Feature level image fusion reconstruction method based on ultrasonic and contrast images

Abstract

The invention discloses a feature level image fusion reconstruction method based on ultrasonic and angiography images, which comprises the following steps: respectively carrying out bit plane decomposition on an ultrasonic image and an angiography image to generate eight binary bit planes of each image; calculating the weight of each bit plane image, and reconstructing the image according to the weight to generate a new image; flattening the weighted reconstructed image into one-dimensional data, and carrying out standardization processing; extracting the first two main features with the highest contrast in the image as principal components through principal feature analysis, and performing low-dimensional representation; and through low-dimensional representation after principal feature analysis, inverse transformation is carried out to restore the image, and a principal feature analysis fusion reconstruction image of the ultrasound image and the angiography image is obtained. According to the method, feature levels of the ultrasonic image and the angiography image are fused, so that the defects of the ultrasonic image in the aspect of displaying internal details of the uterus and the problem that the edge contour of the angiography image is blurred are successfully overcome, and the processed image can provide clearer and more detailed image information.

Description

Technical Field

[0001] The present invention belongs to the technical fields of image processing and artificial intelligence, and particularly relates to a feature-level image fusion and reconstruction method based on ultrasonic and contrast images, which is applied to medical image analysis. Background Art

[0002] Contrast-enhanced ultrasound is an advanced medical imaging technology that uses special contrast agents to enhance the contrast of ultrasound images, thereby better showing the structure and blood flow of the lesion site. Compared with traditional ultrasound examinations, contrast-enhanced ultrasound has higher diagnostic accuracy and a wider range of applications. In recent years, with the continuous progress of medical technology, contrast-enhanced ultrasound has been widely used in clinical diagnosis. On the other hand, with the continuous progress of artificial intelligence technology and image processing technology, more and more technologies have been applied to the processing of medical images.

[0003] However, in the application of medical images, there are still some technical problems to be solved, such as the problem of noise interference in medical images. Due to the noise interference in medical images, it not only poses requirements for doctors' diagnostic judgments, but also brings challenges to the practical application of image processing technology in medical images.

[0004] The noise in medical images not only affects the quality of the images, but also makes it difficult to extract detailed information, increasing the complexity of diagnosis. Especially in ultrasonic and contrast images, ultrasonic images are insufficient in showing the internal details of organs in the body, while contrast images may have blurred edges. These problems may affect the quality of medical images and thus the diagnostic accuracy of doctors. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the related technologies to some extent.

[0006] An object of the present invention is to provide a feature-level image fusion and reconstruction method based on ultrasonic and contrast images, which combines the advantages of ultrasonic images and contrast images by using a feature-level fusion method, extracts and fuses image features through an image processing algorithm, and overcomes the deficiencies of ultrasonic images in showing details and the blurring of the edge contours of contrast images.

[0007] To achieve the above object, on the one hand, the present invention provides a feature-level image fusion and reconstruction method based on ultrasonic and contrast images, including:

[0008] S100. Perform bit-plane decomposition on the ultrasonic image and the contrast image respectively to generate 8 binary bit-plane images of each image;

[0009] S200. Calculate the weights of each bit-plane image, and perform dynamic weighted fusion on the corresponding bit-planes of the two images to generate a weighted fusion image;

[0010] S300. Flatten the weighted fusion image into one-dimensional data and perform normalization processing;

[0011] S400. Extract the first two main features with the highest contrast in the weighted fusion image as the principal components through principal feature analysis and perform low-dimensional representation;

[0012] S500. Through the low-dimensional representation after principal feature analysis, perform inverse transformation to restore the image, and obtain the principal feature analysis fusion reconstruction image of the ultrasonic image and the contrast-enhanced image.

[0013] A further preferred technical solution of the present invention is that step S100 is specifically:

[0014] Decompose the RGB channels of each pixel of the ultrasonic image and the contrast-enhanced image bit by bit, and the pixel value is P=(P R ,P G ,P B ), where P R ,P G ,P B are the pixel values of the red, green, and blue channels respectively;

[0015] The pixel value of each channel is represented as an 8-bit binary number. Let P C ={p7,p6,…,p0}, where p i represents the value of the pixel value at the i-th bit, and is expressed as:

[0016] Pc = p7·2 7 +p6·2 6 +…+p1·2 1 +p0·2 0 ;

[0017] where p7 is the highest bit and p0 is the lowest bit;

[0018] Split each pixel value into 8 bit positions {p7,p6,…,p0}, and each bit plane B q corresponds to the q-th bit of all pixels in the image, and generates 8 binary bit planes of each image.

[0019] Preferably, step S200 calculates the weights of each bit plane image, and performs dynamic weighted fusion on the corresponding bit planes of the two images to generate a weighted fusion image; including:

[0020] S210. Calculate the contrast of each bit plane of the two images and convert it into a weight, so that the bit plane with high contrast obtains a larger weight, while the bit plane with low contrast obtains a smaller weight;

[0021] S220. Sum the weighted results of each bit plane to obtain a weighted fusion image.

[0022] Preferably, the specific method of step S210 is as follows:

[0023] Let B q be the q-th bit plane after image decomposition. The contrast of each bit plane in the image is calculated by variance. Let Var(B q ) be the variance of the q-th bit plane, and the calculation formula is:

[0024]

[0025] where M and N are the height and width of the image respectively, and μ q is the mean of the q-th bit plane, and B q (i, j) represents the pixel value of the q-th bit plane of the image at position (i, j);

[0026] According to the calculated variance Var(B q ), assign higher weights to the bit planes with higher contrast. The formula for weight assignment is:

[0027]

[0028] where W q is the weight of the q-th bit plane, and ω is a constant used to prevent the variance from being zero.

[0029] Preferably, the specific method of step S220 is as follows:

[0030] Sum the weighted results of each bit plane to obtain a weighted image L, which is expressed as:

[0031]

[0032] where L(i, j) represents the pixel value of the weighted image L at position (i, j).

[0033] Preferably, step S300 flattens the weighted fusion image into one-dimensional data and performs normalization processing. The specific steps are as follows:

[0034] First, perform bit plane decomposition on the weighted fusion image. The decomposed bit plane images are flattened to obtain a one-dimensional vector for each bit plane image. For the i-th image A i , its flattened vector is expressed as:

[0035] A i = [A i (1, 1), A i (1, 2), …, A i(M, N)] T ;

[0036] where M and N are the height and width of the image, respectively;

[0037] Then, the one-dimensional vector is normalized so that the mean of each feature is zero and the variance is 1, which is expressed as:

[0038]

[0039] where μ is the mean of each feature and σ is the standard deviation of each feature.

[0040] Preferably, in step S400, principal component analysis extracts the two main features with the highest contrast in the weighted fusion image as the principal components through a feature extraction method, calculates the covariance matrix of the image, and finds its eigenvectors to extract the main feature directions;

[0041] First, select the most important k principal components, and through linear transformation, obtain a new coordinate system such that the projection variance of the original data in this coordinate system is the largest, and each principal feature corresponds to an eigenvector;

[0042] Calculate the covariance matrix of the data, then solve the eigenvalues and eigenvectors of this covariance matrix, select the eigenvector with the largest eigenvalue as the principal component, and project the data onto these principal features to obtain the data representation after dimensionality reduction:

[0043] Z i = A i ′·P k ;

[0044] where Z i is the data representation after dimensionality reduction, and P k is the matrix containing the eigenvectors of the first k principal components.

[0045] Preferably, in step S500, through inverse transformation, the analyzed data is restored to the original image space, which is expressed as:

[0046]

[0047] where A″ i is the principal component analysis fusion reconstruction image restored to the original image space, which is expressed as the data after inverse transformation; is the transpose of the eigenvector matrix P k .

[0048] On the other hand, the present invention provides a non-transitory computer-readable storage medium, on which computer instructions are stored, and the computer instructions cause the computer to execute the above-mentioned feature-level image fusion and reconstruction method based on ultrasound and contrast image features.

[0049] On the other hand, the present invention provides an electronic device, including: a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus. The processor calls the logical instructions in the memory to execute the above-mentioned feature-level image fusion and reconstruction method based on ultrasonic and contrast images.

[0050] On yet another aspect, the present invention provides a computer program product. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer executes the above-mentioned feature-level image fusion and reconstruction method based on ultrasonic and contrast images.

[0051] Beneficial effects: By performing feature-level fusion of ultrasonic images and contrast images, the present invention effectively overcomes the deficiencies of ultrasonic images in displaying details and the problem of blurred edge contours of contrast images, thereby improving the clarity and detail performance of the images. The fused image can provide more accurate in-vivo organ image information, helping doctors better observe the lesions and their surrounding structures, and thus making more accurate diagnostic judgments.

[0052] During the fusion process, the present invention can effectively reduce noise interference and remove unnecessary irrelevant information, making the finally output image more accurate and meeting the diagnostic needs of doctors. At the same time, through feature-level image fusion, clearer and more reliable input image data are provided, enhancing the accuracy and robustness of the computer-aided diagnosis system in medical image analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a flowchart of the feature-level image fusion and reconstruction method based on ultrasonic and contrast images of the present invention;

[0054] Figure 2 is a flowchart of the decomposition and weight assignment reconstruction of ultrasonic and contrast images of the present invention;

[0055] Figure 3 is a flowchart of the main feature analysis and inverse transformation recovery of ultrasonic and contrast images of the present invention;

[0056] Figure 4 is a decomposition diagram of uterine ultrasonic and contrast images in Example 1;

[0057] Figure 5 is a weight assignment reconstruction diagram of uterine ultrasonic and contrast images in Example 1;

[0058] Figure 6 is a main feature analysis and fusion reconstruction diagram of uterine ultrasonic and contrast images in Example 1. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] 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 with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. They should not be construed as limiting 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 scope of protection of the present invention. In the description of the present invention, it should be understood that the terms used are only for the purpose of description and should not be construed as indicating or implying relative importance.

[0060] The following combines Figures 1 - 6 to describe the feature-level image fusion and reconstruction method based on ultrasonic and contrast images provided by the present invention.

[0061] Embodiment 1: This embodiment provides a feature-level image fusion and reconstruction method based on ultrasonic and contrast images. To illustrate the technical solutions of the present invention in detail, in this embodiment, the ultrasonic and contrast images of the uterus are taken as examples for detailed description.

[0062] Overall, as Figure 1 shown, the reconstruction method includes the following steps:

[0063] S100. Perform bit-plane decomposition on the ultrasonic image and the contrast image respectively to generate 8 binary bit-plane images of each image;

[0064] S200. Calculate the weights of each bit-plane image, and perform dynamic weighted fusion on the corresponding bit-planes of the two images to generate a weighted fusion image;

[0065] S300. Flatten the weighted fusion image into one-dimensional data and perform normalization processing;

[0066] S400. Extract the first two main features with the highest contrast in the weighted fusion image as the principal components through principal feature analysis, and perform low-dimensional representation;

[0067] S500. Through the low-dimensional representation after principal feature analysis, perform inverse transformation to restore the image, and obtain the principal feature analysis fusion and reconstruction image of the ultrasonic image and the contrast image.

[0068] Step S100 is specifically as follows:

[0069] Decompose each pixel of the ultrasonic image and the contrast image bit by bit for the RGB channels. The pixel value is P = (P R , P G , P B ), where P R , P G , P B are the pixel values of the red, green, and blue channels respectively;

[0070] The pixel value of each channel is represented as an 8-bit binary number. Let P C ={p7, p6, …, p0}, where p i represents the value of the pixel value at the i-th bit, and is expressed as:

[0071] Pc = p7·2 7 + p6·2 6 + … + p1·2 1 + p0·2 0 ;

[0072] where p7 is the highest bit and p0 is the lowest bit;

[0073] Each pixel value is split into 8 bit positions {p7, p6, …, p0}, and each bit plane B q corresponds to the q-th bit of all pixels in the image, generating 8 binary bit planes of each image.

[0074] The decomposed image is as Figure 4 shown.

[0075] After the 8-bit decomposition is completed, each bit plane needs to be weighted next. Specifically, the importance of each bit plane (i.e., the contribution to the overall characteristics of the image) is evaluated according to its contrast (variance). As Figure 2 shown, step S200 includes:

[0076] S210, calculating the contrast of each bit plane:

[0077] Let B q be the q-th bit plane after image decomposition. The contrast of each bit plane in the image is calculated by variance. Let Var(B q ) be the variance of the q-th bit plane, and the calculation formula is:

[0078]

[0079] where M and N are the height and width of the image respectively, and μ q is the mean value of the q-th bit plane, and B q (i, j) represents the pixel value of the q-th bit plane of the image at the position (i, j);

[0080] S220, calculating the weight according to the contrast:

[0081] According to the calculated variance Var(B q ), assign a higher weight to the bit plane with higher contrast. The formula for weight assignment is expressed as:

[0082]

[0083] where W q is the weight of the q-th bit plane, and ω is a constant used to prevent the case where the variance is zero.

[0084] S230. Reconstruct the image according to the weights:

[0085] Sum the weighted results of each bit plane to obtain a weighted image L, expressed as:

[0086]

[0087] where L(i, j) represents the pixel value of the weighted image L at the position (i, j).

[0088] By weighted reconstruction of the image, the main features of the high-contrast bit planes can be retained, while removing noise and minor details in the planes. This step aims to extract and enhance the main visual information of the image and reduce unnecessary redundancy and noise. The weighted reconstructed uterine image is as Figure 5 shown. In this weighting method, the high-contrast bit planes contribute more, thus retaining the main features of the image (such as contours and details), while the low-contrast bit planes (usually containing noise) have less impact on the final image.

[0089] To further extract the main features of the image, principal feature analysis is performed on the 8-bit plane images. Before that, the weighted reconstructed image needs to be flattened into one-dimensional data and normalized. The specific steps of step S300 are as follows:

[0090] First, perform bit plane decomposition on the weighted fusion image, and flatten the decomposed bit plane images to obtain a one-dimensional vector for each image. For the i-th image A i , its flattened vector is expressed as:

[0091] A i = [A i (1, 1), A i (1, 2), …, A i (M, N)] T ;

[0092] where M and N are the height and width of the image respectively;

[0093] Then normalize the one-dimensional vector so that the mean of each feature is zero and the variance is 1, expressed as:

[0094]

[0095] where μ is the mean of each feature and σ is the standard deviation of each feature.

[0096] Through principal feature analysis, the first two main features with the highest contrast in the image can be extracted. As Figure 3 shown, the specific method of step S400 is as follows:

[0097] First, select the most important k principal components, which can capture the essential information of the image. Through linear transformation, a new coordinate system is obtained such that the projection variance of the original data in this coordinate system is the largest. Each principal feature corresponds to an eigenvector, and these eigenvectors represent the main change directions of the data;

[0098] Each row of the data matrix is a flattened vector of the image. Calculate the covariance matrix of the data Then solve the eigenvalues and eigenvectors of this covariance matrix. Select the eigenvectors with the largest eigenvalues as the principal components, and project the data onto these principal features to obtain the data representation after dimensionality reduction:

[0099] Z i =A i ′·P k ;

[0100] where, Z i is the data representation after dimensionality reduction, and P k is the matrix containing the eigenvectors of the first k principal components.

[0101] Through the low-dimensional representation after principal feature analysis, perform inverse transformation to restore the image, retain the most important features, and present the core structure of the image. What is obtained in the process of step S500 is a "simplified" image, which removes the information at the detail level and retains the main content of the image. This approach not only effectively reduces the complexity of the image but also enables focusing on the key information in the image. Finally, through inverse transformation, the analyzed data is restored to the original image space, expressed as:

[0102]

[0103] where, A″ i is the principal feature analysis fusion reconstruction image restored to the original image space, expressed as the data after inverse transformation; is the transpose of the eigenvector matrix P k .

[0104] The restored image is as Figure 6 shown, and it can be seen that the restored image can be reconstructed according to the principal components, retaining the most important image features.

[0105] Embodiment 2: This embodiment provides a non-transitory computer-readable storage medium, on which computer instructions are stored. These computer instructions cause the computer to execute a method for feature-level image fusion and reconstruction based on ultrasonic and contrast image features. The method includes the following steps:

[0106] S100. Perform bit-plane decomposition on the ultrasonic image and the contrast-enhanced image respectively to generate 8 binary bit-planes of each image;

[0107] S200. Calculate the weights of each bit-plane image, and reconstruct the image according to the weights to generate a new image;

[0108] S300. Flatten the weighted and reconstructed image into one-dimensional data and perform normalization processing;

[0109] S400. Extract the first two main features with the highest contrast in the image as the principal components through principal feature analysis and perform low-dimensional representation;

[0110] S500. Perform inverse transformation through the low-dimensional representation after principal feature analysis to restore the image, and obtain the principal feature analysis fusion reconstruction map of the ultrasonic image and the contrast-enhanced image.

[0111] Embodiment 3: This embodiment provides an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus. The processor can call the logical instructions in the memory to execute the method for ultrasonic and contrast-enhanced image feature-level image fusion reconstruction, and the method includes the following steps:

[0112] S100. Perform bit-plane decomposition on the ultrasonic image and the contrast-enhanced image respectively to generate 8 binary bit-planes of each image;

[0113] S200. Calculate the weights of each bit-plane image, and reconstruct the image according to the weights to generate a new image;

[0114] S300. Flatten the weighted and reconstructed image into one-dimensional data and perform normalization processing;

[0115] S400. Extract the first two main features with the highest contrast in the image as the principal components through principal feature analysis and perform low-dimensional representation;

[0116] S500. Perform inverse transformation through the low-dimensional representation after principal feature analysis to restore the image, and obtain the principal feature analysis fusion reconstruction map of the ultrasonic image and the contrast-enhanced image.

[0117] In addition, when the logical instructions in the above-mentioned memory are implemented in the form of software functional units and sold or used as independent products, they can 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, can 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 can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0118] Embodiment 4: This embodiment provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute an image fusion reconstruction method based on ultrasonic and contrast image feature levels. The method includes the following steps:

[0119] S100. Perform bit-plane decomposition on the ultrasonic image and the contrast image respectively to generate 8 binary bit-planes of each image;

[0120] S200. Calculate the weights of each bit-plane image, and reconstruct the image according to the weights to generate a new image;

[0121] S300. Flatten the weighted reconstructed image into one-dimensional data and perform normalization processing;

[0122] S400. Extract the first two main features with the highest contrast in the image as the main components through principal feature analysis and perform low-dimensional representation;

[0123] S500. Through the low-dimensional representation after principal feature analysis, perform inverse transformation to restore the image to obtain the principal feature analysis fusion reconstruction map of the ultrasonic image and the contrast image.

[0124] 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 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.

[0125] 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. This 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 to enable 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.

[0126] 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 feature-level image fusion reconstruction method based on ultrasound and contrast images, characterized in that: include: S100, performing bit plane decomposition on the ultrasound image and the contrast image respectively to generate 8 binary bit plane images of each image; S200, calculating the weight of each bit plane image, and dynamically weighted fusion the corresponding bit planes of the two images to generate a weighted fused image; S300, flattening the weighted fusion image into one-dimensional data, and performing standardization processing; S400, extracting the first two main features with the highest contrast in the weighted fusion image as main components through main feature analysis, and performing low-dimensional representation; S500, performing inverse transformation to restore the image through the low-dimensional representation after main feature analysis, and obtaining a main feature analysis fusion reconstructed image of the ultrasound image and the contrast image.

2. The method for image fusion and reconstruction based on feature level of ultrasound and contrast imaging according to claim 1, characterized in that: Step S100 is specifically as follows: Decompose the RGB channels of each pixel of the ultrasound image and the contrast image bit by bit, and the pixel value is P = (P R ,P G ,P B ), where P R ,P G ,P B are the pixel values ​​of the red, green, and blue channels respectively; The pixel value of each channel is represented by an 8-bit binary number. C ={p7,p6,…,p0}, where p i Indicates the value of the pixel at the i-th position, expressed as: Pc=p7·2 7 +p6·2 6 +…+p1·2 1 +p0·2 0 ; Among them, p7 is the highest bit and p0 is the lowest bit; Each pixel value is split into 8 bits {p7, p6, ..., p0}, each bit plane B q Eight binary bit planes of each image are generated corresponding to the qth bit of all pixels in the image.

3. The method for image fusion and reconstruction based on feature level of ultrasound and contrast imaging according to claim 1, characterized in that: Step S200 calculates the weight of each bit plane image, and dynamically weights and fuses the corresponding bit planes of the two images to generate a weighted fused image; including: S210, calculating the contrast of each bit plane of the two images, and converting it into a weight, so that a bit plane with high contrast obtains a larger weight, and a bit plane with low contrast obtains a smaller weight; S220 , summing up the weighted results of each bit plane to obtain a weighted fused image.

4. The method for image fusion and reconstruction based on feature level of ultrasound and contrast imaging according to claim 3, characterized in that: The specific method of step S210 is: Assume B q is the qth bit plane after the image is decomposed. The contrast of each bit plane in the image is calculated by variance. Let Var(B q ) is the variance of the q-th bit plane, and the calculation formula is: Where M and N are the height and width of the image, respectively, and μ q is the mean of the qth bit plane, B q (i, j) represents the pixel value of the qth bit plane of the image at position (i, j); According to the calculated variance Var(B q ), higher weights are assigned to bit planes with higher contrast. The weight assignment formula is expressed as: Where W q is the weight of the qth bit plane, and ω is a constant used to prevent the variance from being zero.

5. The method for image fusion and reconstruction based on feature level of ultrasound and contrast imaging according to claim 4, characterized in that: The specific method of step S220 is: The weighted results of each bit plane are summed to obtain the weighted image L, which is expressed as: Among them, L(i,j) represents the pixel value of the weighted image L at position (i,j).

6. The method for image fusion and reconstruction based on feature level of ultrasound and contrast imaging according to claim 1, characterized in that: Step S300 flattens the weighted fusion image into one-dimensional data and performs standardization processing. The specific steps are: First, the weighted fusion image is decomposed into bit planes, and the decomposed bit plane images are flattened to obtain a one-dimensional vector of each bit plane image. For the i-th image A i , whose flattened vector representation is: A i =[A i (1,1),A i (1,2),…,A i (M,N)] T ; Where M and N are the height and width of the image respectively; Then the one-dimensional vector is standardized so that the mean of each feature is zero and the variance is 1, expressed as: Here, μ is the mean of each feature and σ is the standard deviation of each feature.

7. The method for image fusion and reconstruction based on feature level of ultrasound and contrast imaging according to claim 6, characterized in that: In step S400, the main feature analysis extracts the first two main features with the highest contrast of the weighted fusion image as the main components by using the feature extraction method, calculates the covariance matrix of the image, and finds its eigenvector. Extract the main feature directions; First, the most important k principal components are selected and a new coordinate system is obtained through linear transformation so that the projection variance of the original data in the coordinate system is maximized, and each principal feature corresponds to a eigenvector; Calculate the covariance matrix of the data, then solve the eigenvalues ​​and eigenvectors of the covariance matrix, select the eigenvector with the largest eigenvalue as the principal component, and project the data onto these principal features to obtain the data representation after dimensionality reduction: Z i =A i ′·P k ; Among them, Z i is the data representation after dimensionality reduction, P k is a matrix containing the first k principal component eigenvectors.

8. The method for image fusion and reconstruction based on feature level of ultrasound and contrast imaging according to claim 7, characterized in that: In step S500, the analyzed data is restored to the original image space through inverse transformation, which is expressed as: Among them, A″ i ′ is the main feature analysis fusion reconstructed image restored to the original image space, expressed as the data after inverse transformation; is the eigenvector matrix P k The transpose of .

9. A non-transitory computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and the computer instructions enable the computer to execute the image fusion reconstruction method based on ultrasound and contrast image feature level as described in any one of claims 1-8.

10. An electronic device, characterized in that: include: A processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus, and the processor calls the logic instructions in the memory to execute the feature-level image fusion reconstruction method based on ultrasound and contrast images as described in any one of claims 1-8.

Images