A 4D-CT image reconstruction method

Abstract

The present invention relates to the technical field of 4D-CT images, and more specifically, to a 4D-CT image reconstruction method. The acquisition of the breathing curve is obtained from the three-dimensional image body surface data of a 3D scanning device, without the need to use additional equipment to collect breathing motion information. The operation is simpler, and patients will not feel uncomfortable due to wearing additional breathing equipment, improving the patient experience and simplifying the doctor's operation process, reducing the operation error to a certain extent; moreover, for radiotherapy of the chest and abdomen, during the process of collecting breathing data, it can be collected according to the position where the patient needs radiotherapy, and the selected region of interest (ROI) is directly selected in the radiotherapy area. In this way, the obtained breathing phase information can better fit the undulating motion of the area to be radiated, the target area can be set more precisely, which helps to reduce the outer boundary of the target area, increase the radiation dose received by the target area, and reduce the toxic and side effects on normal tissues, thereby improving the accuracy of radiotherapy.

Description

Technical Field

[0001] The present invention relates to the technical field of 4D-CT images, and more particularly, to a method for reconstructing 4D-CT images. Background Art

[0002] In Computed Tomography (CT) imaging, especially during the scanning of upper abdominal organs, artifacts caused by respiratory motion are a common problem. Due to the natural breathing of the human body, abdominal organs move periodically with breathing, which causes the target organ to possibly not be on the predetermined section during CT scanning, resulting in the data collected by the detector being mixed with information from multiple sections rather than the true projection of a single cross-section. This data confusion ultimately manifests as respiratory motion artifacts on the CT image, which may interfere with the doctor's diagnosis. Especially in radiotherapy treatment planning, such artifacts may lead to incorrect positioning of the treatment target area.

[0003] To solve this problem, four-dimensional computed tomography (4D-CT) technology has emerged. 4D-CT technology not only captures information in the spatial dimension but also adds the time dimension, capable of recording and analyzing the motion changes of organs during the respiratory cycle, thereby more accurately reflecting the true position and shape of the organs. When collecting data, this technology synchronously records the patient's respiratory signals. By analyzing these signals, the data at different respiratory phases can be classified and then reconstructed separately, and finally a continuous image sequence reflecting the respiratory motion changes of the organs can be synthesized.

[0004] 4D-CT data acquisition faces two major challenges: one is how to minimize the impact of respiratory motion during scanning, and the other is how to accurately analyze and locate the respiratory phases. To address the first challenge, two strategies are commonly used in clinical practice: respiratory restriction and free breathing. Respiratory restriction requires the patient to try to keep their breathing consistent during scanning, but this is difficult to achieve for patients with poor respiratory control ability. In contrast, the free breathing strategy is more user-friendly. It allows the patient to breathe at their own pace, and then by analyzing the respiratory curve, the data at different respiratory phases can be identified and classified.

[0005] However, the key to the free breathing strategy lies in accurately identifying and locating specific respiratory phases, which is crucial for the accuracy of 4D-CT reconstruction. Currently, 4D-CT systems mainly acquire the respiratory signals of patients in the following ways: 1. Measuring the respiratory volume using a spirometer; 2. Fixing infrared markers on the patient's body surface and using an infrared camera to measure the undulating movement of the surface markers with respiration to obtain the respiratory signal; 3. Fixing an abdominal pressure belt on the patient's abdomen and measuring the pressure difference caused by respiratory movement, and using the pressure difference signal to characterize the respiratory movement. The above methods all require additional devices on the patient's body, which not only cause discomfort to the patient, but also make the operation more troublesome, increase more operation steps, and are prone to more operation errors, resulting in inaccurate 4D-CT image reconstruction. Summary of the Invention

[0006] To overcome the above defects in the prior art, the present invention provides a 4D-CT image reconstruction method that does not require additional devices to obtain respiratory signals. Instead, it directly obtains respiratory signals through a CT machine, making the operation simpler and further improving the accuracy of the data.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is:

[0008] A 4D-CT image reconstruction method includes the following steps:

[0009] S1. Use a 3D scanning device to obtain the surface data of the patient's three-dimensional image, render the three-dimensional image surface data on the display interface, obtain a three-dimensional point through the mouse using the display buffer, and select the obtained three-dimensional point as a reference point; set a cylinder to intercept the three-dimensional surface point cloud within the cylinder range, and this three-dimensional surface point cloud region is the region of interest ROI.

[0010] S2. Obtain the average value of the coordinate values of the three-dimensional points in the ROI region at the current time t in the Z-axis direction as the amplitude value a of the current respiration. t , and obtain the respiration amplitude values at different times in real time. With time as the horizontal axis and the respiration amplitude value as the vertical axis, a respiration curve is formed.

[0011] S3. Use a smoothing algorithm to smooth the respiration curve.

[0012] S4. Identify the peaks and valleys of the respiration curve to obtain the respiration phase information of the patient.

[0013] S5. Add respiration phase information to the patient's three-dimensional image surface data obtained in step S1.

[0014] S6. Group all the three-dimensional image surface data according to the respiration phase.

[0015] S7. Perform 3D reconstruction on the 3D image body surface data at each phase using the selected reconstruction algorithm to generate a 3D CT image at that phase; splice the 3D CT images at different phases in chronological order to obtain a 4D-CT sequence image.

[0016] Based on the above technical means, a 4D-CT image reconstruction method provided by the present invention obtains the breathing curve from the 3D image body surface data of a 3D scanning device, without the need to use additional devices to collect breathing motion information, the operation is simpler, and the patient will not feel uncomfortable due to wearing additional breathing devices, improving the patient experience and simplifying the doctor's operation process, reducing the operation error to a certain extent; moreover, for radiotherapy of the chest and abdomen, during the process of collecting breathing data, it can be collected according to the position where the patient needs radiotherapy, and the selected region of interest (ROI) is directly selected in the radiotherapy area, so that the obtained breathing phase information can better fit the undulating motion of the radiotherapy area, the target area can be set more accurately, which helps to reduce the outer boundary of the target area, increase the radiation dose received by the target area, and reduce the toxic and side effects on normal tissues, thereby improving the accuracy of radiotherapy.

[0017] Further, the specific steps of step S1 include:

[0018] S11. In the display interface of the scanned 3D image model, adjust the best rendering angle for selecting the breathing monitoring area and the scaling size of the 3D image model;

[0019] S12. Obtain the XY-axis coordinates a(xa, ya) of the mouse cursor, and obtain the perspective projection matrix M of the 3D image model;

[0020] S13. Traverse the 3D points B in the 3D point cloud data, project them into 2D points b(xb, yb) on the 2D image of the display interface size according to the perspective projection matrix M of the model, where the 3D points B and the projected 2D points correspond one by one, and the corresponding formula is: b = M×B; calculate the distance d between the 2D coordinate point a(xa, ya) of the cursor and each projected 2D point b(xb, yb), and form a distance set;

[0021] S14. Find the 2D point with the closest projection, and the 3D point corresponding to this 2D point is the 3D center O of the selected breathing monitoring area;

[0022] S15. Set a cylindrical 3D space area, that is, the region of interest (ROI), with the 3D center O of the breathing monitoring area as the center, a set radius R, and a height H;

[0023] S16. Traverse the 3D points in the 3D point cloud data, and determine whether a 3D point falls within a cylindrical 3D space region. If it does, determine that the 3D point is within the breathing monitoring region, and add the point to the point set of the breathing monitoring region to obtain the 3D point set of the ROI region; otherwise, if it does not, do not add it.

[0024] Further, the step S2 includes: calculating the average value of the coordinates of all 3D points within the ROI region in the Z-axis direction at the current time t as the breathing amplitude value at the current time t; obtaining the 3D point set of the ROI region at different times in real time, and calculating the average value of the coordinates in the Z-axis direction at each time to obtain the breathing amplitude values at different times; using time as the horizontal axis and the breathing amplitude value as the vertical axis to form a breathing curve.

[0025] Further, the step S3 includes, assuming y0, y1, y2, y3, y4 are the latest consecutive 5 breathing amplitude values, using the following calculation formula to obtain the smoothed 5 breathing amplitude values y'0, y1', y2', y3', y'4: y'0 = (3.0 * y0 + 2.0 * y1 + y2 - y4) / 5.0

[0026] y1' = (4.0 * y0 + 3.0 * y1 + 2 * y2 + y3) / 10.0

[0027] y2' = (y0 + y1 + y2 + y3 + y4) / 5.0

[0028] y3' = (4.0 * y4 + 3.0 * y3 + 2 * y2 + y1) / 10.0

[0029] y'4 = (3.0 * y4 + 2.0 * y3 + y2 - y0) / 5.0.

[0030] Further, in step S4, starting from obtaining the breathing curve, first identify the first peak. Only after identifying the first peak, can the first valley be identified; as the breathing curve is continuously updated, only after identifying the i-th peak, can the i-th valley be identified, and only after identifying the i-th valley, can the (i + 1)-th peak be identified, where i ≥ 0.

[0031] Further, in step S4, it includes: S41. Obtain at least 5 latest breathing amplitude values of the breathing curve in chronological order, denoted as: p1, p2, p3, p4, p5; and record the type T (peak, valley, nan) of the peak and valley points closest in time to these 5 breathing amplitude values, and identify the peak or valley according to formula (1). Formula (1) is as follows:

[0032]

[0033] Where R(t) = 1 represents that p3 is a peak, R(t) = -1 represents that p3 is a trough, and R(t) = 0 represents that p3 is neither a peak nor a trough.

[0034] Further, during the peak recognition process, the peak recognized at the current moment is used as the temporary peak p m (t1). As the respiratory curve is continuously updated, the peaks and troughs are continuously recognized in the manner of step S41; if a new peak p n (t2) is recognized first, then compare the amplitudes of p m (t1) and p n (t2), and the finally determined peak is judged according to the following formula:

[0035]

[0036] If a trough is recognized first, then p(t) = p m (t1).

[0037] Further, during the trough recognition process, the trough recognized at the current moment is used as the temporary trough v m (t1). As the respiratory curve is continuously updated, the peaks and troughs are continuously recognized in the manner of step S41. If a new trough v n (t2) is recognized first, then compare the amplitudes of v m (t1) and v n (t2), and the finally determined trough is judged according to the following formula:

[0038]

[0039] If a peak is recognized first, then v(t) = v m (t1).

[0040] Further, in step S6, the collected three-dimensional image body surface data is divided according to the respiratory phase information, and the three-dimensional image body surface data under the same phase is grouped into one set.

[0041] Further, the reconstruction algorithm includes a filtered back projection reconstruction algorithm and an iterative reconstruction algorithm.

[0042] Compared with the prior art, the beneficial effects are as follows: A 4D-CT image reconstruction method provided by the present invention does not need to use additional equipment to collect respiratory motion information, the operation is simpler, and patients will not feel uncomfortable due to wearing additional respiratory equipment, improving the patient experience and simplifying the doctor's operation process, reducing the operation error to a certain extent; moreover, it can make the obtained respiratory phase information more conform to the undulating motion of the radiotherapy area to be treated, ensuring that the target area can be set more accurately, helping to reduce the external margin of the target area, increasing the radiation dose received by the target area, reducing the toxic and side effects on normal tissues, and thus improving the accuracy of radiotherapy. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic flowchart of the method of the present invention.

[0044] Figure 2 It is a schematic diagram of the selection of respiratory monitoring points in Embodiment 2.

[0045] Figure 3 It is a schematic diagram of the effect of smoothing the respiratory curve in Embodiment 3.

[0046] Figure 4 It is a schematic diagram of the respiratory curve phase in Embodiment 4. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The present invention will be described below in conjunction with specific embodiments. Among them, the accompanying drawings are only for illustrative purposes, showing only schematic diagrams, rather than physical diagrams, and should not be construed as a limitation to the present invention; in order to better illustrate the embodiments of the present invention, some components in the accompanying drawings will be omitted, enlarged or reduced, and do not represent the size of the actual product; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the accompanying drawings may be omitted.

[0048] In the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", etc. indicating the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, the terms describing the positional relationship in the drawings are only for illustrative purposes and should not be construed as a limitation to the present invention. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to specific circumstances. In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only for descriptive purposes and should not be construed as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In addition, the meaning of "and / or" appearing throughout the text is that it includes three parallel solutions. Taking "A and / or B" as an example, it includes solution A, or solution B, or the solution where A and B are satisfied simultaneously.

[0049] Embodiment 1:

[0050] As Figure 1 shown, a 4D-CT image reconstruction method includes the following steps:

[0051] S1. Use a 3D scanning device to obtain the three-dimensional image body surface data of a patient, render the three-dimensional image body surface data on a display interface, obtain a three-dimensional point through a mouse using a display buffer, and use the selected three-dimensional point as a reference point; set a cylinder to intercept the three-dimensional body surface point cloud within the cylinder range, and this three-dimensional body surface point cloud area is the region of interest (ROI).

[0052] S2. Obtain the average value of the coordinate values of the three-dimensional points in the ROI region at the current time t in the Z-axis direction as the amplitude value a of the current breath, t and obtain the breath amplitude values at different times in real time. Use time as the horizontal axis and the breath amplitude value as the vertical axis to form a breath curve.

[0053] S3. Use a smoothing algorithm to smooth the breath curve.

[0054] S4. Identify the peaks and valleys of the breath curve to obtain the breath phase information of the patient.

[0055] S5. Add the breath phase information to the three-dimensional image body surface data of the patient obtained in step S1.

[0056] S6. Group all the three-dimensional image body surface data according to the breath phase.

[0057] S7. Perform three-dimensional reconstruction on the three-dimensional image surface data at each phase using the selected reconstruction algorithm to generate a three-dimensional CT image at that phase; splice the three-dimensional CT images at different phases in chronological order to obtain a 4D-CT sequence image.

[0058] In step S6, the collected 3D image surface data is divided according to the respiratory phase information, and the 3D image surface data at the same phase is grouped together. Based on the respiratory waveform data, the respiratory cycle is divided into multiple phases, such as the end-expiratory phase, the end-inspiratory phase, and several intermediate phases. Each phase represents a specific stage in the respiratory process.

[0059] In step S7, a suitable reconstruction algorithm may be selected based on specific scanning conditions and requirements, such as a filtered back projection reconstruction algorithm or an iterative reconstruction algorithm. These reconstruction algorithms are all existing reconstruction algorithms and are not the focus of the present invention, so they will not be described in detail in this embodiment.

[0060] The projection data at each phase is reconstructed three-dimensionally using a selected reconstruction algorithm, generating 3D CT images at that phase. These images depict the morphology of the thoracic and abdominal organs at different respiratory phases. The 3D CT images from different phases are spliced together in chronological order to generate a complete 4D-CT image sequence, which demonstrates the dynamic changes of the thoracic and abdominal organs throughout the respiratory cycle. Furthermore, the generated 4D-CT image sequence undergoes quality control to check for artifacts, noise, and other issues, and performs necessary post-processing to improve image quality.

[0061] In this embodiment, the respiratory curve is obtained by using the three-dimensional image surface data of the 3D scanning device. There is no need to use additional equipment to collect respiratory movement information, the operation is simpler, and the patient will not feel uncomfortable due to wearing additional respiratory equipment, which improves the patient experience and simplifies the doctor's operation process, reducing the operation error to a certain extent; moreover, for radiotherapy of the chest and abdomen, in the process of collecting respiratory data, it can be collected according to the position where the patient needs to be radiotreated, and the selected ROI of interest can be directly selected in the radiotherapy area. In this way, the acquired respiratory phase information can be more in line with the ups and downs of the area to be radiotreated, and the target area can be set more accurately, which helps to reduce the target area's external boundary, increase the target area's irradiation dose, and reduce the toxic and side effects on normal tissues, thereby improving the accuracy of radiotherapy.

[0062] During the operation, a 3D scanning device is used to obtain three-dimensional image body surface data, and a breathing curve is obtained to obtain breathing phase information. In the subsequent 4D-CT image reconstruction process, the 3D scanning device is also used to obtain three-dimensional image body surface data for the reconstruction of three-dimensional images. Obtaining breathing phase information and the reconstruction of three-dimensional images using data from the same device not only simplifies the operation process but also avoids systematic errors between different devices and improves the accuracy of the data.

[0063] Embodiment 2:

[0064] In this embodiment, other steps are the same as those in Embodiment 1. This embodiment provides a method for obtaining a breathing curve; specifically, it includes the following steps:

[0065] S11. In the display interface of the scanned three-dimensional image model, adjust the best rendering angle for selecting the breathing monitoring area and the scaling size of the three-dimensional image model.

[0066] S12. Obtain the XY-axis coordinates a(xa, ya) of the mouse cursor and obtain the perspective projection matrix M of the three-dimensional image model.

[0067] S13. Traverse the three-dimensional points B in the three-dimensional point cloud data and project them into two-dimensional points b(xb, yb) on the two-dimensional image of the display interface size according to the perspective projection matrix M of the model. The three-dimensional points B and the projected two-dimensional points correspond one by one, and the corresponding formula is: b = M × B; calculate the distance d between the two-dimensional coordinate point a(xa, ya) of the cursor and each projected two-dimensional point b(xb, yb) to form a distance set.

[0068] Among them, traversing the three-dimensional points B in the three-dimensional point cloud data according to the perspective projection matrix M of the model, the perspective projection matrix M refers to the product of the perspective matrix and the projection matrix in the current perspective. The coordinates a(xa, ya) obtained in step S12 and the current perspective projection matrix M are used. The purpose of step S13 is to project the three-dimensional points onto the current displayed two-dimensional image screen so that the distance relationship can be established between the projected points of the three-dimensional points and the coordinates a(xa, ya) selected by the mouse cursor in the same coordinate system. Projecting the three-dimensional point cloud onto the current display screen needs to consider the perspective of the three-dimensional point cloud model and the scaling size of the model (reflected in the view projection matrix). Projecting the three-dimensional point cloud model onto the current screen image is to traverse the three-dimensional points B in the three-dimensional point cloud and project B onto the two-dimensional point b(xb, yb) on the two-dimensional screen image of the display window according to the perspective projection matrix M in S12. The three-dimensional points B and the two-dimensional points b correspond one by one: the corresponding relationship is b = M × B; finally, calculate the distance between the two-dimensional points a and b. After the traversal is completed, a distance set between the projected points and the point a is obtained, and the distance set is used in step S14 to find the three-dimensional center point O of the selected breathing area.

[0069] S14. Find the two-dimensional point of the projection with the closest distance. The three-dimensional point corresponding to this two-dimensional point is the three-dimensional center O of the selected breathing monitoring area.

[0070] S15. Set a cylindrical three-dimensional space area, that is, the region of interest ROI, with the three-dimensional center O of the breathing monitoring area as the center, a set radius R, and a height H; as Figure 2 shown, it is the effect of selecting breathing monitoring points, and the red area is the selected small ROI.

[0071] S16. Traverse the three-dimensional points in the three-dimensional point cloud data, and determine whether the three-dimensional point falls within the cylindrical three-dimensional space area. If it is, determine that the three-dimensional point is within the breathing monitoring area, and add this point to the point set of the breathing monitoring area to obtain the three-dimensional point set of the ROI area; otherwise, if not, do not add it.

[0072] To determine whether a three-dimensional point is within the breathing monitoring area, it can be discriminated according to the following formula: Assume that the three-dimensional point coordinates are (xb, yb, zb), the three-dimensional center of the cylinder is O(xo, yo, zo), and the three-dimensional cylindrical area is expressed as:

[0073] (x - xo)×(x - xo)+(y - yo)×(y - yo)<=R×R, and |z - zo|<=H.

[0074] S17. Calculate the average value of the coordinates of all three-dimensional points in the ROI area in the Z-axis direction at the current time t as the breathing amplitude value at the current time t; obtain the three-dimensional point sets of the ROI area at different times in real time, and calculate the average value of the coordinates in the Z-axis direction at each time to obtain the breathing amplitude values at different times; use time as the horizontal axis and the breathing amplitude value as the vertical axis to form a breathing curve.

[0075] After obtaining the three-dimensional point set in the ROI area (n three-dimensional points), calculate the average value of the sum of the Z coordinates of all three-dimensional points, average_z=(z1 + z2+...+zn) / n, and the value of Average_z is the current breathing amplitude value.

[0076] Embodiment 3:

[0077] In this embodiment, other steps are the same as those in Embodiment 1. This embodiment provides a method for smoothing the breathing curve. Specifically, assume that y0, y1, y2, y3, y4 are the latest 5 consecutive breathing amplitude values, and use the following calculation formula to obtain the 5 smoothed breathing amplitude values y'0, y1', y2', y3', y'4:

[0078] y'0=(3.0*y0 + 2.0*y1 + y2 - y4) / 5.0

[0079] y1' = (4.0 * y0 + 3.0 * y1 + 2 * y2 + y3) / 10.0

[0080] y2' = (y0 + y1 + y2 + y3 + y4) / 5.0

[0081] y3' = (4.0 * y4 + 3.0 * y3 + 2 * y2 + y1) / 10.0

[0082] y'4 = (3.0 * y4 + 2.0 * y3 + y2 - y0) / 5.0。

[0083] As Figure 3 shown, the red line is the actual breathing curve, and the black line is the smoothed breathing curve.

[0084] Example 4:

[0085] In this example, a method for identifying the peaks and valleys of a breathing curve is provided. Starting from obtaining the breathing curve, the identification of the first peak is carried out first. Only after the first peak is identified, the identification of the first valley is carried out; as the breathing curve is continuously updated, only after the i-th peak is identified, the identification of the i-th valley is carried out, and only after the i-th valley is identified, the identification of the (i + 1)-th peak is carried out, where i ≥ 0; the peaks and valleys of the continuously updated breathing curve are identified according to such rules. The breathing curve of a real patient is not very smooth, and there is curve jitter. The currently identified peak or valley is used as a temporary peak or valley. If a higher peak or a lower valley appears in the future in the same cycle, the currently identified temporary peak or valley is updated.

[0086] Specifically, it includes the following steps,

[0087] S41. Obtain at least 5 latest breathing amplitude values of the breathing curve in chronological order, and record them as: p1, p2, p3, p4, p5; and record the types T (peak, valley, nan) of the identified peaks and valley points that are nearest in time to these 5 breathing amplitude values. The identification of peaks or valleys is carried out according to formula (1), and formula (1) is as follows:

[0088]

[0089] In the formula, R(t) = 1 represents that p3 is a peak, R(t) = -1 represents that p3 is a valley, and R(t) = 0 represents that p3 is neither a peak nor a valley.

[0090] It should be noted that during the identification process, the first valley needs to be identified after the first peak is identified; the identification of the first peak and the first valley is carried out according to step S41, but the identification of the first peak and the first valley does not require the comparison of the nearest neighbor types.

[0091] S42. During the peak recognition process, the peak recognized at the current moment is used as the temporary peak p m (t1). As the respiratory curve is continuously updated, continue to recognize peaks and valleys in the manner of step S41; if a new peak p n (t2) is recognized first, then compare the amplitudes of p m (t1) and p n (t2), and the finally determined peak is judged according to the following formula:

[0092]

[0093] If a valley is recognized first, then p(t) = p m (t1).

[0094] S43. During the valley recognition process, the valley recognized at the current moment is used as the temporary valley v m (t1). As the respiratory curve is continuously updated, continue to recognize peaks and valleys in the manner of step S41. If a new valley v n (t2) is recognized first, then compare the amplitudes of v m (t1) and v n (t2), and the finally determined valley is judged according to the following formula:

[0095]

[0096] If a peak is recognized first, then v(t) = v m (t1).

[0097] When a peak point or a valley point is recognized in real time, the software immediately sends a signal through the communication interface to mark and classify the CT image sequence of this scanning phase for subsequent 4D-CT reconstruction.

[0098] This embodiment can recognize the peak points and valley points of the respiratory curve in real time from the complex mode of free breathing, so as to extract clear respiratory phase markers. Through the peak-valley recognition method, the start and end of breathing can be recognized, and key phases such as the end of exhalation (valley) and the end of inhalation (peak) can also be accurately judged, so as to classify the data corresponding to these phases for subsequent precise reconstruction.

[0099] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0100] Obviously, the above embodiments of the present invention are only examples given for clearly illustrating the present invention, rather than limitations on the implementation manners of the present invention. For those of ordinary skill in the art, other different forms of changes or alterations can be made based on the above description. It is not necessary and impossible to list all implementation manners here. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the claims of the present invention.

Claims

1. A 4D-CT image reconstruction method, characterized in that: The following steps are involved: S1. Use a 3D scanning device to obtain three-dimensional image surface data of the patient, render the three-dimensional image surface data on a display interface, use a mouse to use the display buffer to obtain a three-dimensional point, and use the selected three-dimensional point as a reference point; set a cylinder to intercept the three-dimensional body surface point cloud within the cylinder, and the three-dimensional body surface point cloud area within the cylinder is the ROI; S2. Get the average value of the coordinates of the three-dimensional points in the ROI area in the Z-axis direction at the current time t as the current breathing amplitude value a t , obtain the respiratory amplitude value at different times in real time, with time as the horizontal axis and the respiratory amplitude value as the vertical axis to form a respiratory curve; S3. Smoothing the respiratory curve using a smoothing algorithm; S4. Identify the peaks and troughs of the respiratory curve to obtain the patient's respiratory phase information; S5. Add respiratory phase information to the patient's three-dimensional image surface data obtained in step S1; S6. Grouping all three-dimensional image surface data according to respiratory phase; S7. Perform three-dimensional reconstruction on the three-dimensional image surface data at each phase using the selected reconstruction algorithm to generate a three-dimensional CT image at that phase; splice the three-dimensional CT images at different phases in chronological order to obtain a 4D-CT sequence image.

2. The 4D-CT image reconstruction method according to claim 1, characterized in that: The step S1 specifically includes: S11. In the scanned 3D image model display interface, adjust the optimal rendering angle and scale size of the selected respiratory monitoring area; S12. Get the XY coordinates a(xa, ya) of the mouse cursor and the perspective projection matrix M of the three-dimensional image model; S13. Projecting a 3D point B in the 3D point cloud data into a 2D point b(xb, yb) on a 2D image of the display interface size according to the view projection matrix M of the model, wherein the 3D point B corresponds one-to-one with the projected 2D point, and the corresponding formula is: b=M×B; calculating the distance d between the 2D coordinate point a(xa, ya) of the cursor and each projected 2D point b(xb, yb), and forming a distance set; S14. Find the closest projected 2D point. The 3D point corresponding to this 2D point is the 3D center O of the selected respiratory monitoring area. S15. A cylindrical three-dimensional region of interest (ROI) is set with the three-dimensional center O of the respiratory monitoring area as the center, the radius R, and the height H. S16. Traverse the three-dimensional points in the three-dimensional point cloud data to determine whether the three-dimensional point falls within the cylindrical three-dimensional space area. If so, the three-dimensional point is determined to be within the respiratory monitoring area, and the point is added to the respiratory monitoring area point set to obtain the ROI area three-dimensional point set; otherwise, if it does not fall within the cylindrical three-dimensional space area, it is not added.

3. The 4D-CT image reconstruction method according to claim 2, characterized in that: The step S2 includes: calculating the average value of the coordinates of all three-dimensional points in the ROI area at the current time t in the Z-axis direction as the breathing amplitude value at the current time t; acquiring the three-dimensional point set of the ROI area at different times in real time, and calculating the average value of the coordinates in the Z-axis direction at each time to obtain the breathing amplitude value at different times; forming a breathing curve with time as the horizontal axis and the breathing amplitude value as the vertical axis.

4. The 4D-CT image reconstruction method according to claim 3, characterized in that: The step S3 includes, assuming that y0, y1, y2, y3, and y4 are the latest five consecutive respiratory amplitude values, obtaining the five smoothed respiratory amplitude values y'0, y1', y2', y3', and y'4 using the following calculation formula: y'0=(3.0*y0+2.0*y1+y2-y4) / 5.0 y1'=(4.0*y0+3.0*y1+2*y2+y3) / 10.0 y2'=(y0+y1+y2+y3+y4) / 5.0 y3'=(4.0*y4+3.0*y3+2*y2+y1) / 10.0 y'4=(3.0*y4+2.0*y3+y2-y0) / 5.

0.

5. The 4D-CT image reconstruction method according to claim 1, characterized in that: In step S4, starting from obtaining the breathing curve, the first peak is first identified. Only after the first peak is identified, the first trough is identified. As the breathing curve is continuously updated, only after the i-th peak is identified, the i-th trough is identified. Only after the i-th trough is identified, the i+1-th peak is identified, where i≥0.

6. The 4D-CT image reconstruction method according to claim 5, characterized in that: In step S4, it includes: S41. Obtain at least five latest respiratory amplitude values of the respiratory curve in chronological order, and record them as: p1, p2, p3, p4, p5; and record the type T (peak, valley, nan) of the identified peak and valley points and the nearest neighbors of these five respiratory amplitude values in time, and identify the peak or valley according to formula (1). Formula (1) is as follows: Wherein, R(t)=1 represents that p3 is a peak, R(t)=-1 represents that p3 is a trough, and R(t)=0 represents that p3 is neither a peak nor a trough.

7. The 4D-CT image reconstruction method according to claim 6, characterized in that: In the peak recognition process, the peak recognized at the current moment is used as the temporary peak p m (t1), as the respiratory curve is continuously updated, the peaks and troughs are continuously identified in the manner of step S41; if the new peak p n (t2) is recognized first, then compare p m (t1) and p n The amplitude of (t2) is determined by the following formula: If the trough is identified first, then p(t)=p m (t1).

8. The 4D-CT image reconstruction method according to claim 6, characterized in that: In the process of identifying the trough, the trough identified at the current moment is used as the temporary trough v m (t1), as the breathing curve is continuously updated, the peaks and troughs are continuously identified in the manner of step S41. If the new trough v n (t2) is recognized first, then compare v m (t1) and v n The amplitude of (t2) is determined, and the final trough is determined according to the following formula: If the peak is identified first, then v(t) = v m (t1).

9. The 4D-CT image reconstruction method according to any one of claims 1 to 8, characterized in that: In step S6, the collected three-dimensional image body surface data is divided according to the respiratory phase information, and the three-dimensional image body surface data in the same phase are grouped together.

10. The 4D-CT image reconstruction method according to claim 9, characterized in that: The reconstruction algorithm includes a filtered back-projection reconstruction algorithm and an iterative reconstruction algorithm.

Images