Matrix imaging method and apparatus based on laplacian operator

Abstract

The application belongs to the technical field of ultrasonic imaging, and discloses a matrix imaging method based on a Laplacian operator, which comprises the following steps: performing feature extraction on an original image by using a Laplacian operator to obtain feature information; obtaining a dynamic normalized mean square error between the original image and a reconstructed image based on the feature information; judging whether the dynamic normalized mean square error calculation result is within a preset range; if yes, performing full-focus imaging processing on the corresponding original image to obtain a target image. The intermediate image data is evaluated once before imaging, the original image is selectively subjected to imaging processing, the overall calculation amount of ultrasonic imaging is reduced by screening the original image, and thus the working efficiency of ultrasonic imaging during in-vivo tissue detection is improved.

Description

Technical Field

[0001] This invention relates to the field of ultrasound imaging technology, and more specifically to a matrix imaging method and device based on the Laplacian operator. Background Technology

[0002] When ultrasound waves propagate through human tissues, they are affected by tissues such as muscles and fat, resulting in beam diffraction and scattering, which reduces imaging resolution and contrast. Based on this, three-dimensional ultrasound matrix imaging technology has been developed. This technology uses a two-dimensional matrix probe to simultaneously emit multiple ultrasound waves and records the tissue reflection echoes to form a "reflection matrix". By analyzing the reflection matrix, the diffraction and scattering effects of tissue on sound waves are quantified, and a "transmission matrix" is extracted and digitally compensated to improve the beam focusing quality.

[0003] However, when applying matrix imaging technology to living organisms, the difficulties caused by tissue movement still need further research. For example, in vascular imaging, blood flow causes rapid changes in reflectivity over time, which can affect the imaging effect. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a matrix imaging method and device based on the Laplacian operator to solve the problem that the imaging effect is poor when matrix imaging is applied to living organisms due to tissue movement.

[0005] This invention provides a matrix imaging method based on the Laplacian operator, comprising:

[0006] Feature extraction is performed on the original image using the Laplacian operator to obtain feature information;

[0007] The dynamic normalized mean square error between the original image and the reconstructed image based on feature information is obtained;

[0008] Determine whether the calculated result of the dynamic normalized mean square error is within the preset range;

[0009] If so, the corresponding original image will be subjected to full-focus imaging processing to obtain the target image.

[0010] Optionally, before performing feature extraction on the original image using the Laplacian operator to obtain feature information, the following steps are also included:

[0011] The original image is smoothed using a smoothing filter.

[0012] Optionally, in the step of extracting features from the original image using the Laplacian operator to obtain feature information, the calculation method for the feature information includes:

[0013] I sharp =I+k(II blur );

[0014] Where I is the original image, and k is the sharpening intensity coefficient, I blur The image after smoothing filtering, I sharp These are the image features after processing with the Laplacian operator.

[0015] Optionally, in the step of obtaining the dynamic normalized mean square error between the original image and the reconstructed image based on feature information, the method for calculating the dynamic normalized mean square error includes:

[0016] ;

[0017] ;

[0018] ;

[0019] ;

[0020] ;

[0021] Among them, X i Y represents the value of the i-th pixel in the original image. i This represents the value of the i-th pixel in the reconstructed image, where N is the total number of pixels in the image. Let be the dynamic normalization factor for the i-th pixel, α be a weighting factor for balancing local contrast, β be a weighting factor for balancing gradient magnitude, and γ be a weighting factor for balancing global brightness deviation. It is a positive constant. Let W represent the standard deviation of the brightness of the i-th pixel, and let W represent the size of the sliding window. This represents the average local brightness within the sliding window window(j). μ is the squared gradient magnitude of the i-th pixel. global This represents the global average brightness.

[0022] Optionally, the preset range is 0 to 0.5.

[0023] Optionally, in the step of performing full-focus imaging processing on the corresponding original image to obtain the target image, the full-focus imaging processing method includes:

[0024] ;

[0025] Where M is the number of array elements, V ab This represents the signal amplitude transmitted by element a and received by element b.

[0026] A coordinate system is established with the probe center as the origin, the length direction of the test block as the x-axis, and the depth direction of the test block as the z-axis. The propagation time of the ultrasonic signal emitted from the transmitting element, passing through the spatial point (x, z), and then to the receiving element is:

[0027] ;

[0028] Where, x a Indicates the coordinates of the transmitting array element, x b The coordinates of the receiving array element are represented by , and c is the speed at which the ultrasonic wave propagates in the test block.

[0029] Optionally, after performing full-focus imaging processing on the corresponding original image to obtain the target image, the method further includes:

[0030] Peak signal-to-noise ratio and structural similarity of the target image are calculated separately for quality assessment;

[0031] Retain the target images whose quality assessment results are satisfactory.

[0032] Optionally, the peak signal-to-noise ratio (PSNR) and structural similarity of the target image are calculated separately for the quality assessment step. The method for calculating the PSNR of the target image includes:

[0033] Calculate the mean square error of all pixels in the target image and the original image:

[0034] ;

[0035] Calculate the peak signal-to-noise ratio (PSNR) of the target image and the original image:

[0036] ;

[0037] Where m×n represents the size of the target image, X(c,d) represents the original image, Y(c,d) represents the target image, and MAX... l The maximum pixel value of the target image is given. The condition for passing the quality assessment is that the calculated peak signal-to-noise ratio is greater than 30.

[0038] Optionally, the peak signal-to-noise ratio and structural similarity of the target image are calculated separately for the quality assessment step. The methods for calculating the structural similarity of the target image include:

[0039] ;

[0040] ;

[0041] Where, μ X μ is the average gray level of the original image X. Y It is the average gray level of the reconstructed image Y, σ XIt is the contrast of the original image X, σ Y It represents the contrast of the reconstructed image Y, c1=(k1L) 2 c2=(k2L) 2 L is the dynamic range of pixel values, k1=0.01, k2=0.03, and the condition for passing the quality assessment is that the calculated result of structural similarity is greater than 0.3.

[0042] This invention provides a matrix imaging device based on the Laplacian operator, comprising:

[0043] The memory and processor are interconnected and communicate with each other. The memory stores computer instructions, and the processor executes the aforementioned matrix imaging method based on the Laplacian operator by executing the computer instructions.

[0044] The beneficial effects of this invention are:

[0045] 1. This embodiment provides a matrix imaging method based on the Laplacian operator. When performing ultrasound imaging on living tissue, image features are extracted by the Laplacian operator during the sound wave acquisition and processing stage. In ultrasound-photoacoustic combined imaging in the medical field, edge alignment and fusion of multimodal data can be achieved, enhancing the specificity of tumor detection.

[0046] 2. Based on the image features extracted by the Laplacian operator, a reconstructed image is obtained. The dynamic normalized mean square error is calculated on the reconstructed image to determine whether the current original image can generate a qualified target image. Since the intermediate image data is evaluated before imaging, the original image is selectively processed. By screening the original image, the overall computational load of ultrasound imaging is reduced, thereby improving the working efficiency of ultrasound imaging in live tissue detection. Attached Figure Description

[0047] The features and advantages of the invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the invention in any way. In the drawings:

[0048] Figure 1 A flowchart of a matrix imaging method based on the Laplacian operator is shown in an embodiment of the present invention;

[0049] Figure 2 A structural diagram of a matrix imaging system based on the Laplacian operator is shown in an embodiment of the present invention.

[0050] Figure 3 This diagram illustrates the principle of full-matrix data capture in a full-focusing algorithm according to an embodiment of the present invention.

[0051] Figure 4The diagram illustrates the imaging processing principle of a total focusing algorithm according to an embodiment of the present invention.

[0052] Figure 5 A structural diagram of a matrix imaging device based on the Laplacian operator is shown in an embodiment of the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] like Figure 1 As shown, this embodiment of the invention provides a matrix imaging method based on the Laplacian operator, including:

[0055] Step S10: Use the Laplacian operator to extract features from the original image to obtain feature information.

[0056] As a second-derivative operator, the Laplacian operator can keenly capture dramatic changes in brightness in an image, accurately highlighting high-frequency information such as edges and textures. Whether identifying the edges of subtle lesions in medical imaging or locating microcracks in materials during industrial inspection, it can effectively delineate the contours of key structures and improve the clarity of image details.

[0057] In this embodiment, the ultrasonic matrix probe acquires acoustic wave information, i.e., the original image, and the Laplacian operator is used to extract features from the original image.

[0058] Step S20: Obtain the dynamic normalized mean square error between the original image and the reconstructed image based on feature information.

[0059] In this embodiment, the dynamic normalized mean square error between the original image and the reconstructed image is calculated by setting a normalization factor related to brightness, contrast, and gradient magnitude.

[0060] Step S30: Determine whether the result of the dynamic normalized mean square error calculation is within the preset range.

[0061] If so, proceed to step S40 to perform full-focus imaging processing on the corresponding original image to obtain the target image.

[0062] In this embodiment, it is determined whether the result of the dynamic normalized mean square error calculated in step S20 is within a preset range. This is used to preliminarily assess whether the currently acquired acoustic information can generate a target image of acceptable quality. If the assessment result is acceptable, that is, the dynamic normalized mean square error between the original image and the reconstructed image is within a preset range, then the corresponding original image is subjected to full-focus imaging processing to obtain the target image; otherwise, the current original image is discarded and no imaging processing is performed.

[0063] This embodiment proposes a matrix imaging method based on the Laplacian operator. During ultrasound imaging of living tissue, image features are extracted using the Laplacian operator during the sound wave acquisition and processing stages. In medical ultrasound-photoacoustic combined imaging, this method enables edge alignment and fusion of multimodal data, enhancing the specificity of tumor detection. Based on the image features extracted by the Laplacian operator, a reconstructed image is obtained. The dynamically normalized mean square error is calculated on this reconstructed image to determine whether the current original image can generate a qualified target image. Because intermediate image data is evaluated before imaging, and the original image is selectively processed, the overall computational load of ultrasound imaging is reduced by filtering the original image, thereby improving the efficiency of ultrasound imaging in living tissue detection.

[0064] As an optional implementation, before performing feature extraction on the original image using the Laplacian operator to obtain feature information, the following steps are also included:

[0065] The original image is then subjected to a smoothing filter.

[0066] In this embodiment, as Figure 2 As shown, the two-dimensional matrix ultrasound probe 10 inputs the acquired sound wave information into the smoothing filter 20 for smoothing and filtering processing.

[0067] As an optional implementation, in the step of extracting features from the original image using the Laplacian operator to obtain feature information, the method for calculating the feature information includes:

[0068] I sharp =I+k(II blur );

[0069] Where I is the original image, and k is the sharpening intensity coefficient, I blur The image after smoothing filtering, I sharp These are the image features after processing with the Laplacian operator.

[0070] In this embodiment, as Figure 2 As shown, after the original image data is processed by a smoothing filter, in order to effectively capture image features, the Laplacian operator is introduced to preprocess the image to obtain image features.

[0071] As an optional implementation, in the step of obtaining the dynamic normalized mean square error between the original image and the reconstructed image based on feature information, the method for calculating the dynamic normalized mean square error includes:

[0072] ;

[0073] ;

[0074] ;

[0075] ;

[0076] ;

[0077] Among them, X i Y represents the value of the i-th pixel in the original image. i This represents the value of the i-th pixel in the reconstructed image, where N is the total number of pixels in the image. Let be the dynamic normalization factor for the i-th pixel, α be a weighting factor for balancing local contrast, β be a weighting factor for balancing gradient magnitude, and γ be a weighting factor for balancing global brightness deviation. It is a positive constant. Let W represent the standard deviation of the brightness of the i-th pixel, and let W represent the size of the sliding window. This represents the average local brightness within the sliding window window(j). μ is the squared gradient magnitude of the i-th pixel. global This represents the global average brightness.

[0078] In this embodiment, Dynamic Normalized Mean Squared Error (DNMSE) is used as an evaluation metric for the correlation between image features and the original image. The DMNSE metric establishes a dynamic mapping relationship between local feature gradients and error weights, enabling enhanced perception of errors in highly variable regions such as edges and textures. Mean Squared Error (X... i -Y i ) 2 It is used to measure the reconstruction error of each pixel, and gives higher weight to larger errors. This is used to dynamically calculate and integrate multiple local features (brightness, contrast, and gradient magnitude), adjusting the contribution of different pixels to the overall error. α, β, and γ are used to balance the contribution of each feature to the normalization factor, controlling the weights of local contrast, gradient magnitude, and global brightness deviation in that order, and can be determined through grid search optimization. ϵ is a positive constant used to avoid a denominator of zero. This represents the standard deviation of the brightness of the i-th pixel, used to reflect local contrast. The squared gradient magnitude of the i-th pixel is used to reflect the edge or texture intensity. (X) i -μ global ) 2 Let the i-th pixel be the global average brightness μ. global The squared difference is used to measure the degree of deviation from the overall brightness of the image. The smaller the DMNSE value, the smaller the error and the higher the similarity.

[0079] In a specific embodiment, the preset range is 0~0.5. If the DMNSE value exceeds 0.5, it is determined that the reconstructed image in the middle does not match the original image.

[0080] As an optional implementation, the step of performing full-focus imaging processing on the corresponding original image to obtain the target image includes the following methods for full-focus imaging processing:

[0081] ;

[0082] Where M is the number of array elements, V ab This represents the signal amplitude transmitted by element a and received by element b.

[0083] A coordinate system is established with the probe center as the origin, the length direction of the test block as the x-axis, and the depth direction of the test block as the z-axis. The propagation time of the ultrasonic signal emitted from the transmitting element, passing through the spatial point (x, z), and then to the receiving element is:

[0084] ;

[0085] Where, x a Indicates the coordinates of the transmitting array element, x b The coordinates of the receiving array element are represented by , and c is the speed at which the ultrasonic wave propagates in the test block.

[0086] In this embodiment, the Total Focusing Method (TFM) is an image post-processing algorithm based on full matrix data, such as... Figure 3 As shown, assume that a linear phased array ultrasound consists of M array elements, all of which act as both transmitters and receivers. Let V be the signal transmitted and received by the first element. 11 The signal V is transmitted by the first array element and received by the second array element. 12 Similarly, the entire matrix data contains M×M A-wave signals. For example... Figure 4As shown, the points inside the test block are discretized and used as virtual focal points. According to the superposition principle, the signal at the virtual focal point can be considered as the linear superposition of each wave at that point, and the amplitude is also the vector sum of each wave. Therefore, the TFM algorithm based on full matrix data can achieve focusing at every point inside the material through virtual focal points, thereby improving imaging resolution. The full matrix data is a symmetric matrix. If the slight differences between different array elements are ignored, the echo signal can be considered equal to the received signal. When only the self-transmitted and self-received signals of each array element are collected, the full matrix becomes a diagonal matrix, and the TFM algorithm is equivalent to synthetic aperture focusing imaging.

[0087] As an optional implementation, after performing full-focus imaging processing on the corresponding original image to obtain the target image, the method further includes:

[0088] Peak signal-to-noise ratio and structural similarity of the target image are calculated separately for quality assessment;

[0089] Retain the target images whose quality assessment results are satisfactory.

[0090] As an optional implementation, the peak signal-to-noise ratio (PSNR) and structural similarity of the target image are calculated separately for the quality assessment step. The method for calculating the PSNR of the target image includes:

[0091] Calculate the mean square error of all pixels in the target image and the original image:

[0092] ;

[0093] Calculate the peak signal-to-noise ratio (PSNR) of the target image and the original image:

[0094] ;

[0095] Where m×n represents the size of the target image, X(c,d) represents the original image, Y(c,d) represents the target image, and MAX... l The maximum pixel value of the target image is given. The condition for passing the quality assessment is that the calculated peak signal-to-noise ratio is greater than 30.

[0096] As an optional implementation, the peak signal-to-noise ratio and structural similarity of the target image are calculated separately for the quality assessment step. The method for calculating the structural similarity of the target image includes:

[0097] ;

[0098] ;

[0099] Where, μ X μ is the average gray level of the original image X. Y It is the average gray level of the reconstructed image Y, σ XIt is the contrast of the original image X, σ Y It represents the contrast of the reconstructed image Y, c1=(k1L) 2 c2=(k2L) 2 L is the dynamic range of pixel values, k1=0.01, k2=0.03, and the condition for passing the quality assessment is that the calculated result of structural similarity is greater than 0.3.

[0100] This invention also provides a matrix imaging device based on the Laplacian operator, such as... Figure 5 As shown, the device may include a processor 51 and a memory 52, wherein the processor 51 and the memory 52 may be connected via a bus or other means. Figure 5 Taking the example of a connection between China and Israel via a bus.

[0101] Processor 51 can be a central processing unit (CPU). Processor 51 can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations of the above types of chips.

[0102] The memory 52, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules. The processor 51 executes various functional applications and data processing by running the non-transitory software programs, instructions, and modules stored in the memory 52, thereby implementing the matrix imaging method based on the Laplacian operator in the above method embodiments.

[0103] The memory 52 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by the processor 51, etc. Furthermore, the memory 52 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory 52 may optionally include memory remotely located relative to the processor 51, and these remote memories may be connected to the processor 51 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0104] The one or more modules are stored in the memory 52, and when executed by the processor 51, they perform the following: Figure 1-4 The matrix imaging method based on the Laplacian operator in the illustrated embodiment.

[0105] For specific details regarding the aforementioned matrix imaging device based on the Laplacian operator, please refer to the relevant documentation. Figures 1 to 4 The relevant descriptions and effects in the illustrated embodiments are for understanding purposes only and will not be repeated here.

[0106] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0107] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A matrix imaging method based on the Laplacian operator, characterized in that, include: Feature extraction is performed on the original image using the Laplacian operator to obtain feature information; The dynamic normalized mean square error between the original image and the reconstructed image based on the feature information is obtained. Determine whether the calculated result of the dynamic normalized mean square error is within the preset range; If so, the corresponding original image is subjected to full-focus imaging processing to obtain the target image; In the step of obtaining the dynamic normalized mean square error between the original image and the reconstructed image based on the feature information, the method for calculating the dynamic normalized mean square error includes: ； ； ； ； ； Among them, X i represents the value of the i-th pixel in the original image, Y i represents the value of the i-th pixel in the reconstructed image, N is the total pixel value of the image, is the dynamic normalization factor of the i-th pixel, α is the weight factor for balancing local contrast, β is the weight factor for balancing gradient magnitude, γ is the weight factor for balancing global brightness deviation, is a positive constant, represents the standard deviation of brightness of the i-th pixel, W represents the size of the sliding window, represents the local brightness average within the sliding window widow(j), is the square of the gradient magnitude of the i-th pixel, μ global represents the global average brightness.

2. The matrix imaging method based on the Laplacian operator according to claim 1, characterized in that, Before performing feature extraction on the original image using the Laplacian operator to obtain feature information, the process also includes: The original image is then subjected to a smoothing filter.

3. The matrix imaging method based on the Laplacian operator according to claim 2, characterized in that, In the step of extracting features from the original image using the Laplacian operator, the methods for calculating the feature information include: I sharp =I+k(I-I blur ); Where I is the original image, and k is the sharpening intensity coefficient, I blur The image after smoothing filtering, I sharp These are the image features after processing with the Laplacian operator.

4. The matrix imaging method based on the Laplacian operator according to claim 1, characterized in that, The preset range is 0~0.

5.

5. The matrix imaging method based on the Laplacian operator according to claim 1, characterized in that, In the step of performing full-focus imaging processing on the corresponding original image to obtain the target image, the full-focus imaging processing method includes: ； Where M is the number of array elements, V ab This represents the signal amplitude transmitted by element a and received by element b. A coordinate system is established with the probe center as the origin, the length direction of the test block as the x-axis, and the depth direction of the test block as the z-axis. The propagation time of the ultrasonic signal emitted from the transmitting element, passing through the spatial point (x, z), and then to the receiving element is: ； Where, x a Indicates the coordinates of the transmitting array element, x b The coordinates of the receiving array element are represented by , and c is the speed at which the ultrasonic wave propagates in the test block under test.

6. The matrix imaging method based on the Laplacian operator according to claim 1, characterized in that, After performing full-focus imaging processing on the corresponding original image to obtain the target image, the process further includes: The peak signal-to-noise ratio and structural similarity of the target image are calculated respectively for quality assessment; The target images whose quality assessment results are satisfactory are retained.

7. The matrix imaging method based on the Laplacian operator according to claim 6, characterized in that, In the quality assessment step, the method for calculating the peak signal-to-noise ratio (PSNR) and structural similarity of the target image includes: Calculate the mean square error of all pixels in the target image and the original image: ； Calculate the peak signal-to-noise ratio (PSNR) of the target image and the original image: ； Where m×n represents the size of the target image, X(c,d) represents the original image, Y(c,d) represents the target image, and MAX... l The maximum pixel value of the target image is defined as the condition for passing the quality assessment, which is that the calculated peak signal-to-noise ratio is greater than 30.

8. The matrix imaging method based on the Laplacian operator according to claim 6, characterized in that, In the quality assessment step, the method for calculating the structural similarity of the target image includes: ； ； Where, μ X μ is the average gray level of the original image X. Y It is the average gray level of the reconstructed image Y, σ X It is the contrast of the original image X, σ Y It represents the contrast of the reconstructed image Y, c1=(k1L) 2 c2=(k2L) 2 L is the dynamic range of pixel values, k1=0.01, k2=0.03, and the condition for passing the quality assessment is that the calculated result of structural similarity is greater than 0.

3.

9. A matrix imaging device based on the Laplacian operator, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the matrix imaging method based on the Laplacian operator as described in any one of claims 1-8.

Images