An intelligent processing method and system for barium meal radiography images of gastroenterology patients

Abstract

The present invention relates to the field of image analysis, and more specifically to a method and system for intelligently processing barium meal angiography images of patients in the Department of Gastroenterology. The method first obtains each frame of a scanned image of the digestive tract of a patient taking a barium sulfate suspension, as well as a respiratory signal from the patient's chest area. The respiratory signal is periodically segmented to obtain the exhalation and inhalation periods of each respiratory cycle, and characteristic pixel points on the digestive tract contour line in each frame of the scanned image in the target respiratory cycle are extracted. The respiratory influence of the target respiratory cycle is then obtained. The position of the pixel points of each frame of the scanned image is adjusted based on the changes in respiratory amplitude data in each sub-period, the respiratory influence of the target respiratory cycle, and the sequence number of each frame of the scanned image in each sub-period. All the adjusted scanned images obtained are then integrated into an adjusted scan video. The present invention can reduce the movement of the digestive tract region in the scanned video caused by respiration, facilitating the observation of the digestive tract structure.

Description

Technical Field

[0001] The present invention relates to the field of image analysis, and in particular to a method and system for intelligently processing barium meal radiography images of patients in a gastroenterology department. Background Art

[0002] Gastrointestinal barium meal contrast examination technology has the advantages of being safe, non-invasive, with few side effects and low cost. It can effectively evaluate the morphology and function of the digestive tract (mainly the esophagus, stomach, duodenum, etc.). The principle of gastrointestinal barium meal contrast examination is mainly to use barium sulfate as a contrast agent. The patient orally takes barium sulfate suspension, which is attached to the surface of the digestive tract mucosa, outlining the contour of the digestive tract under X-rays, helping to observe the digestive tract structures such as the esophagus, stomach, and duodenum.

[0003] In related technologies, digital X-ray equipment such as gastrointestinal machines are usually used to collect scanning videos of the digestive tract of patients taking barium sulfate suspension, thereby assisting relevant personnel in observing the digestive tract structure. However, under the action of the patient's breathing, the patient's digestive tract will move, resulting in large-scale movement characteristics in the digestive tract area in the scanning video, affecting the observation of the digestive tract structure. Summary of the Invention

[0004] In order to solve the technical problem that the patient's digestive tract moves under the action of breathing, resulting in large-scale motion characteristics of the digestive tract area in the scan video, which affects the observation of the digestive tract structure, the purpose of the present invention is to provide a method and system for intelligent processing of barium meal angiography images of patients in the gastroenterology department. The technical solutions adopted are as follows:

[0005] The present invention proposes an intelligent processing method for barium meal radiography images of gastroenterology patients, the method comprising:

[0006] Acquire a scanned image of each frame of the digestive tract of a patient who has taken barium sulfate suspension within a preset time period and a respiratory signal of the patient's chest cavity, wherein the respiratory signal includes respiratory amplitude data at different moments;

[0007] The respiratory signal is periodically segmented to obtain the respiratory cycle of the respiratory signal and the exhalation period and the inhalation period of each respiratory cycle; any respiratory cycle is used as a target respiratory cycle, and the contour line of the digestive tract in each frame scanned image of the target respiratory cycle is extracted, and multiple characteristic pixel points of the contour line are obtained based on the position distribution of each pixel point on the contour line of each frame scanned image of the target respiratory cycle; the respiratory influence degree of the target respiratory cycle is obtained based on the position difference of the characteristic pixel points of the contour line of each frame scanned image between the inhalation period and the exhalation period of the target respiratory cycle;

[0008] Taking an exhalation period or an inhalation period of a target respiratory cycle as a target period, the target period is divided into two subperiods, and according to a change in respiratory amplitude data at each moment of each subperiod, the respiratory influence of the target respiratory cycle, and a sequence number of each frame of the scanned image in each subperiod, the position of a pixel point of each frame of the scanned image is adjusted to obtain an adjusted scanned image of each frame in each subperiod;

[0009] The adjustment scan images of all frames are integrated into an adjustment scan video.

[0010] Furthermore, the respiratory cycle of obtaining the respiratory signal and the exhalation period and the inhalation period of each respiratory cycle include:

[0011] Acquire extreme value points of the respiratory signal, wherein the extreme value points include maximum value points and minimum value points;

[0012] The time period between two adjacent minimum points is regarded as a breathing cycle, wherein each breathing cycle contains a maximum point;

[0013] The time period between the previous minimum point and the maximum point of each respiratory cycle is taken as the inhalation period of each respiratory cycle, and the time period between the maximum point and the next minimum point of each respiratory cycle is taken as the exhalation period of each respiratory cycle.

[0014] Furthermore, the step of obtaining the plurality of characteristic pixel points of the contour line includes:

[0015] Taking any frame of scanned image in the target respiratory cycle as the target scanned image;

[0016] On the contour line of the target scanned image, any pixel point is taken as the target pixel point, and a first local mutation degree of the target pixel point with respect to the abscissa is obtained based on the difference in abscissa between the target pixel point and other pixels excluding the target pixel point in a preset window centered on the target pixel point, wherein the pixel points in the preset window are on the contour line;

[0017] Based on the calculation method of the first local mutation degree of the target pixel point with respect to the horizontal coordinate, the second local mutation degree of the target pixel point with respect to the vertical coordinate is obtained according to the difference in the vertical coordinates between the target pixel point and other pixels other than the target pixel point in a preset window centered at the target pixel point;

[0018] The average of the first local mutation degree and the second local mutation degree is used as the comprehensive mutation degree of the target pixel;

[0019] On the contour line of the target scanned image, pixel points whose comprehensive mutation degree is greater than a preset mutation threshold are taken as feature pixel points.

[0020] Furthermore, obtaining a first local mutation degree of the target pixel point with respect to the horizontal coordinate includes:

[0021] In a preset window of a target pixel, a pixel on one side of the target pixel is used as a first reference pixel, and a pixel on the other side of the target pixel is used as a second reference pixel;

[0022] Taking the difference between the horizontal coordinates of the target pixel and each first reference pixel as the first horizontal coordinate difference value between the target pixel and each first reference pixel, and taking the average of the first horizontal coordinate difference values ​​between the target pixel and all first reference pixels as the first comprehensive difference value of the target pixel with respect to the horizontal coordinate;

[0023] Taking the difference between the horizontal coordinates of the target pixel and each second reference pixel as the second horizontal coordinate difference value between the target pixel and each second reference pixel, and taking the average of the second horizontal coordinate difference values ​​between the target pixel and all second reference pixels as the second comprehensive difference value of the target pixel with respect to the horizontal coordinate;

[0024] A normalization process is performed on the product of the first integrated difference value and the second integrated difference value to obtain a first local mutation degree of the target pixel point with respect to the horizontal coordinate.

[0025] Furthermore, obtaining the respiratory influence of the target respiratory cycle includes:

[0026] The last scanned image frame of the inhalation period of the target respiratory cycle is used as the inhalation termination image of the target respiratory cycle, and the last scanned image frame of the exhalation period of the target respiratory cycle is used as the exhalation termination image of the target respiratory cycle;

[0027] Acquire a first reference point of the inhalation termination image of the target respiratory cycle, wherein the abscissa of the first reference point is equal to the average abscissa of all the characteristic pixel points of the inhalation termination image of the target respiratory cycle, and the ordinate of the first reference point is equal to the average ordinate of all the characteristic pixel points of the inhalation termination image of the target respiratory cycle;

[0028] Acquire a second reference point of the expiration end image of the target respiratory cycle, wherein the abscissa of the second reference point is equal to the average abscissa of all the characteristic pixel points of the expiration end image of the target respiratory cycle, and the ordinate of the second reference point is equal to the average ordinate of all the characteristic pixel points of the expiration end image of the target respiratory cycle;

[0029] The Euclidean distance between the first reference point and the second reference point is used as the respiratory influence degree of the target respiratory cycle.

[0030] Furthermore, obtaining the adjusted scan image of each frame in each sub-period includes:

[0031] Taking any sub-period in the target period as the sub-period to be measured, normalizing the range of the respiratory amplitude data at all moments in the sub-period to be measured to obtain the amplitude variation degree of the sub-period to be measured;

[0032] Obtaining a pixel movement amount of each frame of the scanned image in the sub-period to be measured according to the amplitude variation degree of the sub-period to be measured, the respiratory influence degree of the target respiratory cycle, and the sequence number of each frame of the scanned image in the sub-period to be measured;

[0033] If the sub-period to be measured is the previous sub-period of the inhalation period or the next sub-period of the exhalation period, all pixels within the contour line of each frame of the scanned image of the sub-period to be measured are shifted downward to obtain an adjusted scanned image of each frame of the sub-period to be measured, wherein the shifted distance is the pixel shift amount of each frame of the scanned image of the sub-period to be measured;

[0034] If the sub-period to be measured is the subsequent sub-period of the inhalation period or the previous sub-period of the exhalation period, all pixel points within the contour line of each frame of the scanning image of the sub-period to be measured are moved upward to obtain an adjusted scanning image of each frame of the sub-period to be measured, wherein the moving distance is the pixel movement amount of each frame of the scanning image of the sub-period to be measured.

[0035] Furthermore, obtaining the pixel movement amount of each frame of the scanned image in the sub-period to be measured includes:

[0036] If the sub-period to be measured is the previous sub-period of the target period, the calculation formula for the pixel movement amount of each frame of the scanned image in the sub-period to be measured is:

[0037]

[0038] If the sub-period to be measured is the next sub-period after the target period, the calculation formula for the pixel movement amount of each frame of the scanned image in the sub-period to be measured is:

[0039]

[0040] in, Indicates the first sub-period to be tested The amount of pixel movement in a frame scan image; Indicates the degree of amplitude change of the sub-period to be measured; Indicates the number of all scanned images in the sub-period to be tested; Indicates the first sub-period to be tested The sequence number of the frame scan image; Indicates the respiratory influence of the target respiratory cycle.

[0041] Furthermore, integrating the adjusted scanned images of all frames into the adjusted scanned video includes:

[0042] The adjusted scanned images of all frames are input into the video synthesis software, and the video synthesis software is used to integrate the adjusted scanned images of all frames into the adjusted scanned video.

[0043] Furthermore, extracting the contour line of the digestive tract in each frame of the scanned image of the target respiratory cycle includes:

[0044] The scanned images containing manually annotated digestive tract contours are used as a training set, and the training set is used to train a neural network to obtain a trained neural network;

[0045] Each frame of the scanned image of the target respiratory cycle is input into the trained neural network, and the neural network outputs the contour line of the digestive tract in each frame of the scanned image.

[0046] The present invention also proposes an intelligent processing system for barium meal angiography images of patients in the gastroenterology department. The system includes a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, it implements any one of the steps of a method for intelligent processing of barium meal angiography images of patients in the gastroenterology department.

[0047] The present invention has the following beneficial effects:

[0048] The present invention takes into account that the patient's digestive tract will move under the action of the patient's breathing, resulting in a large-scale movement feature in the digestive tract area in the scanned video, affecting the observation of the digestive tract structure. First, the scanned image of each frame of the digestive tract of the patient who took barium sulfate suspension within a preset time period and the respiratory signal of the patient's chest cavity are obtained. Since the patient's breathing process has a periodic characteristic and there is an obvious difference in the change of the respiratory amplitude data during the inhalation process and the exhalation process, the respiratory signal is first periodically segmented to obtain the exhalation period and the inhalation period of each respiratory cycle. Considering that in a breathing process, the respiration action will cause the digestive tract area in the scanned video to move up and down, the present invention first identifies the characteristic pixel points of the contour line of the digestive tract of each frame of the scanned image of the target respiratory cycle. At the same time, in a breathing process, the inhalation period and The greater the difference in the positions of the characteristic pixel points of the contour lines of each frame of the scanning image between the exhalation periods, the more obvious the movement of the digestive tract area in the scanning video within the target respiratory cycle due to the respiratory effect. Therefore, the acquired respiratory influence degree can be used to reflect the degree to which the scanning video within the target respiratory cycle is affected by the respiratory effect. Since the position of the digestive tract will shift downward under the action of inspiration and upward under the action of exhalation, in order to weaken the movement of the digestive tract area in the scanning video during the target respiratory cycle, the target period is first divided into two sub-periods, and the positions of the pixel points of each frame of the scanning image in each sub-period are adjusted to obtain an adjusted scanning image, and then the adjusted scanning images of all frames are used to construct an adjusted scanning video, so that the digestive tract area in the adjusted scanning video is more stable, which is more conducive to observing the digestive tract structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 A flow chart of an intelligent processing method for barium meal radiography images of gastroenterology patients provided by one embodiment of the present invention;

[0051] Figure 2 A schematic diagram of a respiratory cycle of a respiratory signal and an exhalation period and an inhalation period of each respiratory cycle provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0052] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of a method and system for intelligently processing barium meal radiography images for gastroenterology patients. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0053] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0054] The specific scheme of the method and system for intelligent processing of barium meal angiography images of gastroenterology patients provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0055] See also Figure 1 , which shows a flow chart of a method for intelligently processing barium meal radiography images of gastroenterology patients provided by one embodiment of the present invention, the method comprising:

[0056] Step S1: obtaining scanned images of each frame of the digestive tract of a patient who has taken barium sulfate suspension within a preset time period and a respiratory signal of the patient's chest cavity, wherein the respiratory signal includes respiratory amplitude data at different moments.

[0057] In an embodiment of the present invention, a digital X-ray device such as a gastrointestinal machine is first used to collect a scanning video of the digestive tract of a patient who has taken a barium sulfate suspension within a preset time period. The scanning video is then imported into professional video software such as Premiere software. The imported scanning video is processed using the single-frame image extraction function in the software to extract a scanning image of each continuous frame of the entire scanning video. The patient's digestive tract generally refers to the esophagus, stomach, duodenum, etc. In one embodiment of the present invention, the stomach is used as the digestive tract for analysis.

[0058] At the same time, respiratory monitoring equipment such as a respiratory monitoring belt is used to collect respiratory signals from the chest cavity of patients who take barium sulfate suspension within a preset time period. The respiratory signals include respiratory amplitude data at different times, and the respiratory amplitude data at a certain moment indicates the degree of fluctuation of the patient's chest cavity at that moment.

[0059] The length of the preset time period is usually 15-30 minutes. In one embodiment of the present invention, the length of the preset time period is set to 20 minutes. The specific length of the preset time period can also be set by the implementer according to the specific implementation scenario and is not limited here.

[0060] Step S2: Periodically segment the respiratory signal to obtain the respiratory cycle of the respiratory signal and the exhalation period and inhalation period of each respiratory cycle; take any respiratory cycle as the target respiratory cycle, and extract the contour line of the digestive tract in each frame scan image of the target respiratory cycle, and obtain multiple characteristic pixel points of the contour line according to the position distribution of each pixel point on the contour line of each frame scan image of the target respiratory cycle; obtain the respiratory influence of the target respiratory cycle according to the position difference of the characteristic pixel points of the contour line of each frame scan image between the inhalation period and the exhalation period of the target respiratory cycle.

[0061] Since the patient's breathing process has periodic characteristics, and there are obvious differences in the changes in the respiratory amplitude data during the inhalation process and the exhalation process, during an inhalation process, the respiratory amplitude data of the respiratory signal gradually increases. When the inhalation reaches the limit, the respiratory amplitude data reaches the maximum. Then, during an exhalation process, the respiratory amplitude data of the respiratory signal gradually decreases. When the exhalation reaches the limit, the respiratory amplitude data reaches the minimum. Therefore, the embodiment of the present invention first periodically segments the respiratory signal to obtain the respiratory cycle of the respiratory signal and the exhalation period and inhalation period of each respiratory cycle. Subsequently, the scanned image and respiratory amplitude data of each respiratory cycle can be analyzed.

[0062] Preferably, in one embodiment of the present invention, the method for obtaining the respiratory cycle of the respiratory signal and the exhalation period and the inhalation period of each respiratory cycle specifically includes:

[0063] The existing Newton method is used to obtain the extreme points of the respiratory signal. The extreme points include maximum points and minimum points. In other embodiments of the present invention, other extreme point identification methods can also be used to obtain the extreme points of the respiratory signal. This is not limited here, and the time period between two adjacent minimum points is regarded as a respiratory cycle, where each respiratory cycle contains a maximum point.

[0064] Then the time period between the previous minimum point and the maximum point of each respiratory cycle is taken as the inspiration period of each respiratory cycle, and the time period between the maximum point and the next minimum point of each respiratory cycle is taken as the expiration period of each respiratory cycle. Figure 2 , which shows a schematic diagram of the respiratory cycle of a respiratory signal and the exhalation period and inhalation period of each respiratory cycle provided by an embodiment of the present invention, wherein time period L represents a respiratory cycle of the respiratory signal, time period L1 represents the inhalation period of the respiratory cycle, and time period L2 represents the exhalation period of the respiratory cycle.

[0065] Then, any respiratory cycle is analyzed and any respiratory cycle is used as the target respiratory cycle. Under the action of breathing, the patient's digestive tract organs will undergo positional displacement in the body, thereby presenting the movement of the digestive tract area in the scanning video, resulting in changes in the position of the digestive tract area in the scanning image of each frame in the target respiratory cycle. Therefore, the embodiment of the present invention first extracts the contour line of the digestive tract in each frame of the scanning image of the target respiratory cycle, which facilitates the subsequent extraction of characteristic pixel points on the contour line, thereby accurately analyzing the degree of change in the position of the digestive tract caused by the breathing action of the target respiratory cycle.

[0066] Preferably, in one embodiment of the present invention, the method for acquiring the contour line of the digestive tract in each frame of the scanned image of the target respiratory cycle specifically includes:

[0067] Scanned images containing manually annotated digestive tract contours are used as training sets, and the training sets are used to train a neural network to obtain a trained neural network. Among them, the basic architecture of the neural network adopts the CNN architecture, and the loss function used in training adopts the cross entropy loss function.

[0068] Each frame of the scanned image of the target respiratory cycle is then input into the trained neural network, and the neural network outputs the contour line of the digestive tract in each frame of the scanned image. It should be noted that the contour line of the digestive tract in the embodiment of the present invention is a closed curve, and the area contained in the contour line is the digestive tract area.

[0069] Since the contour of the digestive tract is generally an irregular curve shape, the contour line of each frame of the scanned image of the target respiratory cycle contains a large number of pixels with obvious position characteristics, such as inflection points and corner points. Therefore, an embodiment of the present invention analyzes the position distribution of each pixel point on the contour line of each frame of the scanned image of the target respiratory cycle, thereby extracting the characteristic pixel points on the contour line of each frame of the scanned image. Subsequently, based on the position changes of the characteristic pixel points of the contour line of each frame of the scanned image between the inhalation period and the exhalation period of the target respiratory cycle, the degree of position change of the digestive tract caused by the respiratory action of the target respiratory cycle can be analyzed.

[0070] Preferably, in one embodiment of the present invention, the method for acquiring multiple characteristic pixel points of the contour line of each frame of the scanned image of the target respiratory cycle specifically includes:

[0071] First, any frame of the scanned image of the target respiratory cycle is used as the target scanned image, and any pixel point on the contour line of the target scanned image is used as the target pixel point. According to the difference in the horizontal coordinates between the target pixel point and other pixel points except the target pixel point in the preset window centered on the target pixel point, the first local mutation degree of the target pixel point with respect to the horizontal coordinate is obtained. Among them, the pixel points in the preset window are on the contour line. The greater the first local mutation degree is, the more obvious the direction change of the horizontal coordinates of the local pixel points of the target pixel point on the contour line is, and further the more likely the target pixel point is to be a feature pixel point such as an inflection point or a corner point. In one embodiment of the present invention, the length of the preset window is set to 11, that is, the preset window includes the 10 pixel points on the contour line closest to the target pixel point and the target pixel point itself. The specific length of the preset window can also be set by the implementer according to the specific real-time scenario and is not limited here.

[0072] Preferably, in one embodiment of the present invention, the method for obtaining the first local mutation degree of the target pixel point with respect to the horizontal coordinate specifically includes:

[0073] In a preset window of a target pixel, a pixel located on one side of the target pixel is used as a first reference pixel, and a pixel located on the other side of the target pixel is used as a second reference pixel.

[0074] The difference in the horizontal coordinate between the target pixel point and each first reference pixel point is used as the first horizontal coordinate difference value between the target pixel point and each first reference pixel point, and the average of the first horizontal coordinate difference values ​​between the target pixel point and all first reference pixel points is used as the first comprehensive difference value of the target pixel point with respect to the horizontal coordinate.

[0075] The difference in the horizontal coordinates between the target pixel point and each second reference pixel point is used as the second horizontal coordinate difference value between the target pixel point and each second reference pixel point, and the average of the second horizontal coordinate difference values ​​between the target pixel point and all second reference pixel points is used as the second comprehensive difference value of the target pixel point with respect to the horizontal coordinate.

[0076] When both the first comprehensive difference value and the second comprehensive difference value are positive and the larger they are, or when both the first comprehensive difference value and the second comprehensive difference value are negative and the smaller they are, it means that the horizontal coordinates of the local pixel points of the target pixel point have undergone obvious direction changes, and the target pixel point is more likely to be a feature pixel point. Therefore, the product value of the first comprehensive difference value and the second comprehensive difference value can be normalized to obtain the first local mutation degree of the target pixel point with respect to the horizontal coordinate.

[0077] In one embodiment of the present invention, the normalization processing can be specifically, for example, maximum and minimum value normalization processing, and the normalization in subsequent steps can all adopt maximum and minimum value normalization processing. In other embodiments of the present invention, other normalization methods can be selected according to the specific range of numerical values, which will not be repeated here.

[0078] As an example, in one embodiment of the present invention, the expression of the first local mutation degree of the target pixel point with respect to the horizontal coordinate can be specifically, for example, as follows:

[0079]

[0080]

[0081]

[0082] in, Indicates the first local mutation degree of the target pixel with respect to the horizontal coordinate; Represents the first comprehensive difference value of the target pixel with respect to the horizontal coordinate; Represents the second comprehensive difference value of the target pixel with respect to the horizontal coordinate; Indicates the horizontal coordinate of the target pixel; Indicates the target pixel The horizontal coordinate of the first reference pixel; Indicates the target pixel and the a first horizontal coordinate difference value between first reference pixel points; Indicates the target pixel The horizontal coordinate of a second reference pixel; Indicates the target pixel and the a second horizontal coordinate difference value between the second reference pixel points; represents the number of first reference pixels or the number of second reference pixels, where the number of first reference pixels is equal to the number of second reference pixels; Represents the normalization function.

[0083] The analysis logic is the same as the above. The vertical coordinate is analyzed. Based on the calculation method of the first local mutation degree of the target pixel point with respect to the horizontal coordinate, the second local mutation degree of the target pixel point with respect to the vertical coordinate is obtained according to the difference in the vertical coordinates between the target pixel point and other pixels except the target pixel point in the preset window centered on the target pixel point. The greater the second local mutation degree, the more obvious the change in the direction of the vertical coordinates of the local pixel points of the target pixel point on the contour line, and the more likely the target pixel point is to be a feature pixel point.

[0084] As an example, in one embodiment of the present invention, the expression of the second local mutation degree of the target pixel point with respect to the vertical coordinate can be specifically, for example, as follows:

[0085]

[0086]

[0087]

[0088] in, Indicates the second local mutation degree of the target pixel with respect to the vertical coordinate; Represents the first comprehensive difference value of the target pixel with respect to the vertical coordinate; Represents the second comprehensive difference value of the target pixel with respect to the vertical coordinate; Indicates the vertical coordinate of the target pixel; Indicates the target pixel The vertical coordinate of the first reference pixel; Indicates the target pixel and the a first vertical coordinate difference value between first reference pixel points; Indicates the target pixel The vertical coordinate of the second reference pixel; Indicates the target pixel and the a second vertical coordinate difference value between the second reference pixel points; represents the number of first reference pixels or the number of second reference pixels, where the number of first reference pixels is equal to the number of second reference pixels; Represents the normalization function.

[0089] Then, the average value of the first local mutation degree and the second local mutation degree is used as the comprehensive mutation degree of the target pixel point. The larger the comprehensive mutation degree, the more likely the target pixel point is a feature pixel point. The same method as above can be used to obtain the comprehensive mutation degree of each pixel point on the contour line of the target scanned image. Then, on the contour line of the target scanned image, the pixel point with a comprehensive mutation degree greater than the preset mutation threshold is used as the feature pixel point. Then, the same method as above can be used to obtain the feature pixel point on the contour line of each frame scanned image of the target respiratory cycle, wherein the preset mutation threshold is set to 0.6. The specific value of the preset mutation threshold can also be set by the implementer according to the specific real-time scenario and is not limited here.

[0090] In the target respiratory cycle, the greater the difference in the positions of the characteristic pixel points of the contour lines of each frame of the scanned image between the inhalation period and the exhalation period, the more obvious the movement of the digestive tract region in the scanned video within the target respiratory cycle due to respiration, and the greater the degree to which the scanned video within the target respiratory cycle is affected by respiration. Therefore, the position difference in the characteristic pixel points of the contour lines of each frame of the scanned image between the inhalation period and the exhalation period of the target respiratory cycle can be analyzed, and the degree to which the digestive tract region in the scanned video within the target respiratory cycle is affected by respiration can be reflected by the obtained respiratory influence degree. Subsequently, the pixel point positions in the scanned image can be adjusted based on the respiratory influence degree to reduce the excessive movement amplitude of the digestive tract region in the video.

[0091] Preferably, in one embodiment of the present invention, the method for obtaining the respiratory influence of the target respiratory cycle specifically includes:

[0092] The last frame of the scanned image during the inhalation period of the target respiratory cycle is used as the inhalation termination image of the target respiratory cycle, and the last frame of the scanned image during the exhalation period of the target respiratory cycle is used as the exhalation termination image of the target respiratory cycle. Since the position of the digestive tract will shift downward during the inhalation process, and the position of the digestive tract will shift upward during the exhalation process, and the position difference between the digestive tract area in the inhalation termination image and the exhalation termination image in the target respiratory cycle is the largest, the degree to which the digestive tract area in the scanned video within the target respiratory cycle is affected by respiration can be accurately analyzed based on the position difference of the characteristic pixel points between the inhalation termination image and the exhalation termination image.

[0093] A first reference point of the inspiratory termination image of the target respiratory cycle is obtained, wherein the abscissa of the first reference point is equal to the average abscissa of all characteristic pixel points of the inspiratory termination image of the target respiratory cycle, and the ordinate of the first reference point is equal to the average ordinate of all characteristic pixel points of the inspiratory termination image of the target respiratory cycle.

[0094] Obtain a second reference point of the expiration end image of the target respiratory cycle, wherein the horizontal coordinate of the second reference point is equal to the average of the horizontal coordinates of all characteristic pixel points of the expiration end image of the target respiratory cycle, and the vertical coordinate of the second reference point is equal to the average of the vertical coordinates of all characteristic pixel points of the expiration end image of the target respiratory cycle.

[0095] The larger the distance between the first reference point and the second reference point, the more obvious the movement of the digestive tract region in the scanned video within the target respiratory cycle caused by respiration, which further indicates that the scanned video within the target respiratory cycle is more affected by respiration. Therefore, the Euclidean distance between the first reference point and the second reference point can be used as the respiratory influence degree of the target respiratory cycle.

[0096] Step S3: Taking the exhalation period or the inhalation period of the target respiratory cycle as the target period, the target period is divided into two sub-periods, and the positions of the pixel points of each frame of the scanned image are adjusted according to the changes in the respiratory amplitude data at each moment of each sub-period, the respiratory influence of the target respiratory cycle, and the serial number of each frame of the scanned image in each sub-period, to obtain an adjusted scanned image of each frame in each sub-period.

[0097] Since the position of the digestive tract will shift downward under the action of inhalation and will shift upward under the action of exhalation, the exhalation period or the inhalation period of the target respiratory cycle is first taken as the target period, and the target period is divided into two sub-periods. At this time, there are two sub-periods in the exhalation period and the inhalation period. Subsequently, the pixel points of the scanned image in different sub-periods of the target period can be moved and adjusted in different directions, thereby reducing the movement phenomenon of the digestive tract area in the scanned video in the target respiratory cycle caused by respiration, making the digestive tract area in the scanned video more stable.

[0098] The greater the degree of change in the respiratory amplitude data at each moment in a sub-period, and the greater the respiratory influence of the target respiratory cycle, the more obvious the movement phenomenon of the digestive tract area in the scanning video caused by respiration in the sub-period. At the same time, since the position of the digestive tract area in the scanning video continues to decrease or increase during an inhalation process or an exhalation process, in order to make the digestive tract area in the scanning video more stable, it is also necessary to adjust the position of the pixel points of each frame of the scanning image in combination with the serial number of each frame of the scanning image in the sub-period to obtain an adjusted scanning image of each frame of each sub-period. The adjusted scanning image can then be used to construct an adjusted scanning video, thereby ensuring that the digestive tract area in the adjusted scanning video is more stable, which is more conducive to the observation of the digestive tract structure.

[0099] Preferably, in one embodiment of the present invention, the method for acquiring the adjusted scan image of each frame in each sub-period specifically includes:

[0100] First, any sub-period in the target time period is taken as the sub-period to be measured, and the range of the respiratory amplitude data at all moments in the sub-period to be measured is normalized to obtain the amplitude change degree of the sub-period to be measured. The greater the amplitude change degree, the greater the degree of change in the respiratory amplitude data in the sub-period to be measured, and further the greater the degree to which the digestive tract in the sub-period to be measured is affected by respiration.

[0101] In one embodiment of the present invention, the range of the respiratory amplitude data at all moments in the two sub-periods of the target period can be calculated respectively, and then the range of the respiratory amplitude data at all moments in the sub-period to be measured is used as the numerator, the sum of the ranges of the respiratory amplitude data at all moments in the two sub-periods of the target period is used as the denominator, and the ratio is used as the amplitude change degree of the sub-period to be measured, thereby achieving normalization processing of the range of the respiratory amplitude data at all moments in the sub-period to be measured.

[0102] Then, based on the amplitude change degree of the sub-period to be measured, the respiratory influence of the target respiratory cycle, and the serial number of each frame of the scanned image of the sub-period to be measured, the pixel movement amount of each frame of the scanned image of the sub-period to be measured is obtained. Subsequently, based on the pixel movement amount, the pixel points in the digestive tract area of ​​each frame of the scanned image of the sub-period to be measured can be moved.

[0103] Preferably, in one embodiment of the present invention, the method for obtaining the pixel movement amount of each frame of the scanned image in the sub-period to be measured specifically includes:

[0104] If the sub-period to be measured is the previous sub-period of the target period, then within the sub-period to be measured, the pixel points of the scanned image that are closer in time sequence will have a greater subsequent movement. The calculation formula for the pixel movement of each frame of the scanned image in the sub-period to be measured is:

[0105]

[0106] in, Indicates the first sub-period to be tested The amount of pixel movement in a frame scan image; Indicates the degree of amplitude change of the sub-period to be measured; Indicates the number of all scanned images in the sub-period to be tested; Indicates the first sub-period to be tested The sequence number of the frame scan image; Indicates the respiratory influence of the target respiratory cycle; Indicates the round-up symbol.

[0107] If the sub-period to be measured is the next sub-period after the target period, then within the sub-period to be measured, the pixel points of the scanned images that are closer in time sequence will have a greater subsequent movement. At the same time, in order to ensure that the digestive tract area in the video is more stable during the transition from inspiration to expiration, the pixel movement of each frame of the scanned image at the end of the inspiration period and the beginning of the expiration period is calculated as follows:

[0108]

[0109] in, Indicates the first sub-period to be tested The amount of pixel movement in a frame scan image; Indicates the degree of amplitude change of the sub-period to be measured; Indicates the number of all scanned images in the sub-period to be tested; Indicates the first sub-period to be tested The sequence number of the frame scan image; Indicates the respiratory influence of the target respiratory cycle; Indicates the rounding symbol.

[0110] Furthermore, if the sub-period to be measured is the previous sub-period of the inhalation period or the next sub-period of the exhalation period, all pixel points within the contour line of each frame of the scanning image of the sub-period to be measured are moved downward to obtain an adjusted scanning image of each frame of the sub-period to be measured, wherein the moving distance is the pixel movement amount of each frame of the scanning image of the sub-period to be measured.

[0111] If the sub-period to be measured is the next sub-period of the inhalation period or the previous sub-period of the exhalation period, all pixel points within the contour line of each frame of the scanned image of the sub-period to be measured are moved upward to obtain an adjusted scanned image of each frame of the sub-period to be measured, wherein the moving distance is the pixel movement amount of each frame of the scanned image of the sub-period to be measured.

[0112] By the same method as above, the adjusted scan image of each frame in each sub-period of the target period and the adjusted scan image of each frame in the target respiratory cycle can be obtained, and then the adjusted scan image of each frame in each respiratory cycle can be obtained by the same method.

[0113] Step S4: Integrate the adjusted scan images of all frames into an adjusted scan video.

[0114] After processing the scanned image of each frame by the above method to obtain the adjusted scanned image of each frame, the adjusted scanned images of all frames can be integrated into an adjusted scanned video, thereby reducing the phenomenon of excessive movement of the digestive tract area in the scanned video caused by the patient's breathing, making the digestive tract area in the adjusted scanned video more stable, and more conducive to the observation of the digestive tract structure.

[0115] Preferably, in one embodiment of the present invention, the method for adjusting the acquisition of the scanned video specifically includes:

[0116] Input the adjusted scanned images of all frames into the video synthesis software, and use the video synthesis software to integrate the adjusted scanned images of all frames into the adjusted scanned video. The video synthesis software can be selected from Premiere software, etc., which is not limited here.

[0117] One embodiment of the present invention provides an intelligent processing system for barium meal angiography images of patients in the gastroenterology department. The system includes a memory, a processor, and a computer program, wherein the memory is used to store the corresponding computer program, and the processor is used to run the corresponding computer program. When the computer program runs in the processor, it can implement the method described in steps S1 to S4.

[0118] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0119] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

Claims

1. An intelligent processing method for barium meal radiography images of gastroenterology patients, characterized by: The method comprises: Acquire a scanned image of each frame of the digestive tract of a patient who has taken barium sulfate suspension within a preset time period and a respiratory signal of the patient's chest cavity, wherein the respiratory signal includes respiratory amplitude data at different moments; The respiratory signal is periodically segmented to obtain the respiratory cycle of the respiratory signal and the exhalation period and the inhalation period of each respiratory cycle; any respiratory cycle is used as a target respiratory cycle, and the contour line of the digestive tract in each frame scanned image of the target respiratory cycle is extracted, and multiple characteristic pixel points of the contour line are obtained based on the position distribution of each pixel point on the contour line of each frame scanned image of the target respiratory cycle; the respiratory influence degree of the target respiratory cycle is obtained based on the position difference of the characteristic pixel points of the contour line of each frame scanned image between the inhalation period and the exhalation period of the target respiratory cycle; Taking an exhalation period or an inhalation period of a target respiratory cycle as a target period, the target period is divided into two subperiods, and according to a change in respiratory amplitude data at each moment of each subperiod, the respiratory influence of the target respiratory cycle, and a sequence number of each frame of the scanned image in each subperiod, the position of a pixel point of each frame of the scanned image is adjusted to obtain an adjusted scanned image of each frame in each subperiod; The adjustment scan images of all frames are integrated into an adjustment scan video.

2. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 1, characterized in that: The respiratory cycle for obtaining the respiratory signal and the exhalation period and the inhalation period of each respiratory cycle include: Acquire extreme value points of the respiratory signal, wherein the extreme value points include maximum value points and minimum value points; The time period between two adjacent minimum points is regarded as a breathing cycle, wherein each breathing cycle contains a maximum point; The time period between the previous minimum point and the maximum point of each respiratory cycle is taken as the inhalation period of each respiratory cycle, and the time period between the maximum point and the next minimum point of each respiratory cycle is taken as the exhalation period of each respiratory cycle.

3. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 1, characterized in that: The method of obtaining a plurality of characteristic pixel points of the contour line comprises: Taking any frame of scanned image in the target respiratory cycle as the target scanned image; On the contour line of the target scanned image, any pixel point is taken as the target pixel point, and a first local mutation degree of the target pixel point with respect to the abscissa is obtained based on the difference in abscissa between the target pixel point and other pixels excluding the target pixel point in a preset window centered on the target pixel point, wherein the pixel points in the preset window are on the contour line; Based on the calculation method of the first local mutation degree of the target pixel point with respect to the horizontal coordinate, the second local mutation degree of the target pixel point with respect to the vertical coordinate is obtained according to the difference in the vertical coordinates between the target pixel point and other pixels other than the target pixel point in a preset window centered at the target pixel point; The average of the first local mutation degree and the second local mutation degree is used as the comprehensive mutation degree of the target pixel; On the contour line of the target scanned image, pixel points whose comprehensive mutation degree is greater than a preset mutation threshold are taken as feature pixel points.

4. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 3, characterized in that: Obtaining a first local mutation degree of the target pixel point with respect to the horizontal coordinate includes: In a preset window of a target pixel, a pixel on one side of the target pixel is used as a first reference pixel, and a pixel on the other side of the target pixel is used as a second reference pixel; Taking the difference between the horizontal coordinates of the target pixel and each first reference pixel as the first horizontal coordinate difference value between the target pixel and each first reference pixel, and taking the average of the first horizontal coordinate difference values ​​between the target pixel and all first reference pixels as the first comprehensive difference value of the target pixel with respect to the horizontal coordinate; Taking the difference between the horizontal coordinates of the target pixel and each second reference pixel as the second horizontal coordinate difference value between the target pixel and each second reference pixel, and taking the average of the second horizontal coordinate difference values ​​between the target pixel and all second reference pixels as the second comprehensive difference value of the target pixel with respect to the horizontal coordinate; A normalization process is performed on the product of the first integrated difference value and the second integrated difference value to obtain a first local mutation degree of the target pixel point with respect to the horizontal coordinate.

5. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 1, characterized in that: The respiratory influence degree of obtaining the target respiratory cycle includes: The last scanned image frame of the inhalation period of the target respiratory cycle is used as the inhalation termination image of the target respiratory cycle, and the last scanned image frame of the exhalation period of the target respiratory cycle is used as the exhalation termination image of the target respiratory cycle; Acquire a first reference point of the inhalation termination image of the target respiratory cycle, wherein the abscissa of the first reference point is equal to the average abscissa of all the characteristic pixel points of the inhalation termination image of the target respiratory cycle, and the ordinate of the first reference point is equal to the average ordinate of all the characteristic pixel points of the inhalation termination image of the target respiratory cycle; Acquire a second reference point of the expiration end image of the target respiratory cycle, wherein the abscissa of the second reference point is equal to the average abscissa of all the characteristic pixel points of the expiration end image of the target respiratory cycle, and the ordinate of the second reference point is equal to the average ordinate of all the characteristic pixel points of the expiration end image of the target respiratory cycle; The Euclidean distance between the first reference point and the second reference point is used as the respiratory influence degree of the target respiratory cycle.

6. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 1, characterized in that: The obtaining of the adjusted scanned image of each frame in each sub-period comprises: Taking any sub-period in the target period as the sub-period to be measured, normalizing the range of the respiratory amplitude data at all moments in the sub-period to be measured to obtain the amplitude variation degree of the sub-period to be measured; Obtaining a pixel movement amount of each frame of the scanned image in the sub-period to be measured according to the amplitude variation degree of the sub-period to be measured, the respiratory influence degree of the target respiratory cycle, and the sequence number of each frame of the scanned image in the sub-period to be measured; If the sub-period to be measured is the previous sub-period of the inhalation period or the next sub-period of the exhalation period, all pixels within the contour line of each frame of the scanned image of the sub-period to be measured are shifted downward to obtain an adjusted scanned image of each frame of the sub-period to be measured, wherein the shifted distance is the pixel shift amount of each frame of the scanned image of the sub-period to be measured; If the sub-period to be measured is the subsequent sub-period of the inhalation period or the previous sub-period of the exhalation period, all pixel points within the contour line of each frame of the scanning image of the sub-period to be measured are moved upward to obtain an adjusted scanning image of each frame of the sub-period to be measured, wherein the moving distance is the pixel movement amount of each frame of the scanning image of the sub-period to be measured.

7. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 6, characterized in that: The step of obtaining the pixel movement amount of each frame of the scanned image in the sub-period to be measured comprises: If the sub-period to be measured is the previous sub-period of the target period, the calculation formula for the pixel movement amount of each frame of the scanned image in the sub-period to be measured is: If the sub-period to be measured is the next sub-period after the target period, the calculation formula for the pixel movement amount of each frame of the scanned image in the sub-period to be measured is: in, Indicates the first sub-period to be tested The amount of pixel movement in a frame scan image; Indicates the degree of amplitude change of the sub-period to be measured; Indicates the number of all scanned images in the sub-period to be tested; Indicates the first sub-period to be tested The sequence number of the frame scan image; Indicates the respiratory influence of the target respiratory cycle.

8. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 1, characterized in that: The step of integrating the adjusted scanned images of all frames into the adjusted scanned video comprises: The adjusted scanned images of all frames are input into the video synthesis software, and the video synthesis software is used to integrate the adjusted scanned images of all frames into the adjusted scanned video.

9. The method for intelligently processing barium meal radiography images of gastroenterology patients according to claim 1, characterized in that: The step of extracting the contour of the digestive tract from each frame of the scanned image of the target respiratory cycle includes: The scanned images containing manually annotated digestive tract contours are used as a training set, and the training set is used to train a neural network to obtain a trained neural network; Each frame of the scanned image of the target respiratory cycle is input into the trained neural network, and the neural network outputs the contour line of the digestive tract in each frame of the scanned image.

10. An intelligent processing system for barium meal radiography images of gastroenterology patients, the system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images