU. Therefore, for bone segmentation, a relatively high HU value can be selected as the upper threshold, such as 400HU or higher. The lower threshold can be set relatively low, such as -1000HU, to ensure that only the densest part of the bone is selected.

[0069] The CT value of the lung is usually relatively low because the lung tissue is filled with air. The HU value of the lung is roughly between -1000HU and -200HU. For lung segmentation, -500HU or -600HU can be selected as the lower threshold, and the upper threshold can be set at -200HU or higher to include all lung tissues.

[0070] The CT value of soft tissue is usually between -50HU and 100HU. For soft tissue segmentation, a relatively low lower threshold, such as -50HU, and a moderate upper threshold, such as 100HU, can be selected.

[0071] In actual operation, fine-tuning may be required according to the specific CT image and the required segmentation accuracy. For example, according to the research in the literature, when identifying the solid component of a subsolid nodule in the lung, the threshold set at -250HU has the greatest diagnostic value. For lung nodule segmentation, it may be necessary to combine the HU value range of the lung and the characteristics of the nodule to determine the optimal threshold. In addition, factors such as image noise, artifacts, and partial volume effects may need to be considered during the segmentation process, and these factors may affect the threshold selection and the accuracy of the segmentation result. In some cases, it may be necessary to combine multiple segmentation techniques, such as thresholding, region growing, level set segmentation, etc., as well as deep learning methods, to improve the accuracy and robustness of the segmentation.

[0072] In addition, refer toFigure 5 and Figure 6 It can be seen that the pixel value (i.e., CTHU value) of a single - energy CT slice image characterizes the attenuation ability of the substance in the pixel to X - rays. Different single - energy CT images represent the distribution of the attenuation ability of substances to X - rays with different energies. In other words, for the same position pixel, the HU value in different single - energy images is different. The difference in pixel values within the same image pixel region is within a preset threshold, such as the CT scanning part (head, neck, chest, abdomen, limbs), organ type (brain, lung, intestine, bone, muscle, etc.), whether there is a contrast agent or an in - vivo implant, which is used to improve the accuracy of substance composition discrimination. This solution does not limit the pixel values of CT images in each region and can be set according to the actual needs of the applicant.

[0073] Example 1: If the scanned part is the head, the possibility of lung tissue can be excluded; if there is no contrast agent during scanning, the possibility of iodine can be excluded.

[0074] Example 2: If the tissue region is the lung, the candidate tissues within the region are lung, muscle, blood, and fat; if the tissue region is bone, the candidate tissues within the region are bone, fat, and blood.

[0075] S106. Based on the target substance composition recognition model, each pixel in a number of image regions is processed respectively to generate a characteristic parameter value corresponding to each pixel.

[0076] In one implementation, the target substance composition recognition model includes a calculation formula for obtaining the characteristic parameter value. The specific calculation formula is:

[0077] CTHU(keV) = a·e b·keV +c;

[0078] a = ω1·a x1 +ω2·a x2 ;

[0079] b = ω1·b x1 +ω2·b x2 ;

[0080] c = ω1·c x1 +ω2·c x2 ;

[0081] ω1 + ω2 = 1;

[0082] Among them, a, b, and c represent characteristic parameters, e represents the exponential function, keV represents the energy level corresponding to the single - energy CT image, x1 and x2 are reference substances respectively, and ω1 and ω2 are the content information of the reference substances respectively.

[0083] Taking the category corresponding to the image region as a prior condition, the possible composition of the reference substances of the substances in the pixels within the region is constrained. For example, the possible substances of the pixels in the bone tissue region should be a weighted combination of bone and fat, and the possible substances of the pixels in the liver region should be a weighted combination of liver and fat. According to the possible composition of the reference substances in different image regions, the bilinear interpolation method is used to obtain the weight wi corresponding to the reference substances in each pixel. Taking the bone region as an example, the substance composition of a certain pixel within all bone regions can be characterized by a linear combination of two reference substances, bone and fat (Supplement: The physiological basis is that real bone has a porous structure, the framework is bone tissue, and the gaps are filled with bone marrow. Since bone marrow is similar to fat, fat is used as a substitute).

[0084] a = ω1·a bone + ω2·a fat ;

[0085] b = ω1·b bone + ω2·b fat ;

[0086] c = ω1·c bone + ω2·c fat ;

[0087] ω1 + ω2 = 1.

[0088] When performing HU value threshold segmentation on 70keV monoenergetic CT images, the applicant can calculate the weight corresponding to the reference substances in each pixel by the bilinear interpolation method. This method is used in image processing to estimate pixel values at non-integer coordinates, especially when the image is scaled or rotated. In the segmentation of the bone region, the applicant can use bone and fat as two reference substances because real bone has a porous structure, its framework is composed of bone tissue, and the gaps are filled with bone marrow. Since bone marrow is similar to fat, fat can be used as a substitute substance.

[0089] The calculation process of bilinear interpolation is as follows: First, determine the position of the pixel point in the image, which usually involves the scaling ratio of the image. Then, find the four pixel points closest to this position (in the image grid). Next, perform linear interpolation in the x and y directions respectively. This means first interpolating each pair of pixel points in the x direction to obtain two temporary values, and then interpolating these two temporary values in the y direction to obtain the final pixel value. The interpolation weight is determined by the distance and can be determined by a function that calculates the distance.

[0090] In practical applications, bilinear interpolation can be expressed by the following formula: f(x, y) = (1 - u)(1 - v)f(x1, y1) + (1 - u)vf(x1, y2) + u(1 - v)f(x2, y1) + uvf(x2, y2);

[0091] Among them, f(x, y) is the pixel value to be calculated, f(x1, y1), f(x1, y2), f(x2, y1), and f(x2, y2) are four known adjacent pixel values, and u and v are the relative distances from the four known pixels to the pixel point to be calculated.

[0092] In the segmentation of the bone region, the applicant can distinguish bones and fat by adjusting the threshold. For example, the CT value of bones is usually higher than 400 HU, while the CT value of fat is usually between -50 and -100 HU. By setting an appropriate threshold, the applicant can separate bones and fat and use the bilinear interpolation method to calculate the weights of bones and fat in each pixel, so as to more accurately segment the bone region.

[0093] S107. Based on the target substance composition recognition model, process the characteristic parameter values of each pixel in several image regions and the image region types matching the several image regions to generate the substance composition information corresponding to each pixel region.

[0094] In one implementation, the substance composition information includes the electron density of the substance and the equivalent atomic number. For example, a certain region is composed of 80% water + 20% fat. Or, it is known that the object to be measured is pure water and fat, and the corresponding characteristic parameters are (a1, b1, c1) and (a2, b2, c2). For an unknown object, the corresponding characteristic parameters are (a, b, c). If (a, b, c) is between (a1, b1, c1) and (a2, b2, c2), it can be considered that the unknown substance is a certain combination of water and fat.

[0095] In another implementation, the target substance composition recognition model includes a calculation formula for obtaining the substance composition information. The calculation formula is:

[0096]

[0097] Where p x,y,z represents the substance composition information corresponding to the pixel with the coordinate position of x, y, z, ω i represents the content of the i-th reference substance, M i represents the i-th reference substance, and n represents the number of reference substances constituting the substance composition of this pixel.

[0098] This model is used in medical imaging, such as CT or MRI scans, where different tissue types (such as bone, fat, soft tissue, etc.) can be distinguished by their specific density or signal intensity. By identifying and quantifying the material composition of each pixel point in the image, it can help doctors make more accurate diagnoses. The model provides a method for quantifying the proportion of different material components in the image pixels through mathematical formulas, and it determines the material composition information at specific positions in the image by weighting multiple reference materials.

[0099] S108, process the material composition information corresponding to each pixel region to generate the attribute information of the pixel region.

[0100] In one implementation, the attribute information of the pixel region includes the relative electron density, equivalent atomic number, and relative ray tissue ability of each pixel.

[0101] The target material composition recognition model includes a calculation formula for obtaining the relative electron density. The calculation formula is:

[0102]

[0103] where RED x,y,z represents the relative electron density corresponding to the pixel at the coordinate position x, y, z, ω i represents the content of the i-th reference material, i represents the i-th reference material, ρ i represents the density of the i-th reference material, ρ water represents the density of water, RED i represents the electron density of the i-th reference material; the calculation of RED is obtained by multiplying the ratio of the density of each reference material to the density of water by the RED of the material, and then performing a weighted sum of all reference materials.

[0104] In one implementation, the target material composition recognition model includes a calculation formula for obtaining the equivalent atomic number. The calculation formula is:

[0105]

[0106] where Zeff x,y,z represents the equivalent atomic number corresponding to the pixel at the coordinate position x, y, z, ω i represents the content of the i-th reference material, i represents the i-th reference material, Z i represents the equivalent atomic number of the i-th reference material; the calculation of Zeff is obtained by multiplying the 2.94th power of the equivalent atomic number of each reference material by the content of the material, and then performing a weighted sum of all reference materials.

[0107] In one implementation, the target substance composition recognition model includes obtaining a calculation formula for the relative radiographic tissue air kerma, and the calculation formula is:

[0108]

[0109] Where SPR represents the relative radiographic tissue air kerma, and RED represents the electron density of the reference substance. The calculation of SPR depends on the value of RED, and different linear equations are used for calculation according to different ranges of RED. The relationship between SPR and RED has different expressions in different intervals of RED values, and these intervals are: 0 to 0.9, 0.9 to 1.035, 1.035 to 1.4, and greater than or equal to 1.4.

[0110] This model is used in the field of medical imaging, such as CT scans, where different tissue types can be distinguished by their electron density, atomic number, and radiographic tissue air kerma. Through these calculations, it can help identify and distinguish different substance components in the image, thereby enabling more accurate diagnosis and analysis. The model provides a method for quantifying the substance composition attributes of each pixel point in the image through mathematical formulas, and these attributes include relative electron density, effective atomic number, and relative radiographic tissue air kerma. In summary, the target substance composition recognition model is a pixel substance composition analysis method based on physical attributes, which identifies and quantifies different substance components in the image by calculating the RED, Zeff, and SPR of each pixel point.

[0111] In this application, a CT image of a target object is obtained by a server. The CT image of the target object includes at least one set of monoenergetic CT images. Each slice image of the monoenergetic CT images is grouped to generate several monoenergetic CT slice image groups. A training sample set and an initial material composition recognition model are obtained. The training sample set includes monoenergetic CT slice image groups corresponding to known reference materials. The initial material composition recognition model is constructed based on the change trends of CT image values of different reference materials at different energy levels. The initial material composition recognition model includes reference values representing characteristic parameters corresponding to the material composition and their corresponding parameter spaces. The initial material composition recognition model is processed based on the training sample set to generate a target material composition recognition model. The CT image of the target object is subjected to semantic segmentation processing to generate several image regions and image region types matching the several image regions. Each pixel in the several image regions is processed based on the target material composition recognition model to generate a characteristic parameter value corresponding to each pixel. Based on the target material composition recognition model, the characteristic parameter values of each pixel in the several image regions and the image region types matching the several image regions are processed to generate material composition information corresponding to each pixel region. The material composition information corresponding to each pixel region is processed to generate attribute information of the pixel region. The attribute information of the pixel region includes the relative electron density, equivalent atomic number, and relative ray tissue ability of each pixel. Based on a new material composition modeling method of the monoenergetic CT image formed by spectral CT, through feature parameter extraction, the ability to distinguish material composition and the accuracy of multi-parameter imaging are improved.

[0112] Optionally, in another embodiment based on the method of the present application, the preprocessing of the training sample set to generate a training sample set with identification information further includes:

[0113] Mark the samples in the training sample set to obtain a first label for each sample;

[0114] Extract a single sample from the training sample set, and form a first training data set according to the sample and the corresponding first label. Extract multiple times to obtain multiple first training data sets;

[0115] Extract two samples from the training sample set, generate a second label for the two samples according to the first labels corresponding to the two samples respectively, and form a second training data set according to the two samples and the corresponding second label. Extract multiple times to obtain multiple second training data sets;

[0116] Construct multiple training data sets according to the multiple first training data sets and the second training data sets.

[0117] In one implementation, a preset number of feature information can be randomly extracted with replacement from the multi-dimensional feature information of each sample each time to form corresponding sub-samples. Multiple sub-samples form a training set. After multiple extractions, multiple training sets are constructed, and the preset number can be customarily set according to actual needs. The training set can be divided into two parts. One part is the single sample x, and it is marked whether the target application will be used next. If so, it can be marked as 1, and if not, it is marked as 0, and the form can be (xi, yi), where yi ∈ {0, 1}. The other part is the triple, that is, by sampling two samples (xi, xj), if the labels of the two samples are the same, it is recorded as 1, and if the labels are different, it is recorded as -1, and the form is (xi, xj, γ), where γ ∈ {1, -1}.

[0118] In one implementation, as Figure 2 shown, the present application also provides a device for identifying the material composition based on the single-energy CT image, including:

[0119] An acquisition module 201, configured to acquire a CT image of a target object, where the CT image of the target object includes at least one set of single-energy CT images; acquire a training sample set and an initial material composition recognition model, where the training sample set includes a set of single-energy CT slice images corresponding to known reference materials, and the initial material composition recognition model is formed based on the change trends of CT image values of different reference materials at different energy levels. The initial material composition recognition model includes reference values for characterizing the characteristic parameters corresponding to the material composition and their corresponding parameter spaces;

[0120] A processing module 202, configured to perform grouping processing on each slice image of the single-energy CT image to generate a plurality of single-energy CT slice image groups; process the initial material composition recognition model based on the training sample set to generate a target material composition recognition model; perform semantic segmentation processing on the CT image of the target object to generate a plurality of image regions and image region types matching the plurality of image regions; process each pixel in the plurality of image regions based on the target material composition recognition model to generate a characteristic parameter value corresponding to each pixel; process the characteristic parameter values of each pixel in the plurality of image regions and the image region types matching the plurality of image regions based on the target material composition recognition model to generate material composition information corresponding to each pixel region; process the material composition information corresponding to each pixel region to generate attribute information of the pixel region, where the attribute information of the pixel region includes the relative electron density, equivalent atomic number, and relative ray tissue ability of each pixel.

[0121] An embodiment of the present application provides an electronic device, as Figure 3As shown, the electronic device 3 includes a first processor 300, a memory 301, a bus 302, and a communication interface 303. The first processor 300, the communication interface 303, and the memory 301 are connected through the bus 302. A computer program that can run on the first processor 300 is stored in the memory 301. When the first processor 300 runs the computer program, it executes the method for identifying the material composition based on the single-energy CT image provided in any of the foregoing embodiments of the present application.

[0122] Among them, the memory 301 may include a high-speed random access memory (RAM: Random Access Memory), and may also include a non-volatile memory, such as at least one disk memory. Through at least one communication interface 303 (which can be wired or wireless), a communication connection is established between this system network element and at least one other network element, and the Internet, wide area network, local area network, metropolitan area network, etc. can be used.

[0123] The bus 302 may be an ISA bus, a PCI bus, an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. Among them, the memory 301 is used to store a program. After receiving an execution instruction, the first processor 300 executes the program. The method for identifying the material composition based on the single-energy CT image disclosed in any of the foregoing embodiments of the present application can be applied to the first processor 300 or implemented by the first processor 300.

[0124] The first processor 300 may be an integrated circuit chip with the ability to process signals. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in the first processor 300 or the instructions in the form of software. The above-mentioned first processor 300 may be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute each method, step, and logic block diagram disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application can be embodied as being executed by a hardware decoding processor or by a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory 301, and the first processor 300 reads the information in the memory 301 and combines its hardware to complete the steps of the above method.

[0125] The electronic device provided by the above embodiment of the present application and the method for identifying the material composition based on a single-energy CT image provided by the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the method adopted, run, or implemented by the application program stored therein.

[0126] The embodiment of the present application provides a computer-readable storage medium, such as Figure 4 shown, the computer-readable storage medium 401 stores a computer program, and when the computer program is read and run by the second processor 402, it implements the method for identifying the material composition based on a single-energy CT image as described above.

[0127] The technical solution of the embodiment of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing an electronic device (which may be an air conditioner, a refrigeration device, a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the method described in the embodiment of the present application. The aforementioned storage medium includes: various media such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disc that can store program codes.

[0128] The computer-readable storage medium provided by the above embodiments of the present application and the method for identifying the material composition based on the single-energy CT image provided by the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0129] The embodiments of the present application provide a computer program product, including a computer program, and the computer program is executed by a third processor to implement the method as described above.

[0130] The computer program product provided by the above embodiments of the present application and the method for identifying the material composition based on the single-energy CT image provided by the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0131] It should be noted that in the present application, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0132] Each embodiment in the present application is described in a related manner. For the same and similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the embodiments of the method for evaluating the material composition identification based on the single-energy CT image, the electronic device, the electronic equipment, and the readable storage medium, since they are basically similar to the embodiments of the above method for identifying the material composition based on the single-energy CT image, the description is relatively simple, and reference can be made to the partial description of the embodiments of the above method for identifying the material composition based on the single-energy CT image for the relevant parts.

[0133] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, the protection scope of the present application should be subject to the scope defined by the claims.

Claims

1. A method for identifying the composition of substances based on single - energy CT images, characterized in that, Comprising: Obtaining a CT image of a target object, wherein the CT image of the target object includes at least one set of monoenergetic CT images; Grouping each slice image of the monoenergetic CT images to generate a number of monoenergetic CT slice image groups; Obtaining a training sample set and an initial material composition recognition model, wherein the training sample set includes monoenergetic CT slice image groups corresponding to known reference materials, and the initial material composition recognition model is constituted based on the variation trends of CT image values of different reference materials at different energy levels. The initial material composition recognition model includes reference values for characterizing characteristic parameters corresponding to the material composition and their corresponding parameter spaces; Processing the initial material composition recognition model based on the training sample set to generate a target material composition recognition model; Performing semantic segmentation processing on the CT image of the target object to generate a number of image regions and image region types matching the number of image regions; Processing each pixel in a number of image regions respectively based on the target substance composition recognition model to generate a characteristic parameter value corresponding to each pixel, and the calculation formula for generating the characteristic parameter value is: CTHU(keV) = a·e b·keV + c; a = ω1·a x1 + ω2·a x2 ; b = ω1·b x1 + ω2·b x2 ; c = ω1·c x1 + ω2·c x2 ; ω1 + ω2 = 1; where a, b, and c represent characteristic parameters, e represents an exponential function, keV represents the energy level corresponding to the single-energy CT image, x1 and x2 are reference substances respectively, and ω1 and ω2 are content information of the reference substances respectively; Processing the characteristic parameter values of each pixel in a number of image regions and the image region types matching the number of image regions based on the target material composition recognition model to generate material composition information corresponding to each pixel region; Processing the material composition information corresponding to each pixel region to generate attribute information of the pixel region, wherein the attribute information of the pixel region includes the relative electron density, equivalent atomic number, and relative ray tissue attenuation coefficient of each pixel.

2. The method according to claim 1, wherein Processing the characteristic parameter values of each pixel in a number of image regions and the image region types matching the number of image regions based on the target material composition recognition model to generate material composition information corresponding to each pixel region, including: The target material composition recognition model includes a calculation formula for obtaining material composition information, and the calculation formula is: Among them, p x,y,z represents the material composition information corresponding to the pixel at the coordinate position x, y, z, ω i represents the content of the i-th reference material, M i represents the i-th reference material, and n represents the number of reference materials constituting the material composition of the pixel.

3. The method according to claim 1, wherein The processing based on the target material composition recognition model includes obtaining parameter information of the relative electron density, equivalent atomic number, and relative ray tissue attenuation coefficient of the material, and the calculation formulas are respectively: The target material composition recognition model includes a calculation formula for obtaining the relative electron density, and the calculation formula is: Among them, RED x,y,z represents the relative electron density corresponding to the pixel at the coordinate positions x, y, z, ω i represents the content of the i-th reference substance, i represents the i-th reference substance, ρ i represents the density of the i-th reference substance, ρ water represents the density of water, RED i represents the electron density of the i-th reference substance; The target substance composition recognition model includes a calculation formula for obtaining the equivalent atomic number, and the calculation formula is: Among them, Zeff x,y,z represents the equivalent atomic number corresponding to the pixel at the coordinate position x, y, z, ω i represents the content of the i-th reference substance, i represents the i-th reference substance, Z i represents the equivalent atomic number of the i-th reference substance; The target material composition recognition model includes a calculation formula for obtaining the relative ray tissue attenuation coefficient, and the calculation formula is: Wherein, SPR represents the relative ray tissue attenuation coefficient, and RED represents the electron density of the reference material.

4. The method according to claim 1, wherein Processing the initial material composition recognition model based on the training sample set to generate a target material composition recognition model, including: Preprocessing the training sample set to generate a training sample set with identification information, wherein the training sample set includes a number of standard material information; Training the initial material composition recognition model based on the training sample set with identification information to generate a target material composition recognition model.

5. The method according to claim 4, characterized in that, Preprocessing the training sample set to generate a training sample set with identification information, including: Performing feature extraction on the training sample set to determine an original feature library; Dividing each feature data set according to the original feature library to generate a training data set and a test data set; Using a classifier to predict each test data set divided from the original feature library to determine a prediction result; Use a preset algorithm to divide each training data set in the original feature library for training to obtain the prediction results of the test set classes; Generate training samples with identification information according to the prediction results and the prediction results of the test set classes.

6. The method according to claim 5, characterized in that, The feature extraction of the training sample set to determine the original feature library includes: Process the reference substance information based on preset processing rules to generate standardized features, where the standardized features are the features of samples or phantoms with clear and known material component compositions; Process the standardized features based on preset feature screening and dimensionality reduction rules to generate original features; Generate an original feature library from a number of original features.

7. A substance composition recognition device based on single-energy CT images, characterized in that The device for implementing the method of claim 1 includes: An acquisition module, configured to acquire a CT image of a target object, where the CT image of the target object includes at least one set of monoenergetic CT images; acquire a training sample set and an initial material composition recognition model, where the training sample set includes a set of monoenergetic CT slice images corresponding to known reference substances, and the initial material composition recognition model is constituted based on the change trends of CT image values of different reference substances at different energy levels. The initial material composition recognition model includes reference values for characterizing the characteristic parameters corresponding to the material components and their corresponding parameter spaces; A processing module, configured to perform grouping processing on each slice image of the monoenergetic CT image to generate a number of monoenergetic CT slice image groups; process the initial material composition recognition model based on the training sample set to generate a target material composition recognition model; perform semantic segmentation processing on the CT image of the target object to generate a number of image regions and image region types matching the number of image regions; process each pixel in a number of image regions based on the target material composition recognition model to generate a characteristic parameter value corresponding to each pixel; process the characteristic parameter values of each pixel in a number of image regions and the image region types matching the number of image regions based on the target material composition recognition model to generate the material composition information corresponding to each pixel region; process the material composition information corresponding to each pixel region to generate the attribute information of the pixel region, where the attribute information of the pixel region includes the relative electron density, equivalent atomic number, and relative ray tissue ability of each pixel.

8. An electronic device, characterized in that, Includes: A first processor; And a memory for storing executable instructions of the first processor; Wherein, the first processor is configured to execute the method for identifying the material composition based on the monoenergetic CT image according to any one of claims 1 to 6 by executing the executable instructions.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a second processor, it implements the method for identifying the material composition based on the monoenergetic CT image according to any one of claims 1 to 6.

Images