WO2019223706A1 - 一种基于饱和度聚类的骨髓白细胞定位方法 - Google Patents

一种基于饱和度聚类的骨髓白细胞定位方法 Download PDF

Info

Publication number
WO2019223706A1
WO2019223706A1 PCT/CN2019/087875 CN2019087875W WO2019223706A1 WO 2019223706 A1 WO2019223706 A1 WO 2019223706A1 CN 2019087875 W CN2019087875 W CN 2019087875W WO 2019223706 A1 WO2019223706 A1 WO 2019223706A1
Authority
WO
WIPO (PCT)
Prior art keywords
white blood
blood cells
bone marrow
saturation
parts
Prior art date
Application number
PCT/CN2019/087875
Other languages
English (en)
French (fr)
Inventor
方凤奇
李强
陆炬
李舜
刘永涛
胡佳佳
黄震
Original Assignee
杭州智微信息科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 杭州智微信息科技有限公司 filed Critical 杭州智微信息科技有限公司
Priority to KR1020207030282A priority Critical patent/KR20200135839A/ko
Priority to JP2020547398A priority patent/JP6994275B2/ja
Priority to RU2020133630A priority patent/RU2755553C1/ru
Priority to AU2019273339A priority patent/AU2019273339B2/en
Priority to US16/979,490 priority patent/US11403481B2/en
Priority to EP19807105.2A priority patent/EP3798972A4/en
Publication of WO2019223706A1 publication Critical patent/WO2019223706A1/zh
Priority to IL277040A priority patent/IL277040A/en

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
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/10—Pre-processing; Data cleansing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4015—Image demosaicing, e.g. colour filter arrays [CFA] or Bayer patterns
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration using local operators
    • G06T5/30—Erosion or dilatation, e.g. thinning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • 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
    • G06T7/155—Segmentation; Edge detection involving morphological operators
    • 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
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • 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
    • G06V10/762—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using clustering, e.g. of similar faces in social networks
    • G06V10/763—Non-hierarchical techniques, e.g. based on statistics of modelling distributions
    • 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
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
    • G06V20/695—Preprocessing, e.g. image segmentation
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • 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
    • G06T2207/20024—Filtering details
    • G06T2207/20032—Median filtering
    • 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
    • G06T2207/20172—Image enhancement details
    • G06T2207/20182—Noise reduction or smoothing in the temporal domain; Spatio-temporal filtering
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30024—Cell structures in vitro; Tissue sections in vitro

Definitions

  • the invention belongs to the field of medical image processing, and particularly relates to a bone marrow white blood cell positioning method based on saturation clustering.
  • the localization of bone marrow white blood cells is mainly based on a threshold algorithm to separate white blood cells from background and red blood cells.
  • a threshold algorithm to separate white blood cells from background and red blood cells.
  • the purpose of the present invention is to provide a bone marrow white blood cell localization method based on saturation clustering.
  • a white blood cell localization algorithm is provided, which aims at the density of white blood cells in the bone marrow and the phenomenon of cell adhesion in blood smears of some patients The problem is to be able to more precisely select the area of white blood cells.
  • a bone marrow leukocyte localization method based on saturation clustering includes the following steps:
  • the K-means algorithm is applied to the S (saturation) channel, and it is divided into three parts, where the first part P1 may be a white blood cell region, and the second part P2 may be a red blood cell region or There are red blood cells and white blood cells.
  • the third part P3 is generally the background area, so you only need to select the P1 or (P1 + P2) part to get the white blood cell area.
  • the average value (H1, H2) of the H channel of the first two parts in (3) is calculated, and the average point (S1, S2) of the first two parts in (3) is calculated, The ratio of the area of the first and second partial areas in (3).
  • the formulas for H1 and H2 are given below:
  • H1 ⁇ (P1. * H) / ⁇ (P1)
  • P1 is a binary map, the pixel values belonging to the first part are 1, and the others are 0.
  • ⁇ (p1) is the sum of the pixel values of P1, and P1.
  • * H represents the result of multiplying pixels at the same position;
  • P2 is a binary map, the pixel values belonging to the first part are 1, and the others are 0.
  • ⁇ (p 2 ) is the sum of the pixel values of P2, and P2.
  • * H represents the result of multiplying pixels at the same position.
  • step (6) according to the recording result in (5), a decision tree algorithm is applied to find a rule to formulate the selection conditions, wherein the loss function of the decision tree algorithm plus the number of leaf nodes is used for cutting Prevent overfitting
  • the result of (6) is applied to remove the irrelevant area and fill the holes of the white blood cell area by morphological processing.
  • the specific process is as follows: First, select the appropriate structural element b pair (6) The binary map of the image is subjected to corrosion operation to remove irrelevant areas; then the expansion operation is performed;
  • the present invention has the following beneficial effects:
  • the algorithm of the present invention is simple, effective, and has a wide application range. Compared with the existing threshold-based algorithms, the algorithm of the present invention has stronger adaptability.
  • Figure 1 is a picture of bone marrow leukocytes
  • Figure 2 is a picture of bone marrow leukocytes with median filtering
  • Figure 3 is a three-part result diagram obtained by applying the K-means algorithm to the S channel;
  • FIG. 4 is a result diagram after selection by applying a decision tree algorithm
  • FIG. 5 is a result diagram of removing irrelevant areas and filling point holes
  • FIG. 6 is a diagram of the white blood cell positioning result after separation
  • FIG. 7 is a schematic block diagram of an apparatus for locating bone marrow leukocytes according to one embodiment.
  • a bone marrow leukocyte localization method based on saturation clustering includes the following steps:
  • V max (R, G, B)
  • RGB values [0,1]
  • the K-means algorithm to the S (saturation) channel and divide it into 3 parts: as shown in Figure 3, where the first part (P1) is likely to be a white blood cell area, and the second part (P2) may be a red blood cell area or both There are red blood cells and white blood cells, and the third part (P3) is generally the background area. Therefore, we only need to select the P1 or (P1 + P2) part to get the white blood cell area. The following is the selection step;
  • H1 ⁇ (P1. * H) / ⁇ (P1)
  • P1 is a binary map, the pixel values belonging to the first part are 1, and the others are 0.
  • ⁇ (p1) is the sum of the pixel values of P1, and P1.
  • * H represents the result of multiplying pixels at the same position;
  • P2 is a binary map, the pixel values belonging to the first part are 1, and the others are 0.
  • ⁇ (p 2 ) is the sum of the pixel values of P2, and P2.
  • * H represents the result of multiplying pixels at the same position;
  • f is the binary graph obtained in (6), Is expansion operation, Corrosive operation.
  • FIG. 7 is a block diagram of an apparatus 700 for locating bone marrow leukocytes according to one embodiment.
  • the device 700 may be a computer, a cloud server, or the like.
  • the device 700 in FIG. 1 includes one or more of the following components: a processor 702, a memory 704, a power supply component 706, a multimedia component 708, and an input / output (I / O) interface 710.
  • the processor 702 is configured to control overall operations of the device 700, such as operations associated with locating bone marrow white blood cells.
  • the processor 702 is configured to execute instructions to perform all or part of the disclosed method.
  • the processor 702 includes a multimedia module configured to facilitate interaction between the multimedia component 708 and the processor 702.
  • the memory 704 is configured to store various types of data to support the operation of the device 700. Examples of such data include instructions, unit images, databases, etc. of any application or method implemented by the device 700.
  • the memory 704 can be implemented using any type of volatile or non-volatile storage device, such as static random access memory (static random access memory), electrically erasable programmable read-only memory (electrically erasable) (Except programmable read-only memory), programmable read-only memory (programmable read-only memory), read-only memory (read-only memory), magnetic memory, flash memory, or disk or optical disk.
  • the power component 706 is configured to provide power to various components of the device 700.
  • the power component 706 includes a power management system, one or more power sources, and any other components associated with the generation, management, and distribution of power in the device 700.
  • the multimedia component 708 includes a screen that provides an output interface between the device 700 and a user of the device 700.
  • the screen may include a liquid crystal display and a press panel.
  • the input / output interface 710 is configured to provide an interface for the processor 702 and peripheral interface modules (such as a keyboard, a click wheel, a button, etc.).
  • peripheral interface modules such as a keyboard, a click wheel, a button, etc.
  • the device 700 may use one or more application specific integrated circuits, digital signal processors, digital signal processing devices, programmable logic devices, field programmable gate arrays, controllers, microcontrollers, microprocessors, or Other electronic components are implemented to perform the disclosed method.
  • the present disclosure also provides a non-transitory computer-readable storage medium including instructions, such as instructions included in the memory 704. These instructions may be executed by the processor 702 of the device 700 for performing the disclosed method of locating bone marrow white blood cells.
  • the non-transitory computer-readable storage medium may be a read-only memory, a random access memory, an optical disk, a magnetic tape, a floppy disk, an optical data storage device, and the like.
  • the aforementioned bone marrow leukocyte localization method based on saturation clustering has the advantages that the algorithm is simple, effective, and has a wide range of applications; compared with the existing threshold-based algorithm, the algorithm of the present invention has stronger adaptability. Secondly, the combination of the K-means algorithm and the decision tree algorithm can be used to more accurately select the area of white blood cells.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Medical Informatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Databases & Information Systems (AREA)
  • Computing Systems (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Geometry (AREA)
  • Image Analysis (AREA)
  • Investigating Or Analysing Biological Materials (AREA)
  • Image Processing (AREA)
  • Micro-Organisms Or Cultivation Processes Thereof (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

一种基于饱和度聚类的骨髓白细胞定位方法，首先对骨髓白细胞图片进行预处理，去除一部分噪点同时平滑图片；运用K均值算法对骨髓白细胞图片的饱和度通道聚类，根据决策树算法选择出白细胞所在的类别；然后通过形态学处理算法去除白细胞的二值图片的无关区域，同时填充白细胞的点洞；最终对白细胞定位。所述方法简单、有效，对比现有基于阈值算法，所述方法适应范围更广，同时结合决策树算法，使得最终效果更加准确。

Description

一种基于饱和度聚类的骨髓白细胞定位方法 技术领域
本发明属于医学图像处理领域，尤其涉及一种基于饱和度聚类的骨髓白细胞定位方法。
背景技术
骨髓中的白细胞种类多样，经染色后不同种类白细胞的颜色差异也较大。相对于外周血，骨髓中的白细胞密度更大，部分患者的血涂片出现细胞黏连现象。因此骨髓白细胞的定位一直都是一个富有挑战性的课题。近年来，专家以及广大技术人员提出了许多有效的解决方案。但是大多数方案只能解决特定的问题，没有提出一个普遍的方案能适用于多数场景。
目前骨髓白细胞的定位主要是基于阈值算法，将白细胞从背景和红细胞中分离出来。例如Wu等人的文章“A novel color image segmentation method and its application to white blood cell image analysis”(Signal Processing,2006 8th International Conference on)运用Ostu阈值算法分割定位白细胞；Ko等人的文章“Automatic white blood cell segmentation using stepwise merging rules and gradient vector flow snake”(Micron,2011,42(7):695-705)首先运用阈值算法大致估计细胞的位置，然后利用均值漂移做进一步优化。同时也有学者提出了其他有效方案，例如Dorini L B等人的文章“White blood cell segmentation using morphological operators and scale-space analysis”(Computer Graphics and Image Processing，2007:294-304.)应用形态学处理的方法分割定位白细胞；此外还有像聚类的方法。但是这些方法均具有一定局限性，例如阈值算法中的Ostu阈值算法，其中一个假设就是背景和前景的区域面积大致相同，实际骨髓数字图片中白细胞比例可能很大，或者没有白细胞，而且白细胞的颜色分布在一个很大区间，甚至和染色较深的红细胞有重叠。因此，虽然阈值可以适用于大多数数字图片，但是一些特殊情况下该方案不能很好的定位白细胞。当白细胞的颜色分布比较分散时，聚类算法也出现同样问题。
发明内容
本发明的目的在于提供一种基于饱和度聚类的骨髓白细胞定位方法，通过该方法提供一种白细胞定位算法，针对骨髓中的白细胞密度更大，部分患者的血涂片出现细胞黏连现象的问题，能够更加精确的选择出白细胞的区域。
本发明是这样实现的，一种基于饱和度聚类的骨髓白细胞定位方法，包括如下步骤：
(1)对骨髓白细胞图片进行中值滤波去除部分噪点；
(2)对骨髓白细胞图片进行颜色变换，将图片从RGB(红绿蓝)通道转换到HSV(颜色，饱和度，亮度)通道；
(3)对S(饱和度)通道应用K均值算法，将其分为3个部分，选择第一部分P1或者第一 二部分P1+P2部分得到白细胞的区域，下面是选择的步骤；
(4)计算(3)中第一二部分H通道的平均值(H1,H2)，根据(3)中第一二部分的均值点(S1,S2)，计算(3)中第一二部分区域的面积比值(ratio)；
(5)统计多张图片中白细胞所在的部分，记录在P1或者P2部分时H1-H2，S1-S2和ratio的值；
(6)根据(5)中的记录结果，应用决策树算法，找出规律制定选择的条件；
(7)对(6)的结果进行形态学处理去除无关区域，同时填充点洞；
(8)对(7)中分离的白细胞进行定位。
作为优选，所述步骤(3)中对S(饱和度)通道应用K均值算法，将其分为3个部分，其中第一部分P1为可能为白细胞区域，第二部分P2可能为红细胞区域或者既有红细胞也有白细胞，第三部分P3一般是背景区域，因此只需要选择P1或者(P1+P2)部分就可以得到白细胞的区域。
进一步地，所述步骤(4)中，计算(3)中第一二部分H通道的平均值(H1,H2)，根据(3)中第一二部分的均值点(S1,S2)，计算(3)中第一二部分区域的面积比值(ratio)，下面给出H1和H2的计算公式：
H1＝∑(P1.*H)/∑(P1)
H2＝∑(P2.*H)/∑(P2)
其中P1是二值图，属于第一部分的像素值为1，其它为0。∑(p1)为P1像素值的和，P1.*H表示相同位置像素相乘的结果；
P2是二值图，属于第一部分的像素值为1，其它为0。∑(p 2)为P2像素值的和，P2.*H表示相同位置像素相乘的结果。
进一步地，所述步骤(6)中，根据(5)中的记录结果，应用决策树算法，找出规律制定选择的条件，其中决策树算法的损失函数加上叶子节点个数，用于剪枝防止过拟合；
进一步地，所述步骤(7)中，对(6)的结果运用形态学处理去除无关区域，填充白细胞区域的点洞，具体过程如下：首先，选择合适的结构元b对(6)中得到的二值图做腐蚀操作，去除无关区域；然后再做膨胀操作；
Figure PCTCN2019087875-appb-000001
Figure PCTCN2019087875-appb-000002
其中f为(6)中得到的二值图，
Figure PCTCN2019087875-appb-000003
是膨胀操作，
Figure PCTCN2019087875-appb-000004
是腐蚀操作；
最后通过形态学重构填充f中的点洞。
g＝f
Figure PCTCN2019087875-appb-000005
Figure PCTCN2019087875-appb-000006
其中
Figure PCTCN2019087875-appb-000007
是一次重构的结果，∩是并。
相比于现有技术的缺点和不足，本发明具有以下有益效果：
1、本发明算法简单、有效、适用范围广。对比现有基于阈值的算法，本发明算法有更强的自适应性。
2、针对不同种类白细胞颜色分布范围广的问题，以及因染色导致红细胞颜色较深的情况。采用本发明专利的K均值算法和决策树算法结合，能够更加精确的选择出白细胞的区域。
附图说明
图1是骨髓白细胞图片；
图2是经中值滤波的骨髓白细胞图片；
图3是对S通道应用K均值算法得到的三部分结果图；
图4是应用决策树算法选择后的结果图；
图5是去除无关区域并填充点洞的结果图；
图6是分离后的白细胞定位结果图；
图7是根据一个实施例的用于定位骨髓白细胞的设备的示意框图。
具体实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
一种基于饱和度聚类的骨髓白细胞定位方法，包括如下步骤：
(1)对图1的骨髓白细胞图片进行中值滤波，结果如图2所示：其中滤波模板大小为(5*5)。
(2)对骨髓白细胞图片进行颜色变换，将(1)中的经中值滤波的图片从RGB(红绿蓝)通道转换到HSV(颜色，饱和度，亮度)通道，其具体公式如下：
V＝max(R，G，B)
Figure PCTCN2019087875-appb-000008
Figure PCTCN2019087875-appb-000009
其中RGB值的范围为[0,1]；
对S(饱和度)通道应用K均值算法，将其分为3个部分：如图3所示，其中第一部分(P1)为可能为白细胞区域，第二部分(P2)可能为红细胞区域或者既有红细胞也有白细胞，第三部分(P3)一般是背景区域。因此我们只需要选择P1或者(P1+P2)部分就可以得到白细胞的区域，下面是选择的步骤；
(3)计算(3)中第一二部分H通道的平均值(H1,H2)，根据(3)中第一二部分的均值点(S1,S2)，计算(3)中第一二部分区域的面积比值(ratio),下面给出H1的计算公式：
H1＝∑(P1.*H)/∑(P1)
H2＝∑(P2.*H)/∑(P2)
其中P1是二值图，属于第一部分的像素值为1，其它为0。∑(p1)为P1像素值的和，P1.*H表示相同位置像素相乘的结果；
P2是二值图，属于第一部分的像素值为1，其它为0。∑(p 2)为P2像素值的和，P2.*H表示相同位置像素相乘的结果；
(4)统计多张图片中白细胞所在的部分，记录在P1或者P2部分时H1-H2，S1-S2和ratio的值；我们在实施过程中一共统计230张图片，其中120张图片中的白细胞在第一部分(P1)，110张图片中的白细胞在第一二部分(P1+P2)，同时我们还收集了部分无白细胞的图片；
(5)根据(5)中的记录结果，应用决策树算法，找出规律制定选择的条件，其中决策树算法的损失函数加上叶子节点个数，用于剪枝防止过拟合，选择后结果如图4所示；
(6)对(6)的结果运用形态学处理去除无关区域，填充白细胞区域的点洞，结果如图5所示。具体过程如下：
首先，选择合适的结构元b对(6)中得到的二值图做腐蚀操作，去除无关区域，然后再做膨胀操作。
Figure PCTCN2019087875-appb-000010
Figure PCTCN2019087875-appb-000011
其中f为(6)中得到的二值图，
Figure PCTCN2019087875-appb-000012
是膨胀操作，
Figure PCTCN2019087875-appb-000013
是腐蚀操作。
最后通过形态学重构填充f中的点洞。
g＝f
Figure PCTCN2019087875-appb-000014
Figure PCTCN2019087875-appb-000015
其中
Figure PCTCN2019087875-appb-000016
是一次重构的结果，∩是并。
(7)对(6)中分离的白细胞进行定位，结果如图6所示。
图7是根据一个实施例的用于定位骨髓白细胞的设备700的框图。其中设备700可以是计算机、云服务器等。图1中设备700包括一个或多个以下组件:处理器702、存储器704、电源组件706、多媒体组件708、输入/输出(I/O)接口710。
处理器702被配置成控制设备700的整体操作，例如与定位骨髓白细胞相关联的操作。处理器702被配置成执行指令来执行所公开的方法的全部或部分。在一些实施例中，处理器702包括多媒体模块，该多媒体模块被配置为促进多媒体组件708和处理器702之间的交互。
存储器704被配置成存储各种类型的数据，以支持设备700的操作。这种数据的例子包括由设备700实现的任何应用或方法的指令、单元图像、数据库等。存储器704可以使用任何类型的易失性或非易失性存储设备或其组合来实现，例如静态随机存取存储器(静态随机存取存储器)、电可擦除可编程只读存储器(电可擦除可编程只读存储器)、可编程只读存储器(可编程只读存储器)、只读存储器(只读存储器)、磁存储器、闪存或磁盘或光盘。
电力组件706被配置为向设备700的各种组件提供电力。功率组件706包括功率管理系统、一个或多个电源以及与设备700中功率的产生、管理和分配相关联的任何其他组件。
多媒体组件708包括在设备700和设备700的用户之间提供输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器和按压面板。
输入/输出接口710被配置为为处理器702和外围接口模块(例如键盘、点击轮、按钮等)提供接口。
在一些实施例中，设备700可以用一个或多个专用集成电路、数字信号处理器、数字信号处理设备、可编程逻辑设备、现场可编程门阵列、控制器、微控制器、微处理器或其他电子组件来实现，以执行所公开的方法。
本公开还提供了包括指令的非暂时性计算机可读存储介质，例如包括在存储器704中的指令。这些指令可由设备700的处理器702执行，用于执行所公开的定位骨髓白细胞的方法。例如，非暂时性计算机可读存储介质可以是只读存储器、随机存取存储器、光盘、磁带、软盘、光学数据存储设备等。
上述基于饱和度聚类的骨髓白细胞定位方法具有的优点为：算法简单、有效、适用范围广；对比现有基于阈值的算法，本发明算法有更强的自适应性。其次应用将K均值算法和决策树算法结合，能够更加精确的选择出白细胞的区域。
以上所述仅为本发明的较佳实施例而已，并不用以限制本发明，凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等，均应包含在本发明的保护范围之内。

Claims (5)

  1. 一种基于饱和度聚类的骨髓白细胞定位方法，其特征在于，包括如下步骤：
    (1)对骨髓白细胞图片进行中值滤波去除部分噪点；
    (2)对骨髓白细胞图片进行颜色变换，将图片从RGB(红绿蓝)通道转换到HSV(颜色，饱和度，亮度)通道；
    (3)对S饱和度通道应用K均值算法，将其分为3个部分，选择第一部分P1或者第一二部分P1+P2部分得到白细胞的区域，下面是选择的步骤；
    (4)计算(3)中第一二部分H通道的平均值(H1,H2)，根据(3)中第一二部分的均值点(S1,S2)，计算(3)中第一二部分区域的面积比值(ratio)；
    (5)统计多张图片中白细胞所在的部分，记录在P1或者P2部分时H1-H2，S1-S2和ratio的值；
    (6)根据(5)中的记录结果，应用决策树算法，找出规律制定选择的条件；
    (7)对(6)的结果进行形态学处理去除无关区域，同时填充点洞；
    (8)对(7)中分离的白细胞进行定位。
  2. 根据权利要求1所述的基于饱和度聚类的骨髓白细胞定位方法，其特征在于，所述步骤步骤(3)中对S(饱和度)通道应用K均值算法，将其分为3个部分，其中第一部分P1为可能为白细胞区域，第二部分P2可能为红细胞区域或者既有红细胞也有白细胞，第三部分P3一般是背景区域，因此只需要选择P1或者(P1+P2)部分就可以得到白细胞的区域。
  3. 如权利要求1所述的基于饱和度聚类的骨髓白细胞定位方法，其特征在于，所述步骤(4)中计算(3)中第一二部分H通道的平均值(H1,H2)，计算(3)中第一二部分的均值点(S1,S2)，计算(3)中第一二部分区域的面积比值(ratio)，下面给出H1的计算公式：
    H1＝∑(P1.*H)/∑(P1)
    H2＝∑(P2.*H)/∑(P2)
    其中P1是二值图，属于第一部分的像素值为1，其它为0；∑(p1)为P1像素值的和，P1*H表示相同位置像素相乘的结果；
    P2是二值图，属于第一部分的像素值为1，其它为0；∑(p 2)为P2像素值的和，P2.*H表示相同位置像素相乘的结果。
  4. 根据权利要求1所述的基于饱和度聚类的骨髓白细胞定位方法，其特征在于，所述步骤(6)中，根据(5)中的记录结果，应用决策树算法，找出规律制定选择的条件，其中决策树算法的损失函数加上叶子节点个数，用于剪枝防止过拟合。
  5. 根据权利要求1所述的基于饱和度聚类的骨髓白细胞定位方法，其特征在于，所述步骤(7)中，对(6)的结果运用形态学处理去除无关区域，填充白细胞区域的点洞，具体过程如下：首先，选择合适的结构元b对(6)中得到的二值图做腐蚀操作，去除无关区域，然后再做膨胀操作，
    Figure PCTCN2019087875-appb-100001
    Figure PCTCN2019087875-appb-100002
    其中f为(6)中得到的二值图，
    Figure PCTCN2019087875-appb-100003
    是膨胀操作，
    Figure PCTCN2019087875-appb-100004
    是腐蚀操作；
    最后通过形态学重构填充f中的点洞；
    g＝f
    Figure PCTCN2019087875-appb-100005
    Figure PCTCN2019087875-appb-100006
    其中
    Figure PCTCN2019087875-appb-100007
    是一次重构的结果，∩是并。
PCT/CN2019/087875 2018-05-22 2019-05-22 一种基于饱和度聚类的骨髓白细胞定位方法 WO2019223706A1 (zh)

Priority Applications (7)

Application Number Priority Date Filing Date Title
KR1020207030282A KR20200135839A (ko) 2018-05-22 2019-05-22 채도 클러스터에 의한 골수 백혈구 위치 결정 방법
JP2020547398A JP6994275B2 (ja) 2018-05-22 2019-05-22 飽和度クラスタリングに基づく骨髄白血球の位置特定方法
RU2020133630A RU2755553C1 (ru) 2018-05-22 2019-05-22 Способ определения местонахождения лейкоцитов костного мозга на основе агрегации насыщения
AU2019273339A AU2019273339B2 (en) 2018-05-22 2019-05-22 Saturation clustering-based method for positioning bone marrow white blood cells
US16/979,490 US11403481B2 (en) 2018-05-22 2019-05-22 Method for localization of bone marrow white blood cells based on saturation clustering
EP19807105.2A EP3798972A4 (en) 2018-05-22 2019-05-22 SATURATION CLUSTER BASED METHOD FOR POSITIONING BONE MARROW WHITE BLOOD BLOOD CELLS
IL277040A IL277040A (en) 2018-05-22 2020-08-31 A method for the location of white blood cells from bone marrow based on saturation grouping

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810495118.4 2018-05-22
CN201810495118.4A CN108805865B (zh) 2018-05-22 2018-05-22 一种基于饱和度聚类的骨髓白细胞定位方法

Publications (1)

Publication Number Publication Date
WO2019223706A1 true WO2019223706A1 (zh) 2019-11-28

Family

ID=64091391

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/087875 WO2019223706A1 (zh) 2018-05-22 2019-05-22 一种基于饱和度聚类的骨髓白细胞定位方法

Country Status (10)

Country Link
US (1) US11403481B2 (zh)
EP (1) EP3798972A4 (zh)
JP (1) JP6994275B2 (zh)
KR (1) KR20200135839A (zh)
CN (1) CN108805865B (zh)
AU (1) AU2019273339B2 (zh)
IL (1) IL277040A (zh)
RU (1) RU2755553C1 (zh)
TW (1) TWI711008B (zh)
WO (1) WO2019223706A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113570628A (zh) * 2021-07-30 2021-10-29 西安科技大学 一种基于活动轮廓模型的白细胞分割方法
CN113902817A (zh) * 2021-11-23 2022-01-07 杭州智微信息科技有限公司 一种基于灰度值的细胞图片拼接方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108805865B (zh) * 2018-05-22 2019-12-10 杭州智微信息科技有限公司 一种基于饱和度聚类的骨髓白细胞定位方法
CN110751196B (zh) * 2019-10-12 2020-09-18 东北石油大学 一种油水两相流透明管壁内类油滴附着物识别方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080212868A1 (en) * 2005-06-15 2008-09-04 Tissue Gnostics Gmbh Process for Segmenting Leukocytes
CN102298700A (zh) * 2011-06-09 2011-12-28 华东师范大学 一种骨髓病理图像中细胞识别与定位方法
CN104484877A (zh) * 2014-12-12 2015-04-01 山东大学 一种基于Meanshift聚类和形态学操作的AML细胞分割方法
CN106780522A (zh) * 2016-12-23 2017-05-31 杭州华卓信息科技有限公司 一种基于深度学习的骨髓液细胞分割方法
CN108805865A (zh) * 2018-05-22 2018-11-13 杭州智微信息科技有限公司 一种基于饱和度聚类的骨髓白细胞定位方法

Family Cites Families (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2002047007A2 (en) * 2000-12-07 2002-06-13 Phase It Intelligent Solutions Ag Expert system for classification and prediction of genetic diseases
RU2303812C2 (ru) * 2004-12-29 2007-07-27 Общество с ограниченной ответственностью "НПФ РЕНАМ" Способ распознавания и подсчета клеток в биологических средах человека и животных и устройство для его осуществления
RU2308745C1 (ru) * 2006-10-09 2007-10-20 Государственное образовательное учреждение высшего профессионального образования Московский инженерно-физический институт (государственный университет) Способ микроскопического исследования образца, содержащего микрообъекты с разнородными зонами
KR101191454B1 (ko) * 2010-05-14 2012-10-16 계명대학교 산학협력단 비모수적 확률 모델과 질감 정보를 이용한 백혈구 분할 방법
CN102279146A (zh) * 2011-03-11 2011-12-14 桂林优利特医疗电子有限公司 基于激光鞘流技术的血液细胞五分类方法
US20130094750A1 (en) * 2011-10-12 2013-04-18 Tolga Tasdizen Methods and systems for segmentation of cells for an automated differential counting system
CN103020639A (zh) * 2012-11-27 2013-04-03 河海大学 一种白细胞自动识别计数方法
CN103077529B (zh) * 2013-02-27 2016-04-06 电子科技大学 基于图像扫描的植物叶片特征分析系统
CN103473739B (zh) * 2013-08-15 2016-06-22 华中科技大学 一种基于支持向量机的白细胞图像精确分割方法与系统
JP6316569B2 (ja) * 2013-11-01 2018-04-25 株式会社ブレイン 物品識別システムとそのプログラム
CN104392460B (zh) * 2014-12-12 2015-11-04 山东大学 一种基于胞核标记分水岭变换的粘连白细胞分割方法
US9836839B2 (en) * 2015-05-28 2017-12-05 Tokitae Llc Image analysis systems and related methods
CN106248559B (zh) * 2016-07-14 2018-10-23 中国计量大学 一种基于深度学习的白细胞五分类方法
CN106327490A (zh) * 2016-08-22 2017-01-11 中国计量大学 一种基于白细胞检测的细胞核分割方法
EP3321851A3 (en) * 2016-11-09 2018-08-01 AmCad BioMed Corporation Cytological image processing device, and method for quantifying characteristics of cytological image
CN107274444A (zh) * 2017-05-15 2017-10-20 北京林业大学 球形类植物的计数方法及装置
CN107730499A (zh) * 2017-10-31 2018-02-23 河海大学 一种基于nu‑支持向量机的白细胞分类方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080212868A1 (en) * 2005-06-15 2008-09-04 Tissue Gnostics Gmbh Process for Segmenting Leukocytes
CN102298700A (zh) * 2011-06-09 2011-12-28 华东师范大学 一种骨髓病理图像中细胞识别与定位方法
CN104484877A (zh) * 2014-12-12 2015-04-01 山东大学 一种基于Meanshift聚类和形态学操作的AML细胞分割方法
CN106780522A (zh) * 2016-12-23 2017-05-31 杭州华卓信息科技有限公司 一种基于深度学习的骨髓液细胞分割方法
CN108805865A (zh) * 2018-05-22 2018-11-13 杭州智微信息科技有限公司 一种基于饱和度聚类的骨髓白细胞定位方法

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
DORINI L B: "White blood cell segmentation using morphological operators and scale-space analysis", COMPUTER GRAPHICS AND IMAGE PROCESSING, 2007, pages 294 - 304, XP031153381
KO ET AL.: "Automatic white blood cell segmentation using stepwise merging rules and gradient vector flow snake", MICRON, vol. 42, no. 7, 2011, pages 695 - 705, XP028374174, DOI: 10.1016/j.micron.2011.03.009
See also references of EP3798972A4
SU , SHIMEI ET AL.: "Segmentation Algorithm for Bone Marrow Cell Image Based on the Wavelet Transform and K-means Clustering", JOURNAL OF ZHENGZHOU UNIVERSITY (ENGINEERING SCIENCE), vol. 36, no. 4, 1 July 2015 (2015-07-01), pages 15 - 18, XP055657348, ISSN: 1671-6833, DOI: 10.3969/j.issn.1671-6833.2015.04.004 *
WU ET AL.: "A novel color image segmentation method and its application to white blood cell image analysis", SIGNAL PROCESSING, 2006 8TH INTERNATIONAL CONFERENCE ON, 2006

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113570628A (zh) * 2021-07-30 2021-10-29 西安科技大学 一种基于活动轮廓模型的白细胞分割方法
CN113570628B (zh) * 2021-07-30 2024-04-02 西安科技大学 一种基于活动轮廓模型的白细胞分割方法
CN113902817A (zh) * 2021-11-23 2022-01-07 杭州智微信息科技有限公司 一种基于灰度值的细胞图片拼接方法

Also Published As

Publication number Publication date
CN108805865A (zh) 2018-11-13
US11403481B2 (en) 2022-08-02
KR20200135839A (ko) 2020-12-03
IL277040A (en) 2020-10-29
EP3798972A4 (en) 2022-03-02
CN108805865B (zh) 2019-12-10
JP6994275B2 (ja) 2022-02-04
TW202004663A (zh) 2020-01-16
JP2021510831A (ja) 2021-04-30
AU2019273339B2 (en) 2021-03-04
US20210004640A1 (en) 2021-01-07
TWI711008B (zh) 2020-11-21
EP3798972A1 (en) 2021-03-31
RU2755553C1 (ru) 2021-09-17
AU2019273339A1 (en) 2020-08-27

Similar Documents

Publication Publication Date Title
WO2019223706A1 (zh) 一种基于饱和度聚类的骨髓白细胞定位方法
AU2018102232A4 (en) Bone marrow cell marking method and system
WO2022089236A1 (zh) 基于人工智能的图像处理方法、装置、计算机设备和存储介质
WO2021217851A1 (zh) 异常细胞自动标注方法、装置、电子设备及存储介质
Arslan et al. A color and shape based algorithm for segmentation of white blood cells in peripheral blood and bone marrow images
CN111145209B (zh) 一种医学图像分割方法、装置、设备及存储介质
CN103473739B (zh) 一种基于支持向量机的白细胞图像精确分割方法与系统
Zhi et al. AdipoCount: a new software for automatic adipocyte counting
Shahzad et al. Robust Method for Semantic Segmentation of Whole‐Slide Blood Cell Microscopic Images
WO2020253508A1 (zh) 异常细胞检测方法、装置及计算机可读存储介质
Isa Automated edge detection technique for Pap smear images using moving K-means clustering and modified seed based region growing algorithm
CN108320289B (zh) 一种基于稀疏表示和形态学操作的骨髓细胞分割方法
Khan et al. Segmentation of developing human embryo in time-lapse microscopy
Li et al. Hybrid supervision learning for pathology whole slide image classification
CN106327490A (zh) 一种基于白细胞检测的细胞核分割方法
CN113850792A (zh) 一种基于计算机视觉的细胞分类计数方法及系统
CN114758136B (zh) 目标去除模型建立方法、装置及可读存储介质
US20220335606A1 (en) Systems and methods for patient tumor-immune phenotyping from immunofluorescence (if) image analysis
Karthika Devi et al. A novel region based thresholding for dental cyst extraction in digital dental X-ray images
Rahali et al. Drosophila image segmentation using marker controlled watershed
Wang et al. Automatic cell segmentation and signal detection in fluorescent in situ hybridization
Zhu et al. Morphological reconstruction based segmentation of lung fields on digital radiographs
Magoulianitis et al. HUNIS: High-Performance Unsupervised Nuclei Instance Segmentation
CN117876384B (zh) 目标对象实例分割、模型训练方法及相关产品
US20240202932A1 (en) Systems and methods for automated video matting

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19807105

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019273339

Country of ref document: AU

Date of ref document: 20190522

Kind code of ref document: A

WWE Wipo information: entry into national phase

Ref document number: 277040

Country of ref document: IL

ENP Entry into the national phase

Ref document number: 2020547398

Country of ref document: JP

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 20207030282

Country of ref document: KR

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2019807105

Country of ref document: EP

Effective date: 20201222