Image quality evaluation method, device and electronic equipment

Abstract

This application discloses an image quality assessment method, apparatus, and electronic device. The method includes: acquiring raw image data of a target object; determining the geometric centroid of a metal implant from the raw image data; dividing the human mask region corresponding to the target object in the raw image data into multiple sub-regions by extending multiple rays radially from the geometric centroid; determining statistical indicators for each sub-region; and determining the quality assessment result of the raw image data based on the statistical indicators. This application solves the technical problem of traditional assessment methods that uniformly process all pixels in the entire medical image, easily leading to confusion between radial artifacts produced by metal and statistical fluctuations generated by normal high-density bone structures, resulting in inaccurate assessment results.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and more specifically, to an image quality evaluation method, apparatus, and electronic device. Background Technology

[0002] In medical computed tomography (CT) or PET / CT imaging, metallic implants in patients (such as dental fillings, spinal fixation devices, etc.) have a much higher density than human tissue, causing strong uneven X-ray attenuation and resulting in radial streaks in the images. This severely affects the visualization of lesion areas and the accuracy of diagnosis. Currently, the clinical assessment of these artifacts mainly relies on the subjective visual judgment of imaging technicians or radiologists, or uses overall statistical measures such as global signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) to quantify and score the entire image. These methods treat all pixels in the image as uniformly distributed samples and cannot distinguish between the directional radial structure unique to artifacts and the local statistical fluctuations of normal high-density tissues (such as the skull and cortical bone), leading to easy misinterpretation of the assessment results.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This application provides an image quality assessment method, apparatus, and electronic device to at least solve the technical problem that traditional assessment methods process all pixels in the entire medical image uniformly, which easily leads to confusion between radial artifacts produced by metal and statistical fluctuations produced by normal high-density bone structures, resulting in inaccurate assessment results.

[0005] According to one aspect of the embodiments of this application, an image quality assessment method is provided, comprising: acquiring original image data of a target object, wherein the original image data includes metal artifacts caused by a metal implant inside the target object; determining the geometric centroid corresponding to the metal implant from the original image data; extending multiple rays radially from the geometric centroid to divide the human mask region corresponding to the target object in the original image data into multiple sub-regions, wherein each sub-region is used to cover the region from the geometric centroid to the boundary of the human mask; determining a statistical index corresponding to each sub-region, and determining the quality assessment result of the original image data based on the statistical index, wherein the statistical index is used to quantify the degree of dispersion of pixel distribution within the sub-region.

[0006] In some embodiments of this application, determining the geometric centroid of the metal implant from the original image data includes: determining the human body mask region corresponding to the target object from the original image data, wherein each pixel value within the human body mask region is used to quantify the attenuation coefficient of the human tissue of the target object to X-rays; determining the metal region corresponding to the metal implant within the human body mask region; and determining the geometric centroid of the metal implant based on the metal region.

[0007] In some embodiments of this application, determining the metal region corresponding to the metal implant within the human body mask region includes: acquiring the standardized image values ​​corresponding to all pixels within the human body mask region; comparing the standardized image values ​​with a first preset threshold to obtain a target pixel, wherein the standardized image value of the target pixel is greater than or equal to the first preset threshold, the first preset threshold being determined based on the energy and intensity of X-rays; performing connectivity analysis on the target pixel, and determining the metal region based on the connectivity analysis results.

[0008] In some embodiments of this application, multiple rays are extended radially from the geometric centroid to divide the human mask region corresponding to the target object in the original image data into multiple sub-regions, including: determining the number of sub-regions; dividing the human mask region into multiple fan-shaped regions equal to the number of divisions with the geometric centroid as the center, wherein the angles of the multiple fan-shaped regions are equal; and determining the multiple fan-shaped regions as sub-regions.

[0009] In some embodiments of this application, determining the number of sub-region divisions includes: determining a first distance between the geometric centroid and the boundary of the human body mask region, wherein the first distance is used to reflect the maximum scanning radius of the sub-region division; obtaining a second preset threshold, wherein the second preset threshold is used to reflect the number of pixel center points crossed on the arc path formed from the geometric centroid to the boundary of the human body mask region; and determining the number of divisions based on the first distance and the second preset threshold.

[0010] In some embodiments of this application, determining the quality assessment result of the original image data based on statistical indicators includes: determining the maximum and minimum values ​​from the statistical indicators corresponding to all sub-regions, wherein the minimum value is used to quantify the inherent tissue noise level of the target object; determining the difference between the maximum and minimum values ​​as a first indicator, wherein the first indicator is used to quantify the degree of stripe fluctuation of the metal artifact of the metal implant; and determining the quality assessment result based on the first indicator.

[0011] In some embodiments of this application, determining the quality assessment result based on a first indicator includes: determining a second indicator corresponding to the metal implant, wherein the second indicator is used to quantify the metal volume of the metal implant; determining a third indicator corresponding to the metal implant, wherein the third indicator is used to quantify the extent of metal artifacts of the metal implant; and determining the quality assessment result based on the first indicator, the second indicator, and the third indicator.

[0012] In some embodiments of this application, determining the second indicator corresponding to the metal implant includes: determining the area of ​​the metal region where the metal implant is located, and using the area as the second indicator.

[0013] In some embodiments of this application, determining a third indicator corresponding to a metal implant includes: determining a target sub-region from multiple sub-regions, wherein the target sub-region includes sub-regions whose statistical indicators meet preset conditions; determining the coverage area of ​​the target sub-region, wherein the coverage area is used to quantify the affected area of ​​the metal artifact; and determining the ratio of the coverage area to the total area of ​​the human body cross-section corresponding to the human body mask area as the third indicator.

[0014] In some embodiments of this application, determining the quality assessment result based on a first indicator, a second indicator, and a third indicator includes: weighting and summing the first indicator, the second indicator, and the third indicator to obtain a comprehensive score, wherein the comprehensive score is used to quantify the degree of influence of metal artifacts of metal implants on the image quality of the original image data in multiple dimensions; and mapping the comprehensive score to a preset grading level to obtain the quality assessment result.

[0015] In some embodiments of this application, when the classification level meets the preset conditions, metal artifact suppression is performed on the original image data; or, when the first index of any sub-region is greater than the third preset threshold, metal artifact suppression is performed on the original image data.

[0016] According to another aspect of the embodiments of this application, an image quality evaluation apparatus is also provided, comprising: an acquisition module for acquiring original image data of a target object, wherein the original image data includes metal artifacts caused by a metal implant inside the target object; a determination module for determining the geometric centroid corresponding to the metal implant from the original image data; a division module for dividing the human body mask region corresponding to the target object in the original image data into multiple sub-regions by extending multiple rays in a radial direction centered on the geometric centroid, wherein each sub-region is used to cover the region from the geometric centroid to the boundary of the human body mask; and an evaluation module for determining a statistical index corresponding to each sub-region and determining the quality evaluation result of the original image data based on the statistical index, wherein the statistical index is used to quantify the dispersion of pixel distribution within the sub-region.

[0017] According to another aspect of the embodiments of this application, an electronic device is also provided, including: a memory and a processor, wherein the memory is used to store program instructions; the processor is connected to the memory and is used to execute the above-described image quality evaluation method.

[0018] According to another aspect of the embodiments of this application, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored computer program, wherein the device on which the non-volatile storage medium is located executes the above-described image quality evaluation method by running the computer program.

[0019] According to another aspect of the embodiments of this application, a computer program product is also provided, including computer instructions that, when executed by a processor, implement the above-described image quality evaluation method.

[0020] In this embodiment, a directional statistical dispersion analysis method is adopted. The geometric centroid of the metal implant is automatically located from the original medical image, and multiple fan-shaped sub-regions covering the human body mask area are divided radially with the centroid as the origin. The statistical dispersion index of the pixel values ​​in each sub-region is then calculated. The image quality is then evaluated by comprehensively considering the dispersion differences between the sub-regions. This achieves the objective quantification of the directional characteristics of metal artifacts, thereby realizing the technical effect of distinguishing radial artifacts from normal high-density tissue noise. This solves the technical problem that traditional evaluation methods process all pixels in the entire medical image uniformly, which easily leads to confusion between radial artifacts generated by metal and statistical fluctuations generated by normal high-density bone structures, resulting in inaccurate evaluation results. Attached Figure Description

[0021] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0022] Figure 1 This is a hardware structure block diagram of a computer terminal for an image quality evaluation method according to an embodiment of this application;

[0023] Figure 2 This is a flowchart of an image quality evaluation method according to an embodiment of this application;

[0024] Figure 3 This is a schematic diagram illustrating the principle of multi-sector radial statistical analysis of metal artifacts in an image quality assessment method according to an embodiment of this application.

[0025] Figure 4 This is a schematic diagram of sub-region division within a human body mask area according to an embodiment of the present application for an image quality evaluation method;

[0026] Figure 5 This is a schematic diagram illustrating the process of determining the quality assessment result of an image quality assessment method according to an embodiment of this application;

[0027] Figure 6 This is a schematic diagram of the structure of an image quality evaluation device according to an embodiment of this application. Detailed Implementation

[0028] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0030] To better understand the embodiments of this application, the technical terms involved in the embodiments of this application are explained below:

[0031] The Hounsfield Unit (HU) is a standardized unit used in medical imaging to quantify the ability of tissues to attenuate X-rays. It is based on water (0 HU), air (-1000 HU), and dense bone (+1000 HU or higher). In this embodiment, the Hounsfield Unit is used to linearly convert the original pixel value into a physically meaningful attenuation coefficient value through Rescale Slope / Intercept, thereby achieving accurate differentiation between metal implants (such as dental fillings) and normal tissues (such as bone and soft tissue) in terms of attenuation characteristics.

[0032] Body Mask: A binary image generated by thresholding and morphological processing of the original medical image. It is used to identify the effective areas of the patient's body tissue in the image and exclude interference from non-human structures such as bed boards and air.

[0033] Geometric centroid (C): The coordinates of the center of mass of a connected region in two-dimensional space. For example, it can be obtained by calculating the zeroth and first moments of all pixels in the region. In the embodiments of this application, the geometric centroid is used to accurately characterize the spatial center position of the metal implant. As the unique origin of radial sector division, it ensures that the artifact directionality analysis is based on the metal source as a symmetry reference, thereby improving the symmetry of sector statistics and the reliability of artifact feature extraction.

[0034] It should be noted that artifacts caused by metal implants (such as dental fillings, spinal screws, etc.) in images exhibit distinct radial stripes. Their energy distribution is not uniformly distributed around the metal source, but rather diffuses at a specific angle along the X-ray penetration path, centered on the metal, forming a directional, high-contrast stripe structure. The geometric centroid accurately reflects the spatial center of gravity of the metal entity, regardless of its shape—whether it's a regular circle, a strip, or a polygon—ensuring that subsequent sub-regions constructed with that point as the origin are perfectly symmetrical around the metal source. Furthermore, each sub-region uniformly covers the possible directions of artifact diffusion at different angles, allowing statistical indicators such as standard deviation and range to sensitively respond to which direction of pixel fluctuation is significantly enhanced. This allows for the perception of the directional propagation characteristics of artifacts in a spatial dimension, thus clearly distinguishing directional artifacts from isotropic bone noise based on statistical characteristics.

[0035] Radial sector: A sector-shaped region (a specific implementation of a sub-region) that is uniformly divided along an angle with the geometric centroid as the vertex. Each sector covers a radial region from the centroid to the boundary of the human mask. In the embodiments of this application, the radial sector is used to discretize the image space into multiple directional channels, so that subsequent statistical analysis can capture the directional characteristics of radial stripes unique to artifacts, thereby overcoming the limitation of traditional global statistical methods that cannot identify spatial directional differences.

[0036] Statistical Dispersion: A statistical indicator that describes the degree of dispersion of a set of data, including but not limited to standard deviation and coefficient of variation. In this embodiment, statistical dispersion is used to quantify the fluctuation of pixel intensity within each radial sector. Metal artifacts, due to their high-contrast radial stripes, can cause a significant increase in the standard deviation of a specific sector, while normal tissue is evenly distributed and has low dispersion.

[0037] In medical computed tomography (CT) or positron emission tomography / computed tomography (PET / CT), metallic implants in patients (such as dental fillings, artificial joints, spinal screws, etc.) produce severe radial streak artifacts due to their extremely high density. Currently, clinical control of image quality mainly relies on subjective scoring by technicians or doctors through visual examination, or automatic assessment using global statistical measures such as global signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). However, manual assessment is greatly influenced by experience, lacks a quantitative and unified standard, and is difficult to apply across different hospitals. Furthermore, statistical measures such as global signal-to-noise ratio (SNR) treat the entire image as isotropic and cannot identify the directional radial streak specific to artifacts. In addition, traditional thresholding methods or global statistical methods often confuse the statistical fluctuations produced by normal high-density bone structures (such as the skull and cortical bone) with the radial artifacts produced by metal, leading to inflated scores.

[0038] To address the aforementioned technical problems, this application provides corresponding solutions, which are detailed below.

[0039] The image quality evaluation method provided in this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal for implementing an image quality evaluation method is shown. Figure 1 As shown, the computer terminal 10 may include one or more processors (shown as 102a, 102b, ..., 102n in the figure) (the processor may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission module 106 for communication functions connected via wired and / or wireless networks. In addition, it may also include: a display, a keyboard, a cursor control device, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, and a BUS bus. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0040] It should be noted that the aforementioned one or more processors and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10. As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0041] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the image quality evaluation method in this embodiment. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned image quality evaluation method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0042] The transmission module 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission module 106 includes a network interface controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission module 106 may be a radio frequency (RF) module, used for wireless communication with the Internet.

[0043] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10.

[0044] It should be noted here that, in some optional embodiments, the above... Figure 1 The computer terminal shown may include hardware elements (including circuitry), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware and software elements. It should be noted that... Figure 1 This is only one instance of a specific particular instance, and is intended to illustrate the types of components that may exist in the aforementioned computer terminal.

[0045] In the above operating environment, this application provides an embodiment of an image quality evaluation method. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0046] Figure 2 This is a flowchart of an image quality evaluation method according to an embodiment of this application, such as... Figure 2 As shown, the method includes the following steps:

[0047] Step S202: Obtain the original image data of the target object, wherein the original image data contains metal artifacts caused by the metal implant inside the target object.

[0048] In step S202 above, the raw image data refers to the raw tomographic image data directly acquired by CT or PET / CT scanning equipment without subsequent reconstruction or filtering. For example, it is stored in DICOM format, and its pixel values ​​are integers (such as 0–4095) of the original output of the equipment. It has not yet been converted into attenuation coefficients with physical meaning, and it retains the original shape and intensity distribution of metal artifacts.

[0049] It should be noted that metal artifacts caused by metal implants refer to the phenomenon in X-ray computed tomography (CT) where the extremely high attenuation coefficient of metallic materials (such as titanium alloys, cobalt-chromium alloys, and amalgam) leads to nonlinear saturation, photon starvation, and scattering effects in the projected data. This results in radial or star-shaped stripe artifacts centered on the metal element in the reconstructed image. Metal artifacts exhibit directional characteristics, meaning the stripes spread along specific angles. For example, dental fillings often produce significant artifacts in the horizontal direction, while spinal screws may form stripes in the anterior-posterior direction.

[0050] In some embodiments of this application, after acquiring the raw image data (using a DICOM sequence as an example), since the pixel values ​​in the raw DICOM file are usually meaningless raw integers (e.g., 0-4095), they can be converted into standardized image values, such as Henle units (HU), using the Rescale Slope / Intercept scaling parameters. The HU value is a standardized physical quantity that directly reflects the attenuation coefficient of tissue to X-rays. Human body regions can be extracted by setting a threshold (preferably -500 HU), and non-human structures such as bed boards can be removed by morphological closing operations to generate a body mask.

[0051] Step S204: Determine the geometric centroid of the metal implant from the original image data.

[0052] In step S204 above, the geometric centroid refers to the center of the weighted average of the coordinates of all pixels in the two-dimensional image of the metal implant. The geometric centroid can accurately reflect the spatial distribution center of irregularly shaped metal regions. Even if the metal is elongated, polygonal, or has blurred edges, its geometric center can be output stably.

[0053] It should be noted that the geometric centroid is the only spatial origin for artifact directionality analysis. For example, when dental fillings are distributed in an irregular triangular pattern, if the image center or the largest pixel is used as a reference, the radial sub-regions will be severely offset, causing the artifact fringes to be dispersed into multiple sub-regions and weakening the directional features. However, with the geometric centroid as the origin, all sub-regions radiate outwards with the metal source as the center of symmetry, allowing the concentrated enhancement of artifact energy at a specific angle (such as the horizontal direction) to be captured with high sensitivity.

[0054] In some embodiments of this application, the geometric centroid can be determined using the image moment method. Specifically, a binarized image is first constructed. Pixels belonging to the metallic connected region have a value of 1, while the rest have a value of 0. This is determined by calculating the zeroth moment. (Area of ​​region) and first moment , Determine the centroid coordinates The calculation formula is as follows:

[0055]

[0056]

[0057] in, This represents the row and column coordinate index of a pixel in the image matrix. The (p,q)-th order moment of an image is a mathematical quantity describing the spatial distribution characteristics of an image region, where p and q are powers of the x and y coordinates, respectively, and are used for weighted summation. It represents the zeroth moment, and the total number of pixels with a value of 1 within the metal connected region, which is the area (number of pixels) of the metal region. Let represent the component of the first moment with respect to the y-direction, and let represent the sum of the products of the y-coordinates of all metal pixels and their pixel values ​​I(x,y). It represents the component of the first moment with respect to the x-direction, and represents the sum of the products of the x-coordinates of all metal pixels and their pixel values ​​I(x,y).

[0058] The above method ensures that when the metal shape is irregular (such as a strip or an irregular polygon), the centroid can accurately reflect its geometric center, thus ensuring the symmetry of the subsequent radial sub-region division.

[0059] To further reduce noise in the original image data, the geometric centroid of the metal implant can be determined from the original image data in the following way: determine the human body mask region corresponding to the target object from the original image data, where each pixel value in the human body mask region is used to quantify the attenuation coefficient of the human tissue of the target object to X-rays; determine the metal region corresponding to the metal implant within the human body mask region; and determine the geometric centroid of the metal implant based on the metal region.

[0060] It should be noted that the human body mask region refers to a binary mask image generated after preprocessing the original image data, where each pixel has a value of 1 or 0, indicating whether the location belongs to or does not belong to human tissue, respectively. The metal region refers to a high-density pixel set caused by a metal implant, identified within the human body mask region by an intensity threshold (e.g., ≥2600 HU). It usually appears as isolated or clustered patches. Identifying the metal region can separate the true source of artifacts from the background, so that subsequent geometric centroid calculations are no longer affected by similar intensity regions such as bones and calcifications.

[0061] Specifically, after selecting the human body mask area based on the threshold, the metal region corresponding to the metal implant can be determined within the human body mask area in the following way: obtain the standardized image values ​​corresponding to all pixels within the human body mask area; compare the standardized image values ​​with a first preset threshold to obtain the target pixel, wherein the standardized image value of the target pixel is greater than or equal to the first preset threshold, which is determined based on the energy and intensity of X-rays; perform connectivity analysis on the target pixel, and determine the metal region based on the connectivity analysis results.

[0062] Standardized image values ​​refer to the values ​​obtained by converting the raw integer pixel values ​​(e.g., 0–4095) in the original CT scan data. This allows image data from different devices and with different scanning parameters (e.g., kVp, mA) to be compared at a uniform physical scale, avoiding threshold drift caused by device differences. For example, using the Rescale Slope and Rescale Intercept parameters provided in the DICOM file, a linear conversion can be made to the Hounsfield Unit (HU), which has a clear physical meaning.

[0063] For example, within the human body mask area, based on a preset first preset threshold... (Preferred value is 2600 HU, selectable range [2000 HU, 3000 HU], used as the intensity discrimination threshold to distinguish between metallic and non-metallic high-density structures) Identify metallic pixel regions, and use connected component analysis to identify each independent metallic connected component, and calculate the geometric centroid of each independent connected component. Subsequent sub-region division and statistical analysis will be conducted separately for each individual metal centroid.

[0064] The first preset threshold can be dynamically fine-tuned based on the current tube voltage (kVp) and tube current (mA).

[0065] In some embodiments of this application, the first preset threshold can be dynamically calibrated based on scanning parameters such as X-ray energy (kVp) and tube current (mA). For example, the cranial cortex may reach 1800 HU under standard scanning, while titanium alloy implants can reach 2900 HU under the same conditions. By setting a threshold higher than the upper limit of bone (such as 2600 HU), false bone detection can be effectively eliminated, and only the area with extremely strong attenuation caused by metal can be retained.

[0066] Step S206: Extend multiple rays radially from the geometric centroid to divide the human mask region corresponding to the target object in the original image data into multiple sub-regions, wherein each sub-region is used to cover the area from the geometric centroid to the boundary of the human mask.

[0067] In step S206 above, a sub-region refers to a spatial region of arbitrary shape cut out from the geometric centroid along multiple radial directions to the boundary of the human body mask. Each sub-region covers the complete radial path from the centroid to the mask boundary, including but not limited to fan-shaped regions, wedge-shaped regions (such as wedge-shaped sub-regions formed by replacing the outer boundary of a fan-shaped region with an irregular curve that fits the actual contour of the human body mask), and radial strip regions (non-rigid strip-shaped regions that extend from the centroid along a certain adaptive direction, with boundaries dynamically guided by image gradient, artifact sensitivity, or edge detection algorithms). It should be noted that the role of the sub-region is to capture the energy diffusion characteristics of artifacts along the X-ray penetration direction on a spatial path basis.

[0068] In some embodiments of this application, the number and direction of rays can be dynamically generated based on the local structural complexity from the metal centroid to the human mask boundary. Specifically, the system traverses 360° around the centroid with a step size of 1°, and calculates the standard deviation gradient (i.e., local intensity fluctuation rate) of pixel values ​​on the path from the centroid to the mask boundary for each direction. In directions with high gradients (i.e., drastic changes in tissue structure or significant artifacts), the system automatically increases the ray density (e.g., emitting one ray every 1.5°) to form a dense, narrow sub-region. In directions with low gradients (e.g., uniform adipose tissue), only one ray is set every 5°–10° to form a wide sub-region. The shape of the sub-region can be, for example, an irregular wedge or arc-shaped strip, and the spatial sampling density is positively correlated with the artifact sensitivity.

[0069] To accommodate different image resolutions, metal sizes, and artifact complexities, sub-regions can be divided as follows: determine the number of sub-regions; divide the human mask area into multiple fan-shaped regions equal to the number of divisions, with the geometric centroid as the center, where the angles of the multiple fan-shaped regions are equal; and define the multiple fan-shaped regions as sub-regions.

[0070] In some embodiments of this application, the number of divisions can be dynamically calculated based on the minimum arc length constraint. Specifically, a first distance between the geometric centroid and the boundary of the human body mask region is determined, wherein the first distance is used to reflect the maximum scanning radius of the sub-region division; a second preset threshold is obtained, wherein the second preset threshold is used to reflect the number of pixel center points crossed on the arc path formed from the geometric centroid to the boundary of the human body mask region; and the number of divisions is determined based on the first distance and the second preset threshold.

[0071] First, calculate the metal's centroid. The maximum Euclidean distance to the boundary pixel of the human body mask (i.e., the first distance) is denoted as the maximum scan radius. Furthermore, to ensure that the sector still contains a sufficient number of pixel samples at the far boundary to maintain the stability of the statistical dispersion calculation (avoiding excessive fluctuations in standard deviation due to insufficient samples), a minimum arc length threshold is preset. (That is, the second preset threshold, preferably 30-50 pixels). Based on this, the number of sectors Determined by the following formula:

[0072]

[0073] in, This indicates the floor function.

[0074] In some embodiments of this application, in order to ensure angular resolution and avoid excessively narrow sectors, the angular resolution can also be... Limited to a preset closed interval (e.g.) This means that if the calculation result exceeds this range, it will be forcibly truncated to the boundary value.

[0075] Furthermore, with the center of mass Using the origin as the starting point, the surrounding image region is divided into... An equiangular radial sector Number of sectors The preferred value is 12.

[0076] Step S208: Determine the statistical index corresponding to each sub-region, and determine the quality assessment result of the original image data based on the statistical index. The statistical index is used to quantify the degree of dispersion of pixel distribution within the sub-region.

[0077] In step S208 above, the statistical index refers to the mathematical quantity used to quantify the dispersion of pixel grayscale distribution in each sub-region, and the quality assessment result refers to the quantitative score (such as level 1-5) or continuous value calculated based on the statistical index of all sub-regions, which is used to objectively characterize the degree of influence of metal artifacts on the diagnostic image.

[0078] In some embodiments of this application, the statistical indicator may be, for example, the standard deviation. Specifically, the sub-region is taken as a sector, and each sector is extracted. Calculate the standard deviation of the pixel values ​​within the range. The calculation formula is:

[0079]

[0080] in, The pixel value within the sector. The mean, This represents the total number of pixels within the sector.

[0081] It should be noted that, in addition to standard deviation, coefficient of variation (CV), entropy, or energy can also be used as indicators to describe the dispersion of pixel distribution within a sector.

[0082] In some embodiments of this application, the quality assessment result of the original image data can be determined based on statistical indicators in the following manner: the maximum and minimum values ​​are determined from the statistical indicators corresponding to all sub-regions respectively, wherein the minimum value is used to quantify the inherent tissue noise level of the target object; the difference between the maximum and minimum values ​​is determined as a first indicator, wherein the first indicator is used to quantify the degree of stripe fluctuation of the metal artifact of the metal implant; and the quality assessment result is determined based on the first indicator.

[0083] It should be noted that the minimum value refers to the minimum dispersion value extracted from the statistical indicators of all sub-regions. It is not global background noise, but rather the intrinsic tissue noise level in the direction least affected by metal artifacts within the local region where the target object is located. Its purpose is to establish an adaptive benchmark to eliminate inherent differences in tissue noise among different patients, different anatomical locations (such as skull vs. abdomen), and different scanning parameters. For example, the intrinsic tissue noise in the skull region may have a standard deviation of up to 100 HU due to the trabecular bone structure, while the abdominal fat region may only have 40 HU. If a fixed threshold (such as >200 HU being considered an artifact) is used, it will lead to misjudgment of the skull region. However, using the minimum value as the benchmark can eliminate the interference of tissue background.

[0084] Specifically, the difference between the maximum and minimum standard deviations of all sub-regions, i.e., the range, can be calculated to obtain the artifact directionality severity index (i.e., the first index):

[0085] In some embodiments of this application, the quality assessment result can be determined by: determining a second indicator corresponding to the metal implant, wherein the second indicator is used to quantify the metal volume of the metal implant; determining a third indicator corresponding to the metal implant, wherein the third indicator is used to quantify the extent of metal artifacts of the metal implant; and determining the quality assessment result based on the first indicator, the second indicator, and the third indicator.

[0086] The second indicator can be determined as follows: determine the area of ​​the metal region where the metal implant is located, and use the area as the second indicator. The third indicator can be determined as follows: determine the target sub-region from multiple sub-regions, wherein the target sub-region includes sub-regions whose statistical indicators meet preset conditions; determine the coverage area of ​​the target sub-region, wherein the coverage area is used to quantitatively represent the affected area of ​​the metal artifact; and determine the ratio of the coverage area to the total area of ​​the human body cross-section corresponding to the human body mask area as the third indicator.

[0087] After obtaining the first, second, and third indicators, the quality assessment results can be determined as follows: the first, second, and third indicators are weighted and summed to obtain a comprehensive score, which is used to quantify the impact of metal artifacts of metal implants on the image quality of the original image data in multiple dimensions; the comprehensive score is then mapped to a preset grading level to obtain the quality assessment results.

[0088] Figure 5 This is a schematic diagram illustrating the process of determining the quality assessment result of an image quality assessment method according to an embodiment of this application, as shown below. Figure 5 As shown, in some embodiments of this application, the quality assessment results are determined through the following steps:

[0089] Specifically, a multi-dimensional weighted model can be constructed to calculate the comprehensive severity score. This model integrates three key dimensions of metal artifacts: the area of ​​the metal region. Artifact directionality severity ( ) and the percentage of affected areas ( ).

[0090] Step S502: Obtain the area of ​​the metal region That is, the second indicator.

[0091] Step S504: Obtain the artifact directionality severity ( ), that is, the first indicator.

[0092] Step S506: Obtain the percentage of affected areas ( ), that is, the third indicator.

[0093] Step S508: Calculate the comprehensive severity score S and map it to a clinical grade of 1 to 5 based on the percentile cutoff point of the reference cohort.

[0094] The formula for calculating the overall severity score S is as follows:

[0095]

[0096] Among them, the weighting coefficients are clinically validated and selected. (Focusing on metal volume). (Focusing on artifact fringe intensity) (Focusing on the scope of impact), the final calculated score The percentile cutoff points based on the reference cohort are mapped to clinical grading levels (i.e., grading levels) from 1 to 5, where level 1 represents "mild" and level 5 represents "severe". It should be noted that the above data is for illustrative purposes only and can be adjusted according to needs.

[0097] In some embodiments of this application, the following steps may also be performed: when the grading level meets the preset conditions, metal artifact suppression is performed on the original image data; or, when the first index of any sub-region is greater than the third preset threshold, metal artifact suppression is performed on the original image data.

[0098] Specifically, the grading level can provide objective evidence for clinical practice, helping doctors determine whether the current image needs to be rescanned or whether the Metal Artifact Suppression (MAR) algorithm must be applied. In addition, the primary indicator can also serve as a front-end trigger. For example, when the range R exceeds a certain threshold, the system automatically calls the back-end advanced metal artifact suppression (MAR) algorithm for repair.

[0099] It should be noted that this method is not only applicable to CT / PET images, but also to industrial X-ray inspection, visible light surface defect inspection, and other scenarios.

[0100] Through steps S202 to S208 above, directional statistical dispersion analysis is adopted. By automatically locating the geometric centroid of the metal implant from the original medical image, and dividing the area into multiple fan-shaped sub-regions covering the human body mask area radially with the centroid as the origin, the statistical dispersion index of the pixel values ​​in each sub-region is calculated. Then, the image quality is evaluated by comprehensively considering the dispersion differences between the sub-regions. This achieves the objective quantification of the directional characteristics of metal artifacts, thereby realizing the technical effect of distinguishing radial artifacts from normal high-density tissue noise. This solves the technical problem that traditional evaluation methods process all pixels in the entire medical image uniformly, which easily leads to confusion between radial artifacts generated by metal and statistical fluctuations generated by normal high-density bone structures, resulting in inaccurate evaluation results.

[0101] Figure 3 This is a schematic diagram illustrating the principle of multi-sector radial statistical analysis of metal artifacts in an image quality assessment method according to an embodiment of this application. Figure 3 As shown in the figure, this diagram explains the core geometric logic of the metal artifact quantization algorithm, where:

[0102] (1) Central positioning: A bright solid dot in the center of the figure represents the identified high-density metal implant (such as dental filling).

[0103] (2) Sector division: Taking the center as the origin, radiating dotted lines are emitted to divide the surrounding image area into 12 sector areas (S1 to S12), with each sector having an angle of 30 degrees.

[0104] (3) Artifact characteristics: In the figure, obvious radial shadows or stripes are drawn in the horizontal sector (such as S3 and S9), representing the area affected by metal artifacts.

[0105] (4) Statistical differences: The figure caption shows that the standard deviation of pixels in sectors S3 and S9 is significantly higher than that in sectors S12 or S6 in the vertical direction. This statistical difference (range) between sectors vividly illustrates how the algorithm captures directional stripe artifacts.

[0106] Figure 4 This is a schematic diagram of sub-region division within a human body mask area according to an embodiment of the image quality evaluation method of this application. Figure 4 Multi-sector radial statistical analysis can be referenced. Figure 3 This will not be elaborated upon here.

[0107] To facilitate understanding of the above image quality evaluation method, some specific embodiments are explained below.

[0108] Taking dental implant artifact assessment as an example, it may include the following steps:

[0109] Step 1: The automatic metal source localization and centroid extraction system reads the patient's scan sequence and automatically identifies a significant high-density signal in the 45th slice image. This is achieved by performing a standardized HU transformation and setting a first threshold. The HU algorithm successfully located the cluster of metal pixels containing the dental filling material using connected component analysis (the measured average HU value is 2940).

[0110] Subsequently, the system accurately calculated the geometric centroid of the pixel cluster. The specific operation is as follows: traverse the set of all pixels within the identified high-density metal connected regions. Accumulate the x-coordinates of all pixels respectively. and ordinate and divide by the total number of pixels (i.e., the area of ​​the connected domain), to obtain the centroid coordinates:

[0111]

[0112] The calculation result is preserved to sub-pixel precision (such as floating-point numbers), and the coordinates are... It is determined as the origin of the spatial coordinates for subsequent radial sector division.

[0113] Step 2: Radial sector partitioning and spatial mapping using the centroid determined in Step 1. Centered on the human body mask area of ​​this layer of image, the system uniformly divides the human body mask area into 12 equiangular sectors. to Each sector has an angular span of 30°. This step establishes a mapping from physical space to directional statistical dimensions.

[0114] Step 3: Targeted Statistical Feature Extraction (Discreteness Calculation) The algorithm traverses 12 sectors and calculates the pixel standard deviation of each sector. The calculation results show that the horizontal fringe artifacts are the most severely affected. and The standard deviations of the sectors were as high as 420 and 435, respectively; while those perpendicular to them and with better image quality... and The standard deviation of the sector is only about 75.

[0115] Step 4: Directional Severity Quantification (Range Calculation) - The system extracts the maximum value of the standard deviation for each direction. and minimum value The range of sector dispersion was calculated. This indicator specifically quantifies the degree of striped fluctuations in artifacts, eliminating background interference from inherent human tissues.

[0116] Step 5: Comprehensive evaluation scoring and report output system extracts three normalized feature components:

[0117] Metal area ( ):based on The area of ​​the metal region is obtained by normalizing the total number of pixels in the image. ,in This is the preset maximum metal area reference value.

[0118] Artifact severity ( ): The range value obtained from step four ( (This is obtained by normalization.)

[0119] The calculation formula is:

[0120]

[0121] in This is the preset maximum range reference value based on experience.

[0122] Percentage of affected areas ( : Calculate the ratio of the area affected by artifacts (i.e. the sector coverage area with a significantly increased standard deviation) to the total area of ​​the human body cross section.

[0123] In the example, the reference base is set as Pixels HU.

[0124] Calculate the normalized components:

[0125] (Assuming the metal area is 500 pixels)

[0126]

[0127] (Assuming the affected area accounts for 15%)

[0128] Substitute into the formula Example Result: In a patient's scan sequence, the system detected a metallic artifact in layer 45, and the calculated composite score was approximately 27 (after normalization). According to the preset percentile grading standard, this score falls in the second interval, and the system automatically classifies it as "Grade 2 (Mild)".

[0129] Figure 6 This is a structural diagram of an image quality evaluation device according to an embodiment of this application, such as... Figure 6 As shown, the device includes:

[0130] The acquisition module 602 is used to acquire the original image data of the target object, wherein the original image data includes metal artifacts caused by metal implants inside the target object;

[0131] The determination module 604 is used to determine the geometric centroid corresponding to the metal implant from the original image data;

[0132] The partitioning module 606 is used to extend multiple rays in the radial direction with the geometric centroid as the center, and divide the human mask area corresponding to the target object in the original image data into multiple sub-regions, wherein each sub-region is used to cover the area from the geometric centroid to the boundary of the human mask.

[0133] Evaluation module 608 is used to determine the statistical indicators corresponding to each sub-region and to determine the quality evaluation result of the original image data based on the statistical indicators. The statistical indicators are used to quantify the degree of dispersion of pixel distribution within the sub-region.

[0134] It should be noted that, Figure 6 The image quality evaluation device shown is used to perform... Figure 2 The image quality evaluation method shown is therefore Figure 2 The relevant explanations in the image quality assessment methods also apply to... Figure 6 The image quality evaluation device shown will not be described in detail here.

[0135] This application also provides an electronic device, which includes a memory and a processor, wherein the memory is used to store program instructions; the processor is connected to the memory and is used to execute the steps of the image quality evaluation method implemented in the various embodiments of this application.

[0136] This application also provides a non-volatile storage medium including a stored computer program, wherein the device containing the non-volatile storage medium executes the steps of the image quality evaluation method in various embodiments of this application by running the computer program.

[0137] This application also provides a computer program product, including computer instructions that, when executed by a processor, implement the steps of the image quality evaluation method in various embodiments of this application.

[0138] This application also provides a computer program that, when executed by a processor, implements the steps of the image quality evaluation method in various embodiments of this application.

[0139] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0140] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0141] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0142] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0143] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0144] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0145] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for evaluating image quality, characterized in that, include: Acquire raw image data of the target object, wherein the raw image data includes metal artifacts caused by metal implants inside the target object; The geometric centroid of the metal implant is determined from the original image data; Multiple rays are extended radially from the geometric centroid to divide the human mask region corresponding to the target object in the original image data into multiple sub-regions, wherein each sub-region is used to cover the area from the geometric centroid to the boundary of the human mask. Statistical indicators are determined for each sub-region, and the quality assessment result of the original image data is determined based on the statistical indicators, wherein the statistical indicators are used to quantify the dispersion of pixel distribution within the sub-region.

2. The method according to claim 1, characterized in that, Determining the geometric centroid of the metal implant from the original image data includes: The human body mask region corresponding to the target object is determined from the original image data, wherein each pixel value within the human body mask region is used to quantify the attenuation coefficient of the human tissue of the target object to X-rays; Within the human body mask area, determine the metal region corresponding to the metal implant; The geometric centroid of the metal implant is determined based on the metal region.

3. The method according to claim 2, characterized in that, Determining the metal region corresponding to the metal implant within the human body mask area includes: Obtain the standardized image values ​​corresponding to all pixels within the human body mask area; The standardized image value is compared with a first preset threshold to obtain the target pixel, wherein the standardized image value of the target pixel is greater than or equal to the first preset threshold, and the first preset threshold is determined based on the energy and intensity of the X-ray; A connectivity analysis is performed on the target pixel, and the metal region is determined based on the connectivity analysis results.

4. The method according to claim 1, characterized in that, Multiple rays extend radially from the geometric centroid, dividing the human mask region corresponding to the target object in the original image data into multiple sub-regions, including: Determine the number of sub-regions to be divided; Using the geometric centroid as the center, the human body mask area is divided into multiple sector areas equal to the number of divisions, wherein the angles of the multiple sector areas are equal; The plurality of sector-shaped regions are defined as the sub-regions.

5. The method according to claim 4, characterized in that, Determining the number of sub-regions includes: A first distance is determined between the geometric centroid and the boundary of the human body mask region, wherein the first distance is used to reflect the maximum scanning radius of the sub-region division; Obtain a second preset threshold, wherein the second preset threshold is used to reflect the number of pixel center points crossed on the arc path formed from the geometric centroid to the boundary of the human body mask area; The number of divisions is determined based on the first distance and the second preset threshold.

6. The method according to claim 1, characterized in that, The quality assessment result of the original image data is determined based on the statistical indicators, including: The maximum and minimum values ​​are determined from the statistical indicators corresponding to all sub-regions, wherein the minimum value is used to quantify the inherent tissue noise level of the target object; The difference between the maximum value and the minimum value is determined as a first index, wherein the first index is used to quantify the degree of stripe fluctuation of the metal artifact of the metal implant; The quality assessment result is determined based on the first indicator.

7. The method according to claim 6, characterized in that, Determining the quality assessment result based on the first indicator includes: A second indicator is determined for the metal implant, wherein the second indicator is used to quantitatively represent the metal volume of the metal implant; A third index is determined for the metal implant, wherein the third index is used to quantify the extent of the metal artifacts of the metal implant; The quality assessment result is determined based on the first indicator, the second indicator, and the third indicator.

8. The method according to claim 7, characterized in that, Determining the second indicator corresponding to the metal implant includes: determining the area of ​​the metal region where the metal implant is located, and using the area as the second indicator.

9. The method according to claim 7, characterized in that, The third indicator for determining the metal implant includes: A target sub-region is determined from the plurality of sub-regions, wherein the target sub-region includes sub-regions whose statistical indicators meet preset conditions; Determine the coverage area of ​​the target sub-region, wherein the coverage area is used to quantify the affected area of ​​the metal artifact; The ratio of the coverage area to the total cross-sectional area of ​​the human body corresponding to the human body mask area is determined as the third indicator.

10. The method according to claim 7, characterized in that, Determining the quality assessment result based on the first indicator, the second indicator, and the third indicator includes: The first indicator, the second indicator, and the third indicator are weighted and summed to obtain a comprehensive score, wherein the comprehensive score is used to quantify the degree of influence of the metal artifacts of the metal implant on the image quality of the original image data in multiple dimensions; The comprehensive score is mapped to a preset grading level to obtain the quality assessment result.

11. The method according to claim 10, characterized in that, The method further includes: If the grading level meets the preset conditions, metal artifact suppression is applied to the original image data; or, If the first index of any sub-region is greater than the third preset threshold, metal artifact suppression is performed on the original image data.

12. An image quality evaluation device, characterized in that, include: An acquisition module is used to acquire the original image data of the target object, wherein the original image data includes metal artifacts caused by metal implants inside the target object; A determining module is used to determine the geometric centroid corresponding to the metal implant from the original image data; The segmentation module is used to extend multiple rays in the radial direction with the geometric centroid as the center, and divide the human body mask area corresponding to the target object in the original image data into multiple sub-regions, wherein each sub-region is used to cover the area from the geometric centroid to the boundary of the human body mask; An evaluation module is used to determine the statistical index corresponding to each sub-region and to determine the quality evaluation result of the original image data based on the statistical index, wherein the statistical index is used to quantify the dispersion of pixel distribution within the sub-region.

13. An electronic device, characterized in that, include: A memory and a processor, the memory being used to store program instructions; the processor being connected to the memory and used to execute the image quality evaluation method according to any one of claims 1 to 11.

14. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored computer program, wherein the device containing the non-volatile storage medium executes the image quality evaluation method according to any one of claims 1 to 11 by running the computer program.

15. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the image quality evaluation method according to any one of claims 1 to 11.

Images