WO2022111367A1 - 基于高内涵成像的细胞耐药性检测方法、介质及电子设备 - Google Patents

基于高内涵成像的细胞耐药性检测方法、介质及电子设备 Download PDF

Info

Publication number
WO2022111367A1
WO2022111367A1 PCT/CN2021/131342 CN2021131342W WO2022111367A1 WO 2022111367 A1 WO2022111367 A1 WO 2022111367A1 CN 2021131342 W CN2021131342 W CN 2021131342W WO 2022111367 A1 WO2022111367 A1 WO 2022111367A1
Authority
WO
WIPO (PCT)
Prior art keywords
drug resistance
imaging
content imaging
cell drug
high content
Prior art date
Application number
PCT/CN2021/131342
Other languages
English (en)
French (fr)
Inventor
黄钢
李秀英
聂生东
Original Assignee
上海健康医学院
上海理工大学
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 上海健康医学院, 上海理工大学 filed Critical 上海健康医学院
Publication of WO2022111367A1 publication Critical patent/WO2022111367A1/zh

Links

Images

Classifications

    • 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
    • G06V20/698—Matching; Classification
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
    • G01N15/10—Investigating individual particles
    • G01N15/14—Optical investigation techniques, e.g. flow cytometry
    • G01N15/1434—Optical arrangements
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
    • G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
    • G01N21/64—Fluorescence; Phosphorescence
    • G01N21/645—Specially adapted constructive features of fluorimeters
    • G01N21/6456—Spatial resolved fluorescence measurements; Imaging
    • G01N21/6458—Fluorescence microscopy
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
    • G01N15/10—Investigating individual particles
    • G01N15/14—Optical investigation techniques, e.g. flow cytometry
    • G01N15/1434—Optical arrangements
    • G01N2015/144—Imaging characterised by its optical setup

Definitions

  • the invention belongs to the field of image processing and automatic identification of convolutional neural networks, in particular to a cell drug resistance detection method, medium and electronic equipment based on high-content imaging.
  • Studying the drug resistance of cells can not only help researchers to effectively identify and screen drug-resistant cell lines in the laboratory, but also help doctors to accurately judge the drug resistance of patients in the clinic.
  • the drug resistance discrimination of cancer cells at home and abroad is mainly based on the current laboratory methods to identify the drug resistance of cell lines, mainly using flow cytometry, MTT method and gene chip.
  • the clinical judgment of drug resistance is mainly based on drug sensitivity test or drug resistance gene detection. These methods are difficult and the efficiency is difficult to guarantee.
  • the purpose of the present invention is to provide a simple, reliable and high-efficiency high-content imaging-based cell drug resistance detection method, medium and electronic equipment in order to overcome the above-mentioned defects in the prior art, which can be suitable for drug resistance in the laboratory. Strain screening or clinical drug resistance identification.
  • a cell drug resistance detection method based on high content imaging comprising the following steps:
  • the preprocessed image is used as the input of the trained convolutional neural network-based drug resistance classification and identification model, and the drug resistance category to which the cells to be detected belong are detected.
  • the fluorescent dyes for fluorescent staining include Hoechst 33342, MitoTracker Deep Red, Concanavalin A Alex 488, Phalloidin Alex 594, DAPI, CY5, FITC or Texas Red.
  • At least two fluorescent dyes are used for fluorescent staining.
  • the imaging parameters are adjusted, and each fluorescence channel is displayed uniformly.
  • the imaging parameters include an imaging objective lens, an imaging mode, a corresponding plate type, a photographing field of view, an autofocus parameter, a focal length offset value and/or a channel exposure time.
  • the preprocessing includes cropping, uniform illumination and bilateral filtering enhancement.
  • the convolutional neural network includes multiple convolutional layers and multiple fully connected layers, and each layer is convolved with a Relu activation function and a pooling layer.
  • random feature loss with a probability of 0.1 is performed after each layer of the pooling layer.
  • the present invention also provides a computer-readable storage medium comprising one or more programs for execution by one or more processors of an electronic device, the one or more programs including for executing a high-content imaging-based method as described above Instructions for cell-based drug resistance detection methods.
  • the present invention also provides an electronic device, comprising:
  • processors one or more processors
  • One or more programs stored in memory comprising instructions for performing the high-content imaging-based cellular drug resistance detection method as described.
  • the present invention has the following beneficial effects:
  • the present invention extracts and analyzes cell features through machine learning based on high-content images, so that the machine has or is close to the ability to accurately identify drug resistance, and improves the recognition accuracy of convolutional neural networks.
  • the present invention is the first to classify the drug resistance of cells based on images, which provides a reliable imaging basis for the identification of drug resistance and the screening of drug-resistant strains in the laboratory or in the clinic. .
  • Fig. 1 is a flowchart of the present invention
  • Figure 2- Figure 6 shows the results of high-content imaging using Hoechst 33342, MitoTracker Deep Red, Concanavalin A Alex 488, and Phalloidin Alex 594 to label the nucleus, mitochondria, endoplasmic reticulum, and actin of PC9 cells.
  • Figure 3 is the imaging result of nucleus;
  • Figure 4 is the imaging result of mitochondria;
  • Figure 5 is the imaging result of endoplasmic reticulum;
  • Figure 6 is the imaging result of actin;
  • Figure 7- Figure 11 shows the results of high-content imaging using Hoechst 33342, MitoTracker Deep Red, Concanavalin A Alex 488, and Phalloidin Alex 594 to label the nucleus, mitochondria, endoplasmic reticulum, and actin of PC9/GR cells: among them, Figure 7 Figure 8 is the imaging result of the nucleus; Figure 9 is the imaging result of mitochondria; Figure 10 is the imaging result of the endoplasmic reticulum; Figure 11 is the imaging result of actin;
  • Figure 12 is a schematic diagram of the constructed classification network structure.
  • the invention provides a cell drug resistance detection method based on high-content imaging, comprising the following steps: performing fluorescent staining on cells to be detected; performing high-content imaging on the dyed cells to be detected to obtain corresponding high-content images; The high-content image is preprocessed; the preprocessed image is used as the input of the trained convolutional neural network-based drug resistance classification and identification model to detect the drug resistance category of the cells to be detected.
  • the training of the drug resistance classification and identification model can be realized using the cell data cultured in the laboratory.
  • the detection method can be used to screen drug-resistant strains in the laboratory or identify clinical drug-resistant conditions, and can be applied to lung cancer cells and the like.
  • This embodiment provides a method for detecting cell drug resistance based on high-content imaging, as shown in FIG. 1 , including the following steps:
  • Step 1 Cultivate the target cells and their drug-resistant strains.
  • a 24-well glass plate was used, and 50,000 cells per well were cultured for 36 hours, and then the cells were allowed to adhere.
  • the culture in this example involved PC9 and PC9/GR cells.
  • Step 2 Select a suitable fluorescent dye for the selected cells for staining.
  • the cultured cells were stained in the following manner:
  • Concanavalin A 488 working solution (prepared to 5mg/ml with 0.1M NaHCO solution), add PBS to dilute to 50ug/mL, add 200uL dye solution to each well and incubate at room temperature for 15 minutes, wash three times and discard the solution;
  • Step 3 Perform high-content imaging at a specific multiple of the stained well plate.
  • fluorescent dyes such as DAPI, CY5, FITC, and Texas Red can also be used.
  • High-content imaging combines automatic microscopy technology and image analysis methods, which can obtain a large amount of information in a single experiment.
  • the technology that displays on the screen to output a complete image provides additional depth and dimension to cellular informatics.
  • the high-content imaging process is specifically:
  • (36) Set the experimental name and experimental group, check the parameter information of each channel, and start image acquisition.
  • Step 4 Perform necessary image preprocessing on the acquired high-content image, including denoising and enhancement.
  • (42) Perform uniform illumination, bilateral filtering to enhance the image, and denoise the cropped image.
  • the CellProfiler software is used, and the obtained image is preprocessed by a custom module.
  • First import the data from the Load Images module use the Crop module to crop the image, use the Correct Illumination Apply module for uniform illumination, use the Enhance Or Suppress Features module for enhancement, use Smooth for denoising, and finally save the processed image by Save Images .
  • Step 5 Build a suitable classification network to identify drug-resistant strains.
  • This method builds a convolutional neural network model and uses it as a model for classification and identification of drug-resistant strains after training.
  • the structure of the convolutional neural network is shown in Figure 12.
  • the convolutional neural network consists of 5 convolutional layers and 2 fully connected layers, and each layer is convoluted with a Relu activation function and a pooling layer.
  • Each layer of the convolutional network can be expressed as the following expression:
  • w is the convolution kernel parameter
  • b is the bias parameter
  • the size of the convolution kernel of the first four convolution layers is 3 ⁇ 3, only the size of the convolution kernel of the last convolution layer is 2 ⁇ 2, and the convolution type is the same convolution , the step size is 1, the activation function is ReLU, and the bias size is set to 1.
  • Each convolutional layer is followed by a 3 ⁇ 3 maximum pooling layer, the padding method of the pooling layer is set to the same, and the stride is 3.
  • the cultured cells are used to train the convolutional neural network, the preprocessed data is input into the network for training, and the network is continuously optimized.
  • the loss function of the entire network is binary cross entropy, and its specific formula is:
  • p(x) and q(x) are the probability distributions of model output and labels, respectively.
  • This embodiment also verifies the robustness of the constructed convolutional neural network.
  • the above functions are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present invention can be embodied in the form of a software product in essence, or the part that contributes to the prior art or the part of the technical solution.
  • the computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.
  • the aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .
  • This embodiment provides an electronic device, including one or more processors, a memory, and one or more programs stored in the memory, where the one or more programs include a program for executing the high Instructions for Intrinsic Imaging Methods for Cell Resistance Detection.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Biomedical Technology (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Chemical & Material Sciences (AREA)
  • Computational Linguistics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Pathology (AREA)
  • Immunology (AREA)
  • Biochemistry (AREA)
  • Analytical Chemistry (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Dispersion Chemistry (AREA)
  • Multimedia (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Investigating Or Analysing Biological Materials (AREA)
  • Apparatus Associated With Microorganisms And Enzymes (AREA)

Abstract

一种基于高内涵成像的细胞耐药性检测方法、介质及电子设备，包括以下步骤：对待检测细胞进行荧光染色；对染色好的待检测细胞进行高内涵成像，获取对应的高内涵图像；对所述高内涵图像进行预处理；将预处理后的图像作为经训练的基于卷积神经网络的耐药性分类识别模型的输入，检测该待检测细胞所属耐药性类别。与现有技术相比，该方法具有简单可靠、效率高等优点。

Description

基于高内涵成像的细胞耐药性检测方法、介质及电子设备 技术领域
本发明属于图像处理和卷积神经网络自动识别领域，尤其是涉及一种基于高内涵成像的细胞耐药性检测方法、介质及电子设备。
背景技术
研究细胞的耐药性判别不但可以在实验室中帮助研究员有效的鉴别和筛选耐药细胞株，而且在临床上也有助于医生准确判断患者的耐药情况。国内外目前进行癌细胞的耐药性判别主要是基于目前实验室鉴定细胞株耐药的方法主要是利用流式细胞仪和MTT法以及利用基因芯片进行判定。临床上判断耐药性主要依据药物敏感试验或耐药基因检测。这些方法难度较大，效率难以保证。
众所周知，图像蕴含着大量信息，近年来众多研究着致力于图像信息的挖掘。在临床中，医生会基于细胞图片进行相关诊断，但是病理阅片工作量巨大。
发明内容
本发明的目的就是为了克服上述现有技术存在的缺陷而提供一种简单可靠、效率高的基于高内涵成像的细胞耐药性检测方法、介质及电子设备，可适用于实验室中进行耐药株筛选或临床耐药情况识别。
本发明的目的可以通过以下技术方案来实现：
一种基于高内涵成像的细胞耐药性检测方法，包括以下步骤：
对待检测细胞进行荧光染色；
对染色好的待检测细胞进行高内涵成像，获取对应的高内涵图像；
对所述高内涵图像进行预处理；
将预处理后的图像作为经训练的基于卷积神经网络的耐药性分类识别模型的输入，检测该待检测细胞所属耐药性类别。
进一步地，所述进行荧光染色的荧光染色剂包括Hoechst 33342、MitoTracker Deep Red、Concanavalin A Alex 488、Phalloidin Alex 594、DAPI、CY5、FITC或Texas Red。
进一步地，采用至少两种荧光染色剂对所述进行荧光染色。
进一步地，所述进行高内涵成像时，对成像参数进行调节，各荧光通道均匀显示。
进一步地，所述成像参数包括成像物镜、成像模式、对应板型、拍摄视野、自动聚焦参数、焦距偏移值和/或通道曝光时间。
进一步地，所述预处理包括裁剪、均匀光照和双边滤波增强处理。
进一步地，所述卷积神经网络包括多个卷积层和多个全连接层，且每层卷积后带有Relu激活函数和一层池化层。
进一步地，每层所述池化层后进行概率为0.1的随机特征丢失。
本发明还提供一种计算机可读存储介质，包括供电子设备的一个或多个处理器执行的一个或多个程序，所述一个或多个程序包括用于执行如所述基于高内涵成像的细胞耐药性检测方法的指令。
本发明还提供一种电子设备，包括：
一个或多个处理器；
存储器；和
被存储在存储器中的一个或多个程序，所述一个或多个程序包括用于执行如所述基于高内涵成像的细胞耐药性检测方法的指令。
与现有技术相比，本发明具有以下有益效果：
1、本发明基于高内涵图像通过机器学习对细胞特征进行提取、分析，使机器拥有或接近准确识别耐药性的能力，提高卷积神经网络的识别精度。
2、本发明基于细胞的结构与功能之间的关系，首次基于图像进行细胞耐药性的分类，为实验室或临床中耐药性的判别、耐药株的筛选提供了可靠的影像学依据。
附图说明
图1为本发明的流程框图；
图2-图6为利用Hoechst 33342，MitoTracker Deep Red，Concanavalin A Alex 488，Phalloidin Alex 594标记PC9细胞的细胞核、线粒体、内质网、肌动蛋白进行高内涵成像的结果：其中，图2为四个通道叠加后的图像；图3为细胞核 的成像结果；图4为线粒体的成像结果；图5为内质网的成像结果；图6为肌动蛋白的成像结果；
图7-图11为利用Hoechst 33342，MitoTracker Deep Red，Concanavalin A Alex 488，Phalloidin Alex 594标记PC9/GR细胞的细胞核、线粒体、内质网、肌动蛋白进行高内涵成像的结果：其中，图7为四个通道叠加后的图像；图8为细胞核的成像结果；图9为线粒体的成像结果；图10为内质网的成像结果；图11为肌动蛋白的成像结果；
图12为搭建的分类网络结构示意图。
具体实施方式
下面结合附图和具体实施例对本发明进行详细说明。本实施例以本发明技术方案为前提进行实施，给出了详细的实施方式和具体的操作过程，但本发明的保护范围不限于下述的实施例。
本发明提供一种基于高内涵成像的细胞耐药性检测方法，包括以下步骤：对待检测细胞进行荧光染色；对染色好的待检测细胞进行高内涵成像，获取对应的高内涵图像；对所述高内涵图像进行预处理；将预处理后的图像作为经训练的基于卷积神经网络的耐药性分类识别模型的输入，检测该待检测细胞所属耐药性类别。耐药性分类识别模型的训练可采用实验室培养的细胞数据实现。该检测方法可用于实现实验室中耐药株的筛选或临床耐药情况识别，可适用于肺癌细胞等。
实施例1
本实施例提供一种基于高内涵成像的细胞耐药性检测方法，如图1所示，包括如下步骤：
步骤1、进行目标细胞及其耐药株的培养。本实施例中，使用24孔玻璃板，每孔5万个细胞培养36小时后，使细胞贴壁，本实施例培养涉及PC9及PC9/GR细胞。
步骤2、对于选定的细胞选取合适的荧光染色剂进行染色。
本实施例中，对于培养好的细胞采用如下方式进行染色：
(21)弃去培养液，用2毫升PBS清洗两次，取Mito Tracker Deep Red FM 储存液(用DMSO配制成1mM)加入PBS稀释至250nM加入每孔，放入培养箱继续培养20分钟；
(22)用PBS清洗3次，每次10分钟，吸走清洗液，用4％的多聚甲醛固定细胞20分钟后弃去染色液，用PBS清洗三次，并用含0.5％的Triton X-100在室温下透化细胞10分钟；
(23)每孔使用200微升PBS稀释1微升的鬼笔环肽594储液，在室温下孵育20分钟进行染色后弃去并用PBS清洗三次；
(24)取Concanavalin A 488工作液(用0.1M的NaHCO3溶液配制成5mg/ml)加入PBS稀释至50ug/mL，每孔加入200uL染液在室温下孵育15分钟后，清洗三次后弃液；
(25)每孔加入200微升DAPI染液，5分钟后吸取染液，用PBS清洗三次后吸走废液，每孔加入200uLPBS，保持细胞湿润放入高内涵成像系统进行成像。
步骤3、对染色好的孔板进行特定倍数的高内涵成像。
如图2-图11所示，在本实施例中采用Hoechst 33342，MitoTracker Deep Red，Concanavalin A Alex 488，Phalloidin Alex 594(鬼笔环肽594)四个通道，每孔取10*10个视野，对已经染好的PC9及PC9/GR细胞使用60倍物镜进行高内涵成像。
在其他实施例中，也可采用采用DAPI、CY5、FITC、Texas Red等荧光染液。
高内涵成像结合了自动显微镜技术和图像分析方法，可以在单次实验中获得大量信息，是一项在实验中利用对多种荧光标记的细胞进行多通道荧光扫描检测后通过计算机将获得的信息显示在屏幕上从而输出一幅完整图像的技术，为细胞信息学提供了额外的深度和维度。
高内涵成像过程具体为：
(31)将染色好的孔板放入物品仓内，选择成像物镜、成像模式以及对应的板型；
(32)选择孔内的拍摄视野以及所需的荧光通道；
(33)调节自动聚焦设置，使视野中能初步成像；
(34)设置焦距偏移值使样品能够清晰成像；
(35)调整各个通道的曝光时间，使各个通道能够均匀显示；
(36)设置实验名称以及实验组别，核对各通道的参数信息，开始图像采集。
步骤4、对获取到的高内涵图像进行必要的图像预处理，主要包括去噪、增强。
(41)将获取的2048*2048的图像通过随机裁剪，提取512*512大小的图像；
(42)对裁剪后的图像进行均匀光照、双边滤波增强图像、去噪。
本实施例中使用CellProfiler软件，自定义模块对所获得的图像进行预处理。先由Load Images模块导入数据，使用Crop模块将图像裁剪，使用Correct Illumination Apply模块进行均匀光照处理，使用Enhance Or Suppress Features模块进行增强，使用Smooth进行去噪处理，最后由Save Images保存处理好的图像。
步骤5、搭建合适的分类网络，进行耐药株识别。
该方法搭建卷积神经网络模型，并经训练后作为耐药株分类识别模型。
卷积神经网络的结构如图12所示。该卷积神经网络由5个卷积层和2个全连接层构成，，并且每层卷积后带有Relu激活函数和一层池化层，每层卷积网络可表示为以下表达式：
Figure PCTCN2021131342-appb-000001
其中
Figure PCTCN2021131342-appb-000002
为激活函数，w为卷积核参数，b为偏置参数。
本实施例的卷积神经网络中，前四个卷积层的卷积核大小均为3×3，只有最后一个卷积层的卷积核大小为2×2，卷积类型为same卷积，步长为1，激活函数为ReLU，偏置大小设为1。每个卷积层后接一个3×3的最大值池化层，池化层padding方式均设置为same，步长为3。为防止过拟合，每层池化层后都会进行0.1的随机特征丢失(dropout＝0.9)，Flatten层后使用softmax函数进行特征分类，此时设定dropout为0.5。
本实施例利用培养的细胞对该卷积神经网络进行训练，将预处理后的数据输入网络中进行训练，并不断优化网络，整个网络的损失函数为二值交叉熵， 其具体的公式为：
CE＝∑p(x)log(q(x))
其中p(x)和q(x)分别为模型输出与标签的概率分布。
本实施例还对构建的卷积神经网络的鲁棒性进行验证。
上述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
实施例2
本实施例提供一种电子设备，包括一个或多个处理器、存储器和被存储在存储器中的一个或多个程序，所述一个或多个程序包括用于执行如实施例1所述基于高内涵成像的细胞耐药性检测方法的指令。
以上详细描述了本发明的较佳具体实施例。应当理解，本领域的普通技术人员无需创造性劳动就可以根据本发明的构思作出诸多修改和变化。因此，凡本技术领域中技术人员依本发明的构思在现有技术的基础上通过逻辑分析、推理或者有限的实验可以得到的技术方案，皆应在由权利要求书所确定的保护范围内。

Claims (10)

  1. 一种基于高内涵成像的细胞耐药性检测方法，其特征在于，包括以下步骤：
    对待检测细胞进行荧光染色；
    对染色好的待检测细胞进行高内涵成像，获取对应的高内涵图像；
    对所述高内涵图像进行预处理；
    将预处理后的图像作为经训练的基于卷积神经网络的耐药性分类识别模型的输入，检测该待检测细胞所属耐药性类别。
  2. 根据权利要求1所述的基于高内涵成像的细胞耐药性检测方法，其特征在于，所述进行荧光染色的荧光染色剂包括Hoechst 33342、MitoTracker Deep Red、Concanavalin A Alex 488、Phalloidin Alex 594、DAPI、CY5、FITC或Texas Red。
  3. 根据权利要求1所述的基于高内涵成像的细胞耐药性检测方法，其特征在于，采用至少两种荧光染色剂对所述进行荧光染色。
  4. 根据权利要求1所述的基于高内涵成像的细胞耐药性检测方法，其特征在于，所述进行高内涵成像时，对成像参数进行调节，各荧光通道均匀显示。
  5. 根据权利要求4所述的基于高内涵成像的细胞耐药性检测方法，其特征在于，所述成像参数包括成像物镜、成像模式、对应板型、拍摄视野、自动聚焦参数、焦距偏移值和/或通道曝光时间。
  6. 根据权利要求1所述的基于高内涵成像的细胞耐药性检测方法，其特征在于，所述预处理包括裁剪、均匀光照和双边滤波增强处理。
  7. 根据权利要求1所述的基于高内涵成像的细胞耐药性检测方法，其特征在于，所述卷积神经网络包括多个卷积层和多个全连接层，且每层卷积后带有Relu激活函数和一层池化层。
  8. 根据权利要求7所述的基于高内涵成像的细胞耐药性检测方法，其特征在于，每层所述池化层后进行概率为0.1的随机特征丢失。
  9. 一种计算机可读存储介质，其特征在于，包括供电子设备的一个或多个处理器执行的一个或多个程序，所述一个或多个程序包括用于执行如权利要求 1-8任一所述基于高内涵成像的细胞耐药性检测方法的指令。
  10. 一种电子设备，其特征在于，包括：
    一个或多个处理器；
    存储器；和
    被存储在存储器中的一个或多个程序，所述一个或多个程序包括用于执行如权利要求1-8任一所述基于高内涵成像的细胞耐药性检测方法的指令。
PCT/CN2021/131342 2020-11-26 2021-11-18 基于高内涵成像的细胞耐药性检测方法、介质及电子设备 WO2022111367A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011352891.9A CN112507821A (zh) 2020-11-26 2020-11-26 基于高内涵成像的细胞耐药性检测方法、介质及电子设备
CN202011352891.9 2020-11-26

Publications (1)

Publication Number Publication Date
WO2022111367A1 true WO2022111367A1 (zh) 2022-06-02

Family

ID=74966578

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/131342 WO2022111367A1 (zh) 2020-11-26 2021-11-18 基于高内涵成像的细胞耐药性检测方法、介质及电子设备

Country Status (2)

Country Link
CN (1) CN112507821A (zh)
WO (1) WO2022111367A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112507821A (zh) * 2020-11-26 2021-03-16 上海健康医学院 基于高内涵成像的细胞耐药性检测方法、介质及电子设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190065817A1 (en) * 2017-08-29 2019-02-28 Konica Minolta Laboratory U.S.A., Inc. Method and system for detection and classification of cells using convolutional neural networks
CN111260638A (zh) * 2020-01-19 2020-06-09 陕西未来健康科技有限公司 基于卷积神经网络的脑脊液细胞和病原体的自动检测计数方法及系统
CN111680575A (zh) * 2020-05-19 2020-09-18 苏州大学 一种人类上皮细胞染色分类装置、设备及存储介质
CN112507821A (zh) * 2020-11-26 2021-03-16 上海健康医学院 基于高内涵成像的细胞耐药性检测方法、介质及电子设备

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6629762B2 (ja) * 2014-05-23 2020-01-15 ベンタナ メディカル システムズ， インコーポレイテッド 画像内の生物学的構造及び／又はパターンの検出のためのシステム及び方法
CN109190441B (zh) * 2018-06-21 2022-11-08 丁彦青 女性生殖道细胞病理智能分类方法、诊断仪及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190065817A1 (en) * 2017-08-29 2019-02-28 Konica Minolta Laboratory U.S.A., Inc. Method and system for detection and classification of cells using convolutional neural networks
CN111260638A (zh) * 2020-01-19 2020-06-09 陕西未来健康科技有限公司 基于卷积神经网络的脑脊液细胞和病原体的自动检测计数方法及系统
CN111680575A (zh) * 2020-05-19 2020-09-18 苏州大学 一种人类上皮细胞染色分类装置、设备及存储介质
CN112507821A (zh) * 2020-11-26 2021-03-16 上海健康医学院 基于高内涵成像的细胞耐药性检测方法、介质及电子设备

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
LI XIU-YING;HUANG GANG: "High-Content Imaging and Its Application in the Study of Cell Phenotype", JOURNAL OF CHINESE ELECTRON MICROSCOPY SOCIETY, vol. 39, no. 4, 15 August 2020 (2020-08-15), pages 451 - 455, XP055932652, ISSN: 1000-6281, DOI: 10.3969/j.issn.1000-6281.2020.04.017 *

Also Published As

Publication number Publication date
CN112507821A (zh) 2021-03-16

Similar Documents

Publication Publication Date Title
Caicedo et al. Evaluation of deep learning strategies for nucleus segmentation in fluorescence images
Wang et al. Label-free detection of rare circulating tumor cells by image analysis and machine learning
Versari et al. Long-term tracking of budding yeast cells in brightfield microscopy: CellStar and the Evaluation Platform
He et al. A new method for CTC images recognition based on machine learning
Wang et al. Cellular phenotype recognition for high-content RNA interference genome-wide screening
Khan et al. Content based image retrieval approaches for detection of malarial parasite in blood images
Parab et al. Red blood cell classification using image processing and CNN
US10769432B2 (en) Automated parameterization image pattern recognition method
JP2021517255A (ja) 顕微鏡による生物学的物質の同定方法
Kvarnström et al. Image analysis algorithms for cell contour recognition in budding yeast
Li et al. FecalNet: Automated detection of visible components in human feces using deep learning
Mota et al. Automated mesenchymal stem cell segmentation and machine learning-based phenotype classification using morphometric and textural analysis
Peng et al. PombeX: robust cell segmentation for fission yeast transillumination images
CN112784767A (zh) 基于白细胞显微图像的细胞实例分割算法
WO2022111367A1 (zh) 基于高内涵成像的细胞耐药性检测方法、介质及电子设备
Du et al. Automatic classification of cells in microscopic fecal images using convolutional neural networks
Manescu et al. A weakly supervised deep learning approach for detecting malaria and sickle cells in blood films
Cuny et al. Cell region fingerprints enable highly precise single-cell tracking and lineage reconstruction
Li et al. The gray mold spore detection of cucumber based on microscopic image and deep learning
Fishman et al. Practical segmentation of nuclei in brightfield cell images with neural networks trained on fluorescently labelled samples
Zheng et al. White blood cell detection using saliency detection and CenterNet: A two‐stage approach
CN114387596A (zh) 细胞病理涂片自动判读系统
de Souza Oliveira et al. A new approach for malaria diagnosis in thick blood smear images
Aggarwal et al. Protein Subcellular Localization Prediction by Concatenation of Convolutional Blocks for Deep Features Extraction from Microscopic Images
Lin et al. Automated malaria cells detection from blood smears under severe class imbalance via importance-aware balanced group softmax

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: 21896868

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21896868

Country of ref document: EP

Kind code of ref document: A1