Method of perspective image reconstruction and readable storage medium

Abstract

The application provides a perspective image reconstruction method and a readable storage medium. The perspective image reconstruction method comprises the following steps: setting a plurality of calculation nodes in series based on a perspective image reconstruction process, and the plurality of calculation nodes work in parallel; inputting a view data to a first calculation node of the plurality of calculation nodes in a time period; and outputting a perspective image by a last calculation node of the plurality of calculation nodes. In this way, each calculation node operates independently and does not interfere with each other, and each calculation node processes one view data at a time, the data amount involved in each execution is small, the scheduling is convenient, the view data can be discarded immediately after being used, and the consumption of memory and display memory is reduced. Thus, the problem that the reconstruction method in the prior art has a long delay, a low refresh frequency and cannot meet the real-time requirement of the perspective image in special occasions is solved.

Description

Technical Field

[0001] This invention relates to the field of medical imaging technology, and in particular to a method for reconstructing fluoroscopic images and a readable storage medium. Background Technology

[0002] CT fluoroscopy is a technique that continuously acquires data from the same location, transmitting the raw data to a memory to continuously reconstruct and display images in real time. CT fluoroscopy is primarily used in image-guided interventional procedures and biopsies, allowing doctors to directly visualize and monitor the entire process of the treatment probe.

[0003] The application scenarios of CT fluoroscopy dictate that the entire CT system needs to have the fastest possible processing speed, the shortest possible time from data acquisition to image display, and the ability to update and display CT images more frequently. Within the entire CT chain, CT reconstruction is the largest module, accounting for approximately 80%, and its performance determines the overall system performance.

[0004] However, existing reconstruction methods have long delays and low refresh rates, which cannot meet the real-time requirements for perspective images in special situations. Summary of the Invention

[0005] The purpose of this invention is to provide a perspective image reconstruction method and a readable storage medium to solve the problems of long delays and low refresh rates in existing reconstruction methods, which cannot meet the real-time requirements of perspective images in special situations.

[0006] To address the aforementioned technical problems, this invention provides a perspective image reconstruction method, comprising: setting up a plurality of computing nodes sequentially connected in series based on a perspective image reconstruction process, wherein the plurality of computing nodes work in parallel; inputting view data to the first of the plurality of computing nodes within a time period; each of the plurality of computing nodes performing calculations based on the input and its own logic, and outputting the calculation result; and the last of the plurality of computing nodes outputting a perspective image.

[0007] Optionally, the plurality of computing nodes include at least one of the following nodes: a reading node for acquiring the view data; an air correction node for correcting the effects of air interference in the view data; a crosstalk correction node for removing crosstalk signals caused by mutual interference between detector units in the view data; a conversion node for converting the density information of the view data into density units of the view data; a fan-beam rearrangement node for rearranging the view data stored in fan-beam format into view data stored in parallel-beam format; a parallel-beam rearrangement node for correcting the spacing between parallel beams in the view data stored in parallel-beam format, so that the parallel beams are arranged at equal intervals; a filtering node for removing data noise; a back-projection node for converting the view data into the perspective image; and a ring removal node for removing ring artifacts on the perspective image.

[0008] Optionally, the plurality of computing nodes include: the back projection node, which is used to obtain the perspective image by performing a weighted summation based on the weights of the view data.

[0009] Optionally, the weight of the view data is calculated based on the angle difference, where the angle difference is the difference between the direction of the emission source and the observation direction of the perspective image when the view data is acquired.

[0010] Optionally, the first preset range, the second preset range, and the third preset range constitute the imaging influence range, and the first preset range, the second preset range, and the third preset range are connected end to end on the number axis.

[0011] Optionally, when the angle difference is within the first preset range, the weight gradually increases from 0 to 1 as the angle difference increases; when the angle difference is within the second preset range, the weight is 1; and when the angle difference is within the third preset range, the weight gradually decreases from 1 to 0 as the angle difference increases.

[0012] Optionally, when the angle difference is not within the imaging influence range, the weight is 0.

[0013] Optionally, the angle difference is calculated based on the scanning time for acquiring one of the view data, the refresh rate of the perspective image, and the total number of view data acquired by the emission source in one rotation.

[0014] Optionally, the perspective image reconstruction method adopts a pipeline approach.

[0015] To address the aforementioned technical problems, the present invention also provides a readable storage medium storing a program, which, when executed, performs the aforementioned perspective image reconstruction method.

[0016] Compared with existing technologies, the perspective image reconstruction method and readable storage medium provided by this invention include: setting up multiple computing nodes sequentially connected in series based on the perspective image reconstruction process, wherein the multiple computing nodes work in parallel; inputting a view data to the first of the multiple computing nodes within a time period; each of the multiple computing nodes performing calculations based on the input and its own logic, and outputting the calculation result; and the last of the multiple computing nodes outputting a perspective image. With this configuration, each computing node operates independently without interfering with the others, and processes only one view data at a time. The amount of data involved in each execution is small, scheduling is convenient, and the view data can be discarded immediately after use, reducing memory and video memory consumption. This solves the problems of long latency and low refresh rate in existing reconstruction methods, which cannot meet the real-time requirements of perspective images in special situations. Attached Figure Description

[0017] Those skilled in the art will understand that the accompanying drawings are provided to better understand the invention and do not constitute any limitation on the scope of the invention. Wherein:

[0018] Figure 1 This is a schematic flowchart of a perspective image reconstruction method according to an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of the structure of a CT scanning device according to an embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram of a computing node structure according to an embodiment of the present invention;

[0021] Figure 4 This is a schematic diagram of a weighted curve image according to an embodiment of the present invention;

[0022] Figure 5 This is a schematic diagram of ViewInterval according to an embodiment of the present invention.

[0023] In the attached image:

[0024] 1-X-ray tube; 2-Object under test; 3-Detector; 4-X-ray tube trajectory;

[0025] 10-Reading node; 11-Air correction node; 12-Crosstalk correction node; 13-Conversion node; 14-Fan beam rearrangement node; 15-Parallel beam rearrangement node; 16-Filtering node; 17-Back projection node; 18-De-ringing node; 21-First preset range; 22-Second preset range; 23-Third preset range. Detailed Implementation

[0026] To make the objectives, advantages, and features of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the drawings are all in a very simplified form and are not drawn to scale, and are only used to facilitate and clarify the explanation of the embodiments of this invention. Furthermore, the structures shown in the drawings are often part of the actual structures. In particular, different figures may emphasize different aspects and may sometimes use different scales.

[0027] As used in this invention, the singular forms “a,” “an,” and “the” include plural objects; the term “or” is generally used to mean “and / or”; the term “a number” is generally used to mean “at least one”; the term “at least two” is generally used to mean “two or more”; furthermore, the terms “first,” “second,” and “third” are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, features defined with "first," "second," and "third" may explicitly or implicitly include one or at least two of those features. The term "proximal" typically refers to the end closer to the operator, and the term "distal" typically refers to the end closer to the patient. "One end" and "the other end," as well as "proximal" and "distal," generally refer to two corresponding parts, including not only endpoints. The terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can be fixed connections, detachable connections, or integral connections; they can be mechanical connections or electrical connections; they can be direct connections or indirect connections through an intermediate medium; they can be internal connections between two elements or interactions between two elements. Furthermore, as used in this invention, the placement of one element on another element generally only indicates a connection, coupling, cooperation, or transmission relationship between the two elements, and the connection, coupling, cooperation, or transmission between the two elements can be direct or indirect through an intermediate element. It should not be construed as indicating or implying a spatial positional relationship between the two elements, i.e., one element can be located arbitrarily inside, outside, above, below, or to one side of another element, unless otherwise explicitly stated. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0028] The core idea of ​​this invention is to provide a perspective image reconstruction method and a readable storage medium to solve the problems of long delays and low refresh rates in existing reconstruction methods, which cannot meet the real-time requirements of perspective images in special situations.

[0029] The following description refers to the accompanying drawings.

[0030] Please refer to Figures 1 to 4 ,in, Figure 1This is a schematic flowchart of a perspective image reconstruction method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a CT scanning device according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a computing node structure according to an embodiment of the present invention; Figure 4 This is a schematic diagram of a weighted curve image according to an embodiment of the present invention; Figure 5 This is a schematic diagram of ViewInterval according to an embodiment of the present invention.

[0031] like Figure 1 As shown, the perspective image reconstruction method provided by the present invention includes the following steps:

[0032] S10 sets up multiple computing nodes connected in series based on the perspective image reconstruction process, wherein the multiple computing nodes work in parallel.

[0033] S20 inputs view data to the first of the plurality of computing nodes within a time period.

[0034] Each of the multiple computing nodes in S30 performs calculations based on the input and its own logic, and outputs the calculation results.

[0035] And, the last output perspective image among the plurality of computing nodes described in S40.

[0036] The view data is acquired during the process of the transmitter rotating around the object under test. When the transmitter is at a preset angle or at a preset time, a view data is acquired.

[0037] In step S20, by inputting one view data at a time, the granularity of the processed data for each computing node is reduced to one view, thereby minimizing the latency of the CT fluoroscopy function. In one embodiment, the overall latency is within 70ms, which is quite effective.

[0038] The input of each of the plurality of computing nodes, except for the first one, is connected to the output of the previous one; the input of the first computing node is used to acquire the view data, and the output of the last computing node is used to output the perspective image.

[0039] The view data can be arranged according to Figure 2 To understand, in Figure 2The image illustrates a CT scanning device, which includes an X-ray tube 1 (also referred to as an emission source) for emitting penetrating X-rays and detectors 3 arranged in a fan shape. A space is provided between the X-ray tube 1 and the detectors 3 for accommodating a subject 2 (e.g., a patient). The X-ray tube 1 and the detectors 3 can rotate relatively fixedly around the subject 2. During rotation, the X-ray tube 1 moves along a trajectory 4. The data simultaneously received by multiple detectors 3 when the emission source is at a preset angle or at a preset time is referred to as the view data. The preset angle or preset time can be set according to actual needs and will not be described in detail here.

[0040] Because the multiple computing nodes perform parallel computations, at any given moment, the first node is processing the computation related to the k-th view data, the second node is processing the computation related to the (k-1)-th view data, and the m-th node is processing the computation related to the (k-m+1)-th view data. Therefore, the computing nodes do not interfere with each other, reducing computational latency and refresh rate while ensuring computational correctness. Here, k is a positive integer, and its maximum value is determined by the total amount of view data at runtime. m is a positive integer, and its maximum value represents the total number of computing nodes.

[0041] Please refer to Figure 3 The plurality of computing nodes include the following nodes: a reading node 10 for acquiring the view data; an air correction node 11 for correcting the effects of air interference in the view data; a crosstalk correction node 12 for removing crosstalk signals caused by mutual influence between detector units in the view data; a conversion node 13 for converting the density information of the view data into density units of the view data; a fan-beam rearrangement node 14 for rearranging the view data stored in fan-beam format into view data stored in parallel-beam format; a parallel-beam rearrangement node 15 for correcting the spacing between parallel beams in the view data stored in parallel-beam format, so that the parallel beams are arranged at equal intervals; a filtering node 16 for removing data noise; a back-projection node 17 for converting the view data into the perspective image; and a ring removal node 18 for removing ring artifacts on the perspective image.

[0042] In other embodiments, some of the computing nodes may be added or removed as needed; that is, the plurality of computing nodes may include at least one of the aforementioned nodes.

[0043] In one embodiment, the plurality of computing nodes includes: the back projection node, which is used to obtain the perspective image by performing a weighted summation based on the weights of the view data.

[0044] Subsequent image processing using weights enables real-time processing of the current view data, providing the possibility of discarding the view data immediately after use. This ensures the continuity of data for each perspective image, reduces latency, and also reduces computational load.

[0045] The weight of the view data is calculated based on the angle difference, which is the difference between the direction of the emission source and the viewing direction of the perspective image when the view data is acquired. It should be understood that in different embodiments, for ease of calculation, the direction of the emission source and the viewing direction of the perspective image can be set according to actual needs. Although this ultimately leads to a change in the absolute value of the angle difference, the trend of the angle difference remains unchanged. For example, if the viewing direction of the perspective image is rotated by an angle, the angle difference will also change by the same angle.

[0046] Considering that image distortion may occur when the angle difference changes to a certain range, the angle difference is divided into four ranges: a first preset range, a second preset range, a third preset range, and a range not belonging to any of the above three. When the angle difference belongs to the first preset range, the weight gradually increases from 0 to 1 as the angle difference increases. When the angle difference belongs to the second preset range, the weight is 1. When the angle difference belongs to the third preset range, the weight gradually decreases from 1 to 0 as the angle difference increases. When the angle difference does not belong to the imaging influence range, the weight is 0. The first preset range, the second preset range, and the third preset range are connected end-to-end on the number axis and belong to [0, 2π]. The imaging influence range is composed of the first preset range, the second preset range, and the third preset range. That is, the angle difference corresponding to the case where the two are closely related is set as the second preset range and given the largest weight (i.e., 1); the cases where the two are not closely related are set as the first preset range and the third preset range, and given corresponding smaller weights, which gradually decrease to 0 according to the change of the angle difference. For cases not falling within the above three categories, a weight of 0 is set. In practice, the view data corresponding to the above situations can also be directly excluded. "Connected end to end in sequence" means that there are no gaps between adjacent ranges and they do not contain each other.

[0047] In one embodiment, the first preset range 21 is [0, θ] * The weight within the first preset range 21 is sin 2 [(π / 2θ * )*θ], where θ represents the angle difference, θ * =(θ total -π / 2) / 2,θtotal θ represents the angular region comprised of all the view data required for the reconstruction of each of the aforementioned perspective images. total The value range of is (π, 2π]; the second preset range 22 is [θ]. * θ total -θ * The third preset range 23 is [θ] total -θ * θ total The weight within the third preset range 23 is cos 2 [(π / 2θ * )*(θ total The corresponding weighting curve can be found by referring to -θ). Figure 4 This needs to be understood. It should be understood that in the first preset range 21 and the third preset range 23, the weight curve can also be selected as other function forms, and the endpoints of the first preset range 21 and the third preset range 23 can also be selected as other values.

[0048] Based on the above rules, the calculation of the angle difference is merely a geometric problem from a mathematical perspective. However, considering the algorithm's response speed and understandability, a preferred solution is as follows: The angle difference is calculated based on the scanning time for acquiring one set of view data, the refresh rate of the fluoroscopic image, and the total number of view data acquired by the transmitter during one rotation. With this configuration, the angle difference can be quickly calculated when the required refresh rate of the fluoroscopic image changes in different application scenarios. Furthermore, the calculation method remains unchanged even when the hardware performance of the corresponding CT scanning device changes, demonstrating good portability.

[0049] In a preferred embodiment, the angle difference is calculated according to the following formula:

[0050]

[0051] Where, θ i,j This represents the angular difference between the i-th view data and the j-th perspective image, where i and j are both parameters representing sequence numbers, and both i and j are integers greater than or equal to 0. i represents the number of the view data, while j represents the number of the perspective image. The specific numbering method can be set differently depending on the embodiment and requirements. The maximum value of i depends on the amount of view data, and the maximum value of j depends on the amount of perspective image data. ViewPerRev represents the total number of view data collected by the emission source in one rotation. ViewInterval is calculated according to the following formula:

[0052]

[0053] Where RotationTime represents the scan time for acquiring one of the view data, and RefreshRate represents the refresh rate of the perspective image.

[0054] ViewInterval can also be followed Figure 5 To understand, in Figure 5 In the diagram, each Image represents a perspective image, and the horizontal axis represents the number of the view data. Each approximately trapezoidal curve in the diagram represents the range of all view data related to a perspective image and their respective weight values. The distance between two adjacent perspective images (measured by the difference in the numbers of the view data) is called ViewInterval. Figure 5 In this context, the subscript of Image represents the number of the perspective image, and n represents the number of the last perspective image.

[0055] Based on the above formula, the weight can be calculated according to the following formula:

[0056]

[0057] Where weight(i,j) represents the weight of the i-th view data relative to the j-th perspective image, and w(θ) is calculated according to the following formula:

[0058]

[0059] Where θ represents the angle difference, θ*=(θ total -π / 2) / 2,θ total θ represents the angular region comprised of all the view data required for the reconstruction of each of the aforementioned perspective images. total The range of values ​​for θ is (π, 2π). The formula for w(θ) refers to the same content as the related descriptions in the previous text; here, the related descriptions in the previous text are simply expressed in mathematical language.

[0060] From the above formula, we can see that j that makes weight(i,j) not equal to 0 belongs to the interval [Image start (i), Image end (i)], where the two endpoints of the interval are calculated according to the following formula:

[0061]

[0062]

[0063] Where floor is the floor function and ceil is the floor function.

[0064] The weight calculation method described above can not only be combined with multiple computing nodes to further increase the computational efficiency of the reconstruction method, but also be combined with other image reconstruction methods. Therefore, the weight calculation method can run relatively independently, which also has the beneficial effect of improving the image reconstruction method.

[0065] The efficiency of the aforementioned perspective image reconstruction method can be improved by adopting a pipeline approach.

[0066] This embodiment also provides a readable storage medium storing a program that, when executed, performs the aforementioned perspective image reconstruction method. The readable storage medium, through program execution, can also solve the problems existing in the prior art.

[0067] In summary, the perspective image reconstruction method and readable storage medium provided in this embodiment include: setting up multiple computing nodes sequentially connected in series based on the perspective image reconstruction process, wherein the multiple computing nodes work in parallel; inputting a view data to the first of the multiple computing nodes within a time period; each of the multiple computing nodes performing calculations based on the input and its own logic, and outputting the calculation result; and the last of the multiple computing nodes outputting a perspective image; wherein the view data is acquired during the rotation of the emission source around the object being measured, and a view data is acquired when the emission source is at a preset angle or at a preset time. With this configuration, each computing node operates independently without interfering with each other, and processes only a single view data. The amount of data involved in each execution is small, scheduling is convenient, and the view data can be discarded immediately after use, reducing memory and video memory consumption; thus solving the problems of long latency and low refresh rate in existing reconstruction methods, which cannot meet the real-time requirements of perspective images in special situations.

[0068] The above description is only a description of preferred embodiments of the present invention and is not intended to limit the scope of the present invention in any way. Any changes or modifications made by those skilled in the art based on the above disclosure shall fall within the protection scope of the present invention.

Claims

1. A perspective image reconstruction method, characterized in that, The method includes: Based on the perspective image reconstruction process, multiple computing nodes are set up in series, wherein the multiple computing nodes work in parallel; Within a time period, a view data is input to the first of the plurality of computing nodes; Each of the plurality of computing nodes performs calculations based on the input and its own logic, and outputs the calculation result; and, The last output perspective image among the plurality of computing nodes; The plurality of computing nodes include at least a back projection node: the back projection node is used to obtain the perspective image by weighted summation based on the weights of the view data, and to calculate the weights of the view data based on the angle difference, wherein the angle difference is the difference between the direction of the emission source and the observation direction of the perspective image when the view data is acquired.

2. The perspective image reconstruction method according to claim 1, characterized in that, The plurality of computing nodes also includes at least one of the following nodes: Read node: Used to retrieve the view data; Air correction node: Used to correct for the effects of air interference in the view data; Crosstalk correction node: Used to remove crosstalk signals in the view data caused by mutual interference between detector units; Transformation node: Used to convert the density information of the view data into density units of the view data; Fan bundle rearrangement node: used to rearrange the view data stored in fan bundle format into the view data stored in parallel bundle format; Parallel bundle rearrangement node: used to correct the spacing between parallel bundles in the view data stored in parallel bundle format, so that the parallel bundles are arranged at equal intervals; Filtering nodes: used to remove data noise; as well as, Ring removal: Used to remove ring-shaped artifacts on the perspective image.

3. The perspective image reconstruction method according to claim 1, characterized in that, The first preset range, the second preset range, and the third preset range constitute the imaging influence range, and the first preset range, the second preset range, and the third preset range are connected end to end on the number axis.

4. The perspective image reconstruction method according to claim 3, characterized in that, When the angle difference is within the first preset range, the weight gradually increases from 0 to 1 as the angle difference increases. When the angle difference is within the second preset range, the weight is 1. When the angle difference is within the third preset range, the weight gradually decreases from 1 to 0 as the angle difference increases.

5. The perspective image reconstruction method according to claim 3, characterized in that, When the angle difference is not within the imaging influence range, the weight is 0.

6. The perspective image reconstruction method according to claim 1, characterized in that, The angle difference is calculated based on the scanning time for acquiring one of the view data, the refresh rate of the perspective image, and the total number of view data acquired by the emission source in one rotation.

7. The perspective image reconstruction method according to claim 1, characterized in that, The perspective image reconstruction method adopts a pipeline approach.

8. A readable storage medium, characterized in that, The readable storage medium stores a program that, when executed, performs the perspective image reconstruction method as described in any one of claims 1 to 7.

Images