Image spatial resolution evaluation method and system, electronic equipment and storage medium

Abstract

The invention discloses an image spatial resolution evaluation method and system, electronic equipment and a storage medium. The image spatial resolution evaluation method comprises the following steps: acquiring a two-dimensional pixel matrix of a to-be-evaluated image; determining a reference center point of the to-be-evaluated image according to the two-dimensional pixel matrix; determining the distance of at least part of pixels in the two-dimensional pixel matrix relative to the reference center point; according to the distance of each pixel, distributing the pixel value of each pixel to a plurality of calibration distance points through interpolation to obtain pixel value statistical data of each calibration distance point; and determining at least one spatial resolution index of the to-be-evaluated image according to the pixel value statistical data of the plurality of calibration distance points. According to the image spatial resolution evaluation scheme provided by the invention, the accuracy, comprehensiveness and anti-noise capability of spatial resolution evaluation are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of image processing, specifically to an image spatial resolution evaluation method and system, electronic device, and storage medium. Background Technology

[0002] Performance evaluation of medical imaging equipment is a crucial step in ensuring the quality of clinical diagnosis. Spatial resolution, especially high contrast resolution, is one of the key indicators for measuring the imaging performance of medical imaging equipment. High contrast resolution reflects the system's ability to distinguish closely adjacent high-contrast objects.

[0003] Currently, the industry generally uses the modulation transfer function (MTF) to objectively quantify high contrast resolution indicators, and the MTF is usually obtained by Fourier transform based on the line spread function (LSF).

[0004] Currently, the ability to evaluate the high contrast resolution of images from medical imaging equipment is limited by the existing generation and processing of LSF and MTF, resulting in incomplete and inaccurate evaluation results, weak noise resistance, and poor repeatability.

[0005] Therefore, there is still room for improvement in the current imaging spatial resolution evaluation scheme.

[0006] The background art description is provided solely for the purpose of understanding the relevant technologies in this field and is not intended as an admission of prior art. Summary of the Invention

[0007] Therefore, the embodiments of this application aim to provide an image spatial resolution evaluation method and system, electronic device and storage medium, which can solve or alleviate at least one of the aforementioned technical problems.

[0008] In a first aspect, embodiments of this application provide an image spatial resolution evaluation method, comprising:

[0009] Obtain the two-dimensional pixel matrix of the image to be evaluated;

[0010] The reference center point of the image to be evaluated is determined based on the two-dimensional pixel matrix;

[0011] Determine the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point;

[0012] Based on the distance of each pixel, the pixel value of each pixel is interpolated and distributed to multiple calibration distance points to obtain the pixel value statistics of each calibration distance point;

[0013] Based on the pixel value statistics of the plurality of calibration distance points, at least one spatial resolution index of the image to be evaluated is determined.

[0014] In some embodiments, the step of allocating the pixel values ​​of each pixel to multiple calibration distance points through interpolation based on the distance of each pixel, and obtaining pixel value statistics for each calibration distance point, includes:

[0015] For each pixel, based on the linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, the pixel value of the pixel is assigned to one or more adjacent calibration distance points, and the pixel values ​​assigned to each calibration distance point are accumulated and statistically analyzed to obtain the pixel value statistics of each calibration distance point.

[0016] In some embodiments, the calibration distance point is an integer distance point relative to the reference center point;

[0017] Specifically, for each pixel, based on a linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, the pixel value of the pixel is allocated to one or more adjacent calibration distance points, and the pixel values ​​allocated to each calibration distance point are accumulated and statistically analyzed to obtain pixel value statistics for each calibration distance point, including:

[0018] The distance of each pixel is decomposed into an integer part d_int and a fractional part d_frac;

[0019] The integer part d_int determines one or two calibration distance points adjacent to the pixel.

[0020] The weighting coefficients assigned to one or two adjacent calibration distance points are determined based on the fractional part d_frac.

[0021] The weighted pixel value assigned to the adjacent one or two calibration distance points is determined based on the pixel value of the pixel and the weighting coefficient.

[0022] The weighted pixel values ​​assigned to each calibration distance point are accumulated and statistically analyzed to obtain the pixel value statistics corresponding to each calibration distance point.

[0023] In some embodiments, determining the weighting coefficients assigned to one or two adjacent calibration distance points based on the fractional part d_frac specifically includes:

[0024] Use 1-d_frac as the first weight coefficient of the first calibration distance point among one or two adjacent calibration distance points;

[0025] When there are two adjacent calibration distance points, d_frac is used as the second weight coefficient of the second calibration distance point among the two adjacent calibration distance points, where the distance between the first calibration distance point is less than the distance between the second calibration distance point.

[0026] In some embodiments, determining the weighted pixel value assigned to one or two adjacent calibration distance points based on the pixel value of the pixel and the weighting coefficient specifically includes:

[0027] The pixel value of the pixel is multiplied by the first weighting coefficient to obtain the first multiplication accumulation, which is then added to the first accumulator corresponding to the first calibration distance point.

[0028] When there are two adjacent calibration distance points, the pixel value of the pixel is multiplied by the second weighting coefficient to obtain the second multiplication accumulation, which is added to the second accumulator corresponding to the second calibration distance point.

[0029] In some embodiments, the step of accumulating and statistically analyzing the weighted pixel values ​​assigned to each calibration distance point to obtain statistical data on the pixel values ​​corresponding to each calibration distance point specifically includes:

[0030] After traversing all pixels in the two-dimensional pixel matrix, the value in the accumulator corresponding to each calibration distance point is used as the pixel value statistics corresponding to that calibration distance point.

[0031] In some embodiments, determining the weighted pixel value assigned to one or two adjacent calibration distance points based on the pixel value of the pixel and the weighting coefficient further includes:

[0032] When the first multiplication accumulation is added to the first accumulator, the count in the first counter corresponding to the first calibration distance point is incremented by 1;

[0033] When the second multiplication accumulation is added to the second accumulator, the count in the second counter corresponding to the second calibration distance point is incremented by 1;

[0034] The step of accumulating and statistically analyzing the weighted pixel values ​​assigned to each calibration distance point to obtain statistical data on the pixel values ​​corresponding to each calibration distance point also includes:

[0035] After traversing all pixels in the two-dimensional pixel matrix, the value in the counter corresponding to each calibration distance point is used as the statistical count of the number of associated pixels corresponding to that calibration distance point.

[0036] In some embodiments, determining the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point includes:

[0037] Obtain the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix;

[0038] Based on the first row and column coordinates and the second row and column coordinates, determine the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction;

[0039] Based on the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction, the Euclidean distance of the pixel relative to the reference center point is calculated, which is used as the distance relative to the reference center point.

[0040] In some embodiments, determining the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point includes:

[0041] Obtain the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix;

[0042] Based on the first row and column coordinates and the second row and column coordinates, the polar coordinates of each pixel with the reference center point as the origin are obtained;

[0043] Obtain the radius value of the pixel in the polar coordinates, as the distance relative to the reference center point.

[0044] In some embodiments, determining at least one spatial resolution index of the image to be evaluated based on the pixel value statistics of the plurality of calibration distance points includes:

[0045] Based on the pixel value statistics of the calibration distance points, determine the line spread function curve of the image to be evaluated;

[0046] The line spread function curve is transformed in the frequency domain to obtain the modulation transfer function curve of the image to be evaluated;

[0047] The modulation transfer function curve is processed to obtain the at least one spatial resolution index.

[0048] In some embodiments, processing the modulation transfer function curve to obtain the at least one spatial resolution index includes:

[0049] The spatial frequency value corresponding to the decrease of the parameter value of the modulation transfer function curve to a preset percentage threshold is used as the spatial resolution index.

[0050] In some embodiments, determining the reference center point of the image to be evaluated based on the two-dimensional pixel matrix includes:

[0051] The pixel with the largest pixel value in the two-dimensional pixel matrix is ​​taken as the reference center point.

[0052] In some embodiments, the spatial resolution index is a high-contrast spatial resolution index.

[0053] In some embodiments, the image to be evaluated is a tomographic image, and the format of the image to be evaluated includes at least one of DICOM format, NIfTI format, and MHD / RAW format.

[0054] In a second aspect, embodiments of this application provide an image spatial resolution evaluation system, comprising:

[0055] The acquisition module is configured to acquire the two-dimensional pixel matrix of the image to be evaluated;

[0056] The center positioning module is configured to determine the reference center point of the image to be evaluated based on the two-dimensional pixel matrix;

[0057] A distance determination module is configured to determine the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point;

[0058] The allocation statistics module is configured to allocate the pixel value of each pixel to multiple calibration distance points through interpolation based on the distance of each pixel, so as to obtain the pixel value statistics of each calibration distance point;

[0059] The evaluation module is configured to determine at least one spatial resolution index of the image to be evaluated based on the pixel value statistics of the plurality of calibration distance points.

[0060] In some embodiments, the allocation statistics module is further configured to, for each pixel, allocate the pixel value of the pixel to the one or more adjacent calibration distance points according to the linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, and perform cumulative statistics on the pixel values ​​allocated to each calibration distance point to obtain pixel value statistics for each calibration distance point.

[0061] In some embodiments, the allocation statistics module includes:

[0062] The distance decomposition submodule is configured to decompose the distance of each pixel into an integer part d_int and a fractional part d_frac;

[0063] The nearest neighbor point determination submodule is configured to determine one or two calibration distance points adjacent to the pixel based on the integer part d_int;

[0064] The weight determination submodule is configured to determine the weight coefficients assigned to one or two adjacent calibration distance points based on the fractional part d_frac.

[0065] The allocation submodule is configured to determine, based on the pixel value of the pixel and the weighting coefficient, the weighted pixel value to be weighted and allocated to one or two adjacent calibration distance points; and

[0066] The statistics submodule is configured to accumulate and statistically analyze the weighted pixel values ​​assigned to each calibration distance point to obtain the pixel value statistics corresponding to each calibration distance point.

[0067] In some embodiments, the statistics submodule includes:

[0068] Multiple accumulators, each corresponding to a calibration distance point, are configured to store the sum of the weighted pixel values ​​assigned to that calibration distance point.

[0069] In some embodiments, the statistics submodule includes:

[0070] Multiple counters, each corresponding to a calibration distance point, are configured to record the count of the weighted pixel values ​​assigned to that calibration distance point.

[0071] In some embodiments, the distance determination module includes:

[0072] The coordinate acquisition submodule is configured to acquire the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix;

[0073] The coordinate difference determination submodule is configured to determine, based on the first row and column coordinates and the second row and column coordinates, the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction; and

[0074] The distance calculation submodule is configured to calculate the Euclidean distance of the pixel relative to the reference center point based on the first coordinate difference and the second coordinate difference.

[0075] In some embodiments, the distance determination module includes:

[0076] The coordinate acquisition submodule is configured to acquire the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix;

[0077] The coordinate transformation submodule is configured to obtain the polar coordinates of each pixel with the reference center point as the origin, based on the first row and column coordinates and the second row and column coordinates; and

[0078] The radius determination submodule is configured to obtain the radius value of the pixel in the polar coordinates as the distance relative to the reference center point.

[0079] In some embodiments, the evaluation module includes:

[0080] The line diffusion submodule is configured to determine the line diffusion function curve of the image to be evaluated based on the pixel value statistics of the calibration distance points;

[0081] A frequency domain transformation submodule is configured to perform frequency domain transformation on the line spread function curve to obtain the modulation transfer function curve of the image to be evaluated; and

[0082] The index generation submodule is configured to process the modulation transfer function curve to obtain the at least one spatial resolution index.

[0083] In some embodiments, the spatial resolution index is a high-contrast spatial resolution index.

[0084] In a third aspect, embodiments of this application provide an electronic device including a processor and a memory, wherein the memory stores computer program instructions, and the processor, when executing the computer program instructions, implements the method described in the first aspect.

[0085] In a fourth aspect, embodiments of this application provide a computer storage medium storing computer program instructions, wherein when the computer program instructions are executed by a processor, the method described in the first aspect is implemented.

[0086] The image spatial resolution evaluation method of this application embodiment obtains a two-dimensional pixel matrix of the image to be evaluated; determines a reference center point of the image to be evaluated based on the two-dimensional pixel matrix; determines the distances of at least some pixels in the two-dimensional pixel matrix relative to the reference center point; allocates the pixel values ​​of each pixel to multiple calibration distance points through interpolation based on the distances of each pixel, obtaining statistical data of pixel values ​​at each calibration distance point; and determines at least one spatial resolution index of the image to be evaluated based on the statistical data of pixel values ​​at the multiple calibration distance points. The image spatial resolution evaluation method of this application embodiment innovatively uses a distance-based interpolation allocation mechanism to statistically analyze the two-dimensional pixel matrix, avoiding the information waste caused by traditional methods that only utilize pixel data in a single direction, achieving full utilization of global pixel information of the image, and improving the accuracy and comprehensiveness of spatial resolution evaluation.

[0087] This application further distributes the pixel value of a pixel to one or more adjacent calibration distance points according to the linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, and performs cumulative statistics on the pixel values ​​assigned to each calibration distance point. By establishing a linear proportional relationship between the pixel distance and the calibration distance points, the mapping problem of pixel values ​​from actual distance to calibration distance points is solved, thereby achieving smooth distribution and effective accumulation of pixel values ​​and improving the continuity and accuracy of statistical data.

[0088] The embodiments of this application further set the calibration distance point as an integer distance point relative to the reference center point, and through distance decomposition and bidirectional weighted accumulation of one or two adjacent integer distance points (single-point accumulation in special cases), the accurate allocation and statistics of the information of each pixel in the two-dimensional pixel matrix based on integer distance relationship are realized. This not only realizes the full utilization and smooth statistics of more comprehensive two-dimensional pixel information, but also has significantly enhanced result stability and noise resistance.

[0089] Optional features and other effects of the embodiments of this application are described in part below, and in part will be apparent from reading this document. Attached Figure Description

[0090] The embodiments of this application will be described in detail with reference to the accompanying drawings. The elements shown are not limited to the scale shown in the drawings, and the same or similar reference numerals in the drawings denote the same or similar elements, wherein:

[0091] Figure 1 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0092] Figure 2 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0093] Figure 3 This illustration schematically shows a two-dimensional weighted statistical diagram of an image spatial resolution evaluation method according to an embodiment of this application;

[0094] Figure 4 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0095] Figure 5 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0096] Figure 6 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0097] Figure 7 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0098] Figure 8 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0099] Figure 9 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0100] Figure 10 A flowchart of an image spatial resolution evaluation method according to an embodiment of this application is shown;

[0101] Figure 11 A block diagram of an image spatial resolution evaluation system according to an embodiment of this application is shown;

[0102] Figure 12 A block diagram of an image spatial resolution evaluation system according to an embodiment of this application is shown;

[0103] Figure 13 A block diagram of an image spatial resolution evaluation system according to an embodiment of this application is shown;

[0104] Figure 14 A block diagram of an image spatial resolution evaluation system according to an embodiment of this application is shown;

[0105] Figure 15 A block diagram of an image spatial resolution evaluation system according to an embodiment of this application is shown; and

[0106] Figure 16 A schematic diagram of an electronic device capable of implementing the image spatial resolution evaluation method of the embodiments of this application is shown. Detailed Implementation

[0107] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to specific embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this application are used to explain this application, but are not intended to limit this application.

[0108] The term "comprising" and its variations as used herein signify open inclusion, i.e., "including but not limited to". Unless otherwise stated, the term "or" means "and / or". The term "based on" means "at least partially based on". The terms "one example embodiment" and "one embodiment" mean "at least one example embodiment". The term "another embodiment" means "at least one additional embodiment". The terms "first", "second", etc., may refer to different or the same objects.

[0109] As mentioned earlier, existing methods for evaluating high contrast resolution in medical images still have room for improvement. Specifically, known methods typically utilize pixel data from only one direction (e.g., the X-axis or Y-axis) for statistical analysis during LSF generation. They calculate the arithmetic mean or sum of pixel values ​​at each distance point along a selected axial direction with the center point as the origin, thus obtaining a one-dimensional LSF curve. However, high-contrast edge information in medical images is contained within a two-dimensional pixel matrix. Known methods only utilize pixel data along the X-axis or Y-axis, ignoring a large number of pixels that are not on the axis but still contain edge information. This is equivalent to discarding most of the effective data, resulting in significant information waste.

[0110] Furthermore, since known methods only use data from a single direction, noise or outliers in individual pixels (such as image noise, artifacts, etc.) can significantly affect the shape of the LSF curve, resulting in large fluctuations in the final calculated MTF curve. This leads to reduced stability and repeatability of the results, and weak noise resistance. This application recognizes that even if anisotropy is evaluated by calculating the MTF along the X and Y axes separately, the evaluation in each direction is still based on single-axis data, which fails to fundamentally solve the problem of insufficient data volume for evaluation in a single direction. Therefore, the evaluation results are still not comprehensive and accurate enough.

[0111] To address the problems existing in the prior art, this application provides an image spatial resolution evaluation method that innovatively makes full use of the two-dimensional pixel information in the image, rather than being limited to pixel data in a single axis.

[0112] The image spatial resolution evaluation method of this application according to several embodiments will be described in detail below with reference to the accompanying drawings.

[0113] In some embodiments, reference Figure 1 A method for evaluating image spatial resolution is provided, which may include steps S110 to S150.

[0114] S110: Obtain the two-dimensional pixel matrix of the image to be evaluated.

[0115] In some embodiments, such as in step S110, a two-dimensional pixel matrix of the image to be evaluated may be obtained. In some embodiments, the image to be evaluated includes tomographic images. In some embodiments, tomographic images include medical tomographic images. In this embodiment, medical tomographic images include, but are not limited to, tomographic images generated by medical imaging equipment such as computed tomography (CT) devices, positron emission tomography (PET) devices, and magnetic resonance imaging (MRI) devices.

[0116] In this embodiment, a two-dimensional pixel matrix can be used to characterize the arrangement of pixel data in a two-dimensional space, where each pixel can have a corresponding pixel value (intensity value) and spatial location. In this embodiment, the pixel matrix can be organized in a row-column format, where each pixel can be uniquely identified in the matrix by its row and column coordinates. In this embodiment, the pixel value is interpreted broadly, covering numerical parameters in medical tomographic images that represent tissue characteristics, radioactivity distribution, or signal intensity, including but not limited to: HU values ​​(Hounsfield units) in CT images, SUV values ​​or radioactivity counts in PET images, and signal intensity values ​​in MRI images.

[0117] In some embodiments, the image to be evaluated, such as a medical tomographic image, may include a high-contrast test structure, such as a phantom image containing line pairs, wires, or other high-contrast targets.

[0118] In the embodiments of this application, the format of the image to be evaluated may include, but is not limited to, commonly used medical image formats such as DICOM format, NIfTI format, and MHD / RAW format.

[0119] In some preferred embodiments, the image to be evaluated includes a DICOM format image, and many embodiments described below will use DICOM format images as examples. It is understood that the embodiments of this application do not limit the specific format of the image to be evaluated, as long as it allows the extraction of a two-dimensional pixel matrix from it. In this embodiment, the pixel matrix can be read from the pixel data tags of the DICOM file, and pixel spacing information can be read from relevant tags (such as PixelSpacing or ImagerPixelSpacing).

[0120] S120: Determine the reference center point of the image to be evaluated based on the two-dimensional pixel matrix.

[0121] In some embodiments, such as S120, the two-dimensional pixel matrix is, for example, the two-dimensional pixel matrix obtained according to the aforementioned S110.

[0122] In some embodiments, the pixel with the highest pixel value in a two-dimensional pixel matrix (or optionally, all pixels within the ROI scanned, as described below) can be used as the reference center point. For explanation, in medical images containing high-contrast structures (such as metal wires), high-density materials like metal wires typically appear as areas with the highest pixel values ​​in the image.

[0123] In some embodiments, the step of determining the maximum pixel value may include: scanning all pixels in a two-dimensional pixel matrix (or optionally scanning all pixels within the ROI, as described below), comparing the pixel values ​​of each pixel and identifying the pixel with the maximum pixel value, and determining the coordinates (row coordinates and column coordinates) of that pixel as the coordinates of a reference center point. In some embodiments, when multiple pixels have the same maximum pixel value, any one of them or a pixel closer to the image center may be selected as the reference center point.

[0124] In other embodiments of this application, the method for determining the reference center point may include, but is not limited to: determining the reference center point based on a region containing a high-contrast structure, such as selecting its center or centroid; automatically extracting the high-contrast structure and determining its center position using a target recognition or image segmentation model; having the user specify the position of the reference center point through manual interaction; or directly selecting the center of the image (or ROI, as described below), etc. It is understood that this application does not limit the method for determining the reference center point.

[0125] In some embodiments, after determining the reference center point, its coordinates in the two-dimensional pixel matrix (denoted as the first row and column coordinates) can be used as the origin for subsequent distance calculations.

[0126] S130: Determine the distances of at least some pixels in the two-dimensional pixel matrix relative to the reference center point.

[0127] In some embodiments, distance calculations can be performed on all pixels in a two-dimensional pixel matrix.

[0128] In alternative embodiments, distance calculations can also be performed on a subset of pixels, such as pixels within a predetermined region (e.g., Region of Interest, ROI). Accordingly, the reference center point mentioned in step S120 above can be determined within the predetermined region (e.g., ROI), but the reference center point can also be located outside the predetermined region (e.g., ROI). Accordingly, in some embodiments, the image spatial resolution evaluation method may further include: after acquiring the pixel matrix of the complete image, determining the Region of Interest (ROI) from the pixel matrix based on preset conditions. In some embodiments, the determination of the ROI may include, but is not limited to, threshold segmentation, edge detection, template matching, image segmentation algorithms, or manual selection. In some embodiments, regions containing high-contrast structures can be segmented by setting pixel value thresholds and selected as ROIs.

[0129] In the following description, several different embodiments of this application will be illustrated by taking the implementation of distance calculation for all pixels in a two-dimensional pixel matrix as an example, but this application is not limited thereto.

[0130] In some embodiments, the distance can be defined and calculated in various ways, including but not limited to Euclidean distance based on a rectangular coordinate system or radius value based on a polar coordinate system.

[0131] In some embodiments, the distance of a pixel relative to a reference center point can be determined based on Euclidean distance. In some embodiments, the reference... Figure 2 Step S130 may include the following steps S131 to S133.

[0132] S131: Obtain the coordinates of the reference center point in the first row and column of the two-dimensional pixel matrix, and the coordinates of each pixel in the second row and column of the two-dimensional pixel matrix.

[0133] In some embodiments, a Cartesian coordinate system, preferably based on a two-dimensional pixel matrix, can be used to establish the coordinate system. In some preferred embodiments of this application, the row direction of the two-dimensional pixel matrix can be used as the first coordinate axis, and the column direction of the two-dimensional pixel matrix can be used as the second coordinate axis. In these embodiments, the row and column coordinates of a reference center point and each pixel in the two-dimensional pixel matrix can be obtained. In this embodiment, the distance to the reference center point determined in S120 can be calculated. Specifically, the coordinates of the reference center point in the two-dimensional pixel matrix can be designated as the first row and column coordinates (x0, y0). In some embodiments, the second row and column coordinates (x, y) or (row, col) of each pixel in the two-dimensional pixel matrix can be obtained. As an explanation and not a limitation, although the calibration distance point for the interpolated pixel values ​​may not be limited to a specific direction, establishing the coordinate system according to rows and columns allows the previously assigned pixel values ​​to be "mapped" to the pixels on coordinate axes, such as the x-axis (row coordinate axis), in a more interpretable manner.

[0134] In other alternative embodiments not shown in this application, other orthogonal coordinate systems may also be used to determine the coordinates, such as x-axis and y-axis that do not extend in the row and column direction of the image. This falls within the protection scope of this application and will not be elaborated here.

[0135] S132: Based on the first row and column coordinates and the second row and column coordinates, determine the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction.

[0136] In some embodiments, the coordinate differences between each pixel and the reference center point in the row and column directions can be calculated for subsequent Euclidean distance determination.

[0137] S133: Based on the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction, calculate the Euclidean distance of the pixel relative to the reference center point, and use it as the distance relative to the reference center point.

[0138] In some embodiments, Euclidean distance includes the straight-line distance between two points in a Cartesian coordinate system. In a specific example, the formula for calculating Euclidean distance is:

[0139]

[0140] Where d represents the Euclidean distance of a pixel relative to the reference center point, in pixels, and Δx and Δy are the first coordinate difference and the second coordinate difference, respectively.

[0141] In some embodiments, standard mathematical library functions may be used, and this application does not impose any restrictions on this.

[0142] In a specific embodiment of this application, such as Figure 3 The diagram illustrates a two-dimensional weighted statistical representation of an image spatial resolution evaluation method according to an embodiment of this application. The reference center point is located at the origin, and the coordinates of pixel P are (2, 1) (i.e., offset by 2 pixels in the X-axis direction and 1 pixel in the Y-axis direction). The Euclidean distance from pixel P to the center point is: Similarly, by performing the above calculation on each pixel in the two-dimensional pixel matrix, the distance of all pixels relative to the reference center point can be obtained.

[0143] In some embodiments, the distance of a pixel relative to a reference center point can be calculated based on polar coordinates. In some embodiments, the reference... Figure 4 The aforementioned S130 may include the following steps S131' to S133'.

[0144] S131': Obtain the coordinates of the reference center point in the first row and column of the two-dimensional pixel matrix, and the coordinates of each pixel in the second row and column of the two-dimensional pixel matrix.

[0145] In the embodiments of this application, the specific description of step S131' can be referred to the description of step S131 above, and will not be repeated here.

[0146] S132': Based on the first row and column coordinates and the second row and column coordinates, obtain the polar coordinates of each pixel with the reference center point as the origin.

[0147] S133': Obtain the radius value of the pixel in polar coordinates, as the distance relative to the reference center point.

[0148] In some embodiments, each pixel can be converted from a Cartesian coordinate system (row and column coordinates) to a polar coordinate system with the reference center point as the origin. In some embodiments, the row and column coordinates of a pixel are represented as (x, y), and its polar coordinates relative to the reference center point (x0, y0) are:

[0149]

[0150] θ = arctan 2(y - y0, χ - x0),

[0151] Wherein, the radius r represents the distance from the point to the origin, and the angle θ represents the angle of the point relative to the reference direction.

[0152] In some embodiments, the polar coordinate radius value of the pixel, radius r, can be used as the distance of the pixel relative to the reference center point.

[0153] In some alternative embodiments, polar coordinate transformation or the aforementioned polar coordinate radius value can be achieved by a lookup table. In this embodiment, as an explanation, the radius values ​​corresponding to commonly used coordinate differences can be pre-calculated and stored. During actual calculations, the results can be obtained by looking up the table. For example, this can skip the explicit calculation step of Euclidean distance and avoid repeated square root calculations, thereby helping to improve computational efficiency.

[0154] In some alternative embodiments, the two-dimensional pixel matrix may have different pixel spacing (anisotropic pixel spacing) in different directions, that is, pixels may have different physical sizes in the row direction and column direction. In these embodiments, when the pixels of the image to be evaluated have different pixel spacing (i.e., anisotropic pixel spacing) in different directions, the difference in pixel spacing can be considered when performing distance calculation. Specifically, the pixel spacing in the row direction and the pixel spacing in the column direction of the two-dimensional pixel matrix can be obtained so as to perform physical space normalization in distance calculation. Specifically, in an exemplary embodiment, as a supplement or alternative to the embodiments shown in steps S131 to S133 and S131' to S133' above, step S130 may include the following steps: A1: Normalize the coordinates of each pixel to physical space coordinates based on the pixel spacing in different directions of the two-dimensional pixel matrix; A2: For each pixel, multiply the row coordinate of the pixel by the pixel spacing in the row direction, and multiply the column coordinate of the pixel by the pixel spacing in the column direction to obtain the physical space coordinates of the pixel; A3: Calculate the physical distance from each pixel to the reference center point based on the physical space coordinates. In some embodiments, for example, the pixel spacing in the row direction and the pixel spacing in the column direction of the two-dimensional pixel matrix can be read from the metadata of the medical image file, such as, but not limited to, the PixelSpacing tag of the DICOM file. In some embodiments, step A3 above can be used to determine the distance of each pixel relative to the reference center point by referring to the Euclidean distance method of S131 to S133 or the polar coordinate method of S131' to S133'.

[0155] S140: Based on the distance of each pixel, the pixel value of each pixel is interpolated and distributed to multiple calibration distance points to obtain the pixel value statistics of each calibration distance point.

[0156] In some embodiments, the pixel values ​​of each pixel in the two-dimensional pixel matrix can be allocated to multiple preset calibration distance points by interpolation according to their distances relative to the reference center point, and the pixel values ​​received by each calibration distance point can be accumulated and statistically analyzed to obtain pixel value statistics indexed by the calibration distance points.

[0157] In this embodiment, the calibration distance points include a pre-defined series of discrete distance values, and each calibration distance point may correspond to a calibration distance value. In some embodiments, pixel values ​​within the range of the calibration distance values ​​may be accumulated at the calibration distance point. As an explanation, by normalizing the statistical data of all calibration distance points, a line spread function with distance as the independent variable can be obtained.

[0158] In some embodiments, the calibration distance points include integer distance points relative to the reference center point. In some embodiments, the calibration distance points include consecutive integer distance points of the calibration distance points relative to the reference center point, wherein the distance values ​​may include consecutive integer values ​​in pixels, such as selecting integer distance points such as 0, 1, 2, 3, 4, etc. (where 0 corresponds to the reference center point).

[0159] In other alternative embodiments of this application, the calibration distance points may also be non-contiguous integer sequences, non-integer sequences, or other distance sequences. It is understood that the selection of calibration distance points only needs to cover the distance range from the center point to the edge in the image to be evaluated; this application does not limit the specific method of selecting calibration distance points. Below, several embodiments of this application will be described using continuous integer distance points as an example.

[0160] In some embodiments, the interpolation assignment preferably includes linear interpolation. In other alternative embodiments of this application, other interpolation methods may also be used, such as but not limited to quadratic interpolation, Gaussian weighted interpolation, etc., and this application does not impose any limitations on this. Below, several embodiments of this application will be described using linear interpolation assignment as an example.

[0161] In some embodiments, a linear interpolation method is used for interpolation assignment, and step S140 may include S140' (unidentified):

[0162] S140': For each pixel, based on the linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, the pixel value of the pixel is assigned to one or more adjacent calibration distance points, and the pixel values ​​assigned to each calibration distance point are accumulated and statistically analyzed to obtain the pixel value statistics of each calibration distance point.

[0163] This application embodiment can proportionally allocate pixel values ​​to one or more adjacent, preferably one or two most adjacent, calibration distance points, based on the positional relationship between the actual distance and adjacent calibration distance points. By way of explanation and not limitation, this application embodiment achieves a smooth mapping from continuous actual distances to discrete calibration distance points through this interpolation allocation mechanism.

[0164] In some embodiments, reference Figure 5 Step S140, more specifically S140' may include the following steps S141 to S145.

[0165] S141: Decompose the distance of each pixel into an integer part d_int and a fractional part d_frac.

[0166] In some embodiments, such as in step S141 above, the distance d of each pixel relative to the reference center point calculated in step S130 can be decomposed into an integer part d_int and a fractional part d_frac.

[0167] In some embodiments, the integer part d_int of each pixel can be determined by rounding, and its fractional part d_frac can be determined by subtraction.

[0168] In an optional embodiment of this application, the integer part d_int of each pixel can be determined by rounding down. Specifically, the integer part d_int can be calculated using the following formula:

[0169]

[0170] d_frac = d - d_int.

[0171] Among them, the floor function (or The symbol () indicates rounding down.

[0172] In one specific embodiment, d = 2.236, then d_int = floor(2.236) = 2, d_frac = 2.236 - 2 = 0.236.

[0173] In other embodiments of this application, other rounding methods may also be used.

[0174] In some embodiments, a floor function (ceiling) can be used. Specifically, the integer part d_int can be calculated using the following formula:

[0175]

[0176] Among them, the ceiling function (or The symbol ) indicates rounding up, and d_int is the smallest integer not less than d.

[0177] In a specific embodiment of the present application, the distance d = 2.236, then d_int = ceiling(2.236) = 3.

[0178] In some embodiments, the integer part d_int of each pixel may be determined by rounding.

[0179] S142: Determine one or two calibration distance points adjacent to the distance of the pixel according to the integer part d_int.

[0180] As an explanation, in the above step S1412, adjacent refers to the adjacent relationship at the distance value level, rather than the adjacent relationship of pixels in the image space.

[0181] In some embodiments, the calibration distance points are integer distance points. Among them, in the embodiment of continuous integer calibration points, the distance difference between one or two adjacent calibration distance points is 1.

[0182] In some embodiments, when the floor function is used to determine d_int, the two calibration distance points adjacent to the distance d are respectively: the first calibration distance point (distance value is d_int), the second calibration distance point (distance value is d_int + 1), where the actual distance d is between the first calibration distance point and the second calibration distance point, that is, d_int ≤ d < d_int + 1. It can be understood that when other rounding methods are used, the determination method of adjacent calibration distance points can be adjusted accordingly.

[0183] In some embodiments, the distance d may exactly equal the distance value of a calibration distance point (i.e., d_frac = 0), and the pixel value of this pixel will be completely assigned to this calibration distance point, which is the single closest adjacent point, and the present application intends to cover this situation.

[0184] In a specific example, the distance d = 2.236, the integer part d_int = 2, and the adjacent calibration distance points determined by the floor function are: the first calibration distance point (distance value is 2), the second calibration distance point (distance value is 3), that is, the pixel value will be assigned to the two calibration distance points with distances of 2 and 3.

[0185] S143: Determine the weight coefficients assigned to one or two adjacent calibration distance points according to the fractional part d_frac.

[0186] In some embodiments, the weight coefficients may reflect the relative position relationship between the actual distance d and the corresponding two calibration distance points. As an explanation rather than a limitation, the closer the actual distance d is to a certain calibration distance point, the greater the weight assigned to this calibration distance point.

[0187] In some embodiments, refer to Figure 6 , step S143 may include the following sub-step S1431.

[0188] S1431: Take 1 - d_frac as the first weight coefficient of the first calibration distance point among one or two adjacent calibration distance points.

[0189] When there are two adjacent calibration distance points, step S143 may further include:

[0190] S1432: Take d_frac as the second weight coefficient of the second calibration distance point among two adjacent calibration distance points.

[0191] In the embodiment of the present application, the distance of the first calibration distance point may be less than the distance of the second calibration distance point. In this embodiment, the first weight coefficient may correspond to the nearer calibration distance point, and the second weight coefficient may correspond to the farther calibration distance point.

[0192] In some embodiments, the first weight coefficient w1 = 1 - d_frac may be assigned to the first calibration distance point (distance value is d_int), where, 0 < w1 < 1. Correspondingly, in some embodiments, the second weight coefficient w2 may be assigned to the second calibration distance point (distance value is d_int + 1), where, 0 < w2 < 1.

[0193] In a specific embodiment of the present application, the distance d = 2.236, its integer part d_int = 2, and the decimal part d_frac = 0.236. Thus, the first weight coefficient w1 = 1 - 0.236 = 0.764 is calculated, and the second weight coefficient w2 = 0.236. Subsequently, 76.4% of the pixel value of this pixel is assigned to the calibration distance point with a distance of 2, and 23.6% of the pixel value of this pixel is assigned to the calibration distance point with a distance of 3. Thus, through linear interpolation, the pixel value of the pixel can be weighted and assigned to these two adjacent integer distance points, thereby achieving a smooth mapping from the actual distance to the discrete distance points.

[0194] As described above, when there is only a single adjacent calibration distance point, that is, the distance d may exactly be equal to the distance value of a calibration distance point (i.e., d_frac = 0). Thus, the pixel value of this pixel will be completely assigned to this calibration distance point, and at this time, the single weight coefficient w = 1. In some cases, this alternative embodiment can be regarded as a special case of step S1431 and be covered by it (in this case 0 < w1 ≤ 1).

[0195] S144: Determine the weighted pixel value that is weighted and assigned to one or two adjacent calibration distance points according to the pixel value of this pixel and the weight coefficient.

[0196] In some embodiments, the pixel value I(x, y) of a pixel can be denoted as V. Then, its first weighted pixel value V1 assigned to the first calibration distance point and its second weighted pixel value V2 assigned to the second calibration distance point are calculated, for example, according to the following formula:

[0197] V1 = V × w1 = V × (1 - d_frac),

[0198] V2 = V × w2 = V × d_frac,

[0199] Where V is the pixel value of the pixel, V1 is the first weighted pixel value, and V2 is the second weighted pixel value.

[0200] In some embodiments, a corresponding accumulator can be set for each calibration distance point. In embodiments where the calibration points are consecutive integer distance points, the calibration distance points include integer distance points 0, 1, 2, 3..., and an accumulator array acc[0], acc[1], acc[2], acc[3]... can be set. In this embodiment, the accumulator corresponding to the first calibration distance point d_int is acc[d_int], and the accumulator corresponding to the second calibration distance point d_int+1 is acc[d_int+1].

[0201] In some embodiments, reference Figure 7 Step S144 may include the following sub-step S1441:

[0202] S1441: Multiply the pixel value of the pixel by the first weighting coefficient to obtain the first multiplication accumulation, and add it to the first accumulator corresponding to the first calibration distance point.

[0203] In some embodiments, the accumulator can be used to store and accumulate the weighted pixel values ​​received at each calibration distance point. In some embodiments, an accumulator can be configured for each calibration distance point and initialized to zero before processing begins. In this embodiment, the first accumulator corresponding to the first calibration distance point d_int can receive and accumulate the first weighted pixel value.

[0204] In a specific example, the accumulation operation of the first accumulator corresponding to the first calibration distance point d_int is implemented as follows:

[0205] acc[d_int] = acc[d_int] + V1.

[0206] In this embodiment, by performing an accumulation operation on all pixels, the first accumulator can eventually contain the sum of all weighted pixel values ​​assigned to the first calibration distance point.

[0207] When there are two adjacent calibration distance points, step S144 may also include the following sub-step S1442:

[0208] S1442: Multiply the pixel value of the pixel by the second weighting coefficient to obtain the second multiplication accumulation, and add it to the second accumulator corresponding to the second calibration distance point.

[0209] In some embodiments, such as S1442 above, the description of the second product and the accumulator can be referred to the relevant description in step S1441 above, and will not be repeated here.

[0210] In a specific example, the accumulation operation of the second accumulator corresponding to the second calibration distance point d_int+1 is implemented by the following formula:

[0211] acc[d_int+1]=acc[d_int+1]+V2.

[0212] In this embodiment, by performing an accumulation operation on all pixels, the second accumulator can eventually contain the sum of all weighted pixel values ​​assigned to the second calibration distance point.

[0213] In some optional embodiments of this application, a counter can also be configured for each calibration distance point, which can be used to record the number of times the calibration distance point is assigned a pixel value. In embodiments where the calibration points are consecutive integer distance points, the calibration distance points include integer distance points 0, 1, 2, 3..., and a counter array count[0], count[1], count[2]... can be set.

[0214] In this embodiment, reference Figure 8 Step S144 may also include the following sub-step S1443:

[0215] S1443: When adding the first multiplication accumulation to the first accumulator, increment the count in the first counter corresponding to the first calibration distance point by 1.

[0216] When there are two adjacent calibration distance points, step S144 may also include the following sub-step S1444:

[0217] S1444: When adding the second multiplication accumulation to the second accumulator, increment the count in the second counter corresponding to the second calibration distance point by 1.

[0218] In these embodiments, when performing an accumulation operation at the corresponding calibration distance point, the counter corresponding to that calibration distance point can be incremented by 1.

[0219] In some embodiments, the counter may or may not participate in pixel value interpolation, and may also be used in conjunction with an accumulator to implement other functions, such as for subsequent calculation of average pixel values ​​and / or for normalization processing, as further explained below.

[0220] S145: Accumulate and statistically analyze the weighted pixel values ​​assigned to each calibration distance point to obtain the pixel value statistics corresponding to each calibration distance point.

[0221] In some embodiments, the above interpolation assignment and accumulation operations can be performed on all pixels in the two-dimensional pixel matrix (or all pixels within the ROI).

[0222] In some embodiments, reference Figure 9 The above step S145 may include the following sub-step S1451.

[0223] S1451: After traversing all pixels in the two-dimensional pixel matrix, use the value in the accumulator corresponding to each calibration distance point as the statistical data of the pixel value corresponding to that calibration distance point.

[0224] In some embodiments, after traversal is completed, the value stored in the accumulator corresponding to each calibration distance point is the pixel value statistics of that calibration distance point. It represents the weighted sum of pixel values ​​of all pixels that are close to that calibration distance point, reflecting the pixel intensity information in the two-dimensional pixel matrix within that distance range.

[0225] When a counter is configured at the calibration distance point, step S145 above may also include an optional sub-step S1452:

[0226] S1452: After traversing all pixels in the two-dimensional pixel matrix, use the value in the counter corresponding to each calibration distance point as the statistical count of the number of associated pixels corresponding to that calibration distance point.

[0227] In this embodiment, when a counter is used, after the traversal is completed, the value in the counter corresponding to each calibration distance point is the statistical count of the associated pixel quantity corresponding to that calibration distance point, which reflects the number of pixels that contribute to the pixel value of that calibration distance point.

[0228] By performing distance-based linear interpolation allocation and cumulative statistics on all pixels in the two-dimensional pixel matrix, this application embodiment obtains statistical data that fully utilizes two-dimensional pixel information. Compared with traditional single-axis statistical methods that only utilize a small number of pixels in a single direction, this method weights and accumulates the information of all pixels according to distance relationships, achieving comprehensive utilization of image information. Since the contributions of a large number of pixels are accumulated and statistically analyzed, the influence of noise or outliers of individual pixels on the final statistical data is effectively averaged and suppressed, significantly enhancing the stability and noise resistance of the statistical results.

[0229] S150: Based on the statistical data of pixel values ​​at multiple calibration distance points, determine at least one spatial resolution index of the image to be evaluated.

[0230] In some embodiments, the spatial resolution index is a high-contrast spatial resolution index. Accordingly, in some embodiments, the spatial resolution evaluation method is a high-contrast (spatial) resolution evaluation method.

[0231] In alternative embodiments of this application, the spatial resolution index may also be a low-contrast spatial resolution index. Accordingly, in some embodiments, the spatial resolution evaluation method is a low-contrast (spatial) resolution evaluation method.

[0232] In some embodiments, step 150 may include the following sub-steps S151 to S153.

[0233] S151: Determine the line spread function curve of the image to be evaluated based on the pixel value statistics of the calibration distance points.

[0234] In some embodiments, a line spread function (LSF) curve with distance as the independent variable can be generated based on the pixel value statistics of each calibrated distance point. For explanation, the LSF describes the diffusion characteristics of a point source or line source in an imaging system and can serve as an intermediate function for evaluating the spatial resolution of the imaging system.

[0235] In some embodiments, a discrete data point sequence can be obtained based on the pixel value statistics acc[k] corresponding to each calibrated distance point k. Specifically, each data point can be defined with the distance value of the calibrated distance point as the abscissa and the corresponding pixel value statistics as the ordinate, thereby obtaining a discrete LSF data point set: {(0,acc[0]), (1,acc[1]), (2,acc[2])……}.

[0236] In some alternative embodiments, a counter can be used to count the number of associated pixels, and the pixel value statistics of each calibration distance point can be normalized to obtain the average pixel value.

[0237] S152: Perform frequency domain transformation on the line spread function curve to obtain the modulation transfer function curve of the image to be evaluated.

[0238] In some embodiments, in step S152 above, the line spread function curve obtained in step S151 can be frequency domain transformed, for example, by a Fourier transform, to obtain a modulation transfer function (MTF) curve. As an explanation, the modulation transfer function can be used to describe the ability of an imaging system to transfer signals of different spatial frequencies. Specifically, a one-dimensional discrete Fourier transform (DFT) can be performed on the LSF curve to obtain a frequency domain function. In this embodiment, the Fourier transform can convert the spatial domain LSF into a complex function in the frequency domain, and its amplitude spectrum is the MTF. In some embodiments, this transformation can be implemented using algorithms including but not limited to the Fast Fourier Transform (FFT).

[0239] In some embodiments, the amplitude values ​​of the frequency domain function at each spatial frequency can be calculated after performing a Fourier transform on the LSF to obtain the MTF curve. This process can be achieved, for example, by the following formula:

[0240] MTF(f) = |FFT(LSF)|,

[0241] Where f is the spatial frequency, and the unit is line pairs per centimeter (lp / cm) or line pairs per millimeter (lp / mm).

[0242] In some embodiments, the LSF curve can be smoothed by applying a window function after it is generated and before the frequency domain transformation, such as, but not limited to, applying a Hann window, a Hamming window, or a Gaussian window.

[0243] In some embodiments, the amplitude value of the frequency domain function can be normalized by dividing the amplitude value at each spatial frequency by the amplitude value at zero frequency (i.e., the DC component) to obtain a normalized MTF curve.

[0244] S153: Process the modulation transfer function curve to obtain at least one spatial resolution index.

[0245] In some embodiments, in step S153 above, the MTF curve obtained in step S152 may be processed to extract at least one spatial resolution index.

[0246] In some embodiments, the spatial resolution metric may include the spatial frequency value corresponding to the point where the parameter value of the modulation transfer function curve drops to a preset percentage threshold. For explanation, a higher spatial frequency value corresponding to the point where the parameter value drops to the preset percentage threshold indicates higher spatial resolution.

[0247] In some embodiments, the preset percentage threshold may be set to 10% or 50%, but other suitable preset percentage thresholds can be selected, and this application does not limit this.

[0248] In other embodiments of this application, at least one spatial resolution index may also include the area under the MTF curve (integral value) and the MTF value at the Nyquist frequency.

[0249] Therefore, traditional methods only utilize pixel data along a single axis, failing to fully leverage all the information in the two-dimensional pixel matrix. This step achieves comprehensive utilization of two-dimensional pixel information by performing distance-based interpolation assignment on all pixels (or all pixels within the ROI). Since the pixel value of each pixel is assigned to corresponding calibration distance points (such as integer distance points) according to its distance relationship, the contributions of a large number of pixels are accumulated and statistically analyzed. Therefore, the final statistical data can more realistically and stably reflect the edge characteristics of high-contrast structures in the image, effectively suppressing the influence of individual pixel noise and significantly improving the stability and noise resistance of the statistical results.

[0250] In the embodiments of this application, reference is made to Figure 11 Furthermore, an image spatial resolution evaluation system 1100 is provided, which may include: an acquisition module 1110, a center positioning module 1120, a distance determination module 1130, an allocation statistics module 1140, and an evaluation module 1150. Among these,

[0251] The acquisition module 1110 is configured to acquire a two-dimensional pixel matrix of the image to be evaluated;

[0252] The center positioning module 1120 is configured to determine the reference center point of the image to be evaluated based on the two-dimensional pixel matrix;

[0253] The distance determination module 1130 is configured to determine the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point;

[0254] The allocation statistics module 1140 is configured to allocate the pixel value of each pixel to multiple calibration distance points by interpolation according to the distance of each pixel, so as to obtain the pixel value statistics of each calibration distance point.

[0255] The evaluation module 1150 is configured to determine at least one spatial resolution index of the image to be evaluated based on the pixel value statistics of the plurality of calibration distance points.

[0256] In some embodiments, reference Figure 12 The distance determination module 1130 may include: a coordinate acquisition submodule 1131, a coordinate difference determination submodule 1132, and a distance calculation submodule 1133.

[0257] The coordinate acquisition submodule 1131 is configured to acquire the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix;

[0258] The coordinate difference determination submodule 1132 is configured to determine, based on the first row and column coordinates and the second row and column coordinates, the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction; and

[0259] The distance calculation submodule 1133 is configured to calculate the Euclidean distance of the pixel relative to the reference center point based on the first coordinate difference and the second coordinate difference.

[0260] In some embodiments, reference Figure 13 The distance determination module 1130 may include: a coordinate acquisition submodule 1134, a coordinate transformation submodule 1135, and a radius determination submodule 1136.

[0261] The coordinate acquisition submodule 1134 is configured to acquire the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix;

[0262] The coordinate transformation submodule 1135 is configured to obtain the polar coordinates of each pixel with the reference center point as the origin, based on the first row and column coordinates and the second row and column coordinates; and

[0263] The radius determination submodule 1136 is configured to obtain the radius value of the pixel in the polar coordinates as the distance relative to the reference center point.

[0264] In some embodiments, the allocation statistics module 1140 is further configured to, for each pixel, allocate the pixel value of the pixel to the multiple adjacent calibration distance points according to the linear proportional relationship between the distance of the pixel and the multiple adjacent calibration distance points, and perform cumulative statistics on the pixel values ​​allocated to each calibration distance point to obtain the pixel value statistics of each calibration distance point.

[0265] In some embodiments, reference Figure 14 The allocation and statistics module 1140 may include: a distance decomposition submodule 1141, a nearest-neighbor point determination submodule 1142, a weight determination submodule 1143, an allocation submodule 1144, and a statistics submodule 1145.

[0266] The distance decomposition submodule 1141 is configured to decompose the distance of each pixel into an integer part d_int and a fractional part d_frac;

[0267] The neighbor point determination submodule 1142 is configured to determine one or two calibration distance points adjacent to the pixel based on the integer part d_int;

[0268] The weight determination submodule 1143 is configured to determine the weight coefficients allocated to the one or two adjacent calibration distance points based on the fractional part d_frac.

[0269] The allocation submodule 1144 is configured to determine, based on the pixel value of the pixel and the weighting coefficient, the weighted pixel value to be weighted and allocated to one or two adjacent calibration distance points; and

[0270] The statistics submodule 1143 is configured to accumulate and statistically analyze the weighted pixel values ​​assigned to each calibration distance point to obtain the pixel value statistics corresponding to each calibration distance point.

[0271] In some embodiments, the statistics submodule 1143 includes:

[0272] Multiple accumulators, each corresponding to a calibration distance point, are configured to store the sum of the weighted pixel values ​​assigned to that calibration distance point.

[0273] In some embodiments, the statistics submodule 1143 includes:

[0274] Multiple counters, each corresponding to a calibration distance point, are configured to record the count of the weighted pixel values ​​assigned to that calibration distance point.

[0275] In some embodiments, reference Figure 15 The evaluation module 1150 may include: a line diffusion submodule 1151, a frequency domain conversion submodule 1152, and an index generation submodule 1153. Among them,

[0276] The line diffusion submodule 1151 is configured to determine the line diffusion function curve of the image to be evaluated based on the pixel value statistics of the calibration distance points;

[0277] The frequency domain conversion submodule 1152 is configured to perform frequency domain conversion on the line spread function curve to obtain the modulation transfer function curve of the image to be evaluated; and

[0278] The index generation submodule 1153 is configured to process the modulation transfer function curve to obtain the at least one spatial resolution index.

[0279] In some embodiments, the spatial resolution index is a high-contrast spatial resolution index.

[0280] The steps, sub-steps, and features of the image spatial resolution evaluation method described in the embodiments of this application can be combined in a non-contradictory manner with the image spatial resolution evaluation system of the embodiments of this application to obtain new embodiments. Conversely, the components, modules, units, or features of the image spatial resolution evaluation system described in the embodiments of this application can also be combined in a non-contradictory manner with the image spatial resolution evaluation method of the embodiments of this application to obtain new embodiments.

[0281] The steps and sub-steps described in the embodiments of this application can be executed independently or separately, or they can be combined or merged without contradiction. Furthermore, the order of steps in the embodiments of this application is not absolutely limited; the execution order of some steps can be adjusted or they can be executed in parallel according to actual needs, provided that the technical solution is not affected.

[0282] In this application embodiment, an electronic device is also provided, which may include a processor and a memory storing a computer program, the processor being configured to execute the method of any embodiment of this application when running the computer program.

[0283] Figure 16 A schematic diagram of an exemplary electronic device 1600 that can implement the methods of embodiments of this application is shown. In some embodiments, it may include more or fewer electronic devices than shown. In some embodiments, it may be implemented using a single or multiple electronic devices. In some embodiments, it may be implemented using cloud-based or distributed electronic devices.

[0284] like Figure 16 As shown, the electronic device 1600 includes a processor 1601, which can perform various appropriate operations and processes based on programs and / or data stored in read-only memory (ROM) 1602 or programs and / or data loaded from storage portion 1608 into random access memory (RAM) 1603. The processor 1601 can be a single-core or multi-core processor, or may include multiple processors. In some embodiments, the processor 1601 may include a general-purpose main processor (such as a CPU) and one or more special coprocessors, such as a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), or other general-purpose or application-specific integrated circuits. Various programs and data required for the operation of the electronic device 1600 are also stored in RAM 1603. The processor 1601, ROM 1602, and RAM 1603 are interconnected via bus 1604. An input / output (I / O) interface 1605 is also connected to bus 1604.

[0285] The processor and memory described above are used together to execute a program stored in the memory. When the program is executed by a computer, it can implement the steps or functions of the methods described in the above embodiments.

[0286] The following components are connected to I / O interface 1605: an input section 1606 including a keyboard, mouse, etc.; an output section 1607 including a display and speakers, etc.; a storage section 1608 including a hard disk, etc.; and a communication section 1609 including a network interface card such as a LAN card and a modem, etc. The communication section 1609 performs communication processing via a network such as the Internet. Drive 1610 is also connected to I / O interface 1605 as needed. Removable media 1611, such as a hard disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1610 as needed so that computer programs read from them can be installed into storage section 1608 as needed.

[0287] Figure 16 The electronic device shown is merely illustrative, but the electronic device according to embodiments of this application may include more than [other components]. Figure 16 The electronic device shown has more or fewer components or has more or fewer components than the one shown. Figure 16 The embodiments shown have the same, partially the same, or different architectures.

[0288] In the embodiments of this application, the electronic device can also be combined with various components to obtain methods, apparatus and systems with the advantages of this application.

[0289] Although not shown, this application also provides a computer-readable storage medium storing a computer program configured to execute the methods of any of the embodiments of this application. The computer program includes various program modules / units constituting the apparatus according to the embodiments of this application. When executed, the computer program, composed of the various program modules / units, can perform the functions corresponding to the various steps in the methods described in the above embodiments. The computer program can also run on electronic devices as described in the embodiments of this application.

[0290] Although not shown, some embodiments also provide a program product comprising a computer program configured to be run to perform the methods of any of the embodiments of this application.

[0291] The storage medium in embodiments of this application includes non-volatile and / or volatile articles that can store information by any method or technology. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0292] Those skilled in the art will understand that the embodiments of this specification can be implemented in various forms, such as methods, systems, or computer program products. Therefore, those skilled in the art will realize that the functional modules / units or controllers and related method steps described in the above embodiments can be implemented in software, hardware, or a combination of software and hardware.

[0293] Unless explicitly stated otherwise, the actions or steps of the methods and procedures described in the embodiments of this application do not necessarily have to be performed in a specific order and can still achieve the desired results. In some implementations, multitasking and parallel processing are also possible or may be advantageous.

[0294] This document describes several embodiments, but for the sake of brevity, the descriptions of the embodiments are not exhaustive, and identical or similar features or parts between the embodiments may be omitted. In this document, "one embodiment," "some embodiments," "example," "specific example," or "some examples" refers to at least one embodiment or example applicable to this application, but not all embodiments. The above terms do not necessarily mean referring to the same embodiment or example. Without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of the different embodiments or examples.

[0295] The exemplary systems and methods of this application have been specifically shown and described with reference to the above embodiments, which are merely examples of the best mode for implementing the systems and methods. Those skilled in the art will understand that various changes can be made to the embodiments of the systems and methods described herein without departing from the spirit and scope of this application as defined in the appended claims when implementing the systems and / or methods.

Claims

1. An image spatial resolution evaluation method characterized by, include: Obtain the two-dimensional pixel matrix of the image to be evaluated; The reference center point of the image to be evaluated is determined based on the two-dimensional pixel matrix; Determine the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point; Based on the distance of each pixel, the pixel value of each pixel is interpolated and distributed to multiple calibration distance points to obtain the pixel value statistics of each calibration distance point; Based on the pixel value statistics of the plurality of calibration distance points, at least one spatial resolution index of the image to be evaluated is determined.

2. The method of claim 1, wherein, The step of allocating the pixel values ​​of each pixel to multiple calibration distance points through interpolation based on the distance of each pixel, and obtaining statistical data of the pixel values ​​at each calibration distance point, includes: For each pixel, based on the linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, the pixel value of the pixel is assigned to one or more adjacent calibration distance points, and the pixel values ​​assigned to each calibration distance point are accumulated and statistically analyzed to obtain the pixel value statistics of each calibration distance point.

3. The method of claim 2, wherein, The calibration distance point is an integer distance point relative to the reference center point; Specifically, for each pixel, based on a linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, the pixel value of the pixel is allocated to one or more adjacent calibration distance points, and the pixel values ​​allocated to each calibration distance point are accumulated and statistically analyzed to obtain pixel value statistics for each calibration distance point, including: The distance of each pixel is decomposed into an integer part d_int and a fractional part d_frac; The integer part d_int determines one or two calibration distance points adjacent to the pixel. The weighting coefficients assigned to one or two adjacent calibration distance points are determined based on the fractional part d_frac. The weighted pixel value assigned to the adjacent one or two calibration distance points is determined based on the pixel value of the pixel and the weighting coefficient. The weighted pixel values ​​assigned to each calibration distance point are accumulated and statistically analyzed to obtain the pixel value statistics corresponding to each calibration distance point.

4. The method of claim 3, wherein, The step of determining the weighting coefficients assigned to one or two adjacent calibration distance points based on the fractional part d_frac specifically includes: Use 1-d_frac as the first weight coefficient of the first calibration distance point among one or two adjacent calibration distance points; When there are two adjacent calibration distance points, d_frac is used as the second weight coefficient of the second calibration distance point among the two adjacent calibration distance points, where the distance between the first calibration distance point is less than the distance between the second calibration distance point.

5. The method of claim 4, wherein, The step of determining the weighted pixel value to be weighted and assigned to one or two adjacent calibration distance points based on the pixel value and the weighting coefficient specifically includes: The pixel value of the pixel is multiplied by the first weighting coefficient to obtain the first multiplication accumulation, which is then added to the first accumulator corresponding to the first calibration distance point. When there are two adjacent calibration distance points, the pixel value of the pixel is multiplied by the second weighting coefficient to obtain the second multiplication accumulation, which is added to the second accumulator corresponding to the second calibration distance point.

6. The method of claim 5, wherein, The step of accumulating and statistically analyzing the weighted pixel values ​​assigned to each calibration distance point to obtain the pixel value statistics corresponding to each calibration distance point specifically includes: After traversing all pixels in the two-dimensional pixel matrix, the value in the accumulator corresponding to each calibration distance point is used as the pixel value statistics corresponding to that calibration distance point.

7. The method of claim 6, wherein, The step of determining the weighted pixel value to be weighted and assigned to one or two adjacent calibration distance points based on the pixel value of the pixel and the weighting coefficient further includes: When the first multiplication accumulation is added to the first accumulator, the count in the first counter corresponding to the first calibration distance point is incremented by 1; When the second multiplication accumulation is added to the second accumulator, the count in the second counter corresponding to the second calibration distance point is incremented by 1; The step of accumulating and statistically analyzing the weighted pixel values ​​assigned to each calibration distance point to obtain statistical data on the pixel values ​​corresponding to each calibration distance point also includes: After traversing all pixels in the two-dimensional pixel matrix, the value in the counter corresponding to each calibration distance point is used as the statistical count of the number of associated pixels corresponding to that calibration distance point.

8. The method according to any one of claims 1 to 7, characterized in that, Determining the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point includes: Obtain the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix; Based on the first row and column coordinates and the second row and column coordinates, determine the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction; Based on the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction, the Euclidean distance of the pixel relative to the reference center point is calculated, which is used as the distance relative to the reference center point.

9. The method according to any one of claims 1 to 7, characterized in that, Determining the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point includes: Obtain the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix; Based on the first row and column coordinates and the second row and column coordinates, the polar coordinates of each pixel with the reference center point as the origin are obtained; Obtain the radius value of the pixel in the polar coordinates, as the distance relative to the reference center point.

10. The method according to any one of claims 1 to 7, characterized in that, Determining at least one spatial resolution index of the image to be evaluated based on the pixel value statistics of the plurality of calibrated distance points includes: Based on the pixel value statistics of the calibration distance points, determine the line spread function curve of the image to be evaluated; The line spread function curve is transformed in the frequency domain to obtain the modulation transfer function curve of the image to be evaluated; The modulation transfer function curve is processed to obtain the at least one spatial resolution index.

11. The method according to claim 10, characterized in that, The process of processing the modulation transfer function curve to obtain the at least one spatial resolution index includes: The spatial frequency value corresponding to the decrease of the parameter value of the modulation transfer function curve to a preset percentage threshold is used as the spatial resolution index.

12. The method according to any one of claims 1 to 7, characterized in that, Determining the reference center point of the image to be evaluated based on the two-dimensional pixel matrix includes: The pixel with the largest pixel value in the two-dimensional pixel matrix is ​​taken as the reference center point.

13. The method according to any one of claims 1 to 7, characterized in that, The spatial resolution index is the high-contrast spatial resolution index.

14. The method according to any one of claims 1 to 7, characterized in that, The image to be evaluated is a tomographic image, and the format of the image to be evaluated includes at least one of DICOM format, NIfTI format, and MHD / RAW format.

15. An image spatial resolution evaluation system, characterized in that, include: The acquisition module is configured to acquire the two-dimensional pixel matrix of the image to be evaluated; The center positioning module is configured to determine the reference center point of the image to be evaluated based on the two-dimensional pixel matrix; A distance determination module is configured to determine the distances of at least a portion of the pixels in the two-dimensional pixel matrix relative to the reference center point; The allocation statistics module is configured to allocate the pixel value of each pixel to multiple calibration distance points through interpolation based on the distance of each pixel, so as to obtain the pixel value statistics of each calibration distance point; The evaluation module is configured to determine at least one spatial resolution index of the image to be evaluated based on the pixel value statistics of the plurality of calibration distance points.

16. The system according to claim 15, characterized in that, The allocation statistics module is further configured to, for each pixel, allocate the pixel value of the pixel to the one or more adjacent calibration distance points according to the linear proportional relationship between the distance of the pixel and one or more adjacent calibration distance points, and perform cumulative statistics on the pixel values ​​allocated to each calibration distance point to obtain the pixel value statistics of each calibration distance point.

17. The system according to claim 16, characterized in that, The allocation statistics module includes: The distance decomposition submodule is configured to decompose the distance of each pixel into an integer part d_int and a fractional part d_frac; The nearest neighbor point determination submodule is configured to determine one or two calibration distance points adjacent to the pixel based on the integer part d_int; The weight determination submodule is configured to determine the weight coefficients assigned to one or two adjacent calibration distance points based on the fractional part d_frac. The allocation submodule is configured to determine, based on the pixel value of the pixel and the weighting coefficient, the weighted pixel value to be weighted and allocated to one or two adjacent calibration distance points; and The statistics submodule is configured to accumulate and statistically analyze the weighted pixel values ​​assigned to each calibration distance point to obtain the pixel value statistics corresponding to each calibration distance point.

18. The system according to claim 17, characterized in that, The statistics submodule includes: Multiple accumulators, each corresponding to a calibration distance point, are configured to store the sum of the weighted pixel values ​​assigned to that calibration distance point.

19. The system according to claim 17 or 18, characterized in that, The statistics submodule includes: Multiple counters, each corresponding to a calibration distance point, are configured to record the count of the weighted pixel values ​​assigned to that calibration distance point.

20. The system according to any one of claims 15 to 18, characterized in that, The distance determination module includes: The coordinate acquisition submodule is configured to acquire the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix; The coordinate difference determination submodule is configured to determine, based on the first row and column coordinates and the second row and column coordinates, the first coordinate difference between the pixel and the reference center point in the row direction and the second coordinate difference in the column direction; and The distance calculation submodule is configured to calculate the Euclidean distance of the pixel relative to the reference center point based on the first coordinate difference and the second coordinate difference.

21. The system according to any one of claims 15 to 18, characterized in that, The distance determination module includes: The coordinate acquisition submodule is configured to acquire the first row and column coordinates of the reference center point in the two-dimensional pixel matrix and the second row and column coordinates of each pixel in the two-dimensional pixel matrix; The coordinate transformation submodule is configured to obtain the polar coordinates of each pixel with the reference center point as the origin, based on the first row and column coordinates and the second row and column coordinates; and The radius determination submodule is configured to obtain the radius value of the pixel in the polar coordinates as the distance relative to the reference center point.

22. The system according to any one of claims 15 to 18, characterized in that, The evaluation module includes: The line diffusion submodule is configured to determine the line diffusion function curve of the image to be evaluated based on the pixel value statistics of the calibration distance points; A frequency domain transformation submodule is configured to perform frequency domain transformation on the line spread function curve to obtain the modulation transfer function curve of the image to be evaluated; and The index generation submodule is configured to process the modulation transfer function curve to obtain the at least one spatial resolution index.

23. The system according to any one of claims 15 to 18, characterized in that, The spatial resolution index is the high-contrast spatial resolution index.

24. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores computer program instructions, and the processor, when executing the computer program instructions, implements the method as described in any one of claims 1 to 14.

25. A computer storage medium, characterized in that, The system stores computer program instructions, which, when executed by a processor, implement the method as described in any one of claims 1 to 14.

Images