Colon adenoma detection method and system based on Mueller matrix analysis

Abstract

The invention provides a colon adenoma detection method and system based on Mueller matrix analysis, and the method comprises the steps: collecting a fresh medical sample containing colon adenoma and normal tissues, and carrying out the calculation decomposition of the fresh medical sample, and obtaining a Mueller matrix and polarization decomposition parameters; according to the polarization information separation and combination characteristics of the Mueller matrix, extracting a preset polarized light characteristic component, and carrying out dimensionless processing and linear combination model solving on the preset polarized light characteristic component to obtain characteristic value distribution of each sample; and verifying the distinguishing capability of characteristic value distribution by using experimental data, optimizing weight parameters of the linear combination model, determining a final linear combination model, and outputting a colon adenoma detection result according to the final linear combination model. By utilizing the synergistic effect of M01, M11 and a depolarization factor, efficient differentiation of colon adenoma and normal tissues is realized after dimensionless treatment.

Description

Technical Field

[0001] The present invention relates to the field of optical detection technology, and in particular, to a method and system for colon adenoma detection based on Mueller matrix analysis. Background Art

[0002] When processing medical images, optical image feature analysis is often involved. The polarization characteristics of light change when interacting with substances, and this change contains rich optical and structural information, which has received extensive attention in the fields of polarization imaging and medical imaging. Polarization images can reflect the microscopic structure and optical properties of samples by recording the polarization state information of light. Especially in the biomedical field, it provides a unique perspective for tissue characterization and lesion detection. For example, polarization imaging has been applied to the early diagnosis of lesions such as skin cancer and colon cancer, and it can identify subtle changes in scattering and absorption in tissues. However, how to fully exploit the high-dimensional information in polarization images to improve the detection accuracy remains one of the main challenges in current technology. The Mueller matrix, as a mathematical tool for systematically describing the change of polarization state, provides 16 complete elements to quantify the conversion characteristics of the polarization state when light interacts with substances. The 16 elements of the Mueller matrix can comprehensively describe the conversion relationship of the polarization state of light among unpolarized light, linearly polarized light, and circularly polarized light, which contains a large amount of key information reflecting the optical and structural characteristics of substances. However, in existing optical detection applications, the use of a single Mueller matrix component or simple polarization information cannot fully reflect the optical differences between complex tissue structures such as tumors and normal tissues, and there are the following deficiencies: A single Mueller matrix component cannot fully reflect the optical differences between complex tissues such as colon adenomas and normal tissues, which is likely to lead to information loss or result deviation; there is a lack of effective combination of Mueller matrix components in the prior art, and the synergistic effect between multiple components has not been fully utilized to improve the detection accuracy; due to the different physical meanings, units, and significant magnitude differences of the components of the Mueller matrix, directly using these data may be interfered by experimental conditions (such as light intensity, sample thickness, etc.), thus affecting the stability and robustness of the analysis. In the study of fresh medical samples, these problems may lead to the lack of reliability and repeatability of detection results. Summary of the Invention

[0003] The present invention provides a method and system for colon adenoma detection based on Mueller matrix analysis to solve the defects existing in the prior art.

[0004] In a first aspect, the present invention provides a method for colon adenoma detection based on Mueller matrix analysis, including: Collecting a fresh medical sample containing colon adenoma and normal tissue, performing computational decomposition on the fresh medical sample to obtain a Mueller matrix and polarization decomposition parameters; According to the polarization information separation and combination characteristics of the Mueller matrix, extract the preset polarized light characteristic components, perform dimensionless processing on the preset polarized light characteristic components and solve the linear combination model to obtain the eigenvalue distribution of each sample; Use experimental data to verify the discrimination ability of the eigenvalue distribution, optimize the weight parameters of the linear combination model, determine the final linear combination model, and output the colon adenoma detection result according to the final linear combination model.

[0005] According to a colon adenoma detection method based on Mueller matrix analysis provided by the present invention, collect fresh medical samples containing colon adenomas and normal tissues, including: Provide linear and circularly polarized light sources by an optical polarization measurement system, use a CCD for signal reception, and collect multiple fresh medical samples.

[0006] According to a colon adenoma detection method based on Mueller matrix analysis provided by the present invention, perform computational decomposition on the fresh medical samples to obtain a Mueller matrix and polarization decomposition parameters, including: Map the fresh medical samples to gray values and store them in an 8-bit space; Crop the multiple polarized images corresponding to the fresh medical samples, and retain the main parts of the multiple polarized images; Based on the optical scattering theory, perform multi-dimensional fitting and polarization resolution on the cropped polarized images to obtain the Mueller matrix; Use the Lu-Chipman decomposition technique to perform polarization decomposition on the Mueller matrix to obtain three polarization decomposition parameters, and the three polarization decomposition parameters include dichroism, depolarization, and phase delay.

[0007] According to a colon adenoma detection method based on Mueller matrix analysis provided by the present invention, according to the polarization information separation and combination characteristics of the Mueller matrix, extract the preset polarized light characteristic components, perform dimensionless processing on the preset polarized light characteristic components and solve the linear combination model to obtain the eigenvalue distribution of each sample, including: Extract the M01 component, the M11 component, and the depolarization factor component; Respectively perform normalization processing using the original variables of each component and the maximum value in the component sample set to obtain the M01 normalized component, the M11 normalized component, and the depolarization factor normalized component; Determine the first weight factor, the second weight factor, and the third weight factor, and construct the linear combination model based on the first weight factor, the M01 normalized component, the second weight factor, the M11 normalized component, the third weight factor, and the depolarization factor normalized component; Substitute the M01 standardized component, the M11 standardized component, and the depolarization factor standardized component into the linear combination model respectively to obtain the eigenvalue distribution of each sample.

[0008] According to a colon adenoma detection method based on Mueller matrix analysis provided by the present invention, experimental data is used to verify the discrimination ability of the eigenvalue distribution, optimize the weight parameters of the linear combination model, determine the final linear combination model, and output the colon adenoma detection result according to the final linear combination model, including: Divide the eigenvalue distribution of each sample into a training set and a test set, use the support vector machine algorithm to perform classification modeling on the eigenvalue data, the initial model is trained using a linear kernel function, and the radial basis kernel function, linear kernel function, and parameter C value of the support vector machine are optimized and adjusted; Use the receiver operating characteristic ROC curve to evaluate the performance of the classification model, calculate the area under the ROC curve AUC, and judge the classification accuracy of the model; Through K-fold cross-validation, divide the sample data into K subsets, take turns using K - 1 of these subsets for training, and the remaining one subset for testing, repeat K times to obtain the average classification performance, and perform result visualization, where negative values are the colon adenoma tissue regions and positive values are the normal colon tissue regions.

[0009] According to a colon adenoma detection method based on Mueller matrix analysis provided by the present invention, use the receiver operating characteristic ROC curve to evaluate the performance of the classification model, calculate the area under the ROC curve AUC, and judge the classification accuracy of the model, including: If it is determined that the classification effect of the model reaches the ideal value, then determine the first weight factor, the second weight factor, and the third weight factor in the eigenvalue calculation linear combination model to obtain the determined linear combination model; If it is determined that the classification effect of the model does not reach the ideal value, then adjust the first weight factor, the second weight factor, and the third weight factor, use the network search method to traverse different weight combinations, determine the weight combination corresponding to the highest AUC, and obtain the determined linear combination model.

[0010] In a second aspect, the present invention also provides a colon adenoma detection system based on Mueller matrix analysis, including: A decomposition module, configured to collect fresh medical samples containing colon adenomas and normal tissues, perform calculation and decomposition on the fresh medical samples to obtain a Mueller matrix and polarization decomposition parameters; A conversion module, configured to extract preset polarization light characteristic components according to the polarization information separation and combination characteristics of the Mueller matrix, perform dimensionless processing and linear combination model solution on the preset polarization light characteristic components to obtain the eigenvalue distribution of each sample; A detection module is used to verify the discrimination ability of the eigenvalue distribution by using experimental data, optimize the weight parameters of the linear combination model, determine the final linear combination model, and output the colon adenoma detection result according to the final linear combination model.

[0011] In a third aspect, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the colon adenoma detection method based on Mueller matrix analysis as described in any one of the above.

[0012] In a fourth aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the colon adenoma detection method based on Mueller matrix analysis as described in any one of the above.

[0013] In a fifth aspect, the present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the colon adenoma detection method based on Mueller matrix analysis as described in any one of the above.

[0014] The colon adenoma detection method and system based on Mueller matrix analysis provided by the present invention realize the efficient discrimination between colon adenomas and normal tissues by analyzing the polarization state conversion characteristics of colon adenoma optical images and using the dimensionless processing of Mueller matrix components and polarization decomposition parameters, and have the characteristics of simple implementation, high recognition efficiency and high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 is one of the flow schematic diagrams of the colon adenoma detection method based on Mueller matrix analysis provided by the present invention; Figure 2 is another flow schematic diagram of the colon adenoma detection method based on Mueller matrix analysis provided by the present invention; Figure 3 is the structural schematic diagram of the colon adenoma detection system based on Mueller matrix analysis provided by the present invention; Figure 4 is the structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] Figure 1 It is one of the schematic flowcharts of the colon adenoma detection method based on Mueller matrix analysis provided by an embodiment of the present invention. As Figure 1 shown, it includes: Step 100: Collect a fresh medical sample containing colon adenoma and normal tissue, perform computational decomposition on the fresh medical sample to obtain a Mueller matrix and polarization decomposition parameters; Step 200: According to the polarization information separation and combination characteristics of the Mueller matrix, extract preset polarized light characteristic components, perform dimensionless processing on the preset polarized light characteristic components and solve the linear combination model to obtain the eigenvalue distribution of each sample; Step 300: Use experimental data to verify the discrimination ability of the eigenvalue distribution, optimize the weight parameters of the linear combination model, determine the final linear combination model, and output the colon adenoma detection result according to the final linear combination model.

[0019] Specifically, the optical detection method provided by the present invention for analyzing the polarization state conversion characteristics of the Mueller matrix utilizes the effects of M 11 , M 01 and the depolarization factor to achieve efficient discrimination between colon adenoma and normal tissue after dimensionless processing. As Figure 2 shown, it includes the following steps: Obtain a fresh medical sample containing colon adenoma and normal tissue through a light polarization measurement system. The information in the image collected in this step is the light intensity information of the sample's scattering of polarized light, which is mapped to gray values and stored in an 8-bit space. Perform the calculation of the Mueller matrix on the collected light intensity information data, and then perform Lu-Chipman decomposition to obtain a 4×4 Mueller matrix and three polarization decomposition parameters - dichroism, depolarization, and phase delay.

[0020] The M 01 component describes the characteristic of the conversion of non-polarized light to linearly polarized light and is related to the directional scattering of the sample. The M 11 component reflects the ratio of depolarized light to total scattered light and is related to the non-sphericity of the scatterer. The depolarization factor quantifies the degree of depolarization of light and can effectively characterize the scattering complexity of the tissue. In colon adenoma tissue, due to the mixed arrangement of cells, the depolarization component is often higher than that of normal tissue. Select the M 11 and M01 The components and depolarization factors are subjected to normalization and dimensionless processing to eliminate the influence of factors such as magnitude and physical meaning on the results. Then, a linear combination model is constructed to synthesize the characteristics of the three variables and calculate the eigenvalues.

[0021] The discrimination ability of the comprehensive eigenvalues is verified using experimental data, and the weight parameters are optimized. The accuracy and sensitivity are evaluated through the Receiver Operating Characteristic (ROC) curve assessment method. Finally, the results are visualized. In the results, negative values represent the colon adenoma tissue area, and positive values represent the normal colon tissue area.

[0022] In one embodiment, step 100 includes: Sampling and data acquisition of fresh medical samples containing colon adenoma and normal tissues are performed through a light polarization measurement system to ensure that the polarization characteristics of the samples can be comprehensively reflected. Specifically, the image information collected in this step is derived from the scattered light intensity distribution of the samples under different polarized light incident conditions, reflecting the optical characteristics and microscopic structure of the tissues. During the acquisition process, the light polarization measurement system provides linear and circular polarized light sources to ensure sufficient diversity and integrity of the optical responses of the samples. The signals are received by a CCD, and the light intensity information after image acquisition is digitally processed and mapped to gray values and stored in an 8-bit space for subsequent image analysis and data calculation.

[0023] To reduce the influence of background noise on the analysis results, improve the credibility of the results and the integrity of the expression of sample characteristics, and improve the calculation efficiency, 32 polarized images collected in the sample need to be preprocessed. Specifically, the images are cropped to retain the main areas containing colonic adenomas and normal tissues, and the background parts unrelated to the target area are removed. After retaining the target area, for each polarized image, a complete 4×4 Mueller matrix is obtained using a standardized calculation method. During the calculation process, based on the optical scattering theory, multi-dimensional fitting and polarization resolution are performed on the light intensity information of each image to obtain a Mueller matrix that can provide detailed information about the microscopic structure and optical properties of the sample. Then, the Lu-Chipman decomposition technique is used to perform polarization decomposition on the Mueller matrix, and three key polarization parameters are extracted: dichroism, depolarization, and phase delay. It should be noted that the depolarization characteristic of the sample refers to the phenomenon that polarized light loses some or all of its polarization information due to scattering, reflection, absorption, or other interactions during propagation, which is one of the important polarization characteristics of the sample. As an important tool for describing the polarization characteristics of the sample, the Mueller matrix contains all the polarization characteristic information of the sample under different polarized light conditions, and the manifestation form of the characteristics is implicit. Therefore, through the Lu-Chipman Mueller matrix decomposition method, the Mueller matrix can be decomposed into characteristic components such as depolarization factors, birefringence parameters, and phase delays, so as to quantitatively extract the depolarization factor of the sample as a parameter to measure the depolarization characteristic of the sample.

[0024] During the polarization decomposition, since matrix inversion operations are involved in the calculation, the matrix may be rank-deficient due to shadow areas or high-brightness spot areas that appear on the sample surface during shooting. This phenomenon will affect the stability of the resolution and the accuracy of the calculation results. To overcome this problem, the system designs an automated anomaly detection and correction mechanism. When abnormal pixels (such as rank-deficient situations caused by high light or shadows) are detected during the calculation process, the software system will automatically record the positions of these pixels and repair them through interpolation or assignment algorithms. The specific operations include: extracting the normal values of adjacent pixels and assigning values to the abnormal pixels to restore the integrity of the matrix as much as possible.

[0025] In one embodiment, step 200 includes: Using the polarization information separation and combination characteristics of the Mueller matrix, extract the polarized light characteristic components from the light intensity data of the sample, and calculate the comprehensive characteristic value for tumor and normal tissue classification through a dimensionless and linear combination model. The Mueller matrix systematically describes the response of the sample under different polarized light states, forming 16 components, each of which contains the optical properties of the sample for a specific polarized state conversion. Among these components, select M11, M01, and the depolarization factor, which are the most representative of the tissue optical properties, as analysis variables: M01 component: Describes the ability to convert unpolarized light into linearly polarized light, which is closely related to the directional scattering characteristics of the sample. Directional scattering reflects the degree of order of the sample structure. In normal tissues with relatively regular cell arrangements, directional scattering is higher, while in colonic adenoma tissues with mixed cell arrangements, directional scattering is usually lower.

[0026] M11 component: Reflects the ratio of depolarized light to total scattered light and is related to the non-sphericity of the scatterer. The higher the non-sphericity, the lower the proportion of the depolarized component in the scattered light. Due to the irregular cell shapes and chaotic arrangements in colonic adenoma tissues, their non-sphericity is usually lower than that of normal tissues.

[0027] Depolarization factor: Quantifies the depolarization ability of the sample to polarized light and is an important parameter for measuring the optical complexity of the sample. In colonic adenoma tissues, due to the mixed arrangement of internal cells, resulting in a complex distribution of scatterers, the depolarization factor is usually higher than that of normal tissues.

[0028] By comprehensively analyzing the above three variables, the optical and structural characteristics of the sample can be more comprehensively characterized. However, due to the different physical meanings, inconsistent units, and significant magnitude differences of the components of the Mueller matrix, directly calculating and analyzing them may introduce biases. Therefore, in order to eliminate the influence of these differences and ensure the comparability between different variables, the following normalization and dimensionless processing formula is adopted:

[0029] where X is the original variable, and max(X) is the maximum value of this variable in the sample set. Through dimensionless processing, all variables are standardized to the range of 0 to 1. After completing the dimensionless processing, a linear weighted model is constructed to calculate the comprehensive characteristics of M11, M01, and the depolarization factor, forming an eigenvalue , which is used to classify tumor and normal tissues. The eigenvalue is calculated using the linear weighted model:

[0030] where , and are weight factors determined in subsequent optimizations, , and are the standardized M11, M01, and depolarization factor respectively. Substituting the dimensionless variables into the eigenvalue calculation formula, the eigenvalue of each sample is obtained. The initial weights are obtained by fitting experimental data and are further optimized in subsequent steps.

[0031] In actual calculations, the three normalized variables are substituted into the above eigenvalue formula to calculate the eigenvalues of each sample one by one. By calculating the eigenvalues of each sample, the eigenvalue distribution of each sample is obtained, which is used to distinguish colon adenoma tissue from normal tissue. In subsequent steps, these eigenvalues will be further used for classification modeling and performance verification.

[0032] In one embodiment, step 300 includes: Verify the discrimination ability of the comprehensive eigenvalues using experimental data, and optimize the weight parameters, and evaluate the accuracy and sensitivity through the ROC curve evaluation method. After completing the eigenvalue calculation, verify the classification ability of the eigenvalues, and optimize the weight factors of the linear model through experiments. The specific steps are as follows: 1. Divide the calculation results of the eigenvalues 𝑇 of each sample into a training set and a test set. The training set is used to fit the classification model, and the test set is used to evaluate the classification effect of the model. Use the support vector machine (SVM) machine learning algorithm to perform classification modeling on the eigenvalue data. The initial model is trained using a linear kernel function, and during the training process, the radial basis kernel function, linear kernel function, and parameter C value of the SVM are adjusted and optimized.

[0033] 2. Use the receiver operating characteristic (ROC) curve to evaluate the performance of the classification model. By calculating the area under the ROC curve (AUC value), judge the accuracy of the model in classifying colon adenomas and normal tissues. In this step, two situations will occur: (1) The classification effect of the model reaches the index for distinguishing colon adenoma from normal tissue. If the classification effect of the model is ideal, then determine the values of the weight factors , and in the eigenvalue calculation formula, so as to determine the linear relationship of the three main characteristics when distinguishing colon adenoma from normal tissue in a fresh medical sample.

[0034] (2) If the classification effect of the model is not ideal, optimize by adjusting the values of the weight factors , and in the eigenvalue calculation formula. Use the grid search method to traverse different weight combinations, and select the weight parameter group with the highest AUC value until the index described in (1) is reached.

[0035] 3. During the model training process, to ensure the robustness of model training, K-fold cross-validation (K = 5) is used to further verify the stability and generalization ability of the model. The sample data is divided into K subsets, and K-1 subsets are used for training in turn, while the remaining one subset is used for testing. This is repeated K times to obtain the average classification performance. Ensure that the model performs well on the current data and adapts to different sample sets, without relying on a specific sample distribution. Finally, the results are visualized. In the results, negative values represent the area of colonic adenoma tissue, and positive values represent the area of normal colonic tissue.

[0036] The colon adenoma detection system based on Mueller matrix analysis provided by the present invention will be described below. The colon adenoma detection system based on Mueller matrix analysis described below can be mutually referred to and corresponding to the colon adenoma detection method based on Mueller matrix analysis described above.

[0037] Figure 3 is a schematic structural diagram of the colon adenoma detection system based on Mueller matrix analysis provided by an embodiment of the present invention, as Figure 3 shown, including: a decomposition module 31, a conversion module 32, and a detection module 33, where: The decomposition module 31 is used to collect fresh medical samples containing colon adenomas and normal tissues, perform computational decomposition on the fresh medical samples to obtain the Mueller matrix and polarization decomposition parameters; the conversion module 32 is used to extract preset polarized light characteristic components according to the polarization information separation and combination characteristics of the Mueller matrix, perform dimensionless processing on the preset polarized light characteristic components, and solve the linear combination model to obtain the eigenvalue distribution of each sample; the detection module 33 is used to verify the discrimination ability of the eigenvalue distribution using experimental data, optimize the weight parameters of the linear combination model, determine the final linear combination model, and output the colon adenoma detection result according to the final linear combination model.

[0038] Figure 4 illustrates a schematic physical structure diagram of an electronic device, as Figure 4As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430, and a communication bus 440. Among them, the processor 410, the communications interface 420, and the memory 430 complete communication with each other through the communication bus 440. The processor 410 may call the logical instructions in the memory 430 to execute a colon adenoma detection method based on Mueller matrix analysis. The method includes: collecting a fresh medical sample containing colon adenoma and normal tissue, performing computational decomposition on the fresh medical sample to obtain a Mueller matrix and polarization decomposition parameters; according to the polarization information separation and combination characteristics of the Mueller matrix, extracting preset polarization light characteristic components, performing dimensionless processing and linear combination model solution on the preset polarization light characteristic components to obtain the eigenvalue distribution of each sample; using experimental data to verify the discrimination ability of the eigenvalue distribution, optimizing the weight parameters of the linear combination model, determining the final linear combination model, and outputting the colon adenoma detection result according to the final linear combination model.

[0039] In addition, when the logical instructions in the above-mentioned memory 430 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0040] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the colon adenoma detection method based on Mueller matrix analysis provided by the above-mentioned various methods. The method includes: collecting a fresh medical sample containing colon adenoma and normal tissue, performing computational decomposition on the fresh medical sample to obtain a Mueller matrix and polarization decomposition parameters; according to the separation and combination characteristics of the polarization information of the Mueller matrix, extracting preset polarization light characteristic components, performing dimensionless processing and solving a linear combination model on the preset polarization light characteristic components to obtain the eigenvalue distribution of each sample; using experimental data to verify the discrimination ability of the eigenvalue distribution, optimizing the weight parameters of the linear combination model, determining the final linear combination model, and outputting the colon adenoma detection result according to the final linear combination model.

[0041] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it realizes the colon adenoma detection method based on Mueller matrix analysis provided by the above-mentioned various methods. The method includes: collecting a fresh medical sample containing colon adenoma and normal tissue, performing computational decomposition on the fresh medical sample to obtain a Mueller matrix and polarization decomposition parameters; according to the separation and combination characteristics of the polarization information of the Mueller matrix, extracting preset polarization light characteristic components, performing dimensionless processing and solving a linear combination model on the preset polarization light characteristic components to obtain the eigenvalue distribution of each sample; using experimental data to verify the discrimination ability of the eigenvalue distribution, optimizing the weight parameters of the linear combination model, determining the final linear combination model, and outputting the colon adenoma detection result according to the final linear combination model.

[0042] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0043] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting colon adenoma based on Mueller matrix analysis, characterized in that: include: collecting fresh medical samples containing colon adenoma and normal tissue, and performing computational decomposition on the fresh medical samples to obtain Mueller matrix and polarization decomposition parameters; According to the polarization information separation and combination characteristics of the Mueller matrix, a preset polarization light characteristic component is extracted, and the preset polarization light characteristic component is dimensionlessly processed and a linear combination model is solved to obtain an eigenvalue distribution of each sample; The distinguishing ability of the eigenvalue distribution is verified by using experimental data, the weight parameters of the linear combination model are optimized, the final linear combination model is determined, and the colon adenoma detection result is output according to the final linear combination model.

2. The method for detecting colon adenoma based on Mueller matrix analysis according to claim 1, characterized in that: Collect fresh medical samples containing colon adenomas and normal tissues, including: The optical polarization measurement system provides linear and circular polarized light sources, and CCD is used for signal reception to collect multiple fresh medical samples.

3. The method for detecting colon adenoma based on Mueller matrix analysis according to claim 2, characterized in that: The fresh medical sample is subjected to computational decomposition to obtain a Mueller matrix and polarization decomposition parameters, including: Mapping the fresh medical sample into grayscale values ​​and storing them in an 8-bit space; cropping the multiple polarization images corresponding to the fresh medical sample to retain a main portion of the multiple polarization images; Based on the optical scattering theory, multi-dimensional fitting and polarization solution are performed on the clipped polarization image to obtain the Mueller matrix; The Mueller matrix is ​​polarization decomposed by using Lu-Chipman decomposition technology to obtain three polarization decomposition parameters, wherein the three polarization decomposition parameters include dichroism, depolarization and phase delay.

4. The method for detecting colon adenoma based on Mueller matrix analysis according to claim 1, characterized in that: According to the polarization information separation and combination characteristics of the Mueller matrix, a preset polarization light characteristic component is extracted, and the preset polarization light characteristic component is dimensionlessly processed and a linear combination model is solved to obtain the characteristic value distribution of each sample, including: Extract M01 component, M11 component and depolarization factor component; The original variables of each component and the maximum value of the component sample set are used for normalization processing to obtain the M01 standardized component, the M11 standardized component and the depolarization factor standardized component; Determine a first weight factor, a second weight factor and a third weight factor, and construct the linear combination model based on the first weight factor, the M01 standardized component, the second weight factor, the M11 standardized component, the third weight factor and the depolarization factor standardized component; The M01 standardized component, the M11 standardized component and the depolarization factor standardized component are respectively substituted into the linear combination model to obtain the eigenvalue distribution of each sample.

5. The method for detecting colon adenoma based on Mueller matrix analysis according to claim 1, characterized in that: The distinguishing ability of the characteristic value distribution is verified by using experimental data, the weight parameters of the linear combination model are optimized, the final linear combination model is determined, and the colon adenoma detection result is output according to the final linear combination model, including: The eigenvalue distribution of each sample is divided into a training set and a test set. The support vector machine algorithm is used to classify and model the eigenvalue data. The initial model is trained using a linear kernel function. The radial basis kernel function, linear kernel function and parameter C value of the support vector machine are optimized and adjusted. The receiver operating characteristic (ROC) curve was used to evaluate the performance of the classification model, and the area under the ROC curve (AUC) was calculated to determine the classification accuracy of the model; Through K-fold cross-validation, the sample data is divided into K subsets, and K-1 subsets are used for training in turn, and the remaining subset is used for testing. This is repeated K times to obtain the average classification performance, and the results are visualized, where negative values ​​represent colon adenoma tissue areas and positive values ​​represent normal colon tissue areas.

6. The method for detecting colon adenoma based on Mueller matrix analysis according to claim 5, characterized in that: The receiver operating characteristic (ROC) curve was used to evaluate the performance of the classification model, and the area under the ROC curve (AUC) was calculated to determine the classification accuracy of the model, including: If it is determined that the classification effect of the model reaches an ideal value, the first weight factor, the second weight factor and the third weight factor in the linear combination model are calculated by determining the eigenvalues ​​to obtain a determined linear combination model; If it is determined that the model classification effect does not reach the ideal value, the first weight factor, the second weight factor and the third weight factor are adjusted, and a network search method is used to traverse different weight combinations to determine the weight combination corresponding to the highest AUC, and obtain a determined linear combination model.

7. A colon adenoma detection system based on Mueller matrix analysis, characterized in that: include: A decomposition module, used for collecting fresh medical samples containing colon adenoma and normal tissue, performing computational decomposition on the fresh medical samples, and obtaining Mueller matrix and polarization decomposition parameters; A conversion module, used to extract a preset polarized light characteristic component according to the polarization information separation and combination characteristics of the Mueller matrix, perform dimensionless processing on the preset polarized light characteristic component and solve a linear combination model to obtain a characteristic value distribution of each sample; The detection module is used to verify the distinguishing ability of the characteristic value distribution using experimental data, optimize the weight parameters of the linear combination model, determine the final linear combination model, and output the colon adenoma detection result according to the final linear combination model.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the colon adenoma detection method based on Mueller matrix analysis as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for detecting colon adenoma based on Mueller matrix analysis as claimed in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for detecting colon adenoma based on Mueller matrix analysis as claimed in any one of claims 1 to 6 is implemented.

Images