CN115423834A - Image entropy threshold segmentation method, device, electronic equipment and storage medium - Google Patents
Image entropy threshold segmentation method, device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN115423834A CN115423834A CN202211039468.2A CN202211039468A CN115423834A CN 115423834 A CN115423834 A CN 115423834A CN 202211039468 A CN202211039468 A CN 202211039468A CN 115423834 A CN115423834 A CN 115423834A
- Authority
- CN
- China
- Prior art keywords
- threshold
- entropy
- image
- probability
- background
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/84—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using probabilistic graphical models from image or video features, e.g. Markov models or Bayesian networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20076—Probabilistic image processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
Description
技术领域technical field
本发明属于优化调度技术领域,特别涉及一种图像的熵阈值分割方法、装置、电子设备及存储介质。The invention belongs to the technical field of optimal scheduling, and in particular relates to an image entropy threshold segmentation method, device, electronic equipment and storage medium.
背景技术Background technique
图像分割是医学图像分析的主要过程之一。它在医学领域的许多应用中都有用,包括量化病变、手术模拟、辅助手术决策、辅助诊断多发性硬化症等。而阈值化是一种流行的图像分割技术,特别是在医学图像处理领域。图像阈值化的主要挑战是根据图像中物体和背景的强度分布确定最佳阈值,通过该阈值来将全图的像素划分为目标和背景两个类别。根据核磁共振图像的成像原理,不同的组织在图像中会呈现不同的强度和分布规律,因此熵阈值法通过可以最大化或最小化一个基于灰度构造出来的代价函数,来寻找最佳的分割阈值。Image segmentation is one of the main processes in medical image analysis. It is useful in many applications in the medical field, including quantifying lesions, surgical simulation, assisting surgical decision-making, aiding in the diagnosis of multiple sclerosis, and more. Thresholding is a popular image segmentation technique, especially in the field of medical image processing. The main challenge of image thresholding is to determine the optimal threshold according to the intensity distribution of objects and background in the image, and divide the pixels of the whole image into two categories: object and background through this threshold. According to the imaging principle of MRI images, different tissues will show different intensities and distribution rules in the image, so the entropy threshold method can maximize or minimize a cost function constructed based on grayscale to find the best segmentation. threshold.
现有技术中,提出利用经典Tsallis熵来对图像进行分割,证明了Tsallis熵用于图像分割特别是医学图像分割的优越性。以及现有技术将Tsallis熵推广到二维,除了像素的灰度外,还考虑了像素与其相邻像素之间的相关关系,进一步的提升了Tsallis熵阈值分割的准确度。In the prior art, it is proposed to use classical Tsallis entropy to segment images, which proves the superiority of Tsallis entropy for image segmentation, especially medical image segmentation. And the existing technology extends the Tsallis entropy to two dimensions. In addition to the gray level of the pixel, the correlation between the pixel and its adjacent pixels is also considered, which further improves the accuracy of the Tsallis entropy threshold segmentation.
但是,上述现有技术难以平衡分割的准确度和效率,为了提升准确度,二维熵阈值方法需要在二维空间搜索最优阈值,容易陷入局部最优且耗时长。However, it is difficult for the above existing technologies to balance the accuracy and efficiency of segmentation. In order to improve the accuracy, the two-dimensional entropy threshold method needs to search for the optimal threshold in two-dimensional space, which is easy to fall into local optimum and takes a long time.
发明内容Contents of the invention
本说明书实施例的目的是提供一种图像的熵阈值分割方法、装置、电子设备及存储介质。The purpose of the embodiments of this specification is to provide an image entropy threshold segmentation method, device, electronic equipment and storage medium.
为解决上述技术问题,本申请实施例通过以下方式实现的:In order to solve the above technical problems, the embodiments of the present application are implemented in the following ways:
第一方面,本申请提供一种图像的熵阈值分割方法,该方法包括:In a first aspect, the present application provides a method for entropy threshold segmentation of an image, the method comprising:
利用贝叶斯后验估计,根据每个阈值下待分割图像的目标类概率和背景类概率,确定不同阈值的分割下,每个像素属于目标类和背景类的后验概率;Using Bayesian posterior estimation, according to the target class probability and background class probability of the image to be segmented under each threshold, determine the posterior probability of each pixel belonging to the target class and background class under the segmentation of different thresholds;
计算每个阈值下待分割图像全图像素的Tsallis-BE熵的值,其中,Tsallis-BE熵由后验概率和Tsaliis熵结合构成;Calculate the value of the Tsallis-BE entropy of the full picture pixel of the image to be segmented under each threshold, where the Tsallis-BE entropy is composed of the combination of the posterior probability and the Tsaliis entropy;
从所有Tsallis-BE熵的值中找出最大熵值,将最大熵值对应的阈值确定为最优阈值。Find the maximum entropy value from all Tsallis-BE entropy values, and determine the threshold corresponding to the maximum entropy value as the optimal threshold.
在其中一个实施例中,该方法还包括:In one embodiment, the method also includes:
根据不同阈值将待分割图像分割为目标区域和背景区域;Segment the image to be segmented into a target area and a background area according to different thresholds;
计算每个阈值下分割出的目标区域的目标像素数量和背景区域的背景像素数量;Calculate the number of target pixels in the target area segmented under each threshold and the number of background pixels in the background area;
根据不同阈值下的目标像素数量和背景像素数量,确定每个阈值下待分割图像的目标类概率和背景类概率。According to the number of target pixels and the number of background pixels under different thresholds, determine the target class probability and background class probability of the image to be segmented under each threshold.
在其中一个实施例中,根据不同阈值下的目标像素数量和背景像素数量,确定每个阈值下待分割图像的目标类概率和背景类概率,包括:In one of the embodiments, according to the number of target pixels and the number of background pixels under different thresholds, the target class probability and the background class probability of the image to be segmented under each threshold are determined, including:
po,t(i)=P[f(s)=i|s∈So,t]p o,t (i)=P[f(s)=i|s∈S o,t ]
pb,t(i)=P[f(s)=i|s∈Sb,t]p b,t (i)=P[f(s)=i|s∈S b,t ]
S=So,t∪Sb,t S=S o,t ∪S b,t
其中,po,t(i)为在阈值t时目标像素值为i的目标类概率,pb,t(i)为在阈值t时背景像素值为i的背景类概率;S为待分割图像的全部像素的集合;So,t为阈值t时的目标像素集合;Sb,t为阈值t时的背景像素集合。Among them, p o,t (i) is the target class probability of the target pixel value i at the threshold t, p b,t (i) is the background class probability of the background pixel value i at the threshold t; S is the target class to be segmented The set of all pixels of the image; S o,t is the target pixel set at the threshold t; S b,t is the background pixel set at the threshold t.
在其中一个实施例中,假设待分割图像中像素灰度值的最小值和最大值分别为t_min和t_max,则将阈值遍历的区间设置为[t_min+2,t_max+2]。In one of the embodiments, assuming that the minimum value and the maximum value of the gray value of the pixel in the image to be segmented are t_min and t_max respectively, the interval traversed by the threshold value is set to [t_min+2, t_max+2].
在其中一个实施例中,利用贝叶斯后验估计,根据每个阈值下待分割图像的目标类概率和背景类概率,确定不同阈值的分割下,每个像素属于目标类和背景类的后验概率,包括:In one of the embodiments, using Bayesian posterior estimation, according to the target class probability and background class probability of the image to be segmented under each threshold, the posterior probability of each pixel belonging to the target class and background class under different thresholds is determined. probabilities, including:
假设γ(t)表示像素在阈值t时属于目标的先验概率,则1-γ(t)表示像素在阈值t时属于背景的先验概率,po,t(i)为在阈值t时目标像素值为i的目标类概率,pb,t(i)为在阈值t时背景像素值为i的背景类概率;Suppose γ(t) represents the prior probability that the pixel belongs to the target at the threshold t, then 1-γ(t) represents the prior probability that the pixel belongs to the background at the threshold t, and p o,t (i) is the prior probability at the threshold t The target class probability of the target pixel value i, p b,t (i) is the background class probability of the background pixel value i at the threshold t;
一个像素值为i的像素属于目标类和背景类的后验概率可以由贝叶斯规则分别表示为:The posterior probability of a pixel with pixel value i belonging to the target class and the background class can be expressed by Bayesian rule as:
其中,pt(i)可以通过以下公式计算:Among them, p t (i) can be calculated by the following formula:
pt(i)=γ(t)po,t(i)+(1-γ(t))pb,t(i)。p t (i)=γ(t)p o,t (i)+(1-γ(t))p b,t (i).
在其中一个实施例中,目标类和背景类的Tsallis-BE熵的表达式分别为:In one of the embodiments, the expressions of the Tsallis-BE entropy of the target class and the background class are respectively:
其中,L为待分割图像中像素灰度最高的值。Among them, L is the highest pixel gray value in the image to be segmented.
在其中一个实施例中,从所有Tsallis-BE熵的值中找出最大熵值topt,包括:In one of the embodiments, the maximum entropy value t opt is found from all Tsallis-BE entropy values, including:
第二方面,本申请提供一种图像的熵阈值分割装置,该装置包括:In a second aspect, the present application provides an image entropy threshold segmentation device, which includes:
概率确定模块,用于利用贝叶斯后验估计,根据每个阈值下待分割图像的目标类概率和背景类概率,确定不同阈值的分割下,每个像素属于目标类和背景类的后验概率;The probability determination module is used to use the Bayesian posterior estimation, according to the target class probability and the background class probability of the image to be segmented under each threshold, to determine the posteriori of each pixel belonging to the target class and the background class under the segmentation of different thresholds probability;
熵值确定模块,用于计算每个阈值下待分割图像全图像素的Tsallis-BE熵的值,其中,Tsallis-BE熵由后验概率和Tsaliis熵结合构成;The entropy value determination module is used to calculate the value of the Tsallis-BE entropy of the full image pixel of the image to be segmented under each threshold, wherein the Tsallis-BE entropy is composed of a combination of posterior probability and Tsaliis entropy;
查找模块,用于从所有Tsallis-BE熵的值中找出最大熵值,将最大熵值对应的阈值确定为最优阈值。The search module is used to find out the maximum entropy value from all Tsallis-BE entropy values, and determine the threshold corresponding to the maximum entropy value as the optimal threshold.
第三方面,本申请提供一种电子设备,包括存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序,处理器执行程序时实现如第一方面的图像的熵阈值分割方法。In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and operable on the processor. When the processor executes the program, the entropy threshold segmentation method of an image as in the first aspect is implemented. .
第四方面,本申请提供一种可读存储介质,其上存储有计算机程序,该程序被处理器执行时实现如第一方面的图像的熵阈值分割方法。In a fourth aspect, the present application provides a readable storage medium on which a computer program is stored, and when the program is executed by a processor, the entropy threshold segmentation method of an image as in the first aspect is implemented.
由以上本说明书实施例提供的技术方案可见,该方案:利用贝叶斯后验概率来给出每个像素属于目标和背景的不确定性,改进了Tsallis熵对图像本身信息利用不充分的缺点,同时又避免了引入更多图像特征信息所带来的维度增加的问题,仍然只需要在一维空间求解最优阈值,做到了高精度分割的同时,降低算法的耗时,提升了熵阈值方法的性能。It can be seen from the technical solution provided by the above embodiments of this specification that this solution: uses Bayesian posterior probability to give the uncertainty of each pixel belonging to the target and background, and improves the shortcoming of insufficient utilization of Tsallis entropy for the information of the image itself , while avoiding the problem of dimension increase caused by the introduction of more image feature information, it is still only necessary to solve the optimal threshold in one-dimensional space, achieving high-precision segmentation while reducing the time-consuming algorithm and improving the entropy threshold performance of the method.
附图说明Description of drawings
为了更清楚地说明本说明书实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本说明书中记载的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动性的前提下,还可以根据这些附图获得其他的附图。In order to more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments described in this specification. Those skilled in the art can also obtain other drawings based on these drawings without any creative effort.
图1为本申请提供的图像的熵阈值分割方法的流程示意图;Fig. 1 is a schematic flow chart of the entropy threshold segmentation method of an image provided by the present application;
图2为本申请提供的待分割图像原图及分割结果图;Fig. 2 is the original image of the image to be segmented and the segmented result figure provided by the application;
图3为本申请提供的图像的熵阈值分割装置的结构示意图;FIG. 3 is a schematic structural diagram of an image entropy threshold segmentation device provided by the present application;
图4为本申请提供的电子设备的结构示意图。FIG. 4 is a schematic structural diagram of an electronic device provided by the present application.
具体实施方式detailed description
为了使本技术领域的人员更好地理解本说明书中的技术方案,下面将结合本说明书实施例中的附图,对本说明书实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本说明书一部分实施例,而不是全部的实施例。基于本说明书中的实施例,本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例,都应当属于本说明书保护的范围。In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described The embodiments are only some of the embodiments in this specification, not all of them. Based on the embodiments in this specification, all other embodiments obtained by persons of ordinary skill in the art without creative efforts shall fall within the protection scope of this specification.
以下描述中,为了说明而不是为了限定,提出了诸如特定系统结构、技术之类的具体细节,以便透彻理解本申请实施例。然而,本领域的技术人员应当清楚,在没有这些具体细节的其它实施例中也可以实现本申请。在其它情况中,省略对众所周知的系统、装置、电路以及方法的详细说明,以免不必要的细节妨碍本申请的描述。In the following description, specific details such as specific system structures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
在不背离本申请的范围或精神的情况下,可对本申请说明书的具体实施方式做多种改进和变化,这对本领域技术人员而言是显而易见的。由本申请的说明书得到的其他实施方式对技术人员而言是显而易见得的。本申请说明书和实施例仅是示例性的。It will be obvious to those skilled in the art that various modifications and changes can be made to the specific embodiments described in the present application without departing from the scope or spirit of the present application. Other embodiments will be apparent to those skilled in the art from the description of this application. The specification and examples in this application are exemplary only.
关于本文中所使用的“包含”、“包括”、“具有”、“含有”等等,均为开放性的用语,即意指包含但不限于。As used herein, "comprising", "comprising", "having", "comprising" and so on are all open terms, meaning including but not limited to.
现有技术难以平衡分割的准确度和效率,为了提升准确度,二维熵阈值方法需要在二维空间搜索最优阈值,容易陷入局部最优且耗时长。It is difficult for the existing technology to balance the accuracy and efficiency of segmentation. In order to improve the accuracy, the two-dimensional entropy threshold method needs to search for the optimal threshold in two-dimensional space, which is easy to fall into local optimum and takes a long time.
基于上述缺陷,本申请提出一种图像的熵阈值分割方法,将基于贝叶斯后验估计的灰度概率分布引入Tsallis的结构中,使得本方法仅需要在一维空间搜索就能实现高准确度和高效率的分割。Based on the above defects, this application proposes an image entropy threshold segmentation method, which introduces the gray probability distribution based on Bayesian posterior estimation into the structure of Tsallis, so that this method can achieve high accuracy only by searching in one-dimensional space. Degree and efficient division.
本申请提出的图像的熵阈值分割方法,是针对脑部核磁共振图像的成像特点和特定灰度分布,能有效的分割大脑图像中白质(white matter,WM)、灰质(gray matter,GM)和脑脊液(cerebrospinal fluid,CSF)等不同组织的熵阈值分割方法。The image entropy threshold segmentation method proposed in this application is aimed at the imaging characteristics and specific gray scale distribution of brain MRI images, and can effectively segment white matter (white matter, WM), gray matter (gray matter, GM) and Entropy threshold segmentation method for different tissues such as cerebrospinal fluid (CSF).
下面结合附图和实施例对本发明进一步详细说明。The present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments.
参照图1,其示出了适用于本申请实施例提供的图像的熵阈值分割方法的流程示意图。该图像的熵阈值分割方法的主要任务是确定一个最佳阈值t以区分目标像素和背景像素。Referring to FIG. 1 , it shows a schematic flowchart of an entropy threshold segmentation method applicable to an image provided by an embodiment of the present application. The main task of the image entropy threshold segmentation method is to determine an optimal threshold t to distinguish target pixels from background pixels.
如图1所示,一种图像的熵阈值分割方法,可以包括:As shown in Figure 1, an entropy threshold segmentation method of an image may include:
S110、利用贝叶斯后验估计,根据每个阈值下待分割图像的目标类概率和背景类概率,确定不同阈值的分割下,每个像素属于目标类和背景类的后验概率。S110. Using Bayesian posterior estimation, according to the target class probability and the background class probability of the image to be segmented under each threshold, determine the posterior probability of each pixel belonging to the target class and the background class under segmentation with different thresholds.
具体的,待分割图像可以为核磁共振图像,示例性的,该待分割图像为典型人脑的核磁共振切片图像,该实例中白质(WM)和灰质(GM)被视为目标区域,而脑脊液(CSF)被视为背景区域。Specifically, the image to be segmented may be an MRI image. Exemplarily, the image to be segmented is an MRI slice image of a typical human brain. In this example, white matter (WM) and gray matter (GM) are regarded as target regions, while cerebrospinal fluid (CSF) are considered as background regions.
其中,每个阈值下待分割图像的目标类概率和背景类概率,可以通过下述方式进行确定。Wherein, the target class probability and the background class probability of the image to be segmented under each threshold can be determined in the following manner.
一个实施例中,该方法还包括:In one embodiment, the method also includes:
根据不同阈值将待分割图像分割为目标区域和背景区域;Segment the image to be segmented into a target area and a background area according to different thresholds;
计算每个阈值下分割出的目标区域的目标像素数量和背景区域的背景像素数量;Calculate the number of target pixels in the target area segmented under each threshold and the number of background pixels in the background area;
根据不同阈值下的目标像素数量和背景像素数量,确定每个阈值下待分割图像的目标类概率和背景类概率,包括:According to the number of target pixels and the number of background pixels under different thresholds, determine the target class probability and background class probability of the image to be segmented under each threshold, including:
po,t(i)=P[f(s)=i|s∈So,t]p o,t (i)=P[f(s)=i|s∈S o,t ]
pb,t(i)=P[f(s)=i|s∈Sb,t]p b,t (i)=P[f(s)=i|s∈S b,t ]
S=So,t∪Sb,t S=S o,t ∪S b,t
其中,po,t(i)为在阈值t时目标像素值为i的目标类概率(或称为目标像素的概率密度函数),pb,t(i)为在阈值t时背景像素值为i的背景类概率(或称为背景像素的概率密度函数);S为待分割图像的全部像素的集合;So,t为阈值t时的目标像素集合;Sb,t为阈值t时的背景像素集合。Among them, p o,t (i) is the target class probability (or called the probability density function of the target pixel) of the target pixel value i at the threshold t, p b,t (i) is the background pixel value at the threshold t is the background class probability of i (or called the probability density function of background pixels); S is the set of all pixels of the image to be segmented; S o,t is the target pixel set when the threshold t; S b,t is the threshold t background pixel collection.
假设目标和背景的分布是高斯分布,分别用和其中,mo(t)、mb(t)、分别是目标像素和背景像素的平均值和标准差,则po,t(i)和pb,t(i)分别可以表示为:Assuming that the distribution of the target and the background is a Gaussian distribution, use with Among them, m o (t), m b (t), are the mean and standard deviation of the target pixel and the background pixel respectively, then p o,t (i) and p b,t (i) can be expressed as:
具体的,根据不同阈值将核磁共振图像分割为目标区域和背景区域,计算每个阈值下分割出的目标区域像素数量和背景区域像素数量,得到目标类和背景类的概率,用来衡量每个像素属于不同类别的不确定性。Specifically, the MRI image is segmented into the target area and the background area according to different thresholds, the number of pixels in the target area and the number of pixels in the background area are calculated under each threshold, and the probability of the target class and the background class is obtained, which is used to measure each Pixels belong to different categories of uncertainty.
可以理解的,图像的灰度区间为0~255,为了防止边界溢出问题,将区域缩小2作为确定最佳阈值的搜索区域。It can be understood that the grayscale range of the image is 0-255, and in order to prevent the boundary overflow problem, the area is reduced by 2 as the search area for determining the optimal threshold.
一个实施例中,假设待分割图像中像素灰度值的最小值和最大值分别为t_min和t_max,则将阈值遍历的区间设置为[t_min+2,t_max+2]。In one embodiment, assuming that the minimum value and the maximum value of the pixel gray value in the image to be segmented are t_min and t_max respectively, the interval traversed by the threshold value is set to [t_min+2, t_max+2].
通过上述方式确定每个阈值下待分割图像的目标类概率和背景类概率,再利用贝叶斯后验估计,确定不同阈值的分割下,每个像素属于目标类和背景类的概率分布。Determine the target class probability and background class probability of the image to be segmented under each threshold by the above method, and then use Bayesian posterior estimation to determine the probability distribution of each pixel belonging to the target class and background class under different thresholds.
具体的,假设γ(t)表示像素在阈值t时属于目标的先验概率,则1-γ(t)表示像素在阈值t时属于背景的先验概率。其中,γ(t)通过下式估算:Specifically, assuming that γ(t) represents the prior probability that the pixel belongs to the target at the threshold t, then 1-γ(t) represents the prior probability that the pixel belongs to the background at the threshold t. Among them, γ(t) is estimated by the following formula:
因此,一个像素值为i的像素属于目标类和背景类的后验概率可以由贝叶斯规则分别表示为:Therefore, the posterior probability of a pixel with pixel value i belonging to the target class and the background class can be expressed by Bayesian rule as:
其中,pt(i)可以通过以下公式计算:Among them, p t (i) can be calculated by the following formula:
pt(i)=γ(t)po,t(i)+(1-γ(t))pb,t(i)p t (i)=γ(t)p o,t (i)+(1-γ(t))p b,t (i)
S120、计算每个阈值下待分割图像全图像素的Tsallis-BE熵的值,其中,Tsallis-BE熵由后验概率和Tsaliis熵结合构成。S120. Calculate the value of the Tsallis-BE entropy of the pixels in the entire image of the image to be segmented at each threshold, wherein the Tsallis-BE entropy is formed by combining the posterior probability and the Tsaliis entropy.
具体的,目前,目标类和背景类的Tsaliis熵可以分别表示为:Specifically, at present, the Tsaliis entropy of the target class and the background class can be expressed as:
其中,q是表征不可加性程度的熵指数,pi是像素i的概率;其中,给定阈值t,图像被分为背景区域和目标区域,其中灰度区间分别为{0,1,…,t}和{t+1,…,L-1}, Among them, q is the entropy index that characterizes the degree of non-additivity, p i is the probability of pixel i; where, given the threshold t, the image is divided into background area and target area, where the grayscale intervals are {0,1,… ,t} and {t+1,...,L-1},
后验概率和Tsaliis熵结合后,目标类和背景类的Tsallis-BE熵的表达式分别为:After the combination of posterior probability and Tsaliis entropy, the expressions of Tsallis-BE entropy of target class and background class are respectively:
其中,L为待分割图像中像素灰度最高的值,表示全图所有像素灰度值都在0-L之间。Among them, L is the highest pixel gray value in the image to be segmented, which means that the gray value of all pixels in the whole image is between 0-L.
根据目标类和背景类的Tsallis-BE熵的表达式,分别计算每个阈值下待分割图像全图像素的Tsallis-BE熵的值。According to the expressions of the Tsallis-BE entropy of the target class and the background class, the value of the Tsallis-BE entropy of the whole pixel of the image to be segmented under each threshold is calculated respectively.
S130、从所有Tsallis-BE熵的值中找出最大熵值,将最大熵值对应的阈值确定为最优阈值。S130. Find the maximum entropy value from all Tsallis-BE entropy values, and determine the threshold corresponding to the maximum entropy value as the optimal threshold.
具体的,找到最大熵值对应的阈值,即认为是它将整个图像划分为目标区域和背景区域的最优阈值。Specifically, the threshold corresponding to the maximum entropy value is found, which is considered to be the optimal threshold for dividing the entire image into the target area and the background area.
可选的,从所有Tsallis-BE熵的值中找出最大熵值topt,包括:Optionally, find the maximum entropy value t opt from all Tsallis-BE entropy values, including:
如图2所示,左侧为待分割图像的原图,右侧为采用本申请图像的熵阈值分割方法算出最优阈值,并根据最优阈值分割后的分割结果图。As shown in Figure 2, the left side is the original image of the image to be segmented, and the right side is the image of the segmentation result after the optimal threshold is calculated using the entropy threshold segmentation method of the image of the present application and segmented according to the optimal threshold.
本申请实施例中,使用贝叶斯后验概率代替经典Tsallis熵中每个灰度值的概率,提出了一种新的用于图像分割的Tsallis-BE熵。使用一维熵的形式,取得了媲美二维熵阈值分割的准确度,并且相比二维熵,大大缩短了分割算法所需要的时间。并且本申请利用贝叶斯后验概率的形式,考虑了像素的类不确定性,相比于其他一维熵阈值法,加深了对图像本身的信息的利用。In the embodiment of the present application, Bayesian posterior probability is used to replace the probability of each gray value in classical Tsallis entropy, and a new Tsallis-BE entropy for image segmentation is proposed. Using the form of one-dimensional entropy, the accuracy of threshold segmentation comparable to two-dimensional entropy is achieved, and compared with two-dimensional entropy, the time required for the segmentation algorithm is greatly shortened. In addition, the present application uses the form of Bayesian posterior probability and considers the class uncertainty of pixels. Compared with other one-dimensional entropy threshold methods, it deepens the utilization of information of the image itself.
本申请提出的Tsallis-BE熵结构,与经典Tsallis熵假设每个阈值分割出的目标类和背景类是相互独立分布的不同,本申请考虑两类物质边界处像素的互相影响,利用贝叶斯后验概率来给出每个像素属于目标和背景的不确定性,改进了Tsallis熵对图像本身信息利用不充分的缺点,同时又避免了引入更多图像特征信息所带来的维度增加的问题,仍然只需要在一维空间求解最优阈值,做到了高精度分割的同时,降低算法的耗时,提升了熵阈值方法的性能。The Tsallis-BE entropy structure proposed by this application is different from the classic Tsallis entropy which assumes that the target class and background class segmented by each threshold are independently distributed. The posterior probability is used to give the uncertainty that each pixel belongs to the target and the background, which improves the shortcoming of Tsallis entropy’s insufficient utilization of the information of the image itself, and at the same time avoids the problem of dimension increase caused by introducing more image feature information. , still only need to solve the optimal threshold in one-dimensional space, achieve high-precision segmentation, reduce the time consumption of the algorithm, and improve the performance of the entropy threshold method.
实验验证Experimental verification
在医学图像分割的两个基准数据集Brainweb和MRBS13上做了和经典方法(包括方法一和方法二)的对比实验。定量结果如表1所示,我们采用了三个评价分割效果的通用评价指标DSC(Dice Similariy Coefficient,戴斯相似性系数),JC(Jaccard SimilarityCoefficient,杰卡德相似系数),Accuracy(准确度)来进行评价验证。可以看到,得益于提出的新型熵结构,本申请在两个基准数据集上取得了比方法一和方法二更好的效果。说明,方法一对应文献Albuquerque M P D,Esquef I A,Mello A R G,et al.Imagethresholding using Tsallis entropy[J].Pattern Recognition Letters,2004,25(9):1059-1065中记载的方法,方法二对应文献Ostu N.A threshold selection method fromgray-histogram[J].IEEE Transactions on Systems,Man,and Cybernetics,1979,9(1):62-66中记载的方法。A comparative experiment with classical methods (including method 1 and method 2) was done on two benchmark datasets of medical image segmentation, Brainweb and MRBS13. The quantitative results are shown in Table 1. We used three general evaluation indicators DSC (Dice Similarity Coefficient, Dice Similarity Coefficient), JC (Jaccard Similarity Coefficient, Jaccard Similarity Coefficient), Accuracy (accuracy) to evaluate the segmentation effect. for evaluation verification. It can be seen that thanks to the proposed new entropy structure, this application has achieved better results than Method 1 and Method 2 on the two benchmark datasets. Note that method one corresponds to the method recorded in the literature Albuquerque M P D, Esquef I A, Mello A R G, et al. Imagethresholding using Tsallis entropy[J]. N. A threshold selection method from gray-histogram [J]. The method described in IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62-66.
表1医学图像分割数据集上的定量结果(%)Table 1 Quantitative results (%) on medical image segmentation datasets
参照图3,其示出了根据本申请一个实施例描述的图像的熵阈值分割装置的结构示意图。Referring to FIG. 3 , it shows a schematic structural diagram of an apparatus for entropy threshold segmentation of an image described according to an embodiment of the present application.
如图3所示,图像的熵阈值分割装置300,可以包括:As shown in Figure 3, the image entropy threshold segmentation device 300 may include:
概率确定模块310,用于利用贝叶斯后验估计,根据每个阈值下待分割图像的目标类概率和背景类概率,确定不同阈值的分割下,每个像素属于目标类和背景类的后验概率;The
熵值确定模块320,用于计算每个阈值下待分割图像全图像素的Tsallis-BE熵的值,其中,Tsallis-BE熵由后验概率和Tsaliis熵结合构成;The entropy
查找模块330,用于从所有Tsallis-BE熵的值中找出最大熵值,将最大熵值对应的阈值确定为最优阈值。The
可选的,该图像的熵阈值分割装置300还用于:Optionally, the image entropy threshold segmentation device 300 is also used for:
根据不同阈值将待分割图像分割为目标区域和背景区域;Segment the image to be segmented into a target area and a background area according to different thresholds;
计算每个阈值下分割出的目标区域的目标像素数量和背景区域的背景像素数量;Calculate the number of target pixels in the target area segmented under each threshold and the number of background pixels in the background area;
根据不同阈值下的目标像素数量和背景像素数量,确定每个阈值下待分割图像的目标类概率和背景类概率。According to the number of target pixels and the number of background pixels under different thresholds, determine the target class probability and background class probability of the image to be segmented under each threshold.
可选的,该图像的熵阈值分割装置300还用于:Optionally, the image entropy threshold segmentation device 300 is also used for:
po,t(i)=P[f(s)=i|s∈So,t]p o,t (i)=P[f(s)=i|s∈S o,t ]
pb,t(i)=P[f(s)=i|s∈Sb,t]p b,t (i)=P[f(s)=i|s∈S b,t ]
S=So,t∪Sb,t S=S o,t ∪S b,t
其中,po,t(i)为在阈值t时目标像素值为i的目标类概率,pb,t(i)为在阈值t时背景像素值为i的背景类概率;S为待分割图像的全部像素的集合;So,t为阈值t时的目标像素集合;Sb,t为阈值t时的背景像素集合。Among them, p o,t (i) is the target class probability of the target pixel value i at the threshold t, p b,t (i) is the background class probability of the background pixel value i at the threshold t; S is the target class to be segmented The set of all pixels of the image; S o,t is the target pixel set at the threshold t; S b,t is the background pixel set at the threshold t.
可选的,假设待分割图像中像素灰度值的最小值和最大值分别为t_min和t_max,则将阈值遍历的区间设置为[t_min+2,t_max+2]。Optionally, assuming that the minimum value and maximum value of the gray value of the pixel in the image to be segmented are t_min and t_max respectively, the interval traversed by the threshold value is set to [t_min+2, t_max+2].
可选的,概率确定模块310还用于:Optionally, the
假设γ(t)表示像素在阈值t时属于目标的先验概率,则1-γ(t)表示像素在阈值t时属于背景的先验概率,po,t(i)为在阈值t时目标像素值为i的目标类概率,pb,t(i)为在阈值t时背景像素值为i的背景类概率;Suppose γ(t) represents the prior probability that the pixel belongs to the target at the threshold t, then 1-γ(t) represents the prior probability that the pixel belongs to the background at the threshold t, and p o,t (i) is the prior probability at the threshold t The target class probability of the target pixel value i, p b,t (i) is the background class probability of the background pixel value i at the threshold t;
一个像素值为i的像素属于目标类和背景类的后验概率可以由贝叶斯规则分别表示为:The posterior probability of a pixel with pixel value i belonging to the target class and the background class can be expressed by Bayesian rule as:
其中,pt(i)可以通过以下公式计算:Among them, p t (i) can be calculated by the following formula:
pt(i)=γ(t)po,t(i)+(1-γ(t))pb,t(i)。p t (i)=γ(t)p o,t (i)+(1-γ(t))p b,t (i).
可选的,目标类和背景类的Tsallis-BE熵的表达式分别为:Optionally, the expressions of the Tsallis-BE entropy of the target class and the background class are respectively:
其中,L为待分割图像中像素灰度最高的值。Among them, L is the highest pixel gray value in the image to be segmented.
可选的,查找模块330还用于:Optionally, the
本实施例提供的一种图像的熵阈值分割装置,可以执行上述方法的实施例,其实现原理和技术效果类似,在此不再赘述。An image entropy threshold segmentation device provided in this embodiment can execute the above-mentioned method embodiment, and its implementation principle and technical effect are similar, and will not be repeated here.
图4为本发明实施例提供的一种电子设备的结构示意图。如图4所示,示出了适于用来实现本申请实施例的电子设备400的结构示意图。FIG. 4 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. As shown in FIG. 4 , a schematic structural diagram of an
如图4所示,电子设备400包括中央处理单元(CPU)401,其可以根据存储在只读存储器(ROM)402中的程序或者从存储部分408加载到随机访问存储器(RAM)403中的程序而执行各种适当的动作和处理。在RAM 403中,还存储有设备400操作所需的各种程序和数据。CPU 401、ROM 402以及RAM 403通过总线404彼此相连。输入/输出(I/O)接口405也连接至总线404。As shown in FIG. 4 , the
以下部件连接至I/O接口405:包括键盘、鼠标等的输入部分406;包括诸如阴极射线管(CRT)、液晶显示器(LCD)等以及扬声器等的输出部分407;包括硬盘等的存储部分408;以及包括诸如LAN卡、调制解调器等的网络接口卡的通信部分409。通信部分409经由诸如因特网的网络执行通信处理。驱动器410也根据需要连接至I/O接口406。可拆卸介质411,诸如磁盘、光盘、磁光盘、半导体存储器等等,根据需要安装在驱动器410上,以便于从其上读出的计算机程序根据需要被安装入存储部分408。The following components are connected to the I/O interface 405: an
特别地,根据本公开的实施例,上文参考图1描述的过程可以被实现为计算机软件程序。例如,本公开的实施例包括一种计算机程序产品,其包括有形地包含在机器可读介质上的计算机程序,计算机程序包含用于执行上述图像的熵阈值分割方法的程序代码。在这样的实施例中,该计算机程序可以通过通信部分409从网络上被下载和安装,和/或从可拆卸介质411被安装。In particular, according to an embodiment of the present disclosure, the process described above with reference to FIG. 1 may be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product including a computer program tangibly embodied on a machine-readable medium, the computer program including program code for performing the above-mentioned entropy threshold segmentation method for an image. In such an embodiment, the computer program may be downloaded and installed from a network via
附图中的流程图和框图,图示了按照本发明各种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上,流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分,前述模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意,在有些作为替换的实现中,方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如,两个接连地表示的方框实际上可以基本并行地执行,它们有时也可以按相反的顺序执行,这依所涉及的功能而定。也要注意的是,框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合,可以用执行规定的功能或操作的专用的基于硬件的系统来实现,或者可以用专用硬件与计算机指令的组合来实现。The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or part of code that includes one or more logical functions for implementing specified executable instructions. It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved. It should also be noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or flowchart illustrations, can be implemented by a dedicated hardware-based system that performs the specified functions or operations , or may be implemented by a combination of dedicated hardware and computer instructions.
描述于本申请实施例中所涉及到的单元或模块可以通过软件的方式实现,也可以通过硬件的方式来实现。所描述的单元或模块也可以设置在处理器中。这些单元或模块的名称在某种情况下并不构成对该单元或模块本身的限定。The units or modules involved in the embodiments described in the present application may be implemented by means of software or by means of hardware. The described units or modules may also be provided in a processor. The names of these units or modules do not constitute limitations on the units or modules themselves in some cases.
上述实施例阐明的系统、装置、模块或单元,具体可以由计算机芯片或实体实现,或者由具有某种功能的产品来实现。一种典型的实现设备为计算机。具体的,计算机例如可以为个人计算机、笔记本电脑、行动电话、智能电话、个人数字助理、媒体播放器、导航设备、电子邮件设备、游戏控制台、平板计算机、可穿戴设备或者这些设备中的任何设备的组合。The systems, devices, modules, or units described in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementing device is a computer. Specifically, the computer can be, for example, a personal computer, a notebook computer, a mobile phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any of these devices combination of devices.
作为另一方面,本申请还提供了一种存储介质,该存储介质可以是上述实施例中前述装置中所包含的存储介质;也可以是单独存在,未装配入设备中的存储介质。存储介质存储有一个或者一个以上程序,前述程序被一个或者一个以上的处理器用来执行描述于本申请的图像的熵阈值分割方法。As another aspect, the present application also provides a storage medium, which may be the storage medium contained in the aforementioned device in the above embodiment, or may be a storage medium that exists independently and is not assembled into the device. The storage medium stores one or more programs, and the aforementioned programs are used by one or more processors to execute the entropy threshold segmentation method for images described in this application.
存储介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括,但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带,磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质,可用于存储可以被计算设备访问的信息。按照本文中的界定,计算机可读介质不包括暂存电脑可读媒体(transitory media),如调制的数据信号和载波。Storage media includes permanent and non-permanent, removable and non-removable media. Information storage can be realized by any method or technology. Information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Flash memory or other memory technology, Compact Disc Read-Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cartridge, tape magnetic disk storage or other magnetic storage device or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media excludes transitory computer-readable media, such as modulated data signals and carrier waves.
需要说明的是,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括要素的过程、方法、商品或者设备中还存在另外的相同要素。It should be noted that the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus comprising a set of elements includes not only those elements, but also includes none other elements specifically listed, or also include elements inherent in the process, method, commodity, or apparatus. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.
本说明书中的各个实施例均采用递进的方式描述,各个实施例之间相同相似的部分互相参见即可,每个实施例重点说明的都是与其他实施例的不同之处。尤其,对于系统实施例而言,由于其基本相似于方法实施例,所以描述的比较简单,相关之处参见方法实施例的部分说明即可。Each embodiment in this specification is described in a progressive manner, the same and similar parts of each embodiment can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and for relevant parts, refer to part of the description of the method embodiment.
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211039468.2A CN115423834A (en) | 2022-08-29 | 2022-08-29 | Image entropy threshold segmentation method, device, electronic equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211039468.2A CN115423834A (en) | 2022-08-29 | 2022-08-29 | Image entropy threshold segmentation method, device, electronic equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115423834A true CN115423834A (en) | 2022-12-02 |
Family
ID=84199682
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211039468.2A Pending CN115423834A (en) | 2022-08-29 | 2022-08-29 | Image entropy threshold segmentation method, device, electronic equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115423834A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111553926A (en) * | 2020-04-30 | 2020-08-18 | 中国航空无线电电子研究所 | Threshold segmentation method based on two-dimensional Tsallis gray scale entropy fast iteration |
CN112102350A (en) * | 2020-08-28 | 2020-12-18 | 西安工程大学 | Secondary image segmentation method based on Otsu and Tsallis entropy |
CN114240989A (en) * | 2021-11-30 | 2022-03-25 | 中国工商银行股份有限公司 | Image segmentation method and device, electronic equipment and computer storage medium |
-
2022
- 2022-08-29 CN CN202211039468.2A patent/CN115423834A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111553926A (en) * | 2020-04-30 | 2020-08-18 | 中国航空无线电电子研究所 | Threshold segmentation method based on two-dimensional Tsallis gray scale entropy fast iteration |
CN112102350A (en) * | 2020-08-28 | 2020-12-18 | 西安工程大学 | Secondary image segmentation method based on Otsu and Tsallis entropy |
CN114240989A (en) * | 2021-11-30 | 2022-03-25 | 中国工商银行股份有限公司 | Image segmentation method and device, electronic equipment and computer storage medium |
Non-Patent Citations (1)
Title |
---|
张新明 等: "广义概率Tsallis熵的快速多阈值图像分割", 数据采集与处理, no. 3, 15 May 2016 (2016-05-15), pages 504 - 507 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhang et al. | Brain tumor segmentation based on hybrid clustering and morphological operations | |
Wang et al. | U-net using stacked dilated convolutions for medical image segmentation | |
US9959466B2 (en) | Object tracking apparatus and method and camera | |
CN104036479B (en) | Multi-focus image fusion method based on non-negative matrix factorization | |
CN108364297B (en) | Blood vessel image segmentation method, terminal and storage medium | |
WO2018214769A1 (en) | Image processing method, device and system | |
US12260525B2 (en) | Image processing methods for improving definition of image, electronic devices, and non-transitory computer-readable storage media | |
Cao et al. | A new image edge detection algorithm based on improved Canny | |
CN111223128A (en) | Target tracking method, device, device and storage medium | |
CN112085705B (en) | Image segmentation method and device based on improved goblet sea squirt swarm algorithm | |
Xi et al. | Simultaneous segmentation of fetal hearts and lungs for medical ultrasound images via an efficient multi-scale model integrated with attention mechanism | |
Muhammad et al. | Frequency component vectorisation for image dehazing | |
CN114549833B (en) | Instance segmentation method, device, electronic device and storage medium | |
CN104952071A (en) | Maximum between-cluster variance image segmentation algorithm based on GLSC (gray-level spatial correlation) | |
Wang et al. | Fully automatic measurement of intima-media thickness in ultrasound images of the common carotid artery based on improved Otsu’s method and adaptive wind driven optimization | |
CN115423834A (en) | Image entropy threshold segmentation method, device, electronic equipment and storage medium | |
Liao et al. | An entropy-weighted local intensity clustering-based model for segmenting intensity inhomogeneous images | |
CN116433695B (en) | Mammary gland region extraction method and system of mammary gland molybdenum target image | |
CN116109524B (en) | Magnetic resonance image channel merging method, device, electronic equipment and storage medium | |
Chen et al. | Variational image restoration and segmentation with rician noise | |
Qi et al. | Upi-net: semantic contour detection in placental ultrasound | |
CN115965851A (en) | Learning method and device for generating picture mask based on low-pass filtering | |
Nie et al. | Image segmentation using two-dimensional extension of minimum within-class variance criterion | |
CN116228570A (en) | Breast ultrasound image denoising method based on BM3D model | |
Chang et al. | CUDA‐based acceleration and BPN‐assisted automation of bilateral filtering for brain MR image restoration |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |