WO2023029634A1

WO2023029634A1 - Processing method and apparatus for performing stent positioning on coronary angiography image

Info

Publication number: WO2023029634A1
Application number: PCT/CN2022/097241
Authority: WO
Inventors: 李喆; 曹君; 张碧莹
Original assignee: 乐普（北京）医疗器械股份有限公司
Priority date: 2021-09-03
Filing date: 2022-06-07
Publication date: 2023-03-09
Also published as: CN113723406B; CN113723406A

Abstract

A processing method and apparatus for performing stent positioning on a coronary angiography image. The method comprises: obtaining a first image sequence; performing target detection to obtain a plurality of first recognition frames; marking a recognition frame having a first stent frame confidence level c1 exceeding a first threshold as a first stent frame, and marking a recognition frame having a first endpoint frame confidence level c2 exceeding a second threshold as a first endpoint frame; performing first endpoint reconstruction on a first image of which the number of the first endpoint frames is empty; performing first stent frame reconstruction on the first image of which the number of the first stent frames is empty and the number of the first endpoint frames is greater than or equal to 2; performing second stent reconstruction on the first image of which the number of the first stent frames is empty and the number of the first endpoint frames is less than 2; screening stents at unreasonable positions; screening redundant stents; screening endpoints at unreasonable positions; and using the first image sequence as an angiography image sequence for completing stent positioning. The method improves the positioning efficiency and accuracy of the stent.

Description

一种对冠脉造影图像进行支架定位的处理方法和装置A processing method and device for stent positioning on coronary angiography images

本申请要求于2021年9月3日提交中国专利局、申请号为202111034147.9、发明名称为“一种对冠脉造影图像进行支架定位的处理方法和装置”的中国专利申请的优先权。This application claims the priority of the Chinese patent application submitted to the Chinese Patent Office on September 3, 2021, with the application number 202111034147.9, and the title of the invention is "a processing method and device for stent positioning on coronary angiography images".

技术领域technical field

本发明涉及数据处理技术领域，特别涉及一种对冠脉造影图像进行支架定位的处理方法和装置。The invention relates to the technical field of data processing, in particular to a processing method and device for stent positioning on coronary angiography images.

背景技术Background technique

冠状动脉性心脏病(coronary artery heart disease,CHD)简称冠心病，是指因冠状动脉狭窄、供血不足而引起的心肌机能障碍和(或)器质性病变，故又称缺血性心肌病。通过在冠脉狭窄处放置心脏支架是当前对于此类病症的主要处理手段之一。在放置心脏支架的过程中，主要是通过冠脉造影图像来实时监控支架的位置和在其血管中的姿态。但是，在冠脉造影图像中只有心脏支架两端的两个端点可以被清晰地观测到，并且因为图像明暗和对比度等因素的存在，端点的清晰度也会受到一定程度影响，所以只凭借肉眼是很难在动态的冠脉造影图片中准确的捕捉到心脏支架的位姿变化的。Coronary heart disease (CHD) refers to myocardial dysfunction and (or) organic disease caused by coronary artery stenosis and insufficient blood supply, so it is also called ischemic cardiomyopathy. Placement of cardiac stents in coronary artery stenosis is currently one of the main treatment methods for such diseases. In the process of placing a cardiac stent, the position of the stent and its posture in the blood vessel are mainly monitored in real time through coronary angiography images. However, in the coronary angiography image, only the two endpoints at both ends of the cardiac stent can be clearly observed, and because of factors such as image brightness and contrast, the clarity of the endpoints will also be affected to a certain extent, so only the naked eye can It is difficult to accurately capture the pose change of the heart stent in the dynamic coronary angiography picture.

发明内容Contents of the invention

本发明的目的，就是针对现有技术的缺陷，提供一种对冠脉造影图像进行支架定位的处理方法、装置、电子设备及计算机可读存储介质，通过目标识别网络从造影图像序列的各个造影图像中识别出具体为支架框和端点框的目标识别框，并通过一系列的修正操作对得到的端点框与支架框的进行修正从而输出带有最优支架定位信息的造影图像序列。通过本发明，即可以解决人工定位的不稳定性，又可以提高定位的效率和准确度。The purpose of the present invention is to address the defects of the prior art, to provide a processing method, device, electronic equipment and computer-readable storage medium for stent positioning on coronary angiography images. The target recognition frame, specifically the stent frame and the end point frame, is identified in the image, and the obtained end point frame and the stent frame are corrected through a series of correction operations to output an angiographic image sequence with optimal stent positioning information. The present invention can not only solve the instability of manual positioning, but also improve the efficiency and accuracy of positioning.

为实现上述目的，本发明实施例第一方面提供了一种对冠脉造影图像进行支架定位的处理方法，所述方法包括：In order to achieve the above purpose, the first aspect of the embodiment of the present invention provides a method for processing stent positioning on coronary angiography images, the method comprising:

获取连续的冠脉造影图像生成第一图像序列；所述第一图像序列包括多个第一图像；Acquiring continuous coronary angiography images to generate a first image sequence; the first image sequence includes a plurality of first images;

使用训练成熟的目标检测网络，对所述第一图像进行目标检测处理得到多个第一识别框；所述第一识别框包括第一支架框置信度c ₁和第一端点框置信度c ₂； Using a well-trained target detection network, perform target detection processing on the first image to obtain a plurality of first recognition frames; the first recognition frames include the first support frame confidence c ₁ and the first end point frame confidence c ₂ ;

将所述第一图像中，所述第一支架框置信度c ₁超过预设的第一阈值的所述第一识别框记为第一支架框，所述第一端点框置信度c ₂超过预设的第二阈值的所述第一识别框记为第一端点框； In the first image, the first recognition frame whose confidence degree c ₁ of the first support frame exceeds a preset first threshold is marked as a first support frame, and the first end point frame confidence degree c ₂ The first identified frame exceeding the preset second threshold is marked as a first endpoint frame;

对所述第一图像序列中所述第一端点框的数量为空的所述第一图像，进行第一端点重构处理；performing a first endpoint reconstruction process on the first image in the first image sequence in which the number of the first endpoint frames is empty;

对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像，进行第一支架框重构处理；Performing first stent frame reconstruction processing on the first image in the first image sequence in which the number of the first stent frames is empty and the number of the first end point frames is greater than or equal to 2;

对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像，进行第二支架重构处理；performing a second stent reconstruction process on the first image in the first image sequence in which the number of the first stent frames is empty and the number of the first end point frames is less than 2;

对所述第一图像序列进行不合理位置支架筛查处理；performing unreasonable location bracket screening processing on the first image sequence;

对所述第一图像序列进行冗余支架筛查处理；performing redundant scaffold screening processing on the first image sequence;

对所述第一图像序列进行不合理位置端点筛查处理；performing unreasonable position endpoint screening processing on the first image sequence;

将完成重构与筛查的所述第一图像序列作为完成支架定位的造影图像序列。The first sequence of images completed with reconstruction and screening is used as the sequence of angiography images completed with positioning of the stent.

优选的，所述目标检测网络为YOLOv3神经网络。Preferably, the target detection network is a YOLOv3 neural network.

优选的，所述第一识别框还包括第一中心点坐标、第一宽度w和第一高度h。Preferably, the first identification frame also includes the coordinates of the first center point, the first width w and the first height h.

优选的，所述对所述第一图像序列中所述第一端点框的数量为空的所述第一图像，进行第一端点重构处理，具体包括：Preferably, performing first endpoint reconstruction processing on the first image in the first image sequence whose number of first endpoint frames is empty, specifically includes:

将所述第一图像序列中所述第一端点框的数量为空的所述第一图像记为第一无端点图像；Recording the first image in the first image sequence in which the number of the first endpoint boxes is empty is the first no-endpoint image;

将所述第一图像序列中所述第一端点框的数量不为空且与所述第一无端点图像距离最近的所述第一图像记为第一相邻图像；Recording the first image in the first image sequence whose number of the first endpoint frame is not empty and whose distance is closest to the first image without endpoints is the first adjacent image;

根据所述第一相邻图像的所述第一端点框的所述第一中心点坐标、所述第一宽度w和所述第一高度h，在所述第一无端点图像上的对应位置定位第一区域；所述第一区域的中心点坐标与所述第一相邻图像的所述第一端点框的所述第一中心点坐标对应，所述第一区域的宽度与所述第一相邻图像的所述第一端点框的所述第一宽度w对应，所述第一区域的高度与所述第一相邻图像的所述第一端点框的所述第一高度h对应；According to the first center point coordinates, the first width w, and the first height h of the first endpoint frame of the first adjacent image, the correspondence on the first non-endpoint image positioning the first area; the center point coordinates of the first area correspond to the first center point coordinates of the first end point frame of the first adjacent image, and the width of the first area corresponds to the first end point frame of the first adjacent image. corresponds to the first width w of the first end point frame of the first adjacent image, and the height of the first region corresponds to the first width w of the first end point frame of the first adjacent image. A height h corresponds to;

根据预设的第一宽度微调阈值△w ₁和第一高度微调阈值△h ₁，对所述第一区域进拓展得到第二区域；所述第二区域的中心点坐标与所述第一区域的中心点坐标对应，所述第二区域的宽度为所述第一区域的宽度与所述第一宽度微调阈值△w ₁相加的和，所述第二区域的高度为所述第一区域的高度与所述第一高度微调阈值△h ₁相加的和； According to the preset first width fine-tuning threshold Δw ₁ and first height fine-tuning threshold Δh ₁ , expand the first area to obtain a second area; the center point coordinates of the second area are the same as the first area Corresponding to the coordinates of the central point of , the width of the second area is the sum of the width of the first area and the first width fine-tuning threshold Δw ₁ , and the height of the second area is the sum of the first area The sum of the height and the first height fine-tuning threshold Δh ₁ ;

在所述第二区域中进行端点框重构；将所述第二区域中像素值最小的像素点坐标作为重构出的所述第一端点框的所述第一中心点坐标，并根据所述第一相邻图像的所述第一端点框的所述第一宽度w、所述第一高度h、所述第一支架框置信度c ₁和所述第一端点框置信度c ₂对所述第二区域中重构出的所述第一端点框进行对应参数设置。 Perform endpoint frame reconstruction in the second area; use the coordinate of the pixel point with the smallest pixel value in the second area as the first center point coordinate of the reconstructed first end point frame, and according to The first width w, the first height h, the first bracket frame confidence _c1 and the first end point frame confidence of the first adjacent image c ₂ performing corresponding parameter setting on the first end point frame reconstructed in the second area.

优选的，所述对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像，进行第一支架框重构处理，具体包括：Preferably, in the first image sequence, the number of the first bracket frames is empty and the number of the first end point frames is greater than or equal to 2, the first bracket frame weighting is performed. structure processing, including:

将所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像记为第二图像；recording the first image in which the number of the first bracket frame in the first image sequence is empty and the number of the first end point frame is greater than or equal to 2 as a second image;

在所述第二图像中任选一个所述第一端点框作为当前端点框，并将与之距离最近的所述第一端点框作为当前相邻端点框；Selecting one of the first end point frames in the second image as the current end point frame, and using the first end point frame closest to it as the current adjacent end point frame;

取所述当前端点框的左上角顶点坐标记为第一左上坐标(x ₁₁,y ₁₁)，右下角顶点坐标记为第一右下坐标(x ₁₂,y ₁₂)；并取所述当前相邻端点框的左上角顶点坐标记为第二左上坐标(x ₂₁,y ₂₁)，右下角顶点坐标记为第二右下坐标(x ₂₂,y ₂₂)；并计算以所述第一左上坐标(x ₁₁,y ₁₁)作为左上角顶点坐标，以所述第二右下坐标(x ₂₂,y ₂₂)作为右下角顶点坐标的矩形区域的面积S，S＝|x ₂₂-x ₁₁|×|y ₂₂-y ₁₁|，||为绝对值计算符； Take the vertex coordinates of the upper left corner of the current endpoint frame and mark it as the first upper left coordinate (x ₁₁ , y ₁₁ ), and mark the vertex coordinates of the lower right corner as the first lower right coordinate (x ₁₂ , y ₁₂ ); and take the current phase The vertex coordinates of the upper left corner of the adjacent endpoint frame are marked as the second upper left coordinate (x ₂₁ , y ₂₁ ), and the coordinates of the vertex of the lower right corner are marked as the second lower right coordinate (x ₂₂ , y ₂₂ ); and the calculation is based on the first upper left coordinate (x ₁₁ , y ₁₁ ) as the coordinates of the upper left corner vertex, and the second lower right coordinate (x ₂₂ , y ₂₂ ) as the area S of the rectangular area of the lower right corner vertex coordinates, S=|x ₂₂ -x ₁₁ |× |y ₂₂ -y ₁₁ |, || is an absolute value operator;

在得到的多个所述面积S中，将面积最小的所述面积S对应的所述第一左上坐标(x ₁₁,y ₁₁)作为重构支架框的左上顶点坐标(x _l,y _l)，对应的所述第二右下坐标(x ₂₂,y ₂₂)作为重构支架框的右下顶点坐标(x _r,y _r)；将面积最小的所述面积S对应的两个所述第一端点框的所述第一端点框置信度c ₂进行加权平均计算生成重构支架框置信度； Among the multiple obtained areas S, the first upper left coordinate (x ₁₁ , y ₁₁ ) corresponding to the area S with the smallest area is used as the upper left vertex coordinate (x _l , y _l ) of the reconstructed scaffold frame , the corresponding second lower right coordinates (x ₂₂ , y ₂₂ ) are used as the lower right vertex coordinates (x _r , y _r ) of the reconstructed frame; The confidence degree _c of the first endpoint frame of an endpoint frame is calculated by weighted average to generate the confidence degree of the reconstructed bracket frame;

根据所述左上顶点坐标(x _l,y _l)和所述右下顶点坐标(x _r,y _r)进行支架框重构；将重构的所述第一支架框的所述第一中心点坐标设为由所述左上顶点坐标(x _l,y _l)和所述右下顶点坐标(x _r,y _r)构成的矩形框的中心点坐标；将重构的所述第一支架框的所述第一宽度w设为|x _r-x _l|；将重构的所述第一支架框的所述第一高度h设为|y _r-y _l|；将重构的所述第一支架框的所述第一支架框置信度c ₁设为所述重构支架框置信度；将重构的所述第一支架框的所述第一端点框置信度c ₂设为低于所述第二阈值的概率值。 Reconstruction of the stent frame according to the coordinates of the upper left vertex (x _l , y _l ) and the coordinates of the lower right vertex (x _r , y _r ); the first central point of the reconstructed first stent frame The coordinates are set as the coordinates of the center point of the rectangular frame formed by the coordinates of the upper left vertex (x _l , y _l ) and the coordinates of the lower right vertex (x _r , y _r ); The first width w is set to |x _r -x _l |; the first height h of the reconstructed first stent frame is set to |y _r -y _l |; the reconstructed first The first stent frame confidence level _c1 of a stent frame is set as the reconstructed stent frame confidence level; the first endpoint frame confidence _c2 of the reconstructed first stent frame is set as low at the probability value of the second threshold.

优选的，所述对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像，进行第二支架重构处理，具体包括：Preferably, the second stent reconstruction process is performed on the first image in the first image sequence in which the number of the first stent frames is empty and the number of the first end point frames is less than 2, Specifically include:

将所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像记为第一无支架图像；The number of the first stent frame in the first image sequence is empty and the first image with the number of the first endpoint frame less than 2 is recorded as the first stent-free image;

将所述第一图像序列中所述第一支架框的数量不为空且与所述第一无支架图像距离最近的所述第一图像记为第二相邻图像；Recording the first image in the first image sequence whose number of the first stent frame is not empty and which is the closest to the first stent-free image as a second adjacent image;

根据所述第二相邻图像的所述第一支架框的所述第一中心点坐标、所述第一宽度w和所述第一高度h，在所述第一无支架图像上的对应位置进行支架框重构；根据所述第二相邻图像的所述第一支架框的所述第一中心点坐标、所述第一宽度w、所述第一高度h、所述第一支架框置信度c ₁和所述第一端点框置信度c ₂对重构的所述第一支架框进行对应参数设置。 According to the first central point coordinates, the first width w, and the first height h of the first stent frame of the second adjacent image, the corresponding position on the first stent-free image Carry out frame reconstruction; according to the coordinates of the first center point of the first frame of the second adjacent image, the first width w, the first height h, the first frame Confidence c ₁ and the first endpoint frame confidence c ₂ are used to set corresponding parameters for the reconstructed first bracket frame.

优选的，所述对所述第一图像序列进行不合理位置支架筛查处理，具体包括：Preferably, the screening of brackets at unreasonable positions on the first image sequence specifically includes:

对所述第一图像序列的所有所述第一支架框的所述第一中心点坐标的横坐标平均值进行计算得到第一x平均值；calculating the average value of the abscissa of the coordinates of the first central point of all the first bracket frames in the first image sequence to obtain a first average value of x;

对所述第一图像序列的所有所述第一支架框的所述第一中心点坐标的纵坐标平均值进行计算得到第一y平均值；calculating the average value of the ordinates of the coordinates of the first central point of all the first bracket frames in the first image sequence to obtain a first average value of y;

在xy平面上，根据各个所述第一支架框的所述第一中心点坐标进行对应的中心点标记，并对由所述第一x平均值和所述第一y平均值构成的均值点进行标记；所述均值点的坐标为(第一x平均值,第一y平均值)On the xy plane, mark the corresponding center point according to the coordinates of the first center point of each of the first support frames, and set the mean value point composed of the first mean value of x and the mean value of the first y Marking; the coordinates of the mean point are (the first x average value, the first y average value)

以所述均值点为原点、以平行于x轴的直线为象限横轴、以平行于y轴的直线为象限纵轴，将xy平面划分为四个中心点象限；Taking the mean point as the origin, taking the straight line parallel to the x-axis as the quadrant horizontal axis, and taking the straight line parallel to the y-axis as the quadrant vertical axis, divide the xy plane into four center point quadrants;

对所述四个中心点象限分别进行中心点统计处理，得到对应的象限中心点数量；Carry out center point statistical processing respectively to described four center point quadrants, obtain the corresponding quadrant center point quantity;

将数量为最大值的所述象限中心点数量对应的所述中心点象限作为保留中心点象限；The center point quadrant corresponding to the number of the quadrant center points whose quantity is the maximum value is used as the reserved center point quadrant;

将不属于所述保留中心点象限的中心点对应的所述第一支架框，从所述第一图像序列中移除。removing the first support frame corresponding to the center point not belonging to the quadrant of the reserved center point from the first image sequence.

优选的，所述对所述第一图像序列进行冗余支架筛查处理，具体包括：Preferably, the performing redundant scaffold screening processing on the first image sequence specifically includes:

若所述第一图像序列的所述第一图像中存在多个所述第一支架框，则只保留其中所述第一支架框置信度c ₁为最大值的所述第一支架框。 If there are multiple first stent frames in the first image of the first image sequence, only retain the first stent frame in which the confidence c ₁ of the first stent frame is the maximum value.

优选的，所述对所述第一图像序列进行不合理位置端点筛查处理，具体包括：Preferably, the screening of unreasonable position endpoints on the first image sequence specifically includes:

若所述第一图像序列的所述第一图像中存在多个处于所述第一支架框范围之外的所述第一端点框，则只保留处于所述第一支架框范围内的所述第一端点框。If there are multiple first end point frames outside the range of the first bracket frame in the first image of the first image sequence, only keep all the frames within the range of the first bracket frame Describe the first endpoint box.

本发明实施例第二方面提供了一种实现上述第一方面所述的方法的装置，包括：获取模块、目标检测网络处理模块、定位信息修正模块和输出模块；The second aspect of the embodiment of the present invention provides a device for implementing the method described in the first aspect above, including: an acquisition module, a target detection network processing module, a positioning information correction module, and an output module;

所述获取模块用于获取连续的冠脉造影图像生成第一图像序列；所述第一图像序列包括多个第一图像；The acquisition module is used to acquire continuous coronary angiography images to generate a first image sequence; the first image sequence includes a plurality of first images;

所述目标检测网络处理模块用于使用训练成熟的目标检测网络，对所述第一图像进行目标检测处理得到多个第一识别框；所述第一识别框包括第一支架框置信度c ₁和第一端点框置信度c ₂； The target detection network processing module is used to use a well-trained target detection network to perform target detection processing on the first image to obtain a plurality of first recognition frames; the first recognition frames include a first support frame confidence c ₁ and the first endpoint box confidence c ₂ ;

所述定位信息修正模块用于将所述第一图像中，所述第一支架框置信度c ₁超过预设的第一阈值的所述第一识别框记为第一支架框，所述第一端点框置信度c ₂超过预设的第二阈值的所述第一识别框记为第一端点框；并对所述第一图像序列中所述第一端点框的数量为空的所述第一图像，进行第一端点重构处理；并对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像，进行第一支架框重构处理；并对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像，进行第二支架重构处理；并对所述第一图像序列进行不合理位置支架筛查处理；并对所述第一图像序列进行冗余支架筛查处理；并对所述第一图像序列进行不合理位置端点筛查处理； The positioning information correction module is used to mark the first recognition frame whose confidence level _c1 of the first support frame exceeds a preset first threshold in the first image as a first support frame, and the first support frame The first recognition frame whose confidence degree _c of an endpoint frame exceeds the preset second threshold is marked as a first endpoint frame; and the number of the first endpoint frame in the first image sequence is empty The first image of the first end point reconstruction process; and the number of the first bracket frame in the first image sequence is empty and the number of the first end point frame is greater than or equal to 2 For the first image, perform first bracket frame reconstruction processing; and for the first frame in the first image sequence, the number of the first bracket frame is empty and the number of the first end point frame is less than 2. performing a second bracket reconstruction process on an image; and performing an unreasonable position bracket screening process on the first image sequence; and performing a redundant bracket screening process on the first image sequence; and performing a redundant bracket screening process on the first image sequence; The image sequence is screened for unreasonable position endpoints;

所述输出模块用于将完成重构与筛查的所述第一图像序列作为完成支架定位的造影图像序列。The output module is configured to use the first sequence of images completed with reconstruction and screening as a sequence of contrast images completed with stent positioning.

本发明实施例第三方面提供了一种电子设备，包括：存储器、处理器和收发器；The third aspect of the embodiment of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

所述处理器用于与所述存储器耦合，读取并执行所述存储器中的指令，以实现上述第一方面所述的方法步骤；The processor is configured to be coupled with the memory, read and execute instructions in the memory, so as to implement the method steps described in the first aspect above;

所述收发器与所述处理器耦合，由所述处理器控制所述收发器进行消息收发。The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

本发明实施例第四方面提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机指令，当所述计算机指令被计算机执行时，使得所述计算机执行上述第一方面所述的方法的指令。The fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer executes the above-mentioned first aspect. method directive.

本发明实施例提供了一种对冠脉造影图像进行支架定位的处理方法、装置、电子设备及计算机可读存储介质，通过目标识别网络从造影图像序列的各个造影图像中识别出具体为支架框和端点框的目标识别框，并通过一系列的修正操作对得到的端点框与支架框的进行修正从而输出带有最优支架定位信息的造影图像序列。通过本发明，即解决了人工定位的不稳定性，又提高了定位的效率和准确度。An embodiment of the present invention provides a processing method, device, electronic device, and computer-readable storage medium for stent positioning on coronary angiography images, and recognizes specifically the stent frame from each angiography image of the angiography image sequence through a target recognition network and the target recognition frame of the end point frame, and correct the obtained end point frame and the stent frame through a series of correction operations to output an angiographic image sequence with optimal stent positioning information. The invention not only solves the instability of manual positioning, but also improves the efficiency and accuracy of positioning.

附图说明Description of drawings

图1为本发明实施例一提供的一种对冠脉造影图像进行支架定位的处理方法示意图；FIG. 1 is a schematic diagram of a processing method for stent positioning on coronary angiography images provided by Embodiment 1 of the present invention;

图2为本发明实施例二提供的一种对冠脉造影图像进行支架定位的处理装置的模块结构图；2 is a block diagram of a processing device for stent positioning on coronary angiography images provided by Embodiment 2 of the present invention;

图3为本发明实施例三提供的一种电子设备的结构示意图。FIG. 3 is a schematic structural diagram of an electronic device provided by Embodiment 3 of the present invention.

具体实施方式Detailed ways

为了使本发明的目的、技术方案和优点更加清楚，下面将结合附图对本发明作进一步地详细描述，显然，所描述的实施例仅仅是本发明一部份实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例，都属于本发明保护的范围。In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments . Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

本发明实施例一提供的一种对冠脉造影图像进行支架定位的处理方法，如图1为本发明实施例一提供的一种对冠脉造影图像进行支架定位的处理方法示意图所示，本方法主要包括如下步骤： Embodiment 1 of the present invention provides a method for processing stent positioning on coronary angiography images, as shown in FIG. 1 , a schematic diagram of a method for processing stent positioning on coronary angiography images provided by Embodiment 1 of the present invention. The method mainly includes the following steps:

步骤1，获取连续的冠脉造影图像生成第一图像序列； Step 1, acquiring continuous coronary angiography images to generate a first image sequence;

其中，第一图像序列包括多个第一图像。Wherein, the first image sequence includes a plurality of first images.

这里，第一图像序列中的第一图像为冠脉造影图像。Here, the first image in the first image sequence is a coronary angiography image.

步骤2，使用训练成熟的目标检测网络，对第一图像进行目标检测处理得到多个第一识别框； Step 2, using a well-trained target detection network to perform target detection processing on the first image to obtain a plurality of first recognition frames;

其中，目标检测网络为YOLOv3神经网络；第一识别框包括第一中心点坐标、第一宽度w、第一高度h、第一支架框置信度c ₁和第一端点框置信度c ₂。 Wherein, the target detection network is a YOLOv3 neural network; the first recognition frame includes the coordinates of the first center point, the first width w, the first height h, the confidence degree c ₁ of the first support frame, and the confidence degree c ₂ of the first end point frame.

此处，目标检测网络的结构借鉴了YOLOv3的网络结构；关于YOLOv3神经网络，详细原理可参看由美国华盛顿大学(University of Washington)的Joseph Redmon和Ali Farhadi于2018年联合发表的论文《YOLOv3:An Incremental Improvement》；YOLOv3神经网络被视为是YOLO(You only look once)神经网络的第三版；对二者进行比较，YOLOv3神经网络在小目标的识别精度上有着显著提升效果；Here, the structure of the target detection network draws on the network structure of YOLOv3; for the detailed principles of the YOLOv3 neural network, please refer to the paper "YOLOv3:An Incremental Improvement"; YOLOv3 neural network is regarded as the third version of YOLO (You only look once) neural network; comparing the two, YOLOv3 neural network has a significant improvement in the recognition accuracy of small targets;

这里，YOLOv3神经网络的输入为具体的二维灰度图，输出则为多个不同大小尺寸的目标识别框(第一识别框)，每个目标识别框均带有一组参数：中心点二维坐标参数(第一中心点坐标)、识别框宽度参数(第一宽度w)、识别框高度参数(第一高度h)和2个识别框置信度参数(第一支架框置信度c ₁和第一端点框置信度c ₂)；其中中心点二维坐标参数、识别框宽度参数、识别框高度参数可确定目标框的位置和形状；每个识别框置信度参数为对应的目标类别的置信概率：第一支架框置信度c ₁即为当前识别框为支架识别框的置信概率，第一端点框置信度c ₂即为当前识别框为支架端点识别框的置信概率。 Here, the input of the YOLOv3 neural network is a specific two-dimensional grayscale image, and the output is multiple target recognition frames (first recognition frames) of different sizes. Each target recognition frame has a set of parameters: center point two-dimensional Coordinate parameters (the coordinates of the first center point), recognition frame width parameters (first width w), recognition frame height parameters (first height h) and two recognition frame confidence parameters (the first support frame confidence c ₁ and the second An endpoint frame confidence c ₂ ); where the two-dimensional coordinate parameters of the central point, the recognition frame width parameters, and the recognition frame height parameters can determine the position and shape of the target frame; each recognition frame confidence parameter is the confidence of the corresponding target category Probability: the confidence degree c ₁ of the first support frame is the confidence probability that the current recognition frame is a support recognition frame, and the confidence degree c ₂ of the first end point frame is the confidence probability that the current recognition frame is a support end point recognition frame.

步骤3，将第一图像中，第一支架框置信度c ₁超过预设的第一阈值的第一识别框记为第一支架框，第一端点框置信度c ₂超过预设的第二阈值的第一识别框记为第一端点框。 Step 3: In the first image, the first recognition frame whose confidence degree c ₁ of the first support frame exceeds the preset first threshold is marked as the first support frame, and the confidence degree c ₂ of the first endpoint frame exceeds the preset first threshold The first identified box of the two thresholds is denoted as the first endpoint box.

此处，第一、第二阈值均为预先设定的***参数，其中第一阈值为支架框置信度阈值，若第一支架框置信度c ₁低于第一阈值则意味着当前的第一识别框的类型不为支架框，反之则说明当前的第一识别框的类型为支架框；第二阈值为端点框置信度阈值，若第一端点框置信度c ₂低于第二阈值则意味着当前的第一识别框的类型不为端点框，反之则说明当前的第一识别框的类型为端点框。 Here, the first and second thresholds are both preset system parameters, where the first threshold is the confidence threshold of the stent frame, and if the confidence c ₁ of the first stent frame is lower than the first threshold, it means that the current first The type of the recognition frame is not a stent frame, otherwise it means that the type of the current first recognition frame is a stent frame; the second threshold is the endpoint frame confidence threshold, if the first endpoint frame confidence c ₂ is lower than the second threshold then It means that the type of the current first recognition frame is not an end point frame, otherwise it means that the type of the current first recognition frame is an end point frame.

正常情况下一张第一图像中只包含一个第一支架框，而且第一支架框的图像范围内会包括两个第一端点框；然而在实际应用中，目标检测网络输出的结果可能会出现以下情况：1)在第一图像中，未能识别出第一端点框和第一支架框；2)在第一图像中，只识别出了多个第一端点框而无第一支架框；3)在第一图像中，识别出了多个第一支架框；4)在第一图像中，识别出的第一端点框未能处于第一支架框的图像范围内。为了进一步提高支架定位的准确度和精度，本发明实施例会通过后续步骤4-10对目标识别网络的输出结果进行修正，从而得到最终的精度更高的带有支架定位信息的造影图像序列。Normally, a first image contains only one first bracket frame, and the image range of the first bracket frame will include two first end point frames; however, in practical applications, the output result of the target detection network may be The following situations occur: 1) In the first image, the first end point frame and the first bracket frame cannot be recognized; 2) In the first image, only a plurality of first end point frames are recognized without the first stent frame; 3) in the first image, multiple first stent frames are identified; 4) in the first image, the identified first end point frame fails to be within the image range of the first stent frame. In order to further improve the accuracy and precision of stent positioning, the embodiment of the present invention corrects the output result of the target recognition network through subsequent steps 4-10, so as to obtain a final contrast image sequence with higher precision stent positioning information.

步骤4，对第一图像序列中第一端点框的数量为空的第一图像，进行第一端点重构处理；Step 4, performing first endpoint reconstruction processing on the first image whose number of first endpoint frames in the first image sequence is empty;

这里，若第一图像序列中存在没有第一端点框的第一图像，通过复制与之相邻的带有第一端点框的第一图像中的端点框信息来实现对当前第一图像的端点框重构；Here, if there is a first image without a first endpoint frame in the first image sequence, the current first image is realized by copying the endpoint frame information in the adjacent first image with the first endpoint frame The endpoint box refactoring;

具体包括：步骤41，将第一图像序列中第一端点框的数量为空的第一图像记为第一无端点图像；Specifically include: step 41, the first image whose number of first endpoint frames in the first image sequence is empty is recorded as the first non-endpoint image;

步骤42，将第一图像序列中第一端点框的数量不为空且与第一无端点图像距离最近的第一图像记为第一相邻图像；Step 42, the number of the first endpoint frame in the first image sequence is not empty and the first image with the closest distance to the first non-endpoint image is recorded as the first adjacent image;

这里，距离最近即为时间最近，与第一无端点图像距离最近即为与第一无端点图像时间最近；需要说明的是，若与第一无端点图像距离最近的第一端点框的数量不为空的第一图像前后各有一张，则优选后者作为第一相邻图像；Here, the shortest distance is the shortest time, and the shortest distance to the first non-end point image is the shortest time to the first non-end point image; There are one before and after the non-empty first image, then the latter is preferred as the first adjacent image;

步骤43，根据第一相邻图像的第一端点框的第一中心点坐标、第一宽度w和第一高度h，在第一无端点图像上的对应位置定位第一区域；Step 43, according to the coordinates of the first center point, the first width w and the first height h of the first endpoint frame of the first adjacent image, locate the first region at the corresponding position on the first non-endpoint image;

其中，第一区域的中心点坐标与第一相邻图像的第一端点框的第一中心点坐标对应，第一区域的宽度与第一相邻图像的第一端点框的第一宽度w对应，第一区域的高度与第一相邻图像的第一端点框的第一高度h对应；Wherein, the center point coordinates of the first area correspond to the first center point coordinates of the first end point frame of the first adjacent image, and the width of the first area corresponds to the first width of the first end point frame of the first adjacent image. Corresponding to w, the height of the first region corresponds to the first height h of the first end point frame of the first adjacent image;

这里，第一区域实际就是在第一无端点图像中对第一相邻图像的某个具体的第一端点框的复制区域；Here, the first area is actually a copy area of a specific first end point frame of the first adjacent image in the first no end point image;

步骤44，根据预设的第一宽度微调阈值△w ₁和第一高度微调阈值△h ₁，对第一区域进拓展得到第二区域； Step 44, expand the first area to obtain the second area according to the preset first width fine-tuning threshold Δw ₁ and first height fine-tuning threshold Δh ₁ ;

其中，第二区域的中心点坐标与第一区域的中心点坐标对应，第二区域的宽度为第一区域的宽度与第一宽度微调阈值△w ₁相加的和，第二区域的高度为第一区域的高度与第一高度微调阈值△h ₁相加的和； Wherein, the center point coordinates of the second area correspond to the center point coordinates of the first area, the width of the second area is the sum of the width of the first area and the first width fine-tuning threshold Δw ₁ , and the height of the second area is The sum of the height of the first area and the first height fine-tuning threshold Δh ₁ ;

这里，第一宽度微调阈值△w ₁和第一高度微调阈值△h ₁为预先设定的***参数，第二区域实际就是以第一区域的中心为中心沿正负x方向各扩展△w ₁/2，沿正负y方向各扩展△h ₁/2的矩形； Here, the first width fine-tuning threshold △w ₁ and the first height fine-tuning threshold △h ₁ are preset system parameters, and the second area actually expands △w ₁ along the positive and negative x directions with the center of the first area as the center /2, expand the rectangle of △h ₁ /2 along the positive and negative y directions;

步骤45，在第二区域中进行端点框重构；Step 45, performing endpoint frame reconstruction in the second area;

具体包括：将第二区域中像素值最小的像素点坐标作为重构出的第一端点框的第一中心点坐标，并根据第一相邻图像的第一端点框的第一宽度w、第一高度h、第一支架框置信度c ₁和第一端点框置信度c ₂对第二区域中重构出的第一端点框进行对应参数设置。 It specifically includes: taking the coordinates of the pixel point with the smallest pixel value in the second region as the coordinates of the first center point of the reconstructed first end point frame, and according to the first width w of the first end point frame of the first adjacent image , the first height h, the first support frame confidence c ₁ , and the first end point frame confidence c ₂ perform corresponding parameter settings for the reconstructed first end point frame in the second region.

这里，因为支架端点在扫描时产生的影像图像颜色总是较周围环境深，又已知第一图像为灰度图，基于灰度图的颜色越深像素值(灰度值)越低的原理，我们可知在第一图像中支架端点的像素值应是周围图像范围中的最小值；所以，这里通过对第二区域内所有像素点的像素值进行查询就能精确定位端点的中心点位置，因此将第二区域中像素值最小的像素点坐标作为重构出的第一端点框的第一中心点坐标；对于重构的第一端点框的其他参数(第一宽度w、第一高度h、第一支架框置信度c ₁和第一端点框置信度c ₂)，则直接根据第一相邻图像中对应的第一端点框的对应参数进行设置即可。 Here, because the color of the image image generated by the end point of the bracket is always darker than the surrounding environment when scanning, and it is known that the first image is a grayscale image, based on the principle that the darker the color of the grayscale image, the lower the pixel value (gray value) , we know that the pixel value of the end point of the bracket in the first image should be the minimum value in the range of the surrounding image; therefore, the center point position of the end point can be precisely located by querying the pixel values of all the pixel points in the second area, Therefore, the pixel point coordinates with the minimum pixel value in the second area are used as the first center point coordinates of the reconstructed first end point frame; for other parameters of the reconstructed first end point frame (the first width w, the first The height h, the confidence degree c ₁ of the first bracket frame, and the confidence degree c ₂ of the first end point frame can be set directly according to the corresponding parameters of the corresponding first end point frame in the first adjacent image.

步骤5，对第一图像序列中第一支架框的数量为空且第一端点框的数量大于或等于2的第一图像，进行第一支架框重构处理； Step 5, performing first frame reconstruction processing on the first image in the first image sequence in which the number of first frame frames is empty and the number of first end frame frames is greater than or equal to 2;

这里，若第一图像序列中存在没有第一支架框但含有2个及2个以上第一端点框的第一图像，则通过对第一图像中相邻端点框构成的矩形的覆盖面积进行统计，并以其中的最小覆盖面积对应的两个第一端点框为参考，来实现对当前第一图像的支架框重构；Here, if there is a first image in the first image sequence that does not have the first bracket frame but contains two or more first end point frames, then the coverage area of the rectangle formed by the adjacent end point frames in the first image is calculated. Statistics, and using the two first end point frames corresponding to the minimum coverage area as a reference to realize the frame reconstruction of the current first image;

具体包括：步骤51，将第一图像序列中第一支架框的数量为空且第一端点框的数量大于或等于2的第一图像记为第二图像；It specifically includes: step 51, recording the first image in which the number of first bracket frames in the first image sequence is empty and the number of first end point frames is greater than or equal to 2 as a second image;

步骤52，在第二图像中任选一个第一端点框作为当前端点框，并将与之距离最近的第一端点框作为当前相邻端点框；Step 52, choose a first end point frame in the second image as the current end point frame, and use the first end point frame closest to it as the current adjacent end point frame;

这里，距离最近即为中心点连线最短，即当前端点框与当前相邻端点框的中心点连线最短；Here, the shortest distance is the shortest line between the center points, that is, the shortest line between the center points of the current endpoint frame and the current adjacent endpoint frame;

步骤53，取当前端点框的左上角顶点坐标记为第一左上坐标(x ₁₁,y ₁₁)，右下角顶点坐标记为第一右下坐标(x ₁₂,y ₁₂)；并取当前相邻端点框的左上角顶点坐标记为第二左上坐标(x ₂₁,y ₂₁)，右下角顶点坐标记为第二右下坐标(x ₂₂,y ₂₂)；并计算以第一左上坐标(x ₁₁,y ₁₁)作为左上角顶点坐标，以第二右下坐标(x ₂₂,y ₂₂)作为右下角顶点坐标的矩形区域的面积S，S＝|x ₂₂-x ₁₁|×|y ₂₂-y ₁₁|，||为绝对值计算符； Step 53, mark the vertex coordinates of the upper left corner of the current endpoint frame as the first upper left coordinate (x ₁₁ , y ₁₁ ), and mark the vertex coordinates of the lower right corner as the first lower right coordinate (x ₁₂ , y ₁₂ ); and take the current adjacent The vertex coordinates of the upper left corner of the endpoint frame are marked as the second upper left coordinate (x ₂₁ , y ₂₁ ), and the coordinates of the vertex of the lower right corner are marked as the second lower right coordinate (x ₂₂ , y ₂₂ ); and the calculation is based on the first upper left coordinate (x ₁₁ ,y ₁₁ ) as the coordinates of the upper left corner vertex, and the second lower right coordinate (x ₂₂ ,y ₂₂ ) as the area S of the rectangular area of the lower right corner vertex coordinates, S=|x ₂₂ -x ₁₁ |×|y ₂₂ -y ₁₁ |, || are absolute value calculators;

步骤54，在得到的多个面积S中，将面积最小的面积S对应的第一左上坐标(x ₁₁,y ₁₁)作为重构支架框的左上顶点坐标(x _l,y _l)，对应的第二右下坐标(x ₂₂,y ₂₂)作为重构支架框的右下顶点坐标(x _r,y _r)；将面积最小的面积S对应的两个第一端点框的第一端点框置信度c ₂进行加权平均计算生成重构支架框置信度； Step 54, among the multiple obtained areas S, use the first upper left coordinate (x ₁₁ , y ₁₁ ) corresponding to the area S with the smallest area as the upper left vertex coordinate (x _l , y _l ) of the reconstructed frame, and the corresponding The second lower right coordinate (x ₂₂ , y ₂₂ ) is used as the lower right vertex coordinate (x _r , y _r ) of the reconstructed bracket frame; the first endpoint of the two first endpoint frames corresponding to the area S with the smallest area Frame confidence c ₂ performs weighted average calculation to generate the frame confidence of the reconstructed bracket;

这里，因为第一端点框的第一端点框置信度c ₂越高，则对应的重构出来的第一支架框的支架框置信度越高，所以对两个第一端点框的第一端点框置信度c ₂进行加权平均计算得到重构支架框置信度； Here, because _the higher the confidence degree c of the first end point frame of the first end point frame is, the higher the confidence degree of the corresponding frame of the reconstructed first end point frame is, so for the two first end point frames The confidence degree _c of the first endpoint frame is calculated by weighted average to obtain the confidence degree of the reconstructed bracket frame;

步骤55，根据左上顶点坐标(x _l,y _l)和右下顶点坐标(x _r,y _r)进行支架框重构；将重构的第一支架框的第一中心点坐标设为由左上顶点坐标(x _l,y _l)和右下顶点坐标(x _r,y _r)构成的矩形框的中心点坐标；将重构的第一支架框的第一宽度w设为|x _r-x _l|+第二宽度微调阈值△w ₂；将重构的第一支架框的第一高度h设为|y _r-y _l|+第二高度微调阈值△h ₂；将重构的第一支架框的第一支架框置信度c ₁设为重构支架框置信度；将重构的第一支架框的第一端点框置信度c ₂设为低于第二阈值的概率值。 Step 55, perform frame reconstruction according to the upper left vertex coordinates (x _l , y _l ) and the lower right vertex coordinates (x _r , y _r ); set the coordinates of the first center point of the reconstructed first stent frame as The coordinates of the center point of the rectangular frame formed by the vertex coordinates (x _l , y _l ) and the lower right vertex coordinates (x _r , y _r ); set the first width w of the reconstructed first bracket frame to x _r -x _l |+ the second width fine-tuning threshold △w ₂ ; set the first height h of the reconstructed first frame to |y _r -y _l |+ the second height fine-tuning threshold △h ₂ ; set the reconstructed first The first stent frame confidence c ₁ of the stent frame is set as the reconstructed stent frame confidence level; the first end point frame confidence c ₂ of the reconstructed first stent frame is set as a probability value lower than the second threshold.

这里，第二宽度微调阈值△w ₂和第二高度微调阈值△h ₂为预先设定的***参数；第一支架框的范围内容包含了对应的两个第一端点框。 Here, the second width fine-tuning threshold Δw ₂ and the second height fine-tuning threshold Δh ₂ are preset system parameters; the scope content of the first bracket box includes the corresponding two first end point boxes.

步骤6，对第一图像序列中第一支架框的数量为空且第一端点框的数量小于2的第一图像，进行第二支架重构处理； Step 6, performing second bracket reconstruction processing on the first image in the first image sequence in which the number of first bracket frames is empty and the number of first end point frames is less than 2;

这里，若第一图像序列中存在没有第一支架框且包含的第一端点框的数量也未达到2个及2个以上的第一图像，通过复制与之相邻的带有第一支架框的第一图像中的支架框信息来实现对当前第一图像的支架框重构；Here, if there is no first stent frame in the first image sequence and the number of first end point frames contained in the first image does not reach 2 or more, by copying the adjacent frame with the first stent The bracket frame information in the first image of the frame is used to realize the bracket frame reconstruction of the current first image;

具体包括：步骤61，将第一图像序列中第一支架框的数量为空且第一端点框的数量小于2的第一图像记为第一无支架图像；It specifically includes: step 61, recording the first image in which the number of first stent frames in the first image sequence is empty and the number of first end point frames is less than 2 as the first stent-free image;

步骤62，将第一图像序列中第一支架框的数量不为空且与第一无支架图像距离最近的第一图像记为第二相邻图像；Step 62, the number of the first bracket frame in the first image sequence is not empty and the first image with the closest distance to the first bracket-free image is recorded as the second adjacent image;

这里，距离最近即为时间最近，与第一无支架图像距离最近即为与第一无支架图像时间最近；需要说明的是，若与第一无支架图像距离最近的第一支架框的数量不为空且与第一无支架图像距离最近的第一图像前后各有一张，则优选后者作为第二相邻图像；Here, the shortest distance is the shortest time, and the shortest distance to the first stent-free image is the shortest time to the first stent-free image; it should be noted that if the number of first stent frames closest to the first stent-free image is empty and there are one before and after the first image closest to the first stent-free image, then the latter is preferred as the second adjacent image;

步骤63，根据第二相邻图像的第一支架框的第一中心点坐标、第一宽度w和第一高度h，在第一无支架图像上的对应位置进行支架框重构；Step 63, according to the coordinates of the first center point, the first width w, and the first height h of the first bracket frame of the second adjacent image, perform bracket frame reconstruction at the corresponding position on the first bracket-free image;

具体包括：根据第二相邻图像的第一支架框的第一中心点坐标、第一宽度w、第一高度h、第一支架框置信度c ₁和第一端点框置信度c ₂对重构的第一支架框进行对应参数设置。 Concretely include: according to the first center point coordinates of the first support frame of the second adjacent image, the first width w, the first height h, the first support frame confidence degree c ₁ and the first end point frame confidence degree c ₂ pairs Corresponding parameter settings are performed on the reconstructed first support frame.

这里，直接根据第二相邻图像中第一支架框的参数(第一中心点坐标、第一宽度w、第一高度h、第一支架框置信度c ₁和第一端点框置信度c ₂)对重构的第一支架框的参数进行设置。 Here, directly according to the parameters of the first support frame in the second adjacent image (the coordinates of the first center point, the first width w, the first height h, the confidence degree c of the first support frame and the confidence degree _c of the first end point frame ₂ ) Setting parameters of the reconstructed first bracket frame.

步骤7，对第一图像序列进行不合理位置支架筛查处理； Step 7, performing unreasonable position bracket screening processing on the first image sequence;

这里，第一图像序列中的各个第一图像中的支架位置应前后关联性较强，不存在前后较大位移的情况；若出现前后位置变化过大的情况则需对不合理位置的第一支架框进行清除；Here, the positions of the brackets in each first image in the first image sequence should have a strong front-rear correlation, and there is no large front-rear displacement; bracket frame for removal;

具体包括：步骤71，对第一图像序列的所有第一支架框的第一中心点坐标的横坐标平均值进行计算得到第一x平均值；It specifically includes: step 71, calculating the average value of the abscissa of the coordinates of the first center point of all the first frame frames of the first image sequence to obtain the first average value of x;

例如，第一图像序列共有5张第一图像，每张第一图像中存在一个第一支架框，5个第一支架框第一中心点坐标依次为(100,200),(102,202),(150,180),(104,204),(106,206)；那么第一x平均值＝(100+102+150+104+106)≈112；For example, there are 5 first images in the first image sequence, and there is a first support frame in each first image, and the coordinates of the first center points of the 5 first support frames are (100, 200), (102, 202), (150, 180) ,(104,204),(106,206); then the first average value of x=(100+102+150+104+106)≈112;

步骤72，对第一图像序列的所有第一支架框的第一中心点坐标的纵坐标平均值进行计算得到第一y平均值；Step 72, calculating the y-coordinate mean value of the first center point coordinates of all the first frame frames of the first image sequence to obtain the first y mean value;

例如，第一图像序列的5个第一支架框第一中心点坐标依次为(100,200),(102,202),(150,180),(104,204),(106,206)；那么第一y平均值＝(200+202+180+204+206)≈198；For example, the coordinates of the first center points of the five first frame frames of the first image sequence are (100,200), (102,202), (150,180), (104,204), (106,206); then the first y average value=(200+ 202+180+204+206)≈198;

步骤73，在xy平面上，根据各个第一支架框的第一中心点坐标进行对应的中心点标记，并对由第一x平均值和第一y平均值构成的均值点进行标记；Step 73, on the xy plane, mark the corresponding center point according to the coordinates of the first center point of each first support frame, and mark the mean point formed by the first x mean value and the first y mean value;

其中，均值点的坐标为(第一x平均值,第一y平均值)；Among them, the coordinates of the mean point are (the first x average value, the first y average value);

例如，第一x平均值为112，第一y为198，那么均值点的坐标为(112,198)；For example, the average value of the first x is 112, and the first y is 198, then the coordinates of the average point are (112,198);

步骤74，以均值点为原点、以平行于x轴的直线为象限横轴、以平行于y轴的直线为象限纵轴，将xy平面划分为四个中心点象限；Step 74, taking the mean point as the origin, taking the straight line parallel to the x-axis as the quadrant horizontal axis, and taking the straight line parallel to the y-axis as the quadrant vertical axis, divide the xy plane into four center point quadrants;

例如，均值点的坐标为(112,198)，那么四个中心点象限从左上角顺时针到左下角的象限坐标分别为：左上中心点象限[x<112,y>198]，右上中心点象限[x>112,y>198]，右下中心点象限[x>112,y<198]，左下中心点象限[x<112,y<198]；For example, the coordinates of the mean point are (112,198), then the quadrant coordinates of the four center point quadrants clockwise from the upper left corner to the lower left corner are: upper left center point quadrant [x<112, y>198], upper right center point quadrant [ x>112, y>198], lower right center quadrant [x>112, y<198], lower left center quadrant [x<112, y<198];

步骤75，对四个中心点象限分别进行中心点统计处理，得到对应的象限中心点数量；Step 75, performing center point statistical processing on the four center point quadrants respectively to obtain the corresponding quadrant center point numbers;

例如，第一图像序列的5个第一支架框第一中心点坐标依次为(100,200),(102,202),(150,180),(104,204),(106,206)；四个中心点象限为：左上中心点象限[x<112,y>198]，右上中心点象限[x>112,y>198]，右下中心点象限[x>112,y<198]，左下中心点象限[x<112,y<198]；For example, the coordinates of the first center points of the five first support frames in the first image sequence are (100,200), (102,202), (150,180), (104,204), (106,206); the four center point quadrants are: upper left center point Quadrant [x<112, y>198], upper right center quadrant [x>112, y>198], lower right center quadrant [x>112, y<198], lower left center quadrant [x<112, y <198];

那么，满足左上中心点象限坐标要求的象限中心点数量为4，对应的中心点分别为：(100,200),(102,202),(104,204),(106,206)；满足右上中心点象限坐标要求的象限中心点数量为0；满足右下中心点象限坐标要求的象限中心点数量为1，对应的中心点为(150,180)；满足左下中心点象限坐标要求的象限中心点数量为0；Then, the number of quadrant center points that meet the quadrant coordinate requirements of the upper left center point is 4, and the corresponding center points are: (100,200), (102,202), (104,204), (106,206); the quadrant center that meets the quadrant coordinate requirements of the upper right center point The number of points is 0; the number of quadrant center points that meet the quadrant coordinate requirements of the lower right center point is 1, and the corresponding center point is (150,180); the number of quadrant center points that meet the quadrant coordinate requirements of the lower left center point is 0;

步骤76，将数量为最大值的象限中心点数量对应的中心点象限作为保留中心点象限；Step 76, taking the center point quadrant corresponding to the number of quadrant center points whose quantity is the maximum value as the reserved center point quadrant;

例如，第一图像序列的5个第一图像中的第一支架框的中心点，满足左上中心点象限坐标要求的象限中心点数量为4，满足右上中心点象限坐标要求的象限中心点数量为0，满足右下中心点象限坐标要求的象限中心点数量为1，满足左下中心点象限坐标要求的象限中心点数量为0；那么，数量为最大值的象限中心点数量应为满足左上中心点象限坐标要求的象限中心点数量4，则保留中心点象限应为上述四个中心点象限中的左上中心点象限；For example, for the center points of the first support frame in the five first images of the first image sequence, the number of quadrant center points meeting the quadrant coordinate requirement of the upper left center point is 4, and the number of quadrant center points meeting the quadrant coordinate requirement of the upper right center point is 0, the number of quadrant center points that meet the quadrant coordinate requirements of the lower right center point is 1, and the number of quadrant center points that meet the quadrant coordinate requirements of the lower left center point is 0; then, the number of quadrant center points with the maximum number should satisfy the upper left center point The number of quadrant center points required by the quadrant coordinates is 4, then the reserved center point quadrant should be the upper left center point quadrant of the above four center point quadrants;

步骤77，将不属于保留中心点象限的中心点对应的第一支架框，从第一图像序列中移除。Step 77: Remove the first support frame corresponding to the center point that does not belong to the quadrant of the reserved center point from the first image sequence.

例如，保留中心点象限为四个中心点象限中的左上中心点象限，那么位于右下中心点象限的中心点对应的第一支架框就属于不合理位置的第一支架框，应从所在的第一图像中删除。For example, if the center point quadrant is reserved as the upper left center point quadrant among the four center point quadrants, then the first support frame corresponding to the center point of the lower right center point quadrant belongs to the first support frame of an unreasonable position, and should be selected from the first support frame where it is located. One image is deleted.

需要说明的是，从第一图像序列中移除了不合理位置的第一支架框之后，可通过上述步骤5或6对移除了不合理位置支架框的第一图像进行支架框重构处理。It should be noted that after removing the first bracket frame at an unreasonable position from the first image sequence, the bracket frame reconstruction process can be performed on the first image from which the bracket frame at an unreasonable position is removed through the

above step

5 or 6 .

步骤8，对第一图像序列进行冗余支架筛查处理； Step 8, performing redundant bracket screening processing on the first image sequence;

具体包括：若第一图像序列的第一图像中存在多个第一支架框，则只保留其中第一支架框置信度c ₁为最大值的第一支架框。 Specifically, it includes: if there are multiple first bracket frames in the first image of the first image sequence, only retaining the first bracket frame in which the confidence c ₁ of the first bracket frame is the maximum value.

这里，对于在第一图像中存在多个第一支架框的情况，将其中最大支架置信度的第一支架框进行保留，也就是保留第一支架框置信度c ₁为最大值的第一支架框，对其他非最大支架置信度的第一支架框进行删除。 Here, for the case that there are multiple first bracket frames in the first image, the first bracket frame with the largest bracket confidence degree is retained, that is, the first bracket frame with the first bracket frame confidence degree c ₁ is the maximum value box, delete other first bracket boxes with non-maximum bracket confidence.

步骤9，对第一图像序列进行不合理位置端点筛查处理；Step 9, performing unreasonable position endpoint screening processing on the first image sequence;

具体包括：若第一图像序列的第一图像中存在多个处于第一支架框范围之外的第一端点框，则只保留处于第一支架框范围内的第一端点框。Specifically, it includes: if there are multiple first end point frames outside the range of the first stent frame in the first image of the first image sequence, only retain the first end point frames within the range of the first stent frame.

这里，对于在第一图像中存在第一端点框未能处于第一支架框图像范围内的这种情况，将其中处于第一支架框范围内的第一端点框进行保留，对其他处于第一支架框范围外的第一端点框进行删除。Here, for the situation that the first end point frame fails to be within the image range of the first bracket frame in the first image, the first end point frame within the range of the first bracket frame is retained, and the other The first end point box outside the range of the first bracket box is deleted.

步骤10，将完成重构与筛查的第一图像序列作为完成支架定位的造影图像序列。In step 10, the first image sequence completed with reconstruction and screening is used as the angiography image sequence completed with stent positioning.

这里，最终输出的造影图像序列是在步骤3得到的带有第一支架框、第一端点框标记的第一图像序列基础上，修正了前文提及的四种异常情况(1、在第一图像中，未能识别出第一端点框和第一支架框；2、在第一图像中，只识别出了多个第一端点框而无第一支架框；3、在第一图像中，识别出了多个第一支架框；4、在第一图像中，识别出的第一端点框未能处于第一支架框的图像范围内)的带有最优支架定位信息的造影图像序列。Here, the final output contrast image sequence is based on the first image sequence with the first stent frame and the first end point frame mark obtained in step 3, and the four abnormalities mentioned above are corrected (1. In an image, the first end point frame and the first stent frame cannot be identified; 2. In the first image, only a plurality of first end point frames are recognized without the first stent frame; 3. In the first In the image, a plurality of first stent frames are identified; 4. In the first image, the identified first end point frame fails to be within the image range of the first stent frame) with optimal stent positioning information Contrast image sequence.

图2为本发明实施例二提供的一种对冠脉造影图像进行支架定位的处理装置的模块结构图，该装置可以为实现本发明实施例方法的终端设备或者服务器，也可以为与上述终端设备或者服务器连接的实现本发明实施例方法的装置，例如该装置可以是上述终端设备或者服务器的装置或芯片***。如图2所示，该装置包括：获取模块201、目标检测网络处理模块202、定位信息修正模块203和输出模块204。Fig. 2 is a module structure diagram of a processing device for stent positioning on coronary angiography images provided by Embodiment 2 of the present invention. The device may be a terminal device or a server for realizing the method of the embodiment of the present invention, or it may be the same as the above-mentioned terminal The device or the device connected to the server to implement the method of the embodiment of the present invention, for example, the device may be the device or the chip system of the above-mentioned terminal device or server. As shown in FIG. 2 , the device includes: an acquisition module 201 , a target detection network processing module 202 , a positioning information correction module 203 and an output module 204 .

获取模块201用于获取连续的冠脉造影图像生成第一图像序列；第一图像序列包括多个第一图像。The acquiring module 201 is configured to acquire continuous coronary angiography images to generate a first image sequence; the first image sequence includes a plurality of first images.

目标检测网络处理模块202用于使用训练成熟的目标检测网络，对第一图像进行目标检测处理得到多个第一识别框；第一识别框包括第一支架框置信度c ₁和第一端点框置信度c ₂。 The target detection network processing module 202 is used to use a well-trained target detection network to perform target detection processing on the first image to obtain a plurality of first recognition frames; the first recognition frames include the first support frame confidence c ₁ and the first endpoint Box confidence c ₂ .

定位信息修正模块203用于将第一图像中，第一支架框置信度c ₁超过预设的第一阈值的第一识别框记为第一支架框，第一端点框置信度c ₂超过预设的第二阈值的第一识别框记为第一端点框；并对第一图像序列中第一端点框的数量为空的第一图像，进行第一端点重构处理得到对应的第一端点框；并对第一图像序列中第一支架框的数量为空且第一端点框的数量大于或等于2的第一图像，进行第一支架框重构处理得到对应的第一支架框；并对第一图像序列中第一支架框的数量为空且第一端点框的数量小于2的第一图像，进行第二支架重构处理得到对应的第一支架框；并对第一图像序列进行不合理位置支架筛查处理；并对第一图像序列进行冗余支架筛查处理；并对第一图像序列进行不合理位置端点筛查处理。 The positioning information correction module 203 is used to record the first recognition frame whose confidence degree _c of the first support frame exceeds the preset first threshold in the first image as the first support frame, and the first end point frame confidence degree _c2 exceeds The first recognition frame of the preset second threshold value is marked as the first endpoint frame; and for the first image in the first image sequence, the number of first endpoint frames is empty, the first endpoint reconstruction process is performed to obtain the corresponding The first end point frame; and for the first image in the first image sequence, the number of the first bracket frame is empty and the number of the first end point frame is greater than or equal to 2, the first bracket frame reconstruction process is performed to obtain the corresponding The first bracket frame; and for the first image in which the number of the first bracket frame in the first image sequence is empty and the number of the first end point frame is less than 2, the second bracket reconstruction process is performed to obtain the corresponding first bracket frame; and performing unreasonable position bracket screening processing on the first image sequence; performing redundant bracket screening processing on the first image sequence; and performing unreasonable position endpoint screening processing on the first image sequence.

输出模块204用于将完成重构与筛查的第一图像序列作为完成支架定位的造影图像序列。The output module 204 is configured to use the first sequence of images completed with reconstruction and screening as a sequence of angiography images completed with positioning of the stent.

本发明实施例提供的一种对冠脉造影图像进行支架定位的处理装置，可以执行上述方法实施例中的方法步骤，其实现原理和技术效果类似，在此不再赘述。An embodiment of the present invention provides a processing device for stent positioning on coronary angiography images, which can execute the method steps in the above method embodiments, and its implementation principle and technical effect are similar, and will not be repeated here.

需要说明的是，应理解以上装置的各个模块的划分仅仅是一种逻辑功能的划分，实际实现时可以全部或部分集成到一个物理实体上，也可以物理上分开。且这些模块可以全部以软件通过处理元件调用的形式实现；也可以全部以硬件的形式实现；还可以部分模块通过处理元件调用软件的形式实现，部分模块通过硬件的形式实现。例如，获取模块可以为单独设立的处理元件，也可以集成在上述装置的某一个芯片中实现，此外，也可以以程序代码的形式存储于上述装置的存储器中，由上述装置的某一个处理元件调用并执行以上确定模块的功能。其它模块的实现与之类似。此外这些模块全部或部分可以集成在一起，也可以独立实现。这里所描述的处理元件可以是一种集成电路，具有信号的处理能力。在实现过程中，上述方法的各步骤或以上各个模块可以通过处理器元件中的硬件的集成逻辑电路或者软件形式的指令完成。It should be noted that it should be understood that the division of each module of the above device is only a division of logical functions, and may be fully or partially integrated into one physical entity or physically separated during actual implementation. And these modules can all be implemented in the form of calling software through processing elements; they can also be implemented in the form of hardware; some modules can also be implemented in the form of calling software through processing elements, and some modules can be implemented in the form of hardware. For example, the acquisition module can be a separate processing element, or it can be integrated into a chip of the above-mentioned device. In addition, it can also be stored in the memory of the above-mentioned device in the form of program code, and a certain processing element of the above-mentioned device can Call and execute the functions of the modules identified above. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, and can also be implemented independently. The processing element described here may be an integrated circuit with signal processing capability. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or an instruction in the form of software.

例如，以上这些模块可以是被配置成实施以上方法的一个或多个集成电路，例如：一个或多个特定集成电路(Application Specific Integrated Circuit，ASIC)，或，一个或多个数字信号处理器(Digital Signal Processor，DSP)，或，一个或者多个现场可编程门阵列(Field Programmable Gate Array，FPGA)等。再如，当以上某个模块通过处理元件调度程序代码的形式实现时，该处理元件可以是通用处理器，例如中央处理器(Central Processing Unit，CPU)或其它可以调用程序代码的处理器。再如，这些模块可以集成在一起，以片上***(System-on-a-chip，SOC)的形式实现。For example, the above modules may be one or more integrated circuits configured to implement the above method, for example: one or more specific integrated circuits (Application Specific Integrated Circuit, ASIC), or, one or more digital signal processors ( Digital Signal Processor, DSP), or, one or more Field Programmable Gate Arrays (Field Programmable Gate Array, FPGA), etc. For another example, when one of the above modules is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (Central Processing Unit, CPU) or other processors that can call program codes. For another example, these modules can be integrated together and implemented in the form of a System-on-a-chip (SOC).

在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。该计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行该计算机程序指令时，全部或部分地产生按照本发明实施例所描述的流程或功能。上述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。上述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，上述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线路(Digital Subscriber Line，DSL))或无线(例如红外、无线、蓝牙、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。上述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。上述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘(solid state disk，SSD))等。In the above embodiments, all or part of them may be implemented by software, hardware, firmware or any combination thereof. When implemented using software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, all or part of the processes or functions described in accordance with the embodiments of the present invention will be generated. The above-mentioned computers may be general-purpose computers, special-purpose computers, computer networks, or other programmable devices. The above-mentioned computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. (such as coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.) to another website site, computer, server or data center. The above-mentioned computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center integrated with one or more available media. The above-mentioned usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, or a magnetic tape), an optical medium (for example, DVD), or a semiconductor medium (for example, a solid state disk (solid state disk, SSD)) and the like.

图3为本发明实施例三提供的一种电子设备的结构示意图。该电子设备可以为前述的终端设备或者服务器，也可以为与前述终端设备或者服务器连接的实现本发明实施例方法的终端设备或服务器。如图3所示，该电子设备可以包括：处理器301(例如CPU)、存储器302、收发器303；收发器303耦合至处理器301，处理器301控制收发器303的收发动作。存储器302中可以存储各种指令，以用于完成各种处理功能以及实现本发明上述实施例中提供的方法和处理过程。优选的，本发明实施例涉及的电子设备还包括：电源304、***总线305以及通信端口306。***总线305用于实现元件之间的通信连接。上述通信端口306用于电子设备与其他外设之间进行连接通信。FIG. 3 is a schematic structural diagram of an electronic device provided by Embodiment 3 of the present invention. The electronic device may be the aforementioned terminal device or server, or may be a terminal device or server connected to the aforementioned terminal device or server to implement the method of the embodiment of the present invention. As shown in FIG. 3 , the electronic device may include: a processor 301 (such as a CPU), a memory 302 , and a transceiver 303 ; Various instructions may be stored in the memory 302 for completing various processing functions and realizing the methods and processing procedures provided in the above-mentioned embodiments of the present invention. Preferably, the electronic device involved in this embodiment of the present invention further includes: a power supply 304 , a system bus 305 and a communication port 306 . The system bus 305 is used to realize the communication connection among the components. The above-mentioned communication port 306 is used for connection and communication between the electronic device and other peripheral devices.

在图3中提到的***总线可以是外设部件互连标准(Peripheral Component Interconnect，PCI)总线或扩展工业标准结构(Extended Industry Standard Architecture，EISA)总线等。该***总线可以分为地址总线、数据总线、控制总线等。为便于表示，图中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。通信接口用于实现数据库访问装置与其他设备(例如客户端、读写库和只读库)之间的通信。存储器可能包含随机存取存储器(Random Access Memory，RAM)，也可能还包括非易失性存储器(Non- Volatile Memory)，例如至少一个磁盘存储器。The system bus mentioned in FIG. 3 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus or the like. The system bus can be divided into address bus, data bus, control bus and so on. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to realize the communication between the database access device and other devices (such as client, read-write library and read-only library). The memory may include random access memory (Random Access Memory, RAM), and may also include non-volatile memory (Non-Volatile Memory), such as at least one disk memory.

上述的处理器可以是通用处理器，包括中央处理器CPU、网络处理器(Network Processor，NP)等；还可以是数字信号处理器DSP、专用集成电路ASIC、现场可编程门阵列FPGA或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。Above-mentioned processor can be general-purpose processor, comprises central processing unit CPU, network processor (Network Processor, NP) etc.; Can also be digital signal processor DSP, application-specific integrated circuit ASIC, field programmable gate array FPGA or other available Program logic devices, discrete gate or transistor logic devices, discrete hardware components.

需要说明的是，本发明实施例还提供一种计算机可读存储介质，该存储介质中存储有指令，当其在计算机上运行时，使得计算机执行上述实施例中提供的方法和处理过程。It should be noted that the embodiments of the present invention also provide a computer-readable storage medium, and instructions are stored in the storage medium, and when the storage medium is run on a computer, the computer executes the methods and processing procedures provided in the above-mentioned embodiments.

本发明实施例还提供一种运行指令的芯片，该芯片用于执行上述实施例中提供的方法和处理过程。The embodiment of the present invention also provides a chip for running instructions, and the chip is used for executing the method and the processing procedure provided in the foregoing embodiments.

专业人员应该还可以进一步意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、计算机软件或者二者的结合来实现，为了清楚地说明硬件和软件的可互换性，在上述说明中已经按照功能一般性地描述了各示例的组成及步骤。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本发明的范围。Professionals should further realize that the units and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the relationship between hardware and software Interchangeability. In the above description, the composition and steps of each example have been generally described according to their functions. Whether these functions are executed by hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but such implementation should not be regarded as exceeding the scope of the present invention.

结合本文中所公开的实施例描述的方法或算法的步骤可以用硬件、处理器执行的软件模块，或者二者的结合来实施。软件模块可以置于随机存储器(RAM)、内存、只读存储器(ROM)、电可编程ROM、电可擦除可编程ROM、寄存器、硬盘、可移动磁盘、CD-ROM、或技术领域内所公知的任意其它形式的存储介质中。The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented by hardware, software modules executed by a processor, or a combination of both. Software modules can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other Any other known storage medium.

以上所述的具体实施方式，对本发明的目的、技术方案和有益效果进行了进一步详细说明，所应理解的是，以上所述仅为本发明的具体实施方式而已，并不用于限定本发明的保护范围，凡在本发明的精神和原则之内，所做的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention and are not intended to limit the scope of the present invention. Protection scope, within the spirit and principles of the present invention, any modification, equivalent replacement, improvement, etc., shall be included in the protection scope of the present invention.

Claims

一种对冠脉造影图像进行支架定位的处理方法，其特征在于，所述方法包括：A method for processing stent positioning on coronary angiography images, characterized in that the method comprises:

获取连续的冠脉造影图像生成第一图像序列；所述第一图像序列包括多个第一图像；Acquiring continuous coronary angiography images to generate a first image sequence; the first image sequence includes a plurality of first images;

使用训练成熟的目标检测网络，对所述第一图像进行目标检测处理得到多个第一识别框；所述第一识别框包括第一支架框置信度c ₁和第一端点框置信度c ₂； Using a well-trained target detection network, perform target detection processing on the first image to obtain a plurality of first recognition frames; the first recognition frames include the first support frame confidence c ₁ and the first end point frame confidence c ₂ ;

将所述第一图像中，所述第一支架框置信度c ₁超过预设的第一阈值的所述第一识别框记为第一支架框，所述第一端点框置信度c ₂超过预设的第二阈值的所述第一识别框记为第一端点框； In the first image, the first recognition frame whose confidence degree c ₁ of the first support frame exceeds a preset first threshold is marked as a first support frame, and the first end point frame confidence degree c ₂ The first identified frame exceeding the preset second threshold is marked as a first endpoint frame;

对所述第一图像序列中所述第一端点框的数量为空的所述第一图像，进行第一端点重构处理；performing a first endpoint reconstruction process on the first image in the first image sequence in which the number of the first endpoint frames is empty;

对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像，进行第一支架框重构处理；Performing first stent frame reconstruction processing on the first image in the first image sequence in which the number of the first stent frames is empty and the number of the first end point frames is greater than or equal to 2;

对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像，进行第二支架重构处理；performing a second stent reconstruction process on the first image in the first image sequence in which the number of the first stent frames is empty and the number of the first end point frames is less than 2;

对所述第一图像序列进行不合理位置支架筛查处理；performing unreasonable location bracket screening processing on the first image sequence;

对所述第一图像序列进行冗余支架筛查处理；performing redundant scaffold screening processing on the first image sequence;

对所述第一图像序列进行不合理位置端点筛查处理；performing unreasonable position endpoint screening processing on the first image sequence;

将完成重构与筛查的所述第一图像序列作为完成支架定位的造影图像序列。The first sequence of images completed with reconstruction and screening is used as the sequence of angiography images completed with positioning of the stent.
根据权利要求1所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述目标检测网络为YOLOv3神经网络。The method for processing stent positioning on coronary angiography images according to claim 1, wherein the target detection network is a YOLOv3 neural network.
根据权利要求1所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述第一识别框还包括第一中心点坐标、第一宽度w和第一高度h。The processing method for positioning a stent on a coronary angiography image according to claim 1, wherein the first identification frame further includes coordinates of a first center point, a first width w, and a first height h.
根据权利要求3所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述对所述第一图像序列中所述第一端点框的数量为空的所述第一图像，进行第一端点重构处理，具体包括：The method for processing stent positioning on coronary angiography images according to claim 3, wherein the first image in which the number of the first endpoint frames in the first image sequence is empty , perform the first endpoint reconstruction process, specifically including:

将所述第一图像序列中所述第一端点框的数量为空的所述第一图像记为第一无端点图像；Recording the first image in the first image sequence in which the number of the first endpoint boxes is empty is the first no-endpoint image;

将所述第一图像序列中所述第一端点框的数量不为空且与所述第一无端点图像距离最近的所述第一图像记为第一相邻图像；Recording the first image in the first image sequence whose number of the first endpoint frame is not empty and whose distance is closest to the first image without endpoints is the first adjacent image;

根据所述第一相邻图像的所述第一端点框的所述第一中心点坐标、所述第一宽度w和所述第一高度h，在所述第一无端点图像上的对应位置定位第一区域；所述第一区域的中心点坐标与所述第一相邻图像的所述第一端点框的所述第一中心点坐标对应，所述第一区域的宽度与所述第一相邻图像的所述第一端点框的所述第一宽度w对应，所述第一区域的高度与所述第一相邻图像的所述第一端点框的所述第一高度h对应；According to the first center point coordinates, the first width w, and the first height h of the first endpoint frame of the first adjacent image, the correspondence on the first non-endpoint image positioning the first area; the center point coordinates of the first area correspond to the first center point coordinates of the first end point frame of the first adjacent image, and the width of the first area corresponds to the first end point frame of the first adjacent image. corresponds to the first width w of the first end point frame of the first adjacent image, and the height of the first region corresponds to the first width w of the first end point frame of the first adjacent image. A height h corresponds to;

根据预设的第一宽度微调阈值△w ₁和第一高度微调阈值△h ₁，对所述第一区域进拓展得到第二区域；所述第二区域的中心点坐标与所述第一区域的中心点坐标对应，所述第二区域的宽度为所述第一区域的宽度与所述第一宽度微调阈值△w ₁相加的和，所述第二区域的高度为所述第一区域的高度与所述第一高度微调阈值△h ₁相加的和； According to the preset first width fine-tuning threshold Δw ₁ and first height fine-tuning threshold Δh ₁ , expand the first area to obtain a second area; the center point coordinates of the second area are consistent with the first area Corresponding to the coordinates of the central point of , the width of the second area is the sum of the width of the first area and the first width fine-tuning threshold Δw ₁ , and the height of the second area is the sum of the first area The sum of the height and the first height fine-tuning threshold Δh ₁ ;

在所述第二区域中进行端点框重构；将所述第二区域中像素值最小的像素点坐标作为重构出的所述第一端点框的所述第一中心点坐标，并根据所述第一相邻图像的所述第一端点框的所述第一宽度w、所述第一高度h、所述第一支架框置信度c ₁和所述第一端点框置信度c ₂对所述第二区域中重构出的所述第一端点框进行对应参数设置。 Perform endpoint frame reconstruction in the second area; use the coordinate of the pixel point with the smallest pixel value in the second area as the first center point coordinate of the reconstructed first end point frame, and according to The first width w, the first height h, the first bracket frame confidence _c1 and the first end point frame confidence of the first adjacent image c ₂ performing corresponding parameter setting on the first end point frame reconstructed in the second area.
根据权利要求3所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像，进行第一支架框重构处理，具体包括：The method for processing stent positioning on coronary angiography images according to claim 3, wherein the number of the first stent frames in the first image sequence is empty and the first endpoint For the first image whose number of frames is greater than or equal to 2, perform frame reconstruction processing on the first bracket, specifically including:

将所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像记为第二图像；recording the first image in which the number of the first bracket frame in the first image sequence is empty and the number of the first end point frame is greater than or equal to 2 as a second image;

在所述第二图像中任选一个所述第一端点框作为当前端点框，并将与之距离最近的所述第一端点框作为当前相邻端点框；Selecting one of the first end point frames in the second image as the current end point frame, and using the first end point frame closest to it as the current adjacent end point frame;

取所述当前端点框的左上角顶点坐标记为第一左上坐标(x ₁₁,y ₁₁)，右下角顶点坐标记为第一右下坐标(x ₁₂,y ₁₂)；并取所述当前相邻端点框的左上角顶点坐标记为第二左上坐标(x ₂₁,y ₂₁)，右下角顶点坐标记为第二右下坐标(x ₂₂,y ₂₂)；并计算以所述第一左上坐标(x ₁₁,y ₁₁)作为左上角顶点坐标，以所述第二右下坐标(x ₂₂,y ₂₂)作为右下角顶点坐标的矩形区域的面积S，S＝|x ₂₂-x ₁₁|×|y ₂₂-y ₁₁|，||为绝对值计算符； Take the vertex coordinates of the upper left corner of the current endpoint frame and mark it as the first upper left coordinate (x ₁₁ , y ₁₁ ), and mark the vertex coordinates of the lower right corner as the first lower right coordinate (x ₁₂ , y ₁₂ ); and take the current phase The vertex coordinates of the upper left corner of the adjacent endpoint frame are marked as the second upper left coordinate (x ₂₁ , y ₂₁ ), and the coordinates of the vertex of the lower right corner are marked as the second lower right coordinate (x ₂₂ , y ₂₂ ); and the calculation is based on the first upper left coordinate (x ₁₁ , y ₁₁ ) as the coordinates of the upper left corner vertex, and the second lower right coordinate (x ₂₂ , y ₂₂ ) as the area S of the rectangular area of the lower right corner vertex coordinates, S=|x ₂₂ -x ₁₁ |× |y ₂₂ -y ₁₁ |, || is an absolute value operator;

在得到的多个所述面积S中，将面积最小的所述面积S对应的所述第一左上坐标(x ₁₁,y ₁₁)作为重构支架框的左上顶点坐标(x _l,y _l)，对应的所述第二右下坐标(x ₂₂,y ₂₂)作为重构支架框的右下顶点坐标(x _r,y _r)；将面积最小的所述面积S对应的两个所述第一端点框的所述第一端点框置信度c ₂进行加权平均计算生成重构支架框置信度； Among the multiple obtained areas S, the first upper left coordinate (x ₁₁ , y ₁₁ ) corresponding to the area S with the smallest area is used as the upper left vertex coordinate (x _l , y _l ) of the reconstructed scaffold frame , the corresponding second lower right coordinates (x ₂₂ , y ₂₂ ) are used as the lower right vertex coordinates (x _r , y _r ) of the reconstructed frame; The confidence degree _c of the first endpoint frame of an endpoint frame is calculated by weighted average to generate the confidence degree of the reconstructed bracket frame;

根据所述左上顶点坐标(x _l,y _l)和所述右下顶点坐标(x _r,y _r)进行支架框重构；将重构的所述第一支架框的所述第一中心点坐标设为由所述左上顶点坐标(x _l,y _l)和所述右下顶点坐标(x _r,y _r)构成的矩形框的中心点坐标；将重构的所述第一支架框的所述第一宽度w设为|x _r-x _l|；将重构的所述第一支架框的所述第一高度h设为|y _r-y _l|；将重构的所述第一支架框的所述第一支架框置信度c ₁设为所述重构支架框置信度；将重构的所述第一支架框的所述第一端点框置信度c ₂设为低于所述第二阈值的概率值。 Reconstruction of the stent frame according to the coordinates of the upper left vertex (x _l , y _l ) and the coordinates of the lower right vertex (x _r , y _r ); the first central point of the reconstructed first stent frame The coordinates are set as the coordinates of the center point of the rectangular frame formed by the coordinates of the upper left vertex (x _l , y _l ) and the coordinates of the lower right vertex (x _r , y _r ); The first width w is set to |x _r -x _l |; the first height h of the reconstructed first stent frame is set to |y _r -y _l |; the reconstructed first The first stent frame confidence level _c1 of a stent frame is set as the reconstructed stent frame confidence level; the first endpoint frame confidence _c2 of the reconstructed first stent frame is set as low at the probability value of the second threshold.
根据权利要求3所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像，进行第二支架重构处理，具体包括：The method for processing stent positioning on coronary angiography images according to claim 3, wherein the number of the first stent frames in the first image sequence is empty and the first endpoint The first image whose number of frames is less than 2 is subjected to a second bracket reconstruction process, which specifically includes:

将所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像记为第一无支架图像；Recording the first image in which the number of the first stent frame in the first image sequence is empty and the number of the first end point frame is less than 2 as a first stent-free image;

将所述第一图像序列中所述第一支架框的数量不为空且与所述第一无支架图像距离最近的所述第一图像记为第二相邻图像；Recording the first image whose number of the first stent frame in the first image sequence is not empty and which is the closest to the first stent-free image as a second adjacent image;

根据所述第二相邻图像的所述第一支架框的所述第一中心点坐标、所述第一宽度w和所述第一高度h，在所述第一无支架图像上的对应位置进行支架框重构；根据所述第二相邻图像的所述第一支架框的所述第一中心点坐标、所述第一宽度w、所述第一高度h、所述第一支架框置信度c ₁和所述第一端点框置信度c ₂对重构的所述第一支架框进行对应参数设置。 According to the first central point coordinates, the first width w, and the first height h of the first stent frame of the second adjacent image, the corresponding position on the first stent-free image Carry out frame reconstruction; according to the coordinates of the first center point of the first frame of the second adjacent image, the first width w, the first height h, the first frame Confidence c ₁ and the first endpoint frame confidence c ₂ are used to set corresponding parameters for the reconstructed first bracket frame.
根据权利要求3所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述对所述第一图像序列进行不合理位置支架筛查处理，具体包括：The method for processing stent positioning on coronary angiography images according to claim 3, characterized in that the screening of stents at unreasonable positions on the first image sequence specifically includes:

对所述第一图像序列的所有所述第一支架框的所述第一中心点坐标的横坐标平均值进行计算得到第一x平均值；calculating the average value of the abscissa of the coordinates of the first central point of all the first bracket frames in the first image sequence to obtain a first average value of x;

对所述第一图像序列的所有所述第一支架框的所述第一中心点坐标的纵坐标平均值进行计算得到第一y平均值；calculating the average value of the ordinates of the coordinates of the first central point of all the first bracket frames in the first image sequence to obtain a first average value of y;

在xy平面上，根据各个所述第一支架框的所述第一中心点坐标进行对应的中心点标记，并对由所述第一x平均值和所述第一y平均值构成的均值点进行标记；所述均值点的坐标为(第一x平均值,第一y平均值)On the xy plane, mark the corresponding center point according to the coordinates of the first center point of each of the first support frames, and set the mean value point composed of the first mean value of x and the mean value of the first y Marking; the coordinates of the mean point are (the first x average value, the first y average value)

以所述均值点为原点、以平行于x轴的直线为象限横轴、以平行于y轴的直线为象限纵轴，将xy平面划分为四个中心点象限；Taking the mean point as the origin, taking the straight line parallel to the x-axis as the quadrant horizontal axis, and taking the straight line parallel to the y-axis as the quadrant vertical axis, divide the xy plane into four center point quadrants;

对所述四个中心点象限分别进行中心点统计处理，得到对应的象限中心点数量；Carry out center point statistical processing respectively to described four center point quadrants, obtain the corresponding quadrant center point quantity;

将数量为最大值的所述象限中心点数量对应的所述中心点象限作为保留中心点象限；The center point quadrant corresponding to the number of the quadrant center points whose quantity is the maximum value is used as the reserved center point quadrant;

将不属于所述保留中心点象限的中心点对应的所述第一支架框，从所述第一图像序列中移除。removing the first support frame corresponding to the center point not belonging to the quadrant of the reserved center point from the first image sequence.
根据权利要求1所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述对所述第一图像序列进行冗余支架筛查处理，具体包括：The method for processing stent positioning on coronary angiography images according to claim 1, wherein said performing redundant stent screening processing on said first image sequence specifically includes:

若所述第一图像序列的所述第一图像中存在多个所述第一支架框，则只保留其中所述第一支架框置信度c ₁为最大值的所述第一支架框。 If there are multiple first stent frames in the first image of the first image sequence, only retain the first stent frame in which the confidence c ₁ of the first stent frame is the maximum value.
根据权利要求1所述的对冠脉造影图像进行支架定位的处理方法，其特征在于，所述对所述第一图像序列进行不合理位置端点筛查处理，具体包括：The method for processing stent positioning on coronary angiography images according to claim 1, wherein the screening of unreasonable position endpoints on the first image sequence specifically includes:

若所述第一图像序列的所述第一图像中存在多个处于所述第一支架框范围之外的所述第一端点框，则只保留处于所述第一支架框范围内的所述第一端点框。If there are multiple first end point frames outside the range of the first bracket frame in the first image of the first image sequence, only keep all the frames within the range of the first bracket frame Describe the first endpoint box.
一种用于实现权利要求1-9任一项所述的对冠脉造影图像进行支架定位的处理方法步骤的装置，其特征在于，所述装置包括：获取模块、目标检测网络处理模块、定位信息修正模块和输出模块；A device for realizing the steps of the processing method for stent positioning on coronary angiography images according to any one of claims 1-9, characterized in that the device comprises: an acquisition module, a target detection network processing module, a positioning Information correction module and output module;

所述获取模块用于获取连续的冠脉造影图像生成第一图像序列；所述第一图像序列包括多个第一图像；The acquisition module is used to acquire continuous coronary angiography images to generate a first image sequence; the first image sequence includes a plurality of first images;

所述目标检测网络处理模块用于使用训练成熟的目标检测网络，对所述第一图像进行目标检测处理得到多个第一识别框；所述第一识别框包括第一支架框置信度c ₁和第一端点框置信度c ₂； The target detection network processing module is used to use a well-trained target detection network to perform target detection processing on the first image to obtain a plurality of first recognition frames; the first recognition frames include a first support frame confidence c ₁ and the first endpoint box confidence c ₂ ;

所述定位信息修正模块用于将所述第一图像中，所述第一支架框置信度c ₁超过预设的第一阈值的所述第一识别框记为第一支架框，所述第一端点框置信度c ₂超过预设的第二阈值的所述第一识别框记为第一端点框；并对所述第一图像序列中所述第一端点框的数量为空的所述第一图像，进行第一端点重构处理；并对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量大于或等于2的所述第一图像，进行第一支架框重构处理；并对所述第一图像序列中所述第一支架框的数量为空且所述第一端点框的数量小于2的所述第一图像，进行第二支架重构处理；并对所述第一图像序列进行不合理位置支架筛查处理；并对所述第一图像序列进行冗余支架筛查处理；并对所述第一图像序列进行不合理位置端点筛查处理； The positioning information correction module is used to mark the first recognition frame whose confidence level _c1 of the first support frame exceeds a preset first threshold in the first image as a first support frame, and the first support frame The first recognition frame whose confidence degree _c of an endpoint frame exceeds the preset second threshold is marked as a first endpoint frame; and the number of the first endpoint frame in the first image sequence is empty The first image of the first end point reconstruction process; and the number of the first bracket frame in the first image sequence is empty and the number of the first end point frame is greater than or equal to 2 For the first image, perform first bracket frame reconstruction processing; and for the first frame in the first image sequence, the number of the first bracket frame is empty and the number of the first end point frame is less than 2. performing a second bracket reconstruction process on an image; and performing an unreasonable position bracket screening process on the first image sequence; and performing a redundant bracket screening process on the first image sequence; and performing a redundant bracket screening process on the first image sequence; The image sequence is screened for unreasonable position endpoints;

所述输出模块用于将完成重构与筛查的所述第一图像序列作为完成支架定位的造影图像序列。The output module is configured to use the first sequence of images completed with reconstruction and screening as a sequence of contrast images completed with stent positioning.
一种电子设备，其特征在于，包括：存储器、处理器和收发器；An electronic device, characterized in that it includes: a memory, a processor, and a transceiver;

所述处理器用于与所述存储器耦合，读取并执行所述存储器中的指令，以实现权利要求1-9任一项所述的方法步骤；The processor is configured to be coupled to the memory, read and execute instructions in the memory, so as to implement the method steps described in any one of claims 1-9;

所述收发器与所述处理器耦合，由所述处理器控制所述收发器进行消息收发。The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.
一种计算机可读存储介质，其特征在于，所述计算机可读存储介质存储有计算机指令，当所述计算机指令被计算机执行时，使得所述计算机执行权利要求1-9任一项所述的方法的指令。A computer-readable storage medium, characterized in that the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer executes the method described in any one of claims 1-9. method directive.