WO2020252900A1

WO2020252900A1 - Fundus oct picture-based lesion extraction method, apparatus and device, and storage medium

Info

Publication number: WO2020252900A1
Application number: PCT/CN2019/102545
Authority: WO
Inventors: 王俊; 高鹏; 吕彬; 谢国彤
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-06-18
Filing date: 2019-08-26
Publication date: 2020-12-24
Also published as: CN110390667A; CN110390667B

Abstract

A fundus optical coherence tomography (OCT) picture-based lesion extraction method, apparatus and device, and a storage medium, being used for performing segmentation processing on a lesion region of a fundus OCT picture, avoiding problems of incomplete groove filling of a lesion region and false positives. The lesion extraction method comprises: acquiring an original fundus OCT picture (101); performing segmentation pre-processing of the original fundus OCT picture to obtain a pre-processed fundus OCT image (102); generating an initial closed curve in the pre-processed fundus OCT image, the interior of the initial closed curve comprising a target fundus lesion (103); determining an initial curve function of the initial closed curve by means of a pre-set morphological serpentine curve evolution algorithm (104); evolving the initial closed curve according to the initial curve function to obtain a target closed curve, a target curve function of the target closed curve representing the contour of the target fundus lesion (105); and performing groove filling of the interior of the target closed curve to obtain a lesion image of the target fundus lesion (106).

Description

基于眼底OCT影像的病灶提取方法、装置、设备及存储介质Lesion extraction method, device, equipment and storage medium based on fundus OCT image

本申请要求于2019年6月18日提交中国专利局、申请号为201910524302.1、发明名称为“基于眼底OCT影像的病灶提取方法、装置、设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在申请中。This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on June 18, 2019, the application number is 201910524302.1, and the invention title is "Methods, devices, equipment and storage media for lesion extraction based on fundus OCT images", all of which The content is incorporated in the application by reference.

技术领域Technical field

本申请涉及区域提取领域，尤其涉及基于眼底OCT影像的病灶提取方法、装置、设备及存储介质。This application relates to the field of region extraction, and in particular to a method, device, equipment, and storage medium for lesion extraction based on fundus OCT images.

背景技术Background technique

随着人工智能的快速发展，人工智能的应用领域也越来越多，人工智能在医疗领域也得到了广泛应用。人工智能在医疗领域的应用经常面临医疗影像数据少，各类医疗影像数据不均衡的问题。With the rapid development of artificial intelligence, there are more and more applications of artificial intelligence, and artificial intelligence has also been widely used in the medical field. The application of artificial intelligence in the medical field often faces the problem of insufficient medical image data and imbalance of various medical image data.

由于光学相干断层扫描技术(optical coherence tomography，OCT)具有无创性、成像安全快速等特点，在临床上被广泛用于眼科疾病的诊断。但因为OCT影像成像条件和缺乏颜色特征信息等因素的限制，导致大量噪声的存在，使得目前多数对OCT影像的眼底病灶分割方法无法获得理想结果。Because optical coherence tomography (OCT) has the characteristics of non-invasiveness, safe and fast imaging, it is widely used in clinical diagnosis of ophthalmic diseases. However, due to the limitations of OCT image imaging conditions and lack of color feature information and other factors, a large amount of noise exists, making most of the current fundus lesion segmentation methods for OCT images unable to obtain ideal results.

现有方案中，发明人意识到即使在分割算法优化基础上，仍然存在部分难以准确分割识别出完整形态的病灶，影像中部分病灶的渐变边界导致分割出的病灶区域多数存在不规则孔洞或者凹槽，此时通常采用形态学闭合运算或Convex Hull凸包等方法进行后处理，而这些技术手段都会带来明显的假阳性或者对病灶凹槽填补不全的问题，无法有效获得预期的准确病变区域，此类缺陷对医生的临床诊断带来了困难和歧义性。In the existing solution, the inventor realized that even based on the optimization of the segmentation algorithm, there are still some lesions that are difficult to accurately segment and identify the complete shape. The gradual boundary of some lesions in the image causes most of the segmented lesion areas to have irregular holes or concaves. At this time, post-processing methods such as morphological closing operations or Convex Hull convex hulls are usually used, and these technical methods will bring obvious false positives or incomplete filling of the lesion grooves, and cannot effectively obtain the expected accurate lesion area Such defects bring difficulties and ambiguities to doctors’ clinical diagnosis.

发明内容Summary of the invention

本申请提供了基于眼底OCT影像的病灶提取方法、装置、设备及存储介质，用于对眼底OCT影像的病灶区域进行分割处理，得到更准确的眼底OCT影像的病灶区域信息，避免了病灶区域凹槽填补不全和假阳性的问题，提高了处理效率。This application provides a focus extraction method, device, equipment, and storage medium based on fundus OCT images, which are used to segment the focus area of the fundus OCT image to obtain more accurate focus area information of the fundus OCT image, and avoid the focus area concave The problem of incomplete slot filling and false positive improves the processing efficiency.

本申请实施例的第一方面提供一种基于眼底OCT影像的病灶提取方法，包括：获取原始眼底光学相干断层扫描技术OCT影像；对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数；根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。The first aspect of the embodiments of the present application provides a focus extraction method based on fundus OCT images, including: acquiring original fundus optical coherence tomography OCT images; performing segmentation preprocessing on the original fundus OCT images to obtain a preprocessed fundus OCT image; an initial closed curve is generated in the preprocessed fundus OCT image, the inside of the initial closed curve includes the target fundus lesion; the initial closed curve is determined by a preset morphological serpentine curve evolution algorithm Curve function; according to the initial curve function to evolve the initial closed curve to obtain a target closed curve, the target curve function of the target closed curve represents the contour of the target fundus lesion; the inside of the target closed curve The groove is filled to obtain the lesion image of the target fundus lesion.

本申请实施例的第二方面提供了一种基于眼底OCT影像的病灶提取装置，包括：第一获取单元，用于获取原始眼底光学相干断层扫描技术OCT影像；分割单元，用于对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；生成单元，用于在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；确定单元，用于通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数；第一演化单元，用于根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；填充单元，用于对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。The second aspect of the embodiments of the present application provides a focus extraction device based on fundus OCT images, including: a first acquisition unit for acquiring original fundus optical coherence tomography OCT images; a segmentation unit for comparing the original The fundus OCT image is segmented and preprocessed to obtain a preprocessed fundus OCT image; a generating unit is configured to generate an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes a target fundus lesion; determine Unit for determining the initial curve function of the initial closed curve by a preset morphological serpentine curve evolution algorithm; the first evolution unit for evolving the initial closed curve according to the initial curve function to obtain a target A closed curve, the target curve function of the target closed curve represents the contour of the target fundus lesion; the filling unit is used to fill grooves inside the target closed curve to obtain a lesion image of the target fundus lesion.

本申请实施例的第三方面提供了一种基于眼底OCT影像的病灶提取设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现上述基于眼底OCT影像的病灶提取方法。The third aspect of the embodiments of the present application provides a focus extraction device based on fundus OCT images, including a memory, a processor, and a computer program stored in the memory and running on the processor. The processor When the computer program is executed, the above-mentioned lesion extraction method based on fundus OCT images is realized.

本申请实施例的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中存储有指令，当所述指令在计算机上运行时，使得计算机执行上述基于眼底OCT影像的病灶提取方法的步骤。The fourth aspect of the embodiments of the present application provides a computer-readable storage medium that stores instructions in the computer-readable storage medium. When the instructions are run on a computer, the computer executes the above-mentioned fundus OCT image-based lesion Steps of extraction method.

本申请实施例提供的技术方案中，获取原始眼底光学相干断层扫描技术OCT影像；对原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；在预处理的眼底OCT图像中生成一个初始闭合曲线，初始闭合曲线的内部包括目标眼底病灶；通过预置的形态学蛇形曲线演化算法确定初始闭合曲线的初始曲线函数；根据初始曲线函数对初始闭合曲线进行演化，得到目标闭合曲线，目标闭合曲线的目标曲线函数表示目标眼底病灶的轮廓；对目标闭合曲线的内部进行凹槽填充，得到目标眼底病灶的病灶图像。本申请实施例，对眼底OCT影像的病灶区域进行分割处理，得到更准确的眼底OCT影像的病灶区域信息，避免了病灶区域凹槽填补不全和假阳性的问题，提高了处理效率。In the technical solution provided by the embodiments of this application, the original fundus optical coherence tomography technology OCT image is acquired; the original fundus OCT image is segmented and preprocessed to obtain the preprocessed fundus OCT image; an initial fundus OCT image is generated from the preprocessed fundus OCT image Closed curve, the inside of the initial closed curve includes the target fundus lesion; the initial curve function of the initial closed curve is determined by the preset morphological serpentine curve evolution algorithm; the initial closed curve is evolved according to the initial curve function to obtain the target closed curve, target The target curve function of the closed curve represents the contour of the target fundus lesion; groove filling is performed on the inside of the target closed curve to obtain the lesion image of the target fundus lesion. In the embodiment of the present application, the lesion area of the fundus OCT image is segmented to obtain more accurate information of the lesion area of the fundus OCT image, which avoids the problems of insufficient filling of grooves in the lesion area and false positives, and improves processing efficiency.

附图说明Description of the drawings

图1为本申请实施例中基于眼底OCT影像的病灶提取方法的一个实施例示意图；FIG. 1 is a schematic diagram of an embodiment of a method for extracting lesions based on fundus OCT images in an embodiment of the application;

图2为本申请实施例中基于眼底OCT影像的病灶提取方法的另一个实施例示意图；2 is a schematic diagram of another embodiment of the method for extracting lesions based on fundus OCT images in an embodiment of the application;

图3为本申请实施例中基于眼底OCT影像的病灶提取装置的一个实施例示意图；Fig. 3 is a schematic diagram of an embodiment of a lesion extraction device based on fundus OCT images in an embodiment of the application;

图4为本申请实施例中基于眼底OCT影像的病灶提取装置的另一个实施例示意图；4 is a schematic diagram of another embodiment of the device for extracting lesions based on fundus OCT images in an embodiment of the application;

图5为本申请实施例中基于眼底OCT影像的病灶提取设备的一个实施例示意图。Fig. 5 is a schematic diagram of an embodiment of a lesion extraction device based on fundus OCT images in an embodiment of the application.

具体实施方式Detailed ways

本申请提供了一种基于眼底OCT影像的病灶提取方法、装置、设备及存储介质，用于对眼底OCT影像的病灶区域进行分割处理，得到更准确的眼底OCT影像的病灶区域信息，避免了病灶区域凹槽填补不全和假阳性的问题，提高了处理效率。This application provides a focus extraction method, device, device and storage medium based on fundus OCT images, which are used to segment the focus area of the fundus OCT image to obtain more accurate focus area information of the fundus OCT image, and avoid the focus The problem of incomplete filling of regional grooves and false positives improves the processing efficiency.

请参阅图1，本申请实施例提供的基于眼底OCT影像的病灶提取方法的流程图，具体包括：Please refer to FIG. 1, the flow chart of the method for extracting lesions based on fundus OCT images provided by the embodiment of the present application specifically includes:

101、获取原始眼底光学相干断层扫描技术OCT影像。101. Obtain OCT images of the original fundus optical coherence tomography technology.

基于眼底OCT影像的病灶提取装置获取原始眼底光学相干断层扫描技术(optical coherence tomography，OCT)影像。其中，该原始眼底OCT影像由OCT设备直接得到，未经过任何处理。The focus extraction device based on fundus OCT images acquires original fundus optical coherence tomography (OCT) images. Among them, the original fundus OCT image is directly obtained by the OCT device without any processing.

目前OCT分为两大类：时域OCT(time domain optical coherence tomography，TD-OCT)和频域OCT(frequency domain optical coherence tomography，FD-OCT)。时域OCT是把在同一时间从组织中反射回来的光信号与参照反光镜反射回来的光信号叠加、干涉，然后成像。频域OCT是参考臂的参照反光镜固定不动，通过改变光源光波的频率来实现信号的干涉。本申请实施例可以通过多种方式获取原始眼底OCT影像，可以通过TD-OCT方式获取，还可以通过FD-OCT方式获取，具体采用何种获取方式此处不做限定。At present, OCT is divided into two categories: time domain optical coherence tomography (TD-OCT) and frequency domain optical coherence tomography (FD-OCT). Time-domain OCT is to superimpose and interfere with the optical signal reflected from the tissue at the same time and the optical signal reflected from the reference mirror, and then image. Frequency domain OCT means that the reference mirror of the reference arm is fixed, and the signal interference is realized by changing the frequency of the light wave of the light source. In the embodiments of the present application, the original fundus OCT image can be obtained in a variety of ways, which can be obtained in a TD-OCT mode, or can also be obtained in an FD-OCT mode. The specific obtaining method is not limited here.

可以理解的是，本申请的执行主体可以为基于眼底OCT影像的病灶提取装置，还可以是终端或者服务器，具体此处不做限定。本申请实施例以基于眼底OCT影像的病灶提取装置为执行主体为例进行说明。It is understandable that the subject of this application may be a focus extraction device based on fundus OCT images, or may also be a terminal or a server, which is not specifically limited here. The embodiment of the present application takes the focus extraction device based on the fundus OCT image as the execution subject as an example for description.

102、对原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像。102. Perform segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image.

基于眼底OCT影像的病灶提取装置对原始眼底OCT图像进行分割预处理，得到预处理的眼底OCT图像。具体的，基于眼底OCT影像的病灶提取装置对原始眼底OCT图像上分割出包含病灶区域的方形图像，得到的预处理的眼底OCT图像。The focus extraction device based on the fundus OCT image performs segmentation and preprocessing on the original fundus OCT image to obtain the preprocessed fundus OCT image. Specifically, the device for extracting a lesion based on the fundus OCT image segments the original fundus OCT image into a square image containing the lesion area to obtain a preprocessed fundus OCT image.

需要说明的是，预处理的眼底OCT图像中，会根据原始眼底OCT图像中的病灶范围，用不同的颜色进行区分，例如，将病灶区域显示为白色，将非病灶区域显示为黑色，还可以是其他组合，只要能将病灶区域和非病灶区域区分开即可，例如，病灶区域用白色显示，非病灶区域用灰色显示，具体此处不做限定。It should be noted that the preprocessed fundus OCT image will be distinguished by different colors according to the lesion range in the original fundus OCT image. For example, the lesion area is displayed as white, and the non-lesion area is displayed as black. It is other combinations, as long as the lesion area can be distinguished from the non-lesion area, for example, the lesion area is displayed in white, and the non-lesion area is displayed in gray, and the details are not limited here.

103、在预处理的眼底OCT图像中生成一个初始闭合曲线，初始闭合曲线的内部包括目标眼底病灶。103. Generate an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes the target fundus lesion.

基于眼底OCT影像的病灶提取装置在预处理的眼底OCT图像中生成一个初始闭合曲线，该初始闭合曲线的内部包括目标眼底病灶。具体的，基于眼底OCT影像的病灶提取装置在预处理的眼底OCT图像中生成初始闭合曲线，该初始闭合曲线可以是预置的闭合曲线中的任意一种，预置的闭合曲线可以是圆形、椭圆形、三角形等几何图形，或者是不规则的闭合形状，具体此处不做限定。The fundus OCT image-based lesion extraction device generates an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes the target fundus lesion. Specifically, the focus extraction device based on fundus OCT images generates an initial closed curve in the preprocessed fundus OCT image. The initial closed curve can be any of the preset closed curves, and the preset closed curve can be a circle. , Ellipse, triangle and other geometric figures, or irregular closed shapes, the specifics are not limited here.

需要说明的是，该初始闭合曲线可以是形状任意，只要保证将目标眼底病灶完全包含在曲线内部即可。It should be noted that the initial closed curve can have any shape, as long as the target fundus lesion is completely contained in the curve.

104、通过预置的形态学蛇形曲线演化算法确定初始闭合曲线的初始曲线函数。104. Determine the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm.

基于眼底OCT影像的病灶提取装置通过预置的形态学蛇形曲线演化算法确定初始闭合曲线的初始曲线函数。具体的，生成初始化水平集函数u(x,y)，初始化水平集函数的表达式如下：The focus extraction device based on fundus OCT images determines the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm. Specifically, the initialization level set function u(x,y) is generated, and the expression of the initialization level set function is as follows:

其中，d[(x,y),C]是点(x,y)到曲线C的最短有向距离；根据初始化水平集函数生成初始闭合曲线方程：

其中，u为目标水平集函数，

为梯度算子，div为散度算子，v为常数，g(I)为边缘停止函数；根据初始闭合曲线方程计算得到预置的形态学蛇形曲线演化算法的零水平集函数

判断零水平集函数是否收敛；若零水平集函数收敛，则确定零水平集函数为初始曲线函数。

Among them, d[(x,y),C] is the shortest directional distance from point (x,y) to curve C; the initial closed curve equation is generated according to the initial level set function:

Among them, u is the target level set function,

Is the gradient operator, div is the divergence operator, v is the constant, g(I) is the edge stop function; calculated according to the initial closed curve equation, the zero level set function of the preset morphological serpentine curve evolution algorithm

Determine whether the zero-level set function converges; if the zero-level set function converges, determine the zero-level set function as the initial curve function.

需要说明的是，在二维平面内，显性表示的函数曲线在演化过程中无法随着目标的合并或者***进行拓扑变化，因此在表示目标拓扑变化方面，显性表示的二维函数曲线就无能为力。为了解决这个问题，Osher和Sethian用隐式参数函数表示演化曲线，进而提出了水平集方法。由于水平集函数能够灵活地表示目标的拓扑变化，可以有效地应用到轮廓提取领域。It should be noted that in the two-dimensional plane, the explicit representation of the function curve cannot undergo topological changes with the merging or splitting of the target during the evolution process. Therefore, the explicit representation of the two-dimensional function curve is Powerless. To solve this problem, Osher and Sethian used implicit parameter functions to express the evolution curve, and then proposed a level set method. Since the level set function can flexibly represent the topological changes of the target, it can be effectively applied to the field of contour extraction.

水平集函数用的是一种隐式函数，跟传统的snake算法相比在思想上差异很大，snake算法曲线演化的时候，是曲线上离散点显示坐标的位置更新移动，只要懂得能量最小化的曲线演化规则即可；然而水平集函数，更新的不是曲线离散点的坐标，而是更新整张图片像素点到曲线的有向距离场。因此水平集函数算法最关键的是理解这个距离场的更新规则。例如，生成一条初始封闭轮廓曲线C，进行水平集图像分割，需要些的第一个函数就是计算图像的每个像素点p(x,y)到曲线的最短距离d，如果该像素点p位于曲线C的内部，那么有向距离为-d，反之为d。这样遍历图像每个像素点，每个像素点都可以求得对应的有向距离u(x,y)。The level set function uses an implicit function. Compared with the traditional snake algorithm, the idea is very different. When the snake algorithm curve evolves, the position of the discrete point on the curve is updated and moved. As long as you know how to minimize energy The rule of curve evolution is sufficient; however, the level set function updates not the coordinates of discrete points on the curve, but updates the directed distance field from the pixels of the entire picture to the curve. Therefore, the key to the level set function algorithm is to understand the update rule of this distance field. For example, to generate an initial closed contour curve C and perform level set image segmentation, the first function needed is to calculate the shortest distance d from each pixel p(x,y) of the image to the curve, if the pixel p is located Inside the curve C, then the directed distance is -d, otherwise it is d. In this way, each pixel of the image is traversed, and the corresponding directed distance u(x,y) can be obtained for each pixel.

水平集函数算法的基本原理是把目标曲线或者曲面作为零水平集嵌入到更高一维的水平集函数中，也即用零平面截取水平集函数得到的封闭曲线或者曲面来代替演化曲线或者曲面，随着水平集函数的变化，演化曲线或者曲面也随之改变，并能适应拓扑的变化。在二维平面，如果隐式表示的闭合曲线为：C(x,y)＝0，根据水平集原理，把此闭合曲线嵌入到三维水平集函数z＝φ(x,y)中，则用z＝0平面截此水平集函数的曲面，即可得到闭合曲线C(x,y)＝0。当三维水平集曲面z＝φ(x,y)在驱动力的作用下发生变化时，其零水平集平面截得的闭合曲线C(x,y)＝0也随之变化。The basic principle of the level set function algorithm is to embed the target curve or surface as a zero level set into a higher one-dimensional level set function, that is, use the zero plane to intercept the closed curve or surface obtained by the level set function to replace the evolution curve or surface As the level set function changes, the evolution curve or surface also changes, and can adapt to changes in topology. In a two-dimensional plane, if the implicitly expressed closed curve is: C(x,y)=0, according to the principle of level set, embed this closed curve into the three-dimensional level set function z=φ(x,y), then use The surface of this level set function is truncated by the plane z=0, and the closed curve C(x,y)=0 can be obtained. When the three-dimensional level set surface z=φ(x,y) changes under the action of the driving force, the closed curve C(x,y)=0 intercepted by the zero level set plane also changes.

105、根据初始曲线函数对初始闭合曲线进行演化，得到目标闭合曲线，目标闭合曲线的目标曲线函数表示目标眼底病灶的轮廓。105. The initial closed curve is evolved according to the initial curve function to obtain the target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion.

基于眼底OCT影像的病灶提取装置根据初始曲线函数对初始闭合曲线进行演化，得到目标闭合曲线，目标闭合曲线的目标曲线函数表示目标眼底病灶的轮廓。具体的，基于眼底OCT影像的病灶提取装置根据初始曲线函数构建能量函数，能量函数包括内能量项和外能量项，内能量项用于使得目标闭合曲线不断向内部紧缩且保持平滑，外能量项用于保证目标闭合曲线紧缩到目标眼底病灶边缘时停止；根据能量方程计算出边缘停止函数，边缘停止函数用于表示初始闭合曲线的受力情况；根据边缘停止函数对初始闭合曲线进行变形；当边缘停止函数为零时，生成目标闭合曲线。The focus extraction device based on the OCT image of the fundus evolves the initial closed curve according to the initial curve function to obtain the target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion. Specifically, the lesion extraction device based on fundus OCT images constructs an energy function according to the initial curve function. The energy function includes an internal energy term and an external energy term. The internal energy term is used to make the target closed curve continuously shrink and remain smooth. The external energy term It is used to ensure that the target closed curve stops when it shrinks to the edge of the target fundus lesion; the edge stop function is calculated according to the energy equation, and the edge stop function is used to express the force of the initial closed curve; the initial closed curve is deformed according to the edge stop function; When the edge stop function is zero, the target closed curve is generated.

例如，设初始闭合曲线为C，初始闭合曲线在驱动力的作用下，随着时间的推移进行演化，可得到演化曲线集C(t)。根据水平集原理，把初始闭合曲线嵌入到三维水平集函数

其零水平集可以表示为

具体的，初始闭合曲线随水平集函数的变化而进行演化。C(t)表示初始闭合曲线，

表示三维水平集函数，C(t ₀)表示在t ₀时刻获得的轮廓曲线，此时的轮廓曲线是一个单连通区域。随着时间的推移，在t ₁时刻获得的轮廓曲线变为C(t ₁)，此时轮廓曲线***为两个单联通区域，也即随着轮廓曲线的演化，可以检测出两个Mura曲线区域。其中，三维水平集函数

的零水平为：

对上式零水平集方程两端分别对t求导整理得到演化方程:

其中，F为轮廓曲线C(t)的边缘停止函数(也即速度函数)。当边缘停止函数不为零的时候，轮廓曲线演化不停止，当边缘停止函数为零时，轮廓曲线停止演化，表示能量函数达到了目标的最小值。因此，可以把水平集解法转化为求解

的过程。具体求解过程如下： For example, assuming that the initial closed curve is C, the initial closed curve evolves over time under the action of the driving force, and the evolution curve set C(t) can be obtained. According to the principle of level set, embed the initial closed curve into the three-dimensional level set function

Its zero level set can be expressed as

Specifically, the initial closed curve evolves as the level set function changes. C(t) represents the initial closed curve,

Represents the three-dimensional level set function, C(t ₀ ) represents the contour curve obtained at t ₀ , and the contour curve at this time is a single connected region. As time goes by, the contour curve obtained at t ₁ becomes C(t ₁ ). At this time, the contour curve splits into two single-connected areas, that is, as the contour curve evolves, two Mura curves can be detected area. Among them, the three-dimensional level set function

The zero level is:

The two ends of the zero level set equation of the above formula are derivable to t respectively to obtain the evolution equation:

Among them, F is the edge stop function (that is, the speed function) of the contour curve C(t). When the edge stop function is not zero, the contour curve evolution does not stop. When the edge stop function is zero, the contour curve stops evolving, indicating that the energy function has reached the target minimum. Therefore, the level set solution method can be transformed into a solution

the process of. The specific solution process is as follows:

通过水平集方法推导，可以得到轮廓曲线内部的平均灰度C ₀和轮廓曲线外部的平均灰度C _b的计算公式如下： Derived by the level set method, the calculation formulas of the average gray C ₀ inside the contour curve and the average gray C _b outside the contour curve can be obtained as follows:

分析可知，水平集函数算法的求解过程可以转换为偏微分方程

的求解过程。在利用欧拉-拉格朗日求解式，并根据梯度下降流，得到最终的偏微分方程如下式：

其中，H(z)为Heaviside函数：

The analysis shows that the solving process of the level set function algorithm can be transformed into a partial differential equation

The solution process. Using the Euler-Lagrangian equation and according to the gradient descent flow, the final partial differential equation is obtained as follows:

Among them, H(z) is the Heaviside function:

δ(z)为Dirac delta函数：

δ(z) is the Dirac delta function:

上述Heaviside函数与Dirac delta函数是一种理论上的表示形式，在实际计算时，一般取其近似值分别如下：The above Heaviside function and Dirac delta function are a theoretical representation. In actual calculations, the approximate values are generally as follows:

106、对目标闭合曲线的内部进行凹槽填充，得到目标眼底病灶的病灶图像。106. Fill the inside of the target closed curve with grooves to obtain a lesion image of the target fundus lesion.

基于眼底OCT影像的病灶提取装置对目标闭合曲线的内部进行凹槽填充，得到目标眼底病灶的病灶图像。具体的，目标闭合曲线最终会自动收敛到目标眼底病灶区域的外轮廓，所以病灶内部封闭的孔洞可以直接被填充修复。实际上，内部封闭孔洞的填充可以通过漫水填充(flood fill)算法实现，以实现对封闭孔洞和开放的凹槽缺陷同时进行修补，修复效果更准确。The focus extraction device based on the OCT image of the fundus fills the inside of the target closed curve to obtain a focus image of the target fundus lesion. Specifically, the target closed curve will eventually automatically converge to the outer contour of the target fundus lesion area, so the closed hole inside the lesion can be directly filled and repaired. In fact, the filling of the internal closed holes can be realized by flood fill algorithm to realize the simultaneous repair of the closed holes and the open groove defects, and the repair effect is more accurate.

本申请实施例，对眼底OCT影像的病灶区域进行分割处理，得到更准确的眼底OCT影像的病灶区域信息，避免了病灶区域凹槽填补不全和假阳性的问题，提高了处理效率。In the embodiment of the present application, the lesion area of the fundus OCT image is segmented to obtain more accurate information of the lesion area of the fundus OCT image, which avoids the problems of insufficient filling of grooves in the lesion area and false positives, and improves processing efficiency.

请参阅图2，本申请实施例提供的基于眼底OCT影像的病灶提取方法的另一个流程图，具体包括：Please refer to FIG. 2, another flowchart of the method for extracting lesions based on fundus OCT images provided by an embodiment of the present application specifically includes:

201、获取原始眼底光学相干断层扫描技术OCT影像。201. Obtain original OCT images of fundus optical coherence tomography.

202、对原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像。202. Perform segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image.

203、在预处理的眼底OCT图像中生成一个初始闭合曲线，初始闭合曲线的内部包括目标眼底病灶。203. Generate an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes the target fundus lesion.

204、通过预置的形态学蛇形曲线演化算法确定初始闭合曲线的初始曲线函数。204. Determine the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm.

其中，u为目标水平集函数，

Among them, u is the target level set function,

205、根据初始曲线函数构建能量函数。205. Construct an energy function according to the initial curve function.

基于眼底OCT影像的病灶提取装置根据初始曲线函数构建能量函数，能量函数包括内能量项和外能量项，内能量项用于使得目标闭合曲线不断向内部紧缩且保持平滑，外能量项用于保证目标闭合曲线紧缩到目标眼底病灶边缘时停止。具体的，获取所述初始曲线函数上的控制点v；确定所述控制点的弹性能量项和弯曲能量项，所述弹性能量项为所述v的一阶导数的模，所述弯曲能量项为所述v的二阶导数的模；确定所述控制点的外能量项，所述外部能量项为所述控制点所在位置的图像局部特征；根据所述弹性能量项、所述弯曲能量项和所述外能量项构建能量函数，能量函数的表达式为：The lesion extraction device based on fundus OCT images constructs an energy function according to the initial curve function. The energy function includes an internal energy term and an external energy term. The internal energy term is used to make the target closed curve continuously shrink and remain smooth, and the external energy term is used to ensure Stop when the target closed curve shrinks to the edge of the target fundus lesion. Specifically, the control point v on the initial curve function is acquired; the elastic energy term and the bending energy term of the control point are determined, the elastic energy term is the modulus of the first derivative of the v, and the bending energy term Is the modulus of the second derivative of v; determines the external energy term of the control point, where the external energy term is the local feature of the image where the control point is located; according to the elastic energy term and the bending energy term Construct an energy function with the external energy term, and the expression of the energy function is:

其中，v(s)＝[x(s),y(s)]，s∈[0,1]， x(s)和y(s)分别表示每个控制点在图像中的坐标位置，s是以傅里叶变换形式描述边界的自变量，v表示初始曲线函数上的控制点，α、β为常数。

Among them, v(s)=[x(s),y(s)], s∈[0,1], x(s) and y(s) respectively represent the coordinate position of each control point in the image, s It is the independent variable describing the boundary in the form of Fourier transform, v represents the control point on the initial curve function, and α and β are constants.

206、根据能量方程计算出边缘停止函数，边缘停止函数用于表示初始闭合曲线的受力情况。206. Calculate the edge stop function according to the energy equation, and the edge stop function is used to represent the force of the initial closed curve.

基于眼底OCT影像的病灶提取装置根据能量方程计算出边缘停止函数，边缘停止函数用于表示初始闭合曲线的受力情况。The focus extraction device based on the fundus OCT image calculates the edge stop function according to the energy equation, and the edge stop function is used to represent the force of the initial closed curve.

其零水平集可以表示为

的零水平为：

对上式零水平集方程两端分别对t求导整理得到演化方程:

其中，F为轮廓曲线C(t)的边缘停止函数(也即速度函数)。当边缘停止函数不为零的时候，轮廓曲线演化不停止，当边缘停止函数为零时，轮廓曲线停止演化，表示能量函数达到了目标的最小值。 For example, assuming that the initial closed curve is C, the initial closed curve evolves over time under the action of the driving force, and the evolution curve set C(t) can be obtained. According to the principle of level set, embed the initial closed curve into the three-dimensional level set function

Its zero level set can be expressed as

The zero level is:

Among them, F is the edge stop function (that is, the speed function) of the contour curve C(t). When the edge stop function is not zero, the contour curve evolution does not stop. When the edge stop function is zero, the contour curve stops evolving, indicating that the energy function has reached the target minimum.

207、根据边缘停止函数对初始闭合曲线进行变形。207. Deform the initial closed curve according to the edge stop function.

基于眼底OCT影像的病灶提取装置根据边缘停止函数对初始闭合曲线进行变形。具体的，改变初始闭合曲线的受力情况，直至受力为0，即改变边缘停止函数的大小，使能量函数趋于最小值，即对初始闭合曲线进行变形。The focus extraction device based on OCT images of the fundus deforms the initial closed curve according to the edge stop function. Specifically, the force condition of the initial closed curve is changed until the force is 0, that is, the size of the edge stop function is changed to make the energy function tend to the minimum value, that is, the initial closed curve is deformed.

208、当边缘停止函数为零时，生成目标闭合曲线。208. When the edge stop function is zero, generate a target closed curve.

当边缘停止函数为零时，能量函数为最小值，基于眼底OCT影像的病灶提取装置生成目标闭合曲线。具体的，能量函数的水平集解法可以转化为求解

的过程。具体求解过程如下： When the edge stop function is zero, the energy function is the minimum value, and the focus extraction device based on the fundus OCT image generates the target closed curve. Specifically, the level set solution method of the energy function can be transformed into a solution

the process of. The specific solution process is as follows:

其中，H(z)为Heaviside函数：

Among them, H(z) is the Heaviside function:

δ(z)为Dirac delta函数：

δ(z) is the Dirac delta function:

209、对目标闭合曲线的内部进行凹槽填充，得到目标眼底病灶的病灶图像。209. Perform groove filling on the inside of the target closed curve to obtain a lesion image of the target fundus lesion.

210、获取医生标注标准图像。210. Obtain a standard image marked by the doctor.

基于眼底OCT影像的病灶提取装置获取医生标注标准图像。例如，在实际操作中，医生会对病灶区域进行医生标注标准，与在原始眼底OCT影像基础上进行病灶区域的划分，其中，医生标注标准(或称医生标注金标准)，指把眼底OCT影像给到专业的医生，让医生把病灶(病变区域)勾勒(或者称为标注)出来，以此作为最终参考标准(医生标注金标准)，以使得目标眼底病灶的病灶图像接近医生标注的真实病变区域结果。The focus extraction device based on the OCT image of the fundus acquires the standard image marked by the doctor. For example, in actual operation, the doctor will mark the diseased area by the doctor, and divide the diseased area on the basis of the original fundus OCT image. Among them, the doctor’s labeling standard (or the doctor’s gold standard) refers to the fundus OCT image Give to a professional doctor, let the doctor outline (or label) the lesion (lesion area) as the final reference standard (doctor's mark gold standard), so that the image of the target fundus lesion is close to the real lesion marked by the doctor Regional results.

211、将目标眼底病灶的病灶图像与医生标注标准图像进行对比，判断目标眼底病灶的病灶图像是否符合要求。211. Compare the lesion image of the target fundus lesion with the standard image marked by the doctor, and determine whether the lesion image of the target fundus lesion meets the requirements.

基于眼底OCT影像的病灶提取装置将目标眼底病灶的病灶图像与医生标注标准图像进行对比，判断目标眼底病灶的病灶图像是否符合要求。The focus extraction device based on the OCT image of the fundus compares the focus image of the target fundus focus with the standard image marked by the doctor to determine whether the focus image of the target fundus focus meets the requirements.

212、若目标眼底病灶的病灶图像符合要求，则将目标眼底病灶的病灶图像作为输出图像并输出。212. If the lesion image of the target fundus lesion meets the requirements, use the lesion image of the target fundus lesion as an output image and output.

若目标眼底病灶的病灶图像符合要求，则基于眼底OCT影像的病灶提取装置将目标眼底病灶的病灶图像作为输出图像并输出。If the lesion image of the target fundus lesion meets the requirements, the lesion extraction device based on the fundus OCT image uses the lesion image of the target fundus lesion as an output image and outputs it.

213、若目标眼底病灶的病灶图像不符合要求，则重新对目标眼底病灶的病灶图像的轮廓曲线进行演化。213. If the lesion image of the target fundus lesion does not meet the requirements, re-evolve the contour curve of the lesion image of the target fundus lesion.

若目标眼底病灶的病灶图像不符合要求，则基于眼底OCT影像的病灶提取装置重新对目标眼底病灶的病灶图像的轮廓曲线进行演化。If the lesion image of the target fundus lesion does not meet the requirements, the lesion extraction device based on the fundus OCT image re-evolves the contour curve of the lesion image of the target fundus lesion.

本申请实施例，对眼底OCT影像的病灶区域进行分割处理，得到更准确的眼底OCT影像的病灶区域信息，填补病灶区域凹槽，提高了处理效率；自动提取出有价值的眼底病灶区域轮廓，避免了人工后处理勾勒病灶边界的繁杂工作，有效地为医生后续的临床疾病诊断提供了整洁准确的病灶形态信息。In the embodiment of the application, the lesion area of the fundus OCT image is segmented to obtain more accurate information of the lesion area of the fundus OCT image, fill the groove of the lesion area, and improve the processing efficiency; automatically extract the contour of the valuable fundus lesion area, It avoids the complicated work of manual post-processing to outline the boundary of the lesion, and effectively provides clean and accurate lesion morphology information for the doctor's follow-up clinical disease diagnosis.

上面对本申请实施例中基于眼底OCT影像的病灶提取方法进行了描述，下面对本申请实施例中基于眼底OCT影像的病灶提取装置进行描述，请参阅图3，本申请实施例中基于眼底OCT影像的病灶提取装置的一个实施例包括：The focus extraction method based on fundus OCT images in the embodiments of this application is described above, and the focus extraction device based on fundus OCT images in the embodiments of this application is described below. Please refer to FIG. 3, the fundus OCT imaging based on the embodiments of this application An embodiment of the lesion extraction device includes:

第一获取单元301，用于获取原始眼底光学相干断层扫描技术OCT影像；The first acquiring unit 301 is configured to acquire original fundus optical coherence tomography technology OCT images;

分割单元302，用于对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；The segmentation unit 302 is configured to perform segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image;

生成单元303，用于在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；A generating unit 303, configured to generate an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes a target fundus lesion;

确定单元304，用于通过预置的形态学蛇形曲线演化算法零所述初始闭合曲线的初始曲线函数；The determining unit 304 is configured to zero the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm;

第一演化单元305，用于根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；The first evolution unit 305 is configured to evolve the initial closed curve according to the initial curve function to obtain a target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion;

填充单元306，用于对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。The filling unit 306 is configured to perform groove filling on the inside of the target closed curve to obtain a lesion image of the target fundus lesion.

请参阅图4，本申请实施例中基于眼底OCT影像的病灶提取装置的另一个实施例包括：Referring to FIG. 4, another embodiment of the device for extracting a lesion based on fundus OCT images in the embodiment of the present application includes:

可选的，第一演化单元305包括：Optionally, the first evolution unit 305 includes:

构建模块3051，用于根据所述初始曲线函数构建能量函数，所述能量函数包括内能量项和外能量项，所述内能量项用于使得目标闭合曲线不断向内部紧缩且保持平滑，所述外能量项用于保证目标闭合曲线紧缩到目标眼底病灶边缘时停止；The construction module 3051 is configured to construct an energy function according to the initial curve function, the energy function includes an internal energy term and an external energy term, and the internal energy term is used to make the target closed curve continuously shrink inward and remain smooth, the The external energy term is used to ensure that the target closed curve stops when it shrinks to the edge of the target fundus lesion;

计算模块3052，用于根据所述能量方程计算出边缘停止函数，所述边缘停止函数用于表示初始闭合曲线的受力情况；The calculation module 3052 is configured to calculate an edge stop function according to the energy equation, and the edge stop function is used to represent the force of the initial closed curve;

变形模块3053，用于根据所述边缘停止函数对所述初始闭合曲线进行变形；A deformation module 3053, configured to deform the initial closed curve according to the edge stop function;

生成模块3054，当所述边缘停止函数为零时，用于生成目标闭合曲线。The generating module 3054 is used to generate the target closed curve when the edge stop function is zero.

可选的，构建模块3051具体用于：Optionally, the building module 3051 is specifically used for:

获取所述初始曲线函数上的控制点v；确定所述控制点的弹性能量项和弯曲能量项，所述弹性能量项为所述v的一阶导数的模，所述弯曲能量项为所述v的二阶导数的模；确定所述控制点的外能量项，所述外部能量项为所述控制点所在位置的图像局部特征；根据所述弹性能量项、所述弯曲能量项和所述外能量项构建能量函数，能量函数的表达式为：

其中，v(s)＝[x(s),y(s)]，s∈[0,1]，x(s)和y(s)分别表示每个控制点在图像中的坐标位置，s是以傅里叶变换形式描述边界的自变量，v表示初始曲线函数上的控制点，α、β为常数。 Obtain the control point v on the initial curve function; determine the elastic energy term and the bending energy term of the control point, where the elastic energy term is the modulus of the first derivative of the v, and the bending energy term is the The modulus of the second derivative of v; determine the external energy term of the control point, where the external energy term is a local feature of the image where the control point is located; according to the elastic energy term, the bending energy term, and the The external energy term constructs the energy function, and the expression of the energy function is:

可选的，构建模块3051具体还用于：Optionally, the building module 3051 is specifically used to:

将所述控制点的梯度算子作为所述图像局部特征；根据所述梯度算子确定所述外部能量项的表达式：

为所述梯度算子。 Use the gradient operator of the control point as the local feature of the image; determine the expression of the external energy term according to the gradient operator:

Is the gradient operator.

可选的，确定单元304具体用于：Optionally, the determining unit 304 is specifically configured to:

生成初始化水平集函数u(x,y)，所述初始化水平集函数的表达式如下：

其中，d[(x,y),C]是点(x,y)到曲线C的最短有向距离；根据所述初始化水平集函数生成初始闭合曲线方程：

其中，u为目标水平集函数，

为梯度算子，div为散度算子，v为常数，g(I)为边缘停止函数；根据所述初始闭合曲线方程计算得到所述预置的形态学蛇形曲线演化算法的零水平集函数

判断所述零水平集函数是否收敛；若所述零水平集函数收敛，则确定所述零水平集函数为所述初始曲线函数。 Generate the initialization level set function u(x,y), the expression of the initialization level set function is as follows:

Among them, d[(x,y),C] is the shortest directional distance from point (x,y) to curve C; the initial closed curve equation is generated according to the initialization level set function:

Among them, u is the target level set function,

Is a gradient operator, div is a divergence operator, v is a constant, and g(I) is an edge stop function; calculated according to the initial closed curve equation, the zero-level set of the preset morphological serpentine curve evolution algorithm function

Determine whether the zero-level set function converges; if the zero-level set function converges, determine that the zero-level set function is the initial curve function.

可选的，基于眼底OCT影像的病灶提取装置还包括：Optionally, the device for extracting lesions based on fundus OCT images further includes:

第二获取单元307，用于获取医生标注标准图像；The second acquiring unit 307 is configured to acquire a standard image marked by the doctor;

判断单元308，用于将所述目标眼底病灶的病灶图像与所述医生标注标准图像进行对比，判断所述目标眼底病灶的病灶图像是否符合要求；The judging unit 308 is configured to compare the lesion image of the target fundus lesion with the doctor-marked standard image, and determine whether the lesion image of the target fundus lesion meets requirements;

输出单元309，若所述目标眼底病灶的病灶图像符合要求，则用于将所述目标眼底病灶的病灶图像作为输出图像并输出。The output unit 309, if the lesion image of the target fundus lesion meets the requirements, is used to output the lesion image of the target fundus lesion as an output image.

第二演化单元310，若所述目标眼底病灶的病灶图像不符合要求，则用于重新对所述目标眼底病灶的病灶图像的轮廓曲线进行演化。The second evolution unit 310, if the lesion image of the target fundus lesion does not meet the requirements, is used to re-evolve the contour curve of the lesion image of the target fundus lesion.

本申请实施例，根据原始眼底OCT影像生成新眼底OCT影像，提高了新眼底OCT影像的真实性，避免和原始眼底OCT影像差异过大，解决了真实数据太少及数据不均衡问题，提高了影像处理效率。同时在样本较小且样本不均衡的情况，可以通过加入生成对抗网络生成的增强影像数据，提高了训练模型的性能及泛化效果。In the embodiment of the application, a new fundus OCT image is generated based on the original fundus OCT image, which improves the authenticity of the new fundus OCT image, avoids excessive differences with the original fundus OCT image, and solves the problem of too little real data and data imbalance, and improves Image processing efficiency. At the same time, when the sample is small and the sample is not balanced, the performance of the training model and the generalization effect can be improved by adding the enhanced image data generated by the generation confrontation network.

上面图3至图4从模块化功能实体的角度对本申请实施例中的基于眼底OCT影像的病灶提取装置进行详细描述，下面从硬件处理的角度对本申请实施例中基于眼底OCT影像的病灶提取设备进行详细描述。The above Figures 3 to 4 describe the fundus OCT image-based lesion extraction device in the embodiment of the present application in detail from the perspective of modular functional entities. The following describes the fundus OCT image-based lesion extraction device in the embodiment of the present application from the perspective of hardware processing Give a detailed description.

图5是本申请实施例提供的一种基于眼底OCT影像的病灶提取设备的结构示意图，该基于眼底OCT影像的病灶提取设备500可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上处理器(central processing units，CPU)501(例如，一个或一个以上处理器)和存储器509，一个或一个以上存储应用程序507或数据506的存储介质508(例如一个或一个以上海量存储设备)。其中，存储器509和存储介质508可以是短暂存储或持久存储。存储在存储介质508的程序可以包括一个或一个以上模块(图示没标出)，每个模块可以包括对基于眼底OCT影像的病灶提取设备中的一系列指令操作。更进一步地，处理器501可以设置为与存储介质508通信，在基于眼底OCT影像的病灶提取设备500上执行存储介质508中的一系列指令操作。5 is a schematic structural diagram of a fundus OCT image-based lesion extraction device provided by an embodiment of the present application. The fundus OCT image-based lesion extraction device 500 may have relatively large differences due to differences in configuration or performance, and may include one or One or more processors (central processing units, CPU) 501 (for example, one or more processors) and a memory 509, one or more storage media 508 for storing application programs 507 or data 506 (for example, one or one storage device with a large amount of ). Among them, the memory 509 and the storage medium 508 may be short-term storage or persistent storage. The program stored in the storage medium 508 may include one or more modules (not shown in the figure), and each module may include a series of command operations in the focus extraction device based on fundus OCT images. Further, the processor 501 may be configured to communicate with the storage medium 508, and execute a series of instruction operations in the storage medium 508 on the fundus OCT image-based lesion extraction device 500.

基于眼底OCT影像的病灶提取设备500还可以包括一个或一个以上电源502，一个或一个以上有线或无线网络接口503，一个或一个以上输入输出接口504，和/或，一个或一个以上操作***505，例如Windows Serve，Mac OS X，Unix，Linux，FreeBSD等等。本领域技术人员可以理解，图5中示出的基于眼底OCT影像的病灶提取设备结构并不构成对基于眼底OCT影像的病灶提取设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。处理器501可以执行上述实施例中第一获取单元301、分割单元302、生成单元303、确定单元304、第一演化单元305、填充单元306、第二获取单元307、判断单元308和第二演化单元310的功能。The lesion extraction device 500 based on fundus OCT images may also include one or more power supplies 502, one or more wired or wireless network interfaces 503, one or more input and output interfaces 504, and/or, one or more operating systems 505 , Such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD and so on. Those skilled in the art can understand that the structure of the fundus OCT image-based lesion extraction device shown in FIG. 5 does not constitute a limitation on the fundus OCT image-based lesion extraction device, and may include more or less components than shown. Or combine certain components, or different component arrangements. The processor 501 can execute the first acquisition unit 301, the segmentation unit 302, the generation unit 303, the determination unit 304, the first evolution unit 305, the filling unit 306, the second acquisition unit 307, the judgment unit 308, and the second evolution in the foregoing embodiment. Function of unit 310.

下面结合图5对基于眼底OCT影像的病灶提取设备的各个构成部件进行具体的介绍：The following is a detailed introduction to each component of the fundus OCT image-based lesion extraction equipment in conjunction with Figure 5:

处理器501是基于眼底OCT影像的病灶提取设备的控制中心，可以按照设置的基于眼底OCT影像的病灶提取方法进行处理。处理器501利用各种接口和线路连接整个基于眼底OCT影像的病灶提取设备的各个部分，通过运行或执行存储在存储器509内的软件程序和/或模块，以及调用存储在存储器509内的数据，执行基于眼底OCT影像的病灶提取设备的各种功能和处理数据，避免病灶区域凹槽填补不全和假阳性的问题，提高处理效率。存储介质508和存储器509都是存储数据的载体，本申请实施例中，存储介质508可以是指储存容量较小，但速度快的内存储器，而存储器509可以是储存容量大，但储存速度慢的外存储器。The processor 501 is the control center of the focus extraction device based on the fundus OCT image, and can perform processing according to the set focus extraction method based on the fundus OCT image. The processor 501 uses various interfaces and lines to connect various parts of the entire fundus OCT image-based lesion extraction device, and by running or executing the software programs and/or modules stored in the memory 509, and calling the data stored in the memory 509, Perform various functions and process data of the focus extraction equipment based on fundus OCT images, avoid the problems of incomplete groove filling and false positives in the focus area, and improve processing efficiency. The storage medium 508 and the memory 509 are both carriers for storing data. In the embodiment of the present application, the storage medium 508 may refer to an internal memory with a small storage capacity but high speed, and the storage 509 may have a large storage capacity but a slow storage speed. External memory.

存储器509可用于存储软件程序以及模块，处理器501通过运行存储在存储器509的软件程序以及模块，从而执行基于眼底OCT影像的病灶提取设备500的各种功能应用以及数据处理。存储器509可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作***、至少一个功能所需的应用程序(比如通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数)等；存储数据区可存储根据基于眼底OCT影像的病灶提取设备的使用所创建的数据(比如初始闭合曲线)等。此外，存储器509可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在本申请实施例中提供的基于眼底OCT影像的病灶提取方法程序和接收到的数据流存储在存储器中，当需要使用时，处理器501从存储器509中调用。The memory 509 may be used to store software programs and modules. The processor 501 executes various functional applications and data processing of the lesion extraction device 500 based on fundus OCT images by running the software programs and modules stored in the memory 509. The memory 509 may mainly include a program storage area and a data storage area. The storage program area may store an operating system and at least one application program required by a function (for example, the initial closed curve is determined by a preset morphological serpentine curve evolution algorithm). The initial curve function), etc.; the storage data area can store data (such as the initial closed curve) created according to the use of the fundus OCT image-based lesion extraction device. In addition, the memory 509 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. The method and program for extracting lesions based on fundus OCT images and the received data stream provided in the embodiment of the present application are stored in the memory, and the processor 501 is called from the memory 509 when needed.

本申请还提供一种计算机可读存储介质，该计算机可读存储介质可以为非易失性计算机可读存储介质，所述计算机可读存储介质中存储有指令，当所述指令在计算机上运行时，使得计算机执行如下基于眼底OCT影像的病灶提取方法的步骤：This application also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium. The computer-readable storage medium stores instructions. When the instructions run on a computer At the time, the computer is made to execute the following steps of the method for extracting lesions based on fundus OCT images:

获取原始眼底光学相干断层扫描技术OCT影像；Obtain original fundus optical coherence tomography OCT images;

对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；Performing segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image;

在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；Generating an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes a target fundus lesion;

通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数；Determining the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm;

根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；Evolving the initial closed curve according to the initial curve function to obtain a target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion;

对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。Groove filling is performed on the inside of the target closed curve to obtain a lesion image of the target fundus lesion.

在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所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。所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计算机可读存储介质传输，例如，所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、双绞线)或无线(例如红外、无线、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。所述计算机可读存储介质可以是计算机能够存储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。所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，光盘)、或者半导体介质(例如固态硬盘(solid state disk，SSD))等。When the computer program instructions are loaded and executed on the computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The 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. For example, the computer instructions may be transmitted from a website, computer, server, or data center. Transmission to another website site, computer, server or data center via wired (such as coaxial cable, optical fiber, twisted pair) or wireless (such as infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that can be stored by a computer or a data storage device such as a server or data center integrated with one or more available media. The usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, and a magnetic tape), an optical medium (for example, an optical disc), or a semiconductor medium (for example, a solid state disk (SSD)).

Claims

一种基于眼底OCT影像的病灶提取方法，包括：A method for extracting lesions based on fundus OCT images, including:

获取原始眼底光学相干断层扫描技术OCT影像；Obtain original fundus optical coherence tomography OCT images;

对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；Performing segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image;

在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；Generating an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes a target fundus lesion;

通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数；Determining the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm;

根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；Evolving the initial closed curve according to the initial curve function to obtain a target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion;

对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。Groove filling is performed on the inside of the target closed curve to obtain a lesion image of the target fundus lesion.
根据权利要求1所述的基于眼底OCT影像的病灶提取方法，所述根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓，包括：The method for extracting lesions based on fundus OCT images according to claim 1, wherein the initial closed curve is evolved according to the initial curve function to obtain a target closed curve, and the target curve function of the target closed curve represents the The outline of the target fundus lesion, including:

根据所述初始曲线函数构建能量函数，所述能量函数包括内能量项和外能量项，所述内能量项用于使得目标闭合曲线不断向内部紧缩且保持平滑，所述外能量项用于保证目标闭合曲线紧缩到目标眼底病灶边缘时停止；An energy function is constructed according to the initial curve function, the energy function includes an internal energy term and an external energy term, the internal energy term is used to make the target closed curve continuously shrink to the inside and remain smooth, and the external energy term is used to ensure Stop when the target closed curve shrinks to the edge of the target fundus lesion;

根据所述能量方程计算出边缘停止函数，所述边缘停止函数用于表示初始闭合曲线的受力情况；Calculate an edge stop function according to the energy equation, where the edge stop function is used to represent the force on the initial closed curve;

根据所述边缘停止函数对所述初始闭合曲线进行变形；Deform the initial closed curve according to the edge stop function;

当所述边缘停止函数为零时，生成目标闭合曲线。When the edge stop function is zero, a target closed curve is generated.
根据权利要求2所述的基于眼底OCT影像的病灶提取方法，所述根据所述初始曲线函数构建能量函数，所述能量函数包括内能量项和外能量项，包括：The method for extracting lesions based on fundus OCT images according to claim 2, wherein said constructing an energy function according to said initial curve function, said energy function including an internal energy term and an external energy term, comprises:

获取所述初始曲线函数上的控制点v；Acquiring the control point v on the initial curve function;

确定所述控制点的弹性能量项和弯曲能量项，所述弹性能量项为所述v的一阶导数的模，所述弯曲能量项为所述v的二阶导数的模；Determining an elastic energy term and a bending energy term of the control point, where the elastic energy term is the modulus of the first derivative of v, and the bending energy term is the modulus of the second derivative of v;

确定所述控制点的外能量项，所述外部能量项为所述控制点所在位置的图像局部特征；Determining an external energy term of the control point, where the external energy term is a local feature of the image where the control point is located;

根据所述弹性能量项、所述弯曲能量项和所述外能量项构建能量函数，能量函数的表达式为：An energy function is constructed according to the elastic energy term, the bending energy term, and the external energy term, and the expression of the energy function is:

其中，v(s)＝[x(s),y(s)]，s∈[0,1]，x(s)和y(s)分别表示每个控制点在图像中的坐标位置，s是以傅里叶变换形式描述边界的自变量，v表示初始曲线函数上的控制点，α、β为常数。Among them, v(s)=[x(s),y(s)], s∈[0,1], x(s) and y(s) respectively represent the coordinate position of each control point in the image, s It is the independent variable describing the boundary in the form of Fourier transform, v represents the control point on the initial curve function, and α and β are constants.
根据权利要求3所述的基于眼底OCT影像的病灶提取方法，所述确定所述控制点的外能量项，所述外部能量项为所述控制点所在位置的图像局部特征，包括：The method for extracting lesions based on fundus OCT images according to claim 3, wherein the determining an external energy item of the control point, the external energy item being a local feature of the image at the location of the control point, comprises:

将所述控制点的梯度算子作为所述图像局部特征；Using the gradient operator of the control point as a local feature of the image;

根据所述梯度算子确定所述外部能量项的表达式：

为所述梯度算子。 Determine the expression of the external energy term according to the gradient operator:

Is the gradient operator.
根据权利要求1所述的基于眼底OCT影像的病灶提取方法，所述通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数，包括：The method for extracting lesions based on fundus OCT images according to claim 1, wherein the determining the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm comprises:

生成初始化水平集函数u(x,y)，所述初始化水平集函数的表达式如下：Generate the initialization level set function u(x,y), the expression of the initialization level set function is as follows:

其中，d[(x,y),C]是点(x,y)到曲线C的最短有向距离；
Among them, d[(x,y),C] is the shortest directional distance from point (x,y) to curve C;

根据所述初始化水平集函数生成初始闭合曲线方程：Generate an initial closed curve equation according to the initialization level set function:

其中，u为目标水平集函数，
为梯度算子，div为散度算子，v为常数，g(I)为边缘停止函数；
Among them, u is the target level set function,
Is a gradient operator, div is a divergence operator, v is a constant, g(I) is an edge stop function;

根据所述初始闭合曲线方程计算得到所述预置的形态学蛇形曲线演化算法的零水平集函数
The zero-level set function of the preset morphological serpentine curve evolution algorithm is calculated according to the initial closed curve equation

判断所述零水平集函数是否收敛；Judging whether the zero level set function converges;

若所述零水平集函数收敛，则确定所述零水平集函数为所述初始曲线函数。If the zero level set function converges, it is determined that the zero level set function is the initial curve function.
根据权利要求1-5中任一所述的基于眼底OCT影像的病灶提取方法，在所述对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像之后，所述方法还包括：According to the method for extracting lesions based on fundus OCT images according to any one of claims 1-5, after the groove filling is performed on the inside of the target closed curve to obtain the lesion image of the target fundus lesion, the Methods also include:

获取医生标注标准图像；Obtain standard images marked by doctors;

将所述目标眼底病灶的病灶图像与所述医生标注标准图像进行对比，判断所述目标眼底病灶的病灶图像是否符合要求；Comparing the lesion image of the target fundus lesion with the standard image marked by the doctor to determine whether the lesion image of the target fundus lesion meets the requirements;

若所述目标眼底病灶的病灶图像符合要求，则将所述目标眼底病灶的病灶图像作为输出图像并输出。If the lesion image of the target fundus lesion meets the requirements, the lesion image of the target fundus lesion is used as an output image and output.
根据权利要求6所述的基于眼底OCT影像的病灶提取方法，所述方法还包括：The method for extracting lesions based on fundus OCT images according to claim 6, the method further comprising:

若所述目标眼底病灶的病灶图像不符合要求，则重新对所述目标眼底病灶的病灶图像的轮廓曲线进行演化。If the lesion image of the target fundus lesion does not meet the requirements, the contour curve of the lesion image of the target fundus lesion is re-evolved.
一种基于眼底OCT影像的病灶提取装置，包括：A focus extraction device based on OCT images of fundus, including:

第一获取单元，用于获取原始眼底光学相干断层扫描技术OCT影像；The first acquisition unit is used to acquire the original OCT image of fundus optical coherence tomography;

分割单元，用于对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；A segmentation unit, configured to perform segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image;

生成单元，用于在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；A generating unit, configured to generate an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes a target fundus lesion;

确定单元，用于通过预置的形态学蛇形曲线演化算法零所述初始闭合曲线的初始曲线函数；A determining unit for zeroing the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm;

第一演化单元，用于根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；The first evolution unit is configured to evolve the initial closed curve according to the initial curve function to obtain a target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion;

填充单元，用于对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。The filling unit is used for filling grooves inside the target closed curve to obtain a lesion image of the target fundus lesion.
根据权利要求8所述的基于眼底OCT影像的病灶提取装置，第一演化单元包括：The device for extracting lesions based on fundus OCT images according to claim 8, wherein the first evolution unit comprises:

构建模块，用于根据所述初始曲线函数构建能量函数，所述能量函数包括内能量项和外能量项，所述内能量项用于使得目标闭合曲线不断向内部紧缩且保持平滑，所述外能量项用于保证目标闭合曲线紧缩到目标眼底病灶边缘时停止；The construction module is used to construct an energy function according to the initial curve function, the energy function includes an internal energy term and an external energy term, and the internal energy term is used to make the target closed curve continuously shrink to the inside and remain smooth. The energy term is used to ensure that the target closed curve stops when it shrinks to the edge of the target fundus lesion;

计算模块，用于根据所述能量方程计算出边缘停止函数，所述边缘停止函数用于表示初始闭合曲线的受力情况；A calculation module, configured to calculate an edge stop function according to the energy equation, and the edge stop function is used to represent the force of the initial closed curve;

变形模块，用于根据所述边缘停止函数对所述初始闭合曲线进行变形；A deformation module, configured to deform the initial closed curve according to the edge stop function;

生成模块，当所述边缘停止函数为零时，用于生成目标闭合曲线。The generating module is used to generate the target closed curve when the edge stop function is zero.
根据权利要求9所述的基于眼底OCT影像的病灶提取装置，构建模块具体用于：According to the device for extracting lesions based on fundus OCT images according to claim 9, the building module is specifically configured to:

获取所述初始曲线函数上的控制点v；Acquiring the control point v on the initial curve function;

确定所述控制点的弹性能量项和弯曲能量项，所述弹性能量项为所述v的一阶导数的模，所述弯曲能量项为所述v的二阶导数的模；Determining an elastic energy term and a bending energy term of the control point, where the elastic energy term is the modulus of the first derivative of v, and the bending energy term is the modulus of the second derivative of v;

确定所述控制点的外能量项，所述外部能量项为所述控制点所在位置的图像局部特征；Determining an external energy term of the control point, where the external energy term is a local feature of the image where the control point is located;

根据所述弹性能量项、所述弯曲能量项和所述外能量项构建能量函数，能量函数的表达式为：An energy function is constructed according to the elastic energy term, the bending energy term, and the external energy term, and the expression of the energy function is:

其中，v(s)＝[x(s),y(s)]，s∈[0,1]，x(s)和y(s)分别表示每个控制点在图像中的坐标位置，s是以傅里叶变换形式描述边界的自变量，v表示初始曲线函数上的控制点，α、β为常数。Among them, v(s)=[x(s),y(s)], s∈[0,1], x(s) and y(s) respectively represent the coordinate position of each control point in the image, s It is the independent variable describing the boundary in the form of Fourier transform, v represents the control point on the initial curve function, and α and β are constants.
根据权利要求10所述的基于眼底OCT影像的病灶提取装置，构建模块具体还用于：According to the device for extracting lesions based on fundus OCT images according to claim 10, the building module is specifically further used for:

将所述控制点的梯度算子作为所述图像局部特征；Using the gradient operator of the control point as a local feature of the image;

根据所述梯度算子确定所述外部能量项的表达式：

为所述梯度算子。 Determine the expression of the external energy term according to the gradient operator:

Is the gradient operator.
根据权利要求8所述的基于眼底OCT影像的病灶提取装置，确定单元具体用于：According to the device for extracting lesions based on fundus OCT images according to claim 8, the determining unit is specifically configured to:

生成初始化水平集函数u(x,y)，所述初始化水平集函数的表达式如下：Generate the initialization level set function u(x,y), the expression of the initialization level set function is as follows:

其中，d[(x,y),C]是点(x,y)到曲线C的最短有向距离；
Among them, d[(x,y),C] is the shortest directional distance from point (x,y) to curve C;

根据所述初始化水平集函数生成初始闭合曲线方程：Generate an initial closed curve equation according to the initialization level set function:

其中，u为目标水平集函数，
为梯度算子，div为散度算子，v为常数，g(I)为边缘停止函数；
Among them, u is the target level set function,
Is a gradient operator, div is a divergence operator, v is a constant, g(I) is an edge stop function;

根据所述初始闭合曲线方程计算得到所述预置的形态学蛇形曲线演化算法的零水平集函数
The zero-level set function of the preset morphological serpentine curve evolution algorithm is calculated according to the initial closed curve equation

判断所述零水平集函数是否收敛；Judging whether the zero level set function converges;

若所述零水平集函数收敛，则确定所述零水平集函数为所述初始曲线函数。If the zero level set function converges, it is determined that the zero level set function is the initial curve function.
根据权利要求8-12中任一所述的基于眼底OCT影像的病灶提取装置，基于眼底OCT影像的病灶提取装置还包括：The device for extracting lesions based on fundus OCT images according to any one of claims 8-12, the device for extracting lesions based on fundus OCT images further comprises:

第二获取单元，用于获取医生标注标准图像；The second acquiring unit is used to acquire a standard image marked by the doctor;

判断单元，用于将所述目标眼底病灶的病灶图像与所述医生标注标准图像进行对比，判断所述目标眼底病灶的病灶图像是否符合要求；A judging unit, configured to compare the lesion image of the target fundus lesion with the doctor-marked standard image, and determine whether the lesion image of the target fundus lesion meets requirements;

输出单元，若所述目标眼底病灶的病灶图像符合要求，则用于将所述目标眼底病灶的病灶图像作为输出图像并输出。The output unit, if the focus image of the target fundus lesion meets the requirement, is used to output the focus image of the target fundus lesion as an output image.
根据权利要求13所述的基于眼底OCT影像的病灶提取装置，所述基于眼底OCT影像的病灶提取装置还包括：The device for extracting lesions based on fundus OCT images according to claim 13, wherein the device for extracting lesions based on fundus OCT images further comprises:

第二演化单元，若所述目标眼底病灶的病灶图像不符合要求，则用于重新对所述目标眼底病灶的病灶图像的轮廓曲线进行演化。The second evolution unit is used to re-evolve the contour curve of the lesion image of the target fundus lesion if the lesion image of the target fundus lesion does not meet the requirements.
一种基于眼底OCT影像的病灶提取设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现如下步骤：A focus extraction device based on fundus OCT images includes a memory, a processor, and a computer program stored on the memory and running on the processor, and the processor implements the following steps when the processor executes the computer program:

获取原始眼底光学相干断层扫描技术OCT影像；Obtain original fundus optical coherence tomography OCT images;

对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；Performing segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image;

在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；Generating an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes a target fundus lesion;

通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数；Determining the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm;

根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；Evolving the initial closed curve according to the initial curve function to obtain a target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion;

对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。Groove filling is performed on the inside of the target closed curve to obtain a lesion image of the target fundus lesion.
根据权利要求15所述的基于眼底OCT影像的病灶提取设备，所述处理器执行所述计算机程序实现所述根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓时，包括以下步骤：The device for extracting lesions based on fundus OCT images according to claim 15, wherein said processor executes said computer program to realize said evolution of said initial closed curve according to said initial curve function to obtain a target closed curve, said When the target curve function of the target closed curve represents the contour of the target fundus lesion, the following steps are included:

根据所述初始曲线函数构建能量函数，所述能量函数包括内能量项和外能量项，所述内能量项用于使得目标闭合曲线不断向内部紧缩且保持平滑，所述外能量项用于保证目标闭合曲线紧缩到目标眼底病灶边缘时停止；An energy function is constructed according to the initial curve function, the energy function includes an internal energy term and an external energy term, the internal energy term is used to make the target closed curve continuously shrink to the inside and remain smooth, and the external energy term is used to ensure Stop when the target closed curve shrinks to the edge of the target fundus lesion;

根据所述能量方程计算出边缘停止函数，所述边缘停止函数用于表示初始闭合曲线的受力情况；Calculate an edge stop function according to the energy equation, where the edge stop function is used to represent the force on the initial closed curve;

根据所述边缘停止函数对所述初始闭合曲线进行变形；Deform the initial closed curve according to the edge stop function;

当所述边缘停止函数为零时，生成目标闭合曲线。When the edge stop function is zero, a target closed curve is generated.
根据权利要求16所述的基于眼底OCT影像的病灶提取设备，所述处理器执行所述计算机程序实现所述根据所述初始曲线函数构建能量函数，所述能量函数包括内能量项和外能量项时，包括以下步骤：The device for extracting lesions based on fundus OCT images according to claim 16, wherein the processor executes the computer program to implement the construction of an energy function according to the initial curve function, the energy function including an internal energy term and an external energy term , Including the following steps:

获取所述初始曲线函数上的控制点v；Acquiring the control point v on the initial curve function;

确定所述控制点的弹性能量项和弯曲能量项，所述弹性能量项为所述v的一阶导数的模，所述弯曲能量项为所述v的二阶导数的模；Determining an elastic energy term and a bending energy term of the control point, where the elastic energy term is the modulus of the first derivative of v, and the bending energy term is the modulus of the second derivative of v;

确定所述控制点的外能量项，所述外部能量项为所述控制点所在位置的图像局部特征；Determining an external energy term of the control point, where the external energy term is a local feature of the image where the control point is located;

根据所述弹性能量项、所述弯曲能量项和所述外能量项构建能量函数，能量函数的表达式为：An energy function is constructed according to the elastic energy term, the bending energy term, and the external energy term, and the expression of the energy function is:

其中，v(s)＝[x(s),y(s)]，s∈[0,1]，x(s)和y(s)分别表示每个控制点在图像中的坐标位置，s是以傅里叶变换形式描述边界的自变量，v表示初始曲线函数上的控制点，α、β为常数。Among them, v(s)=[x(s),y(s)], s∈[0,1], x(s) and y(s) respectively represent the coordinate position of each control point in the image, s It is the independent variable describing the boundary in the form of Fourier transform, v represents the control point on the initial curve function, and α and β are constants.
根据权利要求17所述的基于眼底OCT影像的病灶提取设备，所述处理器执行所述计算机程序实现所述确定所述控制点的外能量项，所述外部能量项为所述控制点所在位置的图像局部特征时，包括以下步骤：The device for extracting lesions based on fundus OCT images according to claim 17, wherein the processor executes the computer program to realize the determination of the external energy item of the control point, and the external energy item is the location of the control point When the local features of the image, include the following steps:

将所述控制点的梯度算子作为所述图像局部特征；Using the gradient operator of the control point as a local feature of the image;

根据所述梯度算子确定所述外部能量项的表达式：

为所述梯度算子。 Determine the expression of the external energy term according to the gradient operator:

Is the gradient operator.
根据权利要求15所述的基于眼底OCT影像的病灶提取设备，所述处理器执行所述计算机程序实现所述通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数时，包括以下步骤：The device for extracting lesions based on fundus OCT images according to claim 15, wherein the processor executes the computer program to realize the time when the initial curve function of the initial closed curve is determined by a preset morphological serpentine curve evolution algorithm , Including the following steps:

生成初始化水平集函数u(x,y)，所述初始化水平集函数的表达式如下：Generate the initialization level set function u(x,y), the expression of the initialization level set function is as follows:

其中，d[(x,y),C]是点(x,y)到曲线C的最短有向距离；
Among them, d[(x,y),C] is the shortest directional distance from point (x,y) to curve C;

根据所述初始化水平集函数生成初始闭合曲线方程：Generate an initial closed curve equation according to the initialization level set function:

其中，u为目标水平集函数，
为梯度算子，div为散度算子，v为常数，g(I)为边缘停止函数；
Among them, u is the target level set function,
Is a gradient operator, div is a divergence operator, v is a constant, g(I) is an edge stop function;

根据所述初始闭合曲线方程计算得到所述预置的形态学蛇形曲线演化算法的零水平集函数
The zero-level set function of the preset morphological serpentine curve evolution algorithm is calculated according to the initial closed curve equation

判断所述零水平集函数是否收敛；Judging whether the zero level set function converges;

若所述零水平集函数收敛，则确定所述零水平集函数为所述初始曲线函数。If the zero level set function converges, it is determined that the zero level set function is the initial curve function.
一种计算机可读存储介质，所述计算机可读存储介质中存储有指令，当所述指令在计算机上运行时，使得计算机执行如下步骤：A computer-readable storage medium in which instructions are stored, and when the instructions are run on a computer, the computer executes the following steps:

获取原始眼底光学相干断层扫描技术OCT影像；Obtain original fundus optical coherence tomography OCT images;

对所述原始眼底OCT影像进行分割预处理，得到预处理的眼底OCT图像；Performing segmentation preprocessing on the original fundus OCT image to obtain a preprocessed fundus OCT image;

在所述预处理的眼底OCT图像中生成一个初始闭合曲线，所述初始闭合曲线的内部包括目标眼底病灶；Generating an initial closed curve in the preprocessed fundus OCT image, and the inside of the initial closed curve includes a target fundus lesion;

通过预置的形态学蛇形曲线演化算法确定所述初始闭合曲线的初始曲线函数；Determining the initial curve function of the initial closed curve through a preset morphological serpentine curve evolution algorithm;

根据所述初始曲线函数对所述初始闭合曲线进行演化，得到目标闭合曲线，所述目标闭合曲线的目标曲线函数表示所述目标眼底病灶的轮廓；Evolving the initial closed curve according to the initial curve function to obtain a target closed curve, and the target curve function of the target closed curve represents the contour of the target fundus lesion;

对所述目标闭合曲线的内部进行凹槽填充，得到所述目标眼底病灶的病灶图像。Groove filling is performed on the inside of the target closed curve to obtain a lesion image of the target fundus lesion.