Automatic measurement method and system for imaging ghosting index of nuclear magnetic resonance equipment, medium and product

Abstract

The invention discloses an automatic measurement method and system for imaging ghosting indexes of nuclear magnetic resonance equipment, a medium and a product. The automatic measurement method comprises the following steps: acquiring an image file in a DICOM (Digital Imaging and Communications in Medicine) format; performing pixel extraction and format conversion on the image file in the DICOM format to obtain a second image in an 8-bit format; performing region automatic identification on the second image to obtain a signal region, a ghosting region and a noise region; respectively calculating to obtain statistical characteristics of the signal area, the ghosting area and the noise area; and calculating a ghosting index according to the statistical characteristics of the signal area, the ghosting area and the noise area. According to the method, the ghost related indexes can be automatically calculated, manual intervention is reduced, and the detection efficiency and consistency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear magnetic resonance equipment imaging technology, and in particular, to an automatic measurement method, system, medium and product for the ghost artifact index of nuclear magnetic resonance equipment imaging. Background Art

[0002] During the operation of the MRI system, various artifacts are extremely likely to appear. Among them, "ghost artifacts" are common and have a relatively obvious impact. Among them, ghost artifacts refer to repeated or blurred artifacts that appear in nuclear magnetic resonance images, usually formed along a specific direction (such as the phase encoding direction). The reasons for their generation include gradient imbalance, patient movement, incomplete data acquisition, or system calibration deviation, etc. Ghost artifacts not only affect the beauty of the image, but also interfere with the accurate judgment of real lesions or structures. Therefore, to ensure the accuracy and contrast of subsequent diagnostic images, it is usually necessary to evaluate ghost artifacts in the pre-scan or daily quality assurance (QA) process, that is, to calculate the ghost artifact index.

[0003] Existing practices mostly rely on manual selection of the background area and the artifact area to calculate the ghost artifact index, which is not only inefficient, but also easily affected by the experience and subjective judgment of the operator, resulting in unstable detection results.

[0004] For example, the Chinese patent application with the publication number CN101322648A proposes a measurement method for automatically calculating nuclear magnetic resonance (MRI) ghost artifacts. This method first uses the iterative threshold method to segment the image, dividing the image into the foreground (signal area), background (noise area) and ghost area. Subsequently, by extracting the pixel information of each area and combining statistical analysis methods, the ghost value index is finally calculated to achieve an automated and objective evaluation of the ghost intensity. However, the iterative threshold method is sensitive to noise. Especially when the noise is strong, the foreground and background areas often overlap each other, resulting in the appearance of a large number of isolated pixel areas, thereby causing a large deviation in the calculation result and affecting the accuracy and stability of the ghost artifact index. Summary of the Invention

[0005] To solve the technical problems existing in the background art, the present invention proposes an automatic measurement method, system, medium and product for the ghost artifact index of nuclear magnetic resonance equipment imaging.

[0006] The automatic measurement method, system, medium and product for the ghost artifact index of nuclear magnetic resonance equipment imaging proposed by the present invention include: Obtain an image file in DICOM format; Perform pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format; Perform automatic region recognition on the second image to obtain the signal area, the ghost area and the noise area; Calculate the statistical features of the signal region, the ghost region, and the noise region respectively; Calculate the ghost index based on the statistical features of the signal region, the ghost region, and the noise region.

[0007] Preferably, perform pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format, specifically including: Parse and display the DICOM format image file, and convert it from high-bit pixels to a first image in 16-bit format and a second image in 8-bit format through the tag window width and window level of the image file.

[0008] Preferably, the conversion formula of the second image is expressed as: ; In the formula, represents the displayed gray value, represents the window width, represents the window level.

[0009] Preferably, perform automatic region recognition on the second image to obtain the signal region, the ghost region, and the noise region, specifically including: Perform binarization processing on the second image using the Otsu threshold method to obtain a first binarized image; Perform boundary tracking on the first binarized image using the Suzuki-Abe boundary tracking algorithm, extract all the closed curves of the signal region, and select the closed curve with the largest area of the signal region as the contour of the signal region; Obtain four-direction noise rectangular regions according to the contour of the signal region and the first binarized image; Obtain the contour of the ghost region according to the contour of the signal region and the first image.

[0010] Preferably, obtain four-direction noise rectangular regions according to the contour of the signal region and the first binarized image, specifically including: According to the contour of the signal region, calculate the minimum circumscribed rectangle of the signal region; extract the four corner points of the minimum circumscribed rectangle of the signal region, and at the same time define the four corner points of the entire first binarized image as reference corner points; calculate the coordinates of the center points between each corner point of the minimum circumscribed rectangle and the corresponding reference corner point respectively, and use the coordinates of each center point as the center point of a noise rectangular region respectively; calculate the distances between each corner point of the minimum circumscribed rectangle and the corresponding reference corner point in the x-axis and y-axis directions respectively, and take half of them as the width and height of the corresponding noise rectangular region to obtain four-direction noise rectangular regions.

[0011] Preferably, obtain the contour of the ghost region according to the contour of the signal region and the first image, specifically including: Calculate the centroid of the contour of the signal region, calculate the direction vectors of each contour point in the contour of the signal region and the centroid, and then magnify all the direction vectors according to a preset ratio to obtain an intermediate signal region; map the intermediate signal region to the corresponding position in the first image to obtain an intermediate image block; remove the intermediate image block from the first image to obtain a ghost region to be recognized. Perform Gaussian filtering on the ghost region to be recognized, and use the Otsu threshold method to perform binary processing on the Gaussian-filtered ghost region to be recognized to obtain a second binary image; perform foreground erosion and then dilation on the second binary image to obtain an intermediate ghost region to be recognized; use the Suzuki-Abe boundary tracking algorithm to perform boundary tracking on the intermediate ghost region to be recognized, extract all the closed curves of the ghost regions in the intermediate ghost region to be recognized, and select the closed curve of the ghost region with the largest area as the contour of the ghost region.

[0012] Preferably, calculate the statistical features of the signal region, the ghost region, and the noise region respectively, specifically including: Determine the signal image block corresponding to the signal region in the first image, and calculate the mean value of the pixel values of the signal image block. Determine the noise image block corresponding to the noise rectangular region in the first image, and calculate the standard deviation of the pixel values of each noise image block; calculate the noise level according to the standard deviations of the pixel values of the four noise image blocks. According to the contour of the ghost region, calculate the centroid of the ghost region, and use this centroid as the center point of the ghost region to calculate the mean value of the ghost region.

[0013] In a second aspect, the present invention also proposes an automatic measurement system for imaging ghost metrics of a nuclear magnetic resonance device, including: An acquisition module for acquiring an image file in DICOM format. A conversion module for performing pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format. An identification module for automatically identifying regions in the second image file to obtain a signal region, a ghost region, and a noise region. A calculation module for calculating the statistical features of the signal region, the ghost region, and the noise region respectively; calculating the ghost metric according to the statistical features of the signal region, the ghost region, and the noise region.

[0014] In a third aspect, the present invention also proposes a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the automatic measurement method for imaging ghost metrics of a nuclear magnetic resonance device described in any item of the first aspect are implemented.

[0015] Fourthly, the present invention also provides a computer program product, including a computer program which, when executed by a processor, implements the steps of the automatic measurement method for imaging ghosting index of a nuclear magnetic resonance device as described in any one of the first aspect.

[0016] In the present invention, for the automatic measurement method, system, medium and product for the imaging ghosting index of a nuclear magnetic resonance device, by performing pixel extraction and format conversion on a DICOM format image file, a second image in 8-bit format is obtained; region automatic recognition is performed on the second image to obtain a signal region, a ghost region and a noise region; statistical features of the signal region, the ghost region and the noise region are respectively calculated; and according to the statistical features of the signal region, the ghost region and the noise region, the ghosting index is calculated. The present invention can automatically calculate ghosting-related indexes, reduce manual intervention, provide objective and standardized ghosting measurement results, and greatly improve the measurement accuracy and consistency of nuclear magnetic resonance image quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic flowchart of the automatic measurement method for the imaging ghosting index of a nuclear magnetic resonance device in an embodiment proposed by the present invention.

[0018] Figure 2 It is a schematic display diagram of a DICOM format image file in an embodiment proposed by the present invention.

[0019] Figure 3 It is a schematic diagram of the signal region in an embodiment proposed by the present invention.

[0020] Figure 4 It is a schematic diagram of the noise region in an embodiment proposed by the present invention.

[0021] Figure 5 It is a schematic diagram of the ghost region to be recognized in an embodiment proposed by the present invention.

[0022] Figure 6 It is a schematic diagram of the ghost region in an embodiment proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] It should be noted that, without conflict, the features of the embodiments and implementation manners in the present invention can be combined with each other. The present invention will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0024] First of all, it should be understood that ghosting usually refers to repeated or blurred artifacts that appear in magnetic resonance images, usually formed along a specific direction (such as the phase encoding direction). The causes of its generation include gradient imbalance, patient movement, incomplete data acquisition, or system calibration deviation, etc. Ghosting not only affects the aesthetics of the image, but also interferes with the accurate judgment of real lesions or structures. The ghosting index refers to a numerical index used to quantify the degree of ghosting in MRI images, usually characterized by calculating the signal intensity ratio or difference between the main image area and the artifact area.

[0025] In a first aspect, referring to Figure 1 , an automatic measurement method, system, medium and product for the ghosting index of magnetic resonance equipment imaging proposed by the present invention include: Obtain an image file in DICOM format; Perform pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format; Automatically identify regions of the second image to obtain a signal region, a ghost region, and a noise region; Calculate the statistical features of the signal region, the ghost region, and the noise region respectively; Calculate the ghosting index according to the statistical features of the signal region, the ghost region, and the noise region.

[0026] The present invention performs pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format; automatically identifies regions of the 8-bit format second image to obtain a signal region, a ghost region, and a noise region; calculates the statistical features of the signal region, the ghost region, and the noise region respectively; and calculates the ghosting index according to the statistical features of the signal region, the ghost region, and the noise region. The present invention can automatically calculate ghost-related indexes, reduce manual intervention, provide objective and standardized ghost measurement results, so as to improve the measurement accuracy and consistency of magnetic resonance imaging quality.

[0027] In this embodiment, performing pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format specifically includes: Parse and display the DICOM format image file; According to the displayed result, convert it from high-bit pixels to a first image in 16-bit format and a second image in 8-bit format through the tag window width and window level of the image file.

[0028] This embodiment is set in this way to ensure the correct acquisition of DICOM image data.

[0029] Specifically, the image is converted into a 16-bit first image and an 8-bit second image. The 8-bit data is the result after window width and window level adjustment. The conversion method is shown in the following formula. Window width (WW) and window level (WL) are used to adjust the gray level to make the image clearer on the display device. Medical image data of 16 bits or 12 bits can be converted into 8 bits (gray values from 0 to 255) for display.

[0030] ; In the formula, represents the displayed gray value, represents the window width, represents the window level.

[0031] The binarization processing method of this embodiment includes the average value method, the bimodal method, and the OTSU threshold method.

[0032] In this embodiment, before automatically identifying regions of the second image to obtain the signal region, the ghost region, and the noise region, it further includes: Performing Gaussian filtering on the second image to remove noise.

[0033] In this embodiment, automatically identifying regions of the second image to obtain the signal region, the ghost region, and the noise region specifically includes: Performing binarization processing on the second image using the Otsu threshold method to obtain a first binarized image; Performing boundary tracking on the first binarized image using the Suzuki-Abe boundary tracking algorithm, extracting all closed curves, and selecting the curve with the largest area as the contour of the signal region; Obtaining four-direction noise rectangular regions according to the contour of the signal region and the first binarized image; Obtaining the contour of the ghost region according to the contour of the signal region and the first image.

[0034] This embodiment uses the Otsu algorithm for foreground and background separation and combines it with the Suzuki-Abe boundary tracking algorithm, which can accurately identify the signal region and the ghost region. Since in quality assurance, the nuclear magnetic resonance equipment scans the water phantom to generate DICOM format image files. The water phantom used usually has a clear shape contour, with an obvious contrast between the foreground and the background, rather than complex textures. The Otsu + Suzuki-Abe scheme adopted in this embodiment is more suitable for this application scenario.

[0035] Among them, obtaining four-direction noise rectangular regions according to the contour of the signal region and the first binarized image specifically includes: According to the contour of the signal region, the minimum circumscribed rectangle of the signal region is calculated; Four corner points of the minimum circumscribed rectangle of the signal region are extracted, namely the upper left, upper right, lower left, and lower right. At the same time, four corner points of the entire first binary image are defined as reference corner points; The coordinates of the center points between the corner points of each minimum circumscribed rectangle and the corresponding reference corner points are calculated respectively, and the coordinates of each center point are used as the center points of a noise rectangle region; The distances between the corner points of each minimum circumscribed rectangle and the corresponding reference corner points in the x-axis and y-axis directions are calculated respectively, and half of them are taken as the width and height of the corresponding noise rectangle region, obtaining noise rectangle regions in four directions.

[0036] In this embodiment, when extracting the noise region, the range of the signal region is appropriately enlarged to reduce interference, ensuring the accuracy of the recognition region; moreover, in order to optimize the selection of the noise region, the minimum circumscribed rectangle of the signal region is calculated, and the noise region is divided between its four corner points and the four corners of the image, ensuring the rationality of the measurement.

[0037] Among them, according to the contour of the signal region and the first image, the contour of the ghost region is obtained, specifically including: The centroid of the contour of the signal region is calculated, then the direction vectors between each contour point in the contour of the signal region and the centroid are calculated, and then all the direction vectors are enlarged according to a preset ratio to obtain an intermediate signal region; the intermediate signal region is mapped to the corresponding position in the first image to obtain an intermediate image block; the intermediate image block is removed from the first image to obtain a ghost region to be recognized; The ghost region to be recognized is subjected to Gaussian filtering, and the Gaussian-filtered ghost region to be recognized is binarized using the Otsu threshold method to obtain a second binary image; The second binary image is subjected to foreground erosion and then dilation to obtain an intermediate ghost region to be recognized; All closed curves of the ghost regions in the intermediate ghost region to be recognized are extracted, and the closed curve of the ghost region with the largest area is selected as the contour of the ghost region.

[0038] In this embodiment, when extracting the ghost region, the range of the signal region is appropriately enlarged to reduce interference, ensuring the accuracy of the recognition region; and during the extraction of the ghost region, an opening operation (morphological processing) is used to remove the influence of isolated pixel regions, thereby improving the accuracy of the ghost region, making the calculated ghost index closer to its true size, and at the same time retaining the main structure of the ghost region to enhance the stability, accuracy, and robustness of the detection.

[0039] The preset ratio in this embodiment is 1.1.

[0040] In this embodiment, the statistical features of the signal region, the ghost region, and the noise region are calculated respectively, specifically including: Determine the signal image blocks corresponding to the signal region in the first image, and calculate the mean value S of the signal image blocks; Determine the noise image blocks corresponding to the noise rectangular region in the first image, and calculate the standard deviation of each noise image block; according to the standard deviation of each noise image block , calculate the noise level ; According to the contour of the ghost region, calculate the centroid of the ghost region, and use this centroid as the center point of the ghost region to calculate the mean value of the ghost region .

[0041] The statistical features of this embodiment include the mean value and the standard deviation.

[0042] In this embodiment, the ghost metrics include the ratio of the ghost to the signal, the ratio of the ghost to the noise, and the signal-to-noise ratio; Among them, the ratio of the ghost to the signal ; in the formula, represents the mean value of the ghost region, represents the mean value of the signal; The ratio of the ghost to the noise ; in the formula, represents the mean value of the ghost region, represents the mean value of the noise; ; in the formula, represents the signal-to-noise ratio, represents the mean value of the signal; represents the mean value of the noise.

[0043] During specific implementation, based on the identified signal region, noise region, and ghost region, calculate the mean value and standard deviation of each region, so as to quantify the intensity of the ghost.

[0044] In summary, the present invention can automatically calculate the ghost-related metrics, reduce manual intervention, improve the detection efficiency and consistency, and adapt to complex imaging environments. Even in the case of low signal-to-noise ratio or poor image quality, it can still stably and accurately identify the signal region and the ghost region.

[0045] In a second aspect, the present invention also proposes an automatic measurement system for imaging ghost metrics of a nuclear magnetic resonance device, including: An acquisition module, configured to acquire an image file in DICOM format; A conversion module, configured to perform pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format; An identification module, configured to perform automatic region identification on the second image file to obtain a signal region, a ghost region, and a noise region; A calculation module is configured to calculate the statistical features of the signal region, the ghost region, and the noise region respectively; and calculate the ghost index according to the statistical features of the signal region, the ghost region, and the noise region.

[0046] In this embodiment, a display module is further included; During the format conversion process, the display module is configured to parse and display the DICOM format image file; the conversion module is configured to determine whether the DICOM format image file is high-bit data greater than 8 bits; if so, perform pixel extraction on the DICOM format image file to obtain a first image, and convert the first image into a second image in 8-bit format.

[0047] The region automatic recognition process of the recognition module in this embodiment includes: Perform binarization processing on the second image using the Otsu threshold method to obtain a first binarized image; Perform boundary tracking on the first binarized image using the Suzuki-Abe boundary tracking algorithm, extract all the closed curves of the signal region, and select the closed curve with the largest area as the contour of the signal region; Obtain four-direction noise rectangular regions according to the contour of the signal region and the first binarized image; Obtain the contour of the ghost region according to the contour of the signal region and the first image.

[0048] Among them, obtaining four-direction noise rectangular regions according to the contour of the signal region and the first binarized image specifically includes: Calculate the minimum bounding rectangle of the signal region according to the contour of the signal region; extract the four corner points of the minimum bounding rectangle of the signal region, and define the four corner points of the entire first binarized image as reference corner points; calculate the coordinates of the center points between each corner point of the minimum bounding rectangle and the corresponding reference corner point respectively, and use the coordinates of each center point as the center point of a noise rectangular region; calculate the distances between each corner point of the minimum bounding rectangle and the corresponding reference corner point in the x-axis and y-axis directions respectively, and take half of them as the width and height of the corresponding noise rectangular region to obtain four-direction noise rectangular regions.

[0049] Among them, obtaining the contour of the ghost region according to the contour of the signal region and the first image specifically includes: Calculate the centroid of the contour of the signal region, calculate the direction vectors between each contour point in the contour of the signal region and the centroid, then scale all the direction vectors according to a preset ratio to obtain an intermediate signal region; map the intermediate signal region to the corresponding position in the first image to obtain an intermediate image block; remove the intermediate image block from the first image to obtain a ghost region to be recognized; Perform Gaussian filtering on the ghost area to be recognized, and use the Otsu threshold method to binarize the ghost area to be recognized after Gaussian filtering to obtain a second binary image; perform foreground erosion and then dilation on the second binary image to obtain an intermediate ghost area to be recognized; use the Suzuki-Abe boundary tracking algorithm to perform boundary tracking on the intermediate ghost area to be recognized, extract all the closed curves of the ghost areas in the intermediate ghost area to be recognized, and select the closed curve with the largest area among them as the contour of the ghost area.

[0050] In a third aspect, the present invention also proposes a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the automatic measurement method for imaging ghost metrics of a nuclear magnetic resonance device as described in any one of the first aspects are implemented.

[0051] In a fourth aspect, the present invention also proposes a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the automatic measurement method for imaging ghost metrics of a nuclear magnetic resonance device as described in any one of the first aspects are implemented.

[0052] The present invention will be described below with reference to specific embodiments.

[0053] Embodiment 1

[0054] An automatic measurement method for imaging ghost metrics of a nuclear magnetic resonance device proposed by the present invention includes: Step 1: Obtain a DICOM format image file obtained by scanning a phantom with a nuclear magnetic resonance device; Step 2: Display the DICOM format image file, as Figure 2 shown; Step 3: Convert the DICOM format image file into a 16-bit first image and an 8-bit second image; among them, the 8-bit second image is the result after window width and window level adjustment; among them, the conversion method is shown in the following formula. The window width (Window Width, WW) and window level (Window Level, WL) are used to adjust the gray level so that the image is clearer on the display device. The 16-bit or 12-bit medical image data can be converted into 8-bit (0-255 gray values) for display; ; In the formula, represents the displayed gray value, represents the window width, represents the window level; Step 4: Perform noise filtering on the second image through Gaussian filtering, and use a Gaussian kernel for processing; Step 5: Use the Otsu threshold method for the filtered image in Step 4 to separate the foreground and background, generating a first binary image; Step 6: Use the Suzuki-Abe algorithm on the first binary image obtained in Step 5 to implement the boundary tracking of the signal area, obtaining all closed curves, and select the one with the largest area as the signal area, as Figure 3 shown; and calculate the mean value S of the 16-bit image in this signal area; Step 7: Calculate its minimum bounding rectangle according to the signal area in Step 6, and obtain its four corner points, namely the upper left, upper right, lower left, and lower right. At the same time, define the four corner points of the entire image as a reference. Subsequently, by traversing each corner point of the bounding rectangle and the corresponding corner points of the image, calculate the center points of both as the center points of the noise rectangle area; then, calculate the distances between the two points in the x-axis and y-axis directions, and take half of them as the width and height of the noise rectangle area, obtaining four azimuthal noise rectangle areas, as Figure 4 ; then calculate the standard deviation of the 16-bit image in this noise area , and then obtain the noise level ; Step 8: To avoid the influence of the signaler on the recognition of the ghost area, moderately enlarge the contour of the signal area in Step 6; specifically, calculate the centroid of the signal area contour, then calculate the direction vector between each contour point and the centroid, then enlarge the signal area by a ratio of 1.1 for this vector, and then cut out the enlarged signal area from the 16-bit image to obtain the ghost area to be recognized; Step 9: Filter the ghost area to be recognized obtained in Step 8, and use the Otsu threshold method for it to separate the foreground and background, generating a second binary image, as Figure 5 shown; where the foreground is the ghost area; to avoid the influence of isolated pixel areas, perform a morphological opening on the second binary image, which specifically includes first eroding the foreground of the image and then dilating it to remove small-area noise and retain the main structure of the ghost area; Step 10: For all the closed curves of the ghost areas obtained in Step 9, select the one with the largest area as the ghost area, calculate its centroid, and use this centroid as the center point of the final ghost area. The length and width of the ghost area are both 5, and calculate the mean value of this ghost area , as Figure 6 shown; Step 11: Calculate the final ghost index for the statistical quantities calculated in Steps 6, 7, and 9, obtaining the ratio of the ghost to the signal , obtaining the ratio of the ghost to the noise , and calculate the signal-to-noise ratio .

[0055] As described above, it is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes should be covered within the protection scope of the present invention.

Claims

1. An automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device, characterized in that Including: Obtain an image file in DICOM format; Perform pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format; Perform automatic region recognition on the second image to obtain a signal region, a ghost region, and a noise region; Calculate the statistical features of the signal region, the ghost region, and the noise region respectively; Calculate a ghost index according to the statistical features of the signal region, the ghost region, and the noise region.

2. The automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device according to claim 1, characterized in that, Performing pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format specifically includes: Parse and display the DICOM format image file, and convert it from high-bit pixels to a first image in 16-bit format and a second image in 8-bit format through the tag window width and window level of the image file.

3. The automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device according to claim 2, characterized in that, The conversion formula for the second image is expressed as: ; where represents the displayed grayscale value, represents the window width, represents the window level.

4. The automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device according to claim 2, wherein Performing automatic region recognition on the second image to obtain a signal region, a ghost region, and a noise region specifically includes: Perform binaryzation processing on the second image using the Otsu threshold method to obtain a first binary image; Perform boundary tracking on the first binary image using the Suzuki-Abe boundary tracking algorithm, extract all the closed curves of the signal region, and select the closed curve with the largest area as the contour of the signal region; Obtain noise rectangular regions in four directions according to the contour of the signal region and the first binary image; Obtain the contour of the ghost region according to the contour of the signal region and the first image.

5. The automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device according to claim 4, wherein, Obtaining noise rectangular regions in four directions according to the contour of the signal region and the first binary image specifically includes: Calculate the minimum bounding rectangle of the signal region according to the contour of the signal region; extract the four corner points of the minimum bounding rectangle of the signal region, and at the same time define the four corner points of the entire first binary image as reference corner points; calculate the coordinates of the center points between the corner points of each minimum bounding rectangle and the corresponding reference corner points respectively, and use the coordinates of each center point as the center point of a noise rectangular region respectively; calculate the distances between the corner points of each minimum bounding rectangle and the corresponding reference corner points in the x-axis and y-axis directions respectively, and take half of them as the width and height of the corresponding noise rectangular region to obtain noise rectangular regions in four directions.

6. The automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device according to claim 4 or 5, characterized in that, Obtaining the contour of the ghost region according to the contour of the signal region and the first image specifically includes: Calculate the centroid of the contour of the signal region, calculate the direction vectors of each contour point in the contour of the signal region and the centroid, and then magnify all the direction vectors according to a preset ratio to obtain an intermediate signal region; map the intermediate signal region to the corresponding position in the first image to obtain an intermediate image block; remove the intermediate image block from the first image to obtain a ghost region to be recognized; Perform Gaussian filtering on the ghost region to be recognized, and perform binaryzation processing on the Gaussian-filtered ghost region to be recognized using the Otsu threshold method to obtain a second binary image; perform foreground erosion and then dilation on the second binary image to obtain an intermediate ghost region to be recognized; perform boundary tracking on the intermediate ghost region to be recognized using the Suzuki-Abe boundary tracking algorithm, extract all the closed curves of the ghost region in the intermediate ghost region to be recognized, and select the closed curve with the largest area among them as the contour of the ghost region.

7. The automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device according to claim 6, characterized in that, Calculate the statistical features of the signal region, the ghost region, and the noise region respectively, specifically including: Determine the signal image blocks corresponding to the signal region in the first image, and calculate the mean of the pixel values of the signal image blocks; Determine the noise image blocks corresponding to the noise rectangular region in the first image, and calculate the standard deviation of the pixel values of each noise image block; calculate the noise level according to the mean of the standard deviations of the pixel values of the four noise image blocks. According to the contour of the ghost region, calculate the centroid of the ghost region, and use this centroid as the center point of the ghost region to calculate the mean of the ghost region.

8. An automatic measurement system for imaging ghosting index of a nuclear magnetic resonance device, characterized in that, Including: An acquisition module for acquiring an image file in DICOM format; A conversion module for performing pixel extraction and format conversion on the DICOM format image file to obtain a second image in 8-bit format; An identification module for automatically identifying regions in the second image to obtain a signal region, a ghost region, and a noise region; A calculation module for calculating the statistical features of the signal region, the ghost region, and the noise region respectively; calculating a ghost index according to the statistical features of the signal region, the ghost region, and the noise region.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device described in any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the automatic measurement method for the imaging ghost index of a nuclear magnetic resonance device described in any one of claims 1-7.

Images