WO2023108412A1 - 目标蛋白快速识别及定量方法 - Google Patents

目标蛋白快速识别及定量方法 Download PDF

Info

Publication number
WO2023108412A1
WO2023108412A1 PCT/CN2021/137871 CN2021137871W WO2023108412A1 WO 2023108412 A1 WO2023108412 A1 WO 2023108412A1 CN 2021137871 W CN2021137871 W CN 2021137871W WO 2023108412 A1 WO2023108412 A1 WO 2023108412A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
target protein
data set
target proteins
target
Prior art date
Application number
PCT/CN2021/137871
Other languages
English (en)
French (fr)
Inventor
牛丽丽
肖杨
石珂珂
庞娜
林争荣
Original Assignee
深圳先进技术研究院
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 深圳先进技术研究院 filed Critical 深圳先进技术研究院
Priority to PCT/CN2021/137871 priority Critical patent/WO2023108412A1/zh
Publication of WO2023108412A1 publication Critical patent/WO2023108412A1/zh

Links

Images

Classifications

    • 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
    • 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
    • G06N3/08—Learning 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
    • G06V10/20—Image preprocessing
    • G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
    • 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
    • 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
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • 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

Definitions

  • the invention relates to the technical fields of medical imaging image diagnosis and CT medical imaging, in particular to a method, device, equipment and storage medium for rapid identification and quantification of target proteins.
  • Protein as an important component of the body, participates in various forms of life activities, and its distribution and role in the body are significantly different due to differences in its types, properties, and functions.
  • protein As a biological macromolecule that plays an important role in life activities, protein is an indispensable part of scientific research that reveals the mystery of life. It participates in the regulation of gene expression, redox in cells, neurotransmission, electron transmission, and learning and memory. Activity. In the fields of biology and neuroscience, the exploration of proteins is essential. Among them, immunofluorescence, immunohistochemistry, optogenetics and other methods can realize the localization and recording of specific target proteins, but the identification and statistics of target proteins in the later stage The workload is huge.
  • the manual counting method not only increases the time of sample analysis, but also reduces the stability and consistency of counting to a certain extent. Cell overlap, background noise, poor staining, etc., parameter adjustment has a greater impact. There is within-group bias.
  • the method of cell identification and counting based on microfluidic chip requires more supporting hardware.
  • Microfluidic impedance counters have unique advantages such as simplicity, low cost, low power consumption, and fewer samples, and can provide on-site solutions. They are not suitable for measuring and processing large samples, and require impedance analyzers, lock-in amplifiers, and huge fluid pumps to work together. , which makes it less portable. Microfluidic optical flow cytometry provides an accurate, high-throughput method for cell counting.
  • microfluidic optical flow cytometers bulky and easy.
  • the method of cell counting based on deep learning adopts related methods of deep learning to predict the number of cells by extracting image features and obtaining cell density. This method has high requirements on the quality of the data set, the algorithm operation speed is slow, the counting accuracy is low, and the model generalization is poor. This method only counts the number of fluorescently stained cells under the microscope, and cannot distinguish between target proteins and non-target proteins and identify them separately.
  • the cell counting method based on the image processing method uses the traditional image processing method to count, and can process the pictures in batches, and count the target protein by fixing the approximate area of the protein.
  • the counting principle of ImageJ software mentioned in manual counting is the same.
  • the robustness of this method is poor, and the anti-noise performance is weak.
  • appropriate thresholds need to be adjusted to obtain reasonable results.
  • this method is easy to identify noise points as targets for counting, resulting in biased results. Therefore, the method is difficult to port to a laboratory environment.
  • the embodiment of the present application provides a method for rapid identification and quantification of a target protein.
  • the method includes: collecting corresponding images through a microscope, and adjusting the background to obtain a high-resolution RGB image; marking the target protein in the image, and Statistically annotate the data set; input the statistical data set into the artificial neural network for detection, output the position of the target protein and the corresponding confidence, integrate the detection results, and obtain the number of target proteins in the area.
  • the collecting corresponding images through a microscope includes: selecting c-Fos staining images, which include activated neurons; immunofluorescence staining of corresponding tissue sections, and when the immunofluorescence staining is successful , to acquire images through a microscope.
  • the adjusting the background to obtain a high-resolution RGB image includes: judging whether the sharpness of the image is less than a threshold, and if so, enhancing the contrast and saturation so that the sharpness of the image reaches the threshold.
  • the labeling of the target protein in the image, and counting the labeled data set includes: inputting the image into Labelme software for labeling, wherein the labeling frame is tangent to the edge of the target cell;
  • the windowing method intercepts the image block containing the label box as a data sample to make a data set.
  • the inputting the statistical data set into the artificial neural network for detection includes: inputting the statistical data set into the artificial neural network, and testing the test model through YOLO, and the test results include True Positive (TP), False Positive (FP), True Negative (TN) and False Negative (FN); by calculating Precision, Recall ( Recall) and average precision (Average Precision, AP) evaluate the prediction results of the model, where the precision rate represents the proportion of correct predictions as positive to all positive predictions; recall rate refers to the proportion of correct predictions as positive to all positive samples.
  • the accuracy The recall rate The average precision
  • the method further includes: judging whether the input terminal is a whole image or a manually marked area, and if the input is a whole image, dividing the image After window processing, input the network model sequentially; if the input end is a manually calibrated area, take the smallest circumscribed rectangle for the polygonal area, and set 0 for the pixels in the non-interest area. At the same time, adjust the confidence through the output end, integrate the results, and calculate Corresponding to the number of candidate boxes.
  • the embodiment of the present application also provides a device for rapid identification and quantification of target proteins, which includes: an acquisition unit for acquiring corresponding images through a microscope, and adjusting the background to obtain high-resolution RGB images; a labeling unit for It is used to mark the target protein in the image, and count the marked data set; the output unit is used to input the statistical data set to the artificial neural network for detection, and output the position of the target protein and the corresponding confidence, which is useful for the detection The results are integrated to obtain the number of target proteins in the region.
  • the embodiment of the present application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and operable on the processor.
  • the processor executes the program, it implements the The method described in any one of the descriptions of the examples.
  • the embodiment of the present application also provides a computer device, a computer-readable storage medium, on which a computer program is stored, and the computer program is used for: when the computer program is executed by a processor, the computer program according to the present application is implemented.
  • a computer device a computer-readable storage medium, on which a computer program is stored, and the computer program is used for: when the computer program is executed by a processor, the computer program according to the present application is implemented.
  • the method for rapid identification and quantification of the target protein uses an image analysis method to count the target protein, and records the quantity of the target protein in a designated area; it is suitable for images under various magnifications.
  • the deep learning network is used to learn the characteristics of the target protein and make a data set to detect the position of the target protein, which brings new ideas for computer-aided counting microfluorescence images.
  • the location of specific brain regions can be located through the brain map, and the number of target proteins in the corresponding regions can be counted.
  • the present invention is not limited to immunofluorescence images, and protein testing methods such as immunohistochemistry, immunocolloidal gold labeling, and fusion gene expression with GFP can realize the identification and statistics of target proteins.
  • the target detection network can also be extended to various types such as faster-rcnn, cascade-rcnn, ssd; images are not limited to optical images, and can be used for other imaging images Applicable; the invention can also be used to process and identify double or multiple stained images.
  • Figure 1 shows a schematic flow chart of the rapid identification and quantification method for the target protein provided by the embodiment of the present application
  • Fig. 2 shows an exemplary structural block diagram of a target protein rapid identification and quantification device 200 according to an embodiment of the present application
  • FIG. 3 shows a schematic structural diagram of a computer system suitable for implementing a terminal device according to an embodiment of the present application
  • Fig. 4 shows a schematic diagram of the precision-recall curve provided by the embodiment of the present application
  • Fig. 5 shows the schematic diagram of the test sample that the embodiment of the present application provides
  • Figure 6 shows a schematic diagram of protein counting visualization provided by the embodiment of the present application.
  • Fig. 7 shows a schematic diagram of the yolo v5 structure provided by the embodiment of the present application.
  • first and second are used for descriptive purposes only, and cannot be interpreted as indicating or implying relative importance or implicitly specifying the quantity of indicated technical features. Thus, features defined as “first” and “second” may explicitly or implicitly include at least one of these features. In the description of the present invention, “plurality” means at least two, such as two, three, etc., unless otherwise specifically defined.
  • the first feature may be in direct contact with the first feature or the first and second feature may be in direct contact with the second feature through an intermediary. touch.
  • “above”, “above” and “above” the first feature on the second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature.
  • “Below”, “beneath” and “beneath” the first feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is less horizontally than the second feature.
  • FIG. 1 shows a schematic flowchart of the method for rapid identification and quantification of the target protein provided in the embodiment of the present application.
  • the method includes:
  • Step 110 collecting corresponding images through a microscope, and adjusting the background to obtain high-resolution RGB images
  • Step 120 labeling the target protein in the image, and counting the labeled data sets
  • Step 130 input the statistical data set into the artificial neural network for detection, output the position of the target protein and the corresponding confidence, integrate the detection results, and obtain the quantity of the target protein in the area.
  • the image analysis method is used to count the target protein, and the number of target protein in the specified area is recorded; it is suitable for images under various magnifications.
  • the deep learning network is used to learn the characteristics of the target protein and make a data set to detect the position of the target protein, which brings new ideas for computer-aided counting microfluorescence images.
  • the location of specific brain regions can be located through the brain map, and the number of target proteins in the corresponding regions can be counted.
  • the present invention is not limited to immunofluorescence images, and protein testing methods such as immunohistochemistry, immunocolloidal gold labeling, and fusion gene expression with GFP can realize the identification and statistics of target proteins.
  • the target detection network can also be extended to various types such as faster-rcnn, cascade-rcnn, ssd; images are not limited to optical images, and can be used for other imaging images Applicable; the invention can also be used to process and identify double or multiple stained images.
  • collecting corresponding images through a microscope in this application includes: selecting c-Fos staining images, which include activated neurons; immunofluorescent staining of corresponding tissue sections, when the immunofluorescence staining is successful Afterwards, images were collected through a microscope. Specifically, c-Fos staining images are selected through indiscriminate selection, wherein the selected images need to have activated neurons, and there is no requirement for staining colors, etc.
  • adjusting the background in this application to obtain a high-resolution RGB image includes: judging whether the sharpness of the image is less than a threshold, and if so, enhancing the contrast and saturation so that the sharpness of the image reaches the threshold.
  • labeling the target protein in the image in the present application and counting the labeled data set includes: inputting the image into Labelme software for labeling, wherein the labeling frame is tangent to the edge of the target cell; Using the windowing method, the image block containing the label box is intercepted as a data sample to make a data set.
  • the labeling frame is tangent to the edge of the target cell.
  • the label is proofread by two experienced technicians. Since the original image is 2048 ⁇ 2048, the resolution is too large, which is not conducive to the parameter calculation of the model. Therefore, the image block containing the label box is intercepted by the window method as a single data sample to make a data set.
  • the size of the window is 512 ⁇ 512, and the stride is 50.
  • inputting the statistical data set into the artificial neural network for detection in the present application includes: inputting the statistical data set into the artificial neural network, and testing the test model through YOLO, and the test result Including true class (True Positive, TP), false positive class (False Positive, FP), true negative class (True Negative, TN) and false negative class (False Negative, FN); by calculating the precision rate (Precision), recall rate (Recall), Average Precision (Average Precision, AP) evaluates the prediction results of the model, where the precision rate represents the proportion of correct predictions as positive to all positive predictions; recall rate refers to the proportion of correct predictions as positive to all positive samples.
  • precision rate represents the proportion of correct predictions as positive to all positive predictions
  • recall rate refers to the proportion of correct predictions as positive to all positive samples.
  • this application uses the yolo-v5 model for training, in which YOLO (you only look once) is a representative of the one-stage algorithm series. It treats the target detection task as a regression problem, and directly obtains the coordinates of the bounding box, the confidence and category probability of the objects contained in the box through all the pixels of the entire image. YOLO detects objects very quickly. There is no complicated detection process. You only need to input the image into the neural network to get the detection result. YOLO can complete the object detection task very quickly.
  • test results are divided into True Positive (TP), False Positive (FP), True Negative (TN) and False Negative (FN), calculated from these four types of situations Precision (Precision), recall (Recall), and average precision (Average Precision, AP) evaluate the prediction results of the model, where the precision rate represents the proportion of correct predictions to all positive predictions; recall rate refers to the correct prediction of Positive is the proportion of all positive samples, where the precision rate recall rate Average precision
  • the method further includes: judging whether the input end is a whole image or a manually marked area, and if the whole image is input, then performing After the windowing process, input the network model in turn; if the input end is a manually calibrated area, take the smallest circumscribed rectangle for the polygon area, and set 0 for the pixels of the non-interest area. At the same time, adjust the confidence level through the output end and integrate the results. Calculate the number of corresponding candidate boxes.
  • FIG. 2 shows an exemplary structural block diagram of a target protein rapid identification and quantification device 200 according to an embodiment of the present application.
  • the device includes:
  • the acquisition unit 210 is configured to acquire corresponding images through a microscope, and adjust the background to obtain high-resolution RGB images;
  • Annotation unit 220 configured to annotate the target protein in the image, and count the annotated data set
  • the output unit 230 is used to input the statistical data set into the artificial neural network for detection, output the position of the target protein and the corresponding confidence, integrate the detection results, and obtain the quantity of the target protein in the region.
  • the units or modules recorded in the device 200 correspond to the steps in the method described with reference to FIG. 1 . Therefore, the operations and features described above for the method are also applicable to the device 200 and the units contained therein, and will not be repeated here.
  • the apparatus 200 may be pre-implemented in the browser of the electronic device or other security applications, and may also be loaded into the browser of the electronic device or its security applications by means of downloading or the like.
  • the corresponding units in the apparatus 200 may cooperate with the units in the electronic device to implement the solutions of the embodiments of the present application.
  • FIG. 3 shows a schematic structural diagram of a computer system 300 suitable for implementing a terminal device or a server according to an embodiment of the present application.
  • a computer system 300 includes a central processing unit (CPU) 301 that can operate according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage section 308 into a random-access memory (RAM) 303 Instead, various appropriate actions and processes are performed.
  • ROM read-only memory
  • RAM random-access memory
  • various programs and data required for the operation of the system 300 are also stored.
  • the CPU 301, ROM 302, and RAM 303 are connected to each other through a bus 304.
  • An input/output (I/O) interface 305 is also connected to the bus 304 .
  • the following components are connected to the I/O interface 305: an input section 306 including a keyboard, a mouse, etc.; an output section 307 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 308 including a hard disk, etc. and a communication section 309 including a network interface card such as a LAN card, a modem, or the like.
  • the communication section 309 performs communication processing via a network such as the Internet.
  • a drive 310 is also connected to the I/O interface 305 as needed.
  • a removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is mounted on the drive 310 as necessary so that a computer program read therefrom is installed into the storage section 308 as necessary.
  • the process described above with reference to FIG. 1 may be implemented as a computer software program.
  • embodiments of the present disclosure include a method for rapid identification and quantification of a target protein comprising a computer program tangibly embodied on a machine-readable medium, the computer program comprising program code for performing the method of FIG. 1 .
  • the computer program may be downloaded and installed from a network via communication portion 309 and/or installed from removable media 311 .
  • 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.
  • 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.
  • 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.
  • 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 set in a processor.
  • a processor includes a first sub-region generating unit, a second sub-region generating unit, and a display region generating unit.
  • the names of these units or modules do not constitute limitations on the units or modules themselves in some cases, for example, the display area generation unit can also be described as "used to generate The cell of the display area of the text".
  • the present application also provides a computer-readable storage medium, which may be the computer-readable storage medium contained in the aforementioned devices in the above-mentioned embodiments; computer-readable storage media stored in the device.
  • the computer-readable storage medium stores one or more programs, and the aforementioned programs are used by one or more processors to execute the text generation method applied to transparent window envelopes described in this application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Software Systems (AREA)
  • Multimedia (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Investigating Or Analysing Biological Materials (AREA)

Abstract

一种目标蛋白快速识别及定量方法、装置、设备及其存储介质，方法包括：通过显微镜采集对应的图像，并调节背景获得高分辨RGB图像（110）；对图像中的目标蛋白进行标注，并统计标注后的数据集（120）；将统计后的数据集输入到人工神经网络进行检测，输出目标蛋白的位置以及对应的置信度，对检测结果进行整合，得到区域内目标蛋白的数量（130）。采用图像分析方法对目标蛋白进行计数，记录指定区域目标蛋白数量，适应多种倍镜下的图像。同时，采用深度学习网络，学习目标蛋白的特征，从而对目标蛋白的位置进行检测，为计算机辅助计数目标蛋白带来新的思路。通过脑图谱可以定位特定脑区位置，并统计相应区域目标蛋白的数量。

Description

目标蛋白快速识别及定量方法 技术领域
本发明涉及医学影像图像诊断和CT医学影像技术领域，具体涉及一种目标蛋白快速识别及定量方法、装置、设备及其存储介质。
背景技术
蛋白质，作为组成机体的重要成分，参与各种形式的生命活动，由于其种类、性质、功能的不同在机体中的分布及作用有明显差异。蛋白质作为在生命活动中起重要作用的生物大分子，是一切揭示生命奥秘科学研究不可忽略的一部分，其参与基因表达的调节，细胞中氧化还原、神经传递、电子传递及学习记忆等多种生命活动。在生物及神经科学领域中，对蛋白的探索必不可少，其中，免疫荧光，免疫组化，光遗传等方法均可以实现对特定目标蛋白进行定位和记录，然而后期对于目标蛋白的识别及统计的工作量巨大。
目前针对荧光细胞计数的方法有以下几种：人工计数。依靠人工经验，通过对图像进行反复对比确认，进行计数统计。或是利用Image J软件进行辅助，对于目标清晰均匀图片此方法便利快捷，然而由于图像之间的质量及目标大小差异与操作人员的主观意识，往往统计的结果差异明显。基于微流控芯片的细胞识别计数。沈鹤柏等在《一种微流控多色荧光细胞计数仪》中提出，利用微流控芯片系统结合荧光光源计数分析计数系统对肿瘤细胞进行分析、计数，实现细胞计数及 形态的分析。并将生化反应的若干步骤包括分析、洗涤、检测等集成在一块或几块微流体芯片上。最终结合磁性细胞分选技术和免疫细胞化学染色技术，用于结肠癌外周血循环肿瘤细胞的检测。基于深度学习的细胞计数。刘晓平在《基于深度学习的荧光显微成像中细胞自动计数方法研究》中提出先通过滤除细胞背景来减少干扰信息，将细胞图像分批输入卷积神经网络提取深层特征，然后利用循环神经网络提取方向特征，特征融合后对荧光细胞进行检测，最后通过拟合细胞密度图得到视野中的细胞总数量。基于图像处理方法的细胞计数。在《荧光细胞计数方法、装置、终端设备及存储介质与流程》中提出，先将图像进行二值化处理，预设细胞面积参数，识别出二值化图像的细胞区域，最后根据细胞区域统计荧光图像的数量。
然而人工计数方法中不但增加了样品分析的时间，而且在一定程度上降低了计数的稳定性和一致性。细胞重叠、背景噪声，染色不佳等情况，参数调节有较大影响。存在组内偏差。基于微流控芯片的细胞识别计数的方法对配套硬件要求较多。微流体阻抗计数器具存简易、低成本、低功耗、少样本等独特优势，并能提供现场解决方案不适合测量和处理大样本，还需要阻抗分析仪、锁定放大器以及庞大的流体泵配合工作，这些使得它具有较差的可移植性。微流控光学流式细胞仪提供了一种准确、高通量的细胞计数方法。然而，除微流控芯片以外，现有研究大多还是使用传统的硬件，如激光、探测器、液压泵、大功率光源和电子器件，这使得微流控光学流式细胞仪存在体积人、易受冲击和难校准等缺点，这些缺点阻碍了它的便携性。基于深度学习的细胞计数的方法采用了深度学习相关方法，通过提取图像特征，获取细胞密度对细胞数量进行预测。该方法对数据集的质量要求较高，算法运算速度较慢，计数准确率低，模型泛化性较差。该方法仅仅计算了显微镜下荧光染色细胞数量，不能对目标蛋白和非目标蛋白进行 区分及分别识别统计。基于图像处理方法的细胞计数的方法利用传统的图像处理方法计数，可以对图片进行批量处理，通过固定蛋白的大致面积，对目标蛋白进行计数。与人工计数中提到的ImageJ软件计数原理相同。该方法的鲁棒性较差，抗噪性能弱，对于不同蛋白需要调整适当的阈值才能得到合理的结果。对于染色效果较差的图像，该方法容易将噪点也识别为目标进行计数，造成结果偏差。因此，该方法难以移植到实验室环境中。
目前的方法多需人工辅助，因此亟需一种高效、便捷、准确率高的方法来完成对目标蛋白的识别统计工作。
发明内容
鉴于现有技术中的上述缺陷或不足，期望提供一种目标蛋白快速识别及定量方法、装置、设备及其存储介质。
第一方面，本申请实施例提供了一种目标蛋白快速识别及定量方法，该方法包括：通过显微镜采集对应的图像，并调节背景获得高分辨RGB图像；对图像中的目标蛋白进行标注，并统计标注后的数据集；将统计后的数据集输入到人工神经网络进行检测，输出目标蛋白的位置以及对应的置信度，对检测结果进行整合，得到区域内目标蛋白的数量。
在其中一个实施例中，所述通过显微镜采集对应的图像，包括：选择c-Fos染色图像，所述图像中包括有激活的神经元；将对应组织切片免疫荧光染色，当免疫荧光染色成功后，通过显微镜采集图像。
在其中一个实施例中，所述调节背景获得高分辨RGB图像，包括：判断图像的清晰度是否小于阈值，若是，则进行对比度和饱和度的增强，以使得图像的清晰度达到阈值。
在其中一个实施例中，所述对图像中的目标蛋白进行标注，并统 计标注后的数据集，包括：将图像输入到Labelme软件中进行标注，其中，标注框与目标细胞边缘相切；采用划窗法将有包含标注框的图像块截取下来作为数据样本制作数据集。
在其中一个实施例中，所述将统计后的数据集输入到人工神经网络进行检测，包括：将统计后的数据集输入到人工神经网络，通过YOLO对测试模型进行检测，所述测试结果包括真正类(True Positive，TP)、假正类(False Positive，FP)、真负类(True Negative，TN)和假负类(False Negative，FN)；通过计算精准率(Precision)、召回率(Recall)、平均精度(Average Precision，AP)对模型的预测结果进行评估，其中精准率代表正确预测为正占全部预测为正的比例；召回率指正确预测为正占全部正样本的比例。
在其中一个实施例中，所述精准率
Figure PCTCN2021137871-appb-000001
所述召回率
Figure PCTCN2021137871-appb-000002
所述平均精度
Figure PCTCN2021137871-appb-000003
在其中一个实施例中，所述将统计后的数据集输入到人工神经网络进行检测之后，该方法还包括：判断输入端为整图还是手动标定区域，若输入整图，则对图像进行划窗处理后依次输入网络模型；若输入端为手动标定区域时，则对多边形区域取最小外接矩形，并对非感兴趣区域像素点置0，同时，通过输出端调整置信度，整合结果，计算对应候选框的个数。
第二方面，本申请实施例还提供了一种目标蛋白快速识别及定量装置，该装置包括：采集单元，用于通过显微镜采集对应的图像，并调节背景获得高分辨RGB图像；标注单元，用于对图像中的目标蛋白进行标注，并统计标注后的数据集；输出单元，用于将统计后的数据集输入到人工神经网络进行检测，输出目标蛋白的位置以及对应的置信度，对检测结果进行整合，得到区域内目标蛋白的数量。
第三方面，本申请实施例还提供了一种计算机设备，包括存储器、 处理器以及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述程序时实现如本申请实施例描述中任一所述的方法。
第四方面，本申请实施例还提供了一种计算机设备一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序用于：所述计算机程序被处理器执行时实现如本申请实施例描述中任一所述的方法。
本发明的有益效果：
本发明提供的目标蛋白快速识别及定量方法，采用图像分析方法对目标蛋白进行计数，记录指定区域目标蛋白数量；适应多种倍镜下的图像。同时，采用深度学习网络，学习目标蛋白的特征，制作数据集，从而对目标蛋白的位置进行检测，为计算机辅助计数显微荧光图像带来新的思路。通过脑图谱可以定位特定脑区位置，并统计相应区域目标蛋白的数量。本发明不限于免疫荧光的图像，对于免疫组化，免疫胶体金标记，与GFP构建融合基因表达融合蛋白等蛋白测试方式均可以实现目标蛋白的识别及统计。不仅可以识别脑组织图像，也可拓展为外周组织细胞识别；目标检测网络也可扩展为faster-rcnn,cascade-rcnn,ssd等多种类型；图像不局限于光学图像，对于其他影像学图像都可适用；该发明还可以用于对双重或多重染色图像进行处理识别。
附图说明
通过阅读参照以下附图所作的对非限制性实施例所作的详细描述，本申请的其它特征、目的和优点将会变得更明显：
图1示出了本申请实施例提供的目标蛋白快速识别及定量方法的流程示意图；
图2示出了根据本申请一个实施例的目标蛋白快速识别及定量装 置200的示例性结构框图；
图3示出了适于用来实现本申请实施例的终端设备的计算机系统的结构示意图；
图4示出了本申请实施例提供的precision-recall曲线的示意图；
图5示出了本申请实施例提供的测试样本的示意图；
图6示出了本申请实施例提供的蛋白质计数可视化的示意图；
图7示出了本申请实施例提供的yolo v5结构的示意图。
具体实施方式
为使本发明的上述目的、特征和优点能够更加明显易懂，下面结合附图对本发明的具体实施方式做详细的说明。在下面的描述中阐述了很多具体细节以便于充分理解本发明。但是本发明能够以很多不同于在此描述的其它方式来实施，本领域技术人员可以在不违背本发明内涵的情况下做类似改进，因此本发明不受下面公开的具体实施例的限制。
在本发明的描述中，需要理解的是，术语“中心”、“纵向”、“横向”、“长度”、“宽度”、“厚度”、“上”、“下”、“前”、“后”、“左”、“右”、“竖直”、“水平”、“顶”、“底”、“内”、“外”、“顺时针”、“逆时针”、“轴向”、“径向”、“周向”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。
此外，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该 特征。在本发明的描述中，“多个”的含义是至少两个，例如两个，三个等，除非另有明确具体的限定。
在本发明中，除非另有明确的规定和限定，术语“安装”、“相连”、“连接”、“固定”等术语应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或成一体；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通或两个元件的相互作用关系，除非另有明确的限定。对于本领域的普通技术人员而言，可以根据具体情况理解上述术语在本发明中的具体含义。
在本发明中，除非另有明确的规定和限定，第一特征在第二特征“上”或“下”可以是第一和第二特征直接接触，或第一和第二特征通过中间媒介间接接触。而且，第一特征在第二特征“之上”、“上方”和“上面”可是第一特征在第二特征正上方或斜上方，或仅仅表示第一特征水平高度高于第二特征。第一特征在第二特征“之下”、“下方”和“下面”可以是第一特征在第二特征正下方或斜下方，或仅仅表示第一特征水平高度小于第二特征。
需要说明的是，当元件被称为“固定于”或“设置于”另一个元件，它可以直接在另一个元件上或者也可以存在居中的元件。当一个元件被认为是“连接”另一个元件，它可以是直接连接到另一个元件或者可能同时存在居中元件。本文所使用的术语“垂直的”、“水平的”、“上”、“下”、“左”、“右”以及类似的表述只是为了说明的目的，并不表示是唯一的实施方式。
请参考图1，图1示出了本申请实施例提供的目标蛋白快速识别及定量方法的流程示意图。
如图1所示，该方法包括：
步骤110，通过显微镜采集对应的图像，并调节背景获得高分辨 RGB图像；
步骤120，对图像中的目标蛋白进行标注，并统计标注后的数据集；
步骤130，将统计后的数据集输入到人工神经网络进行检测，输出目标蛋白的位置以及对应的置信度，对检测结果进行整合，得到区域内目标蛋白的数量。
采用上述技术方案，采用图像分析方法对目标蛋白进行计数，记录指定区域目标蛋白数量；适应多种倍镜下的图像。同时，采用深度学习网络，学习目标蛋白的特征，制作数据集，从而对目标蛋白的位置进行检测，为计算机辅助计数显微荧光图像带来新的思路。通过脑图谱可以定位特定脑区位置，并统计相应区域目标蛋白的数量。本发明不限于免疫荧光的图像，对于免疫组化，免疫胶体金标记，与GFP构建融合基因表达融合蛋白等蛋白测试方式均可以实现目标蛋白的识别及统计。不仅可以识别脑组织图像，也可拓展为外周组织细胞识别；目标检测网络也可扩展为faster-rcnn,cascade-rcnn,ssd等多种类型；图像不局限于光学图像，对于其他影像学图像都可适用；该发明还可以用于对双重或多重染色图像进行处理识别。
在一些实施例中，本申请中的通过显微镜采集对应的图像，包括：选择c-Fos染色图像，所述图像中包括有激活的神经元；将对应组织切片免疫荧光染色，当免疫荧光染色成功后，通过显微镜采集图像。具体地，通过无差别选择c-Fos染色图像，其中，入选的图像需要有激活的神经元，对染色颜色无要求等。
在一些实施例中，本申请中的调节背景获得高分辨RGB图像，包括：判断图像的清晰度是否小于阈值，若是，则进行对比度和饱和度的增强，以使得图像的清晰度达到阈值。
在一些实施例中，本申请中的对图像中的目标蛋白进行标注，并 统计标注后的数据集，包括：将图像输入到Labelme软件中进行标注，其中，标注框与目标细胞边缘相切；采用划窗法将有包含标注框的图像块截取下来作为数据样本制作数据集。
具体地，通过将原始数据直接放入Labelme软件中进行标注，标注框与目标细胞边缘相切。该标注选用两个经验丰富的技术人员进行校对。由于原始图像为2048×2048，分辨率过大，不利于模型的参数计算，因此，采用划窗法将有包含标注框的图像块截取下来作为单一数据样本制作数据集。划窗的大小为512×512，步长为50。
在一些实施例中，本申请中的将统计后的数据集输入到人工神经网络进行检测，包括：将统计后的数据集输入到人工神经网络，通过YOLO对测试模型进行检测，所述测试结果包括真正类(True Positive，TP)、假正类(False Positive，FP)、真负类(True Negative，TN)和假负类(False Negative，FN)；通过计算精准率(Precision)、召回率(Recall)、平均精度(Average Precision，AP)对模型的预测结果进行评估，其中精准率代表正确预测为正占全部预测为正的比例；召回率指正确预测为正占全部正样本的比例。
具体地，参考图7所示，本申请采用yolo-v5模型进行训练，其中YOLO(you only look once)就是one-stage算法系列的代表。它将目标检测任务当做回归问题来处理，直接通过整张图片的所有像素得到bounding box的坐标、box中包含物体的置信度和类别概率。YOLO检测物体非常快，没有复杂的检测流程，只需要将图像输入到神经网络就可以得到检测结果，YOLO可以非常快的完成物体检测任务。
将测试结果分为真正类(True Positive，TP)、假正类(False Positive，FP)、真负类(True Negative，TN)和假负类(False Negative，FN)，由这四类情况计算精准率(Precision)、召回率(Recall)、平均精度(Average Precision，AP)对模型的预测结果进行评估，其中精准率代表正确预 测为正占全部预测为正的比例；召回率指正确预测为正占全部正样本的比例，其中，精准率
Figure PCTCN2021137871-appb-000004
召回率
Figure PCTCN2021137871-appb-000005
平均精度
Figure PCTCN2021137871-appb-000006
进一步地，参考图4、图5以及图6所示，图4中利用划窗采集了512*512大小的图像块共计1839张，训练测试集比例为8:2，在测试集上，到阈值选择为0.5时，AP达到了0.938。图5中给出了测试样本的金标准(左)与预测(右)，图6中给出了蛋白质计数可视化图
在一些实施例中，本申请中的将统计后的数据集输入到人工神经网络进行检测之后，该方法还包括：判断输入端为整图还是手动标定区域，若输入整图，则对图像进行划窗处理后依次输入网络模型；若输入端为手动标定区域时，则对多边形区域取最小外接矩形，并对非感兴趣区域像素点置0，同时，通过输出端调整置信度，整合结果，计算对应候选框的个数。
进一步地，参考图2，图2示出了根据本申请一个实施例的目标蛋白快速识别及定量装置200的示例性结构框图。
如图2所示，该装置包括：
采集单元210，用于通过显微镜采集对应的图像，并调节背景获得高分辨RGB图像；
标注单元220，用于对图像中的目标蛋白进行标注，并统计标注后的数据集；
输出单元230，用于将统计后的数据集输入到人工神经网络进行检测，输出目标蛋白的位置以及对应的置信度，对检测结果进行整合，得到区域内目标蛋白的数量。
应当理解，装置200中记载的诸单元或模块与参考图1描述的方法中的各个步骤相对应。由此，上文针对方法描述的操作和特征同样适用于装置200及其中包含的单元，在此不再赘述。装置200可以预 先实现在电子设备的浏览器或其他安全应用中，也可以通过下载等方式而加载到电子设备的浏览器或其安全应用中。装置200中的相应单元可以与电子设备中的单元相互配合以实现本申请实施例的方案。
下面参考图3，其示出了适于用来实现本申请实施例的终端设备或服务器的计算机系统300的结构示意图。
如图3所示，计算机系统300包括中央处理单元(CPU)301，其可以根据存储在只读存储器(ROM)302中的程序或者从存储部分308加载到随机访问存储器(RAM)303中的程序而执行各种适当的动作和处理。在RAM 303中，还存储有系统300操作所需的各种程序和数据。CPU 301、ROM 302以及RAM 303通过总线304彼此相连。输入/输出(I/O)接口305也连接至总线304。
以下部件连接至I/O接口305：包括键盘、鼠标等的输入部分306；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307；包括硬盘等的存储部分308；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309。通信部分309经由诸如因特网的网络执行通信处理。驱动器310也根据需要连接至I/O接口305。可拆卸介质311，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310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308。
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目标蛋白快速识别及定量方法，其包括有形地包含在机器可读介质上的计算机程序，所述计算机程序包含用于执行图1的方法的程序代码。在这样的实施例中，该计算机程序可以通过通信部分309从网络上被下载和安装，和/或从可拆卸介质311被安装。
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。
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，例如，可以描述为：一种处理器包括第一子区域生成单元、第二子区域生成单元以及显示区域生成单元。其中，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，例如，显示区域生成单元还可以被描述为“用于根据第一子区域和第二子区域生成文本的显示区域的单元”。
作为另一方面，本申请还提供了一种计算机可读存储介质，该计算机可读存储介质可以是上述实施例中前述装置中所包含的计算机可读存储介质；也可以是单独存在，未装配入设备中的计算机可读存储介质。计算机可读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应用于透明窗口信封的文本生成方法。
以上描述仅为本申请的较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本申请中所涉及的发明范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离前述 发明构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本申请中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。

Claims (10)

  1. 一种目标蛋白快速识别及定量方法，其特征在于，该方法包括：
    通过显微镜采集对应的图像，并调节背景获得高分辨RGB图像；
    对图像中的目标蛋白进行标注，并统计标注后的数据集；
    将统计后的数据集输入到人工神经网络进行检测，输出目标蛋白的位置以及对应的置信度，对检测结果进行整合，得到区域内目标蛋白的数量。
  2. 根据权利要求1所述的目标蛋白快速识别及定量方法，其特征在于，所述通过显微镜采集对应的图像，包括：
    选择c-Fos染色图像，所述图像中包括有激活的神经元；
    将对应组织切片免疫荧光染色，当免疫荧光染色成功后，通过显微镜采集图像。
  3. 根据权利要求1所述的目标蛋白快速识别及定量方法，其特征在于，所述调节背景获得高分辨RGB图像，包括：
    判断图像的清晰度是否小于阈值，若是，则进行对比度和饱和度的增强，以使得图像的清晰度达到阈值。
  4. 根据权利要求1所述的目标蛋白快速识别及定量方法，其特征在于，所述对图像中的目标蛋白进行标注，并统计标注后的数据集，包括：
    将图像输入到Labelme软件中进行标注，其中，标注框与目标细胞边缘相切；
    采用划窗法将有包含标注框的图像块截取下来作为数据样本制作数据集。
  5. 根据权利要求1所述的目标蛋白快速识别及定量方法，其特征在于，所述将统计后的数据集输入到人工神经网络进行检测，包括：
    将统计后的数据集输入到人工神经网络，通过YOLO对测试模型进行检测，所述测试结果包括真正类(True Positive，TP)、假正类(False Positive，FP)、真负类(True Negative，TN)和假负类(False Negative，FN)；
    通过计算精准率(Precision)、召回率(Recall)、平均精度(Average Precision，AP)对模型的预测结果进行评估，其中精准率代表正确预测为正占全部预测为正的比例；召回率指正确预测为正占全部正样本的比例。
  6. 根据权利要求5所述的目标蛋白快速识别及定量方法，其特征在于，
    所述精准率
    Figure PCTCN2021137871-appb-100001
    所述召回率
    Figure PCTCN2021137871-appb-100002
    所述平均精度
    Figure PCTCN2021137871-appb-100003
  7. 根据权利要求5所述的目标蛋白快速识别及定量方法，其特征在于，所述将统计后的数据集输入到人工神经网络进行检测之后，该方法还包括：
    判断输入端为整图还是手动标定区域，若输入整图，则对图像进行划窗处理后依次输入网络模型；若输入端为手动标定区域时，则对多边形区域取最小外接矩形，并对非感兴趣区域像素点置0，同时，通过输出端调整置信度，整合结果，计算对应候选框的个数。
  8. 一种目标蛋白快速识别及定量装置，其特征在于，该装置包括：
    采集单元，用于通过显微镜采集对应的图像，并调节背景获得高分辨RGB图像；
    标注单元，用于对图像中的目标蛋白进行标注，并统计标注后的数据集；
    输出单元，用于将统计后的数据集输入到人工神经网络进行检测，输出目标蛋白的位置以及对应的置信度，对检测结果进行整合，得到 区域内目标蛋白的数量。
  9. 一种计算机设备，包括存储器、处理器以及存储在存储器上并可在处理器上运行的计算机程序，其特征在于，所述处理器执行所述程序时实现如权利要求1-7中任一所述的方法。
  10. 一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序用于：
    所述计算机程序被处理器执行时实现如权利要求1-7中任一所述的方法。
PCT/CN2021/137871 2021-12-14 2021-12-14 目标蛋白快速识别及定量方法 WO2023108412A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/137871 WO2023108412A1 (zh) 2021-12-14 2021-12-14 目标蛋白快速识别及定量方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/137871 WO2023108412A1 (zh) 2021-12-14 2021-12-14 目标蛋白快速识别及定量方法

Publications (1)

Publication Number Publication Date
WO2023108412A1 true WO2023108412A1 (zh) 2023-06-22

Family

ID=86774926

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/137871 WO2023108412A1 (zh) 2021-12-14 2021-12-14 目标蛋白快速识别及定量方法

Country Status (1)

Country Link
WO (1) WO2023108412A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117405914A (zh) * 2023-09-21 2024-01-16 湖北微流控科技有限公司 一种离心式微流控试剂定量转移方法及系统
CN117405914B (zh) * 2023-09-21 2024-06-11 湖北微流控科技有限公司 一种离心式微流控试剂定量转移方法及系统

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108596046A (zh) * 2018-04-02 2018-09-28 上海交通大学 一种基于深度学习的细胞检测计数方法及系统
CN111242922A (zh) * 2020-01-13 2020-06-05 上海极链网络科技有限公司 一种蛋白质图像分类方法、装置、设备及介质
US20200342597A1 (en) * 2017-12-07 2020-10-29 Ventana Medical Systems, Inc. Deep-learning systems and methods for joint cell and region classification in biological images
CN112184624A (zh) * 2020-09-01 2021-01-05 燊赛(上海)智能科技有限公司 一种基于深度学习的图片检测方法及系统
CN112434560A (zh) * 2020-10-28 2021-03-02 武汉科技大学 一种基于深度学习的安全装备实时检测方法及装置
CN112784767A (zh) * 2021-01-27 2021-05-11 天津理工大学 基于白细胞显微图像的细胞实例分割算法
US20210216745A1 (en) * 2020-01-15 2021-07-15 DeePathology Ltd. Cell Detection Studio: a system for the development of Deep Learning Neural Networks Algorithms for cell detection and quantification from Whole Slide Images
CN113222991A (zh) * 2021-06-16 2021-08-06 南京农业大学 基于深度学习网络的田间麦穗计数以及小麦产量预测
WO2021189771A1 (zh) * 2020-07-30 2021-09-30 平安科技（深圳）有限公司 玻片数字化信息质量检测方法、装置、设备及介质
CN113724223A (zh) * 2021-08-27 2021-11-30 江南大学 基于光学显微镜的YOLOv3数据集制作方法及系统

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200342597A1 (en) * 2017-12-07 2020-10-29 Ventana Medical Systems, Inc. Deep-learning systems and methods for joint cell and region classification in biological images
CN108596046A (zh) * 2018-04-02 2018-09-28 上海交通大学 一种基于深度学习的细胞检测计数方法及系统
CN111242922A (zh) * 2020-01-13 2020-06-05 上海极链网络科技有限公司 一种蛋白质图像分类方法、装置、设备及介质
US20210216745A1 (en) * 2020-01-15 2021-07-15 DeePathology Ltd. Cell Detection Studio: a system for the development of Deep Learning Neural Networks Algorithms for cell detection and quantification from Whole Slide Images
WO2021189771A1 (zh) * 2020-07-30 2021-09-30 平安科技（深圳）有限公司 玻片数字化信息质量检测方法、装置、设备及介质
CN112184624A (zh) * 2020-09-01 2021-01-05 燊赛(上海)智能科技有限公司 一种基于深度学习的图片检测方法及系统
CN112434560A (zh) * 2020-10-28 2021-03-02 武汉科技大学 一种基于深度学习的安全装备实时检测方法及装置
CN112784767A (zh) * 2021-01-27 2021-05-11 天津理工大学 基于白细胞显微图像的细胞实例分割算法
CN113222991A (zh) * 2021-06-16 2021-08-06 南京农业大学 基于深度学习网络的田间麦穗计数以及小麦产量预测
CN113724223A (zh) * 2021-08-27 2021-11-30 江南大学 基于光学显微镜的YOLOv3数据集制作方法及系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117405914A (zh) * 2023-09-21 2024-01-16 湖北微流控科技有限公司 一种离心式微流控试剂定量转移方法及系统
CN117405914B (zh) * 2023-09-21 2024-06-11 湖北微流控科技有限公司 一种离心式微流控试剂定量转移方法及系统

Similar Documents

Publication Publication Date Title
US10083340B2 (en) Automated cell segmentation quality control
US10176579B2 (en) Tissue object-based machine learning system for automated scoring of digital whole slides
Wang Robust automated tumour segmentation on histological and immunohistochemical tissue images
US8831889B2 (en) Quantification of differences between measured values and statistical validation based on the differences
US8306939B2 (en) Examination value predicting device using electrophoresis waveform, prediction method, and predicting program
US20240079116A1 (en) Automated segmentation of artifacts in histopathology images
Watanabe et al. IPAP: image processor for analytical pathology
Khan et al. A novel system for scoring of hormone receptors in breast cancer histopathology slides
JP4985480B2 (ja) がん細胞を分類する方法、がん細胞を分類するための装置及びがん細胞を分類するためのプログラム
US20210343009A1 (en) Distance-based tissue state determination
US20070206845A1 (en) Granularity analysis in cellular phenotypes
CN113936005A (zh) 一种dna指数计算方法、装置、计算机设备及存储介质
WO2023108412A1 (zh) 目标蛋白快速识别及定量方法
Desale et al. A Deep Learning Framework for Multi-Cancer Detection in Medical Imaging
US20220383629A1 (en) Label-free cell classification and screening system based on hybrid transfer learning
Han et al. eDetect: a fast error detection and correction tool for live cell imaging data analysis
CN114152557B (zh) 基于图像分析的血细胞计数方法和系统
US11520775B2 (en) Adaptive data sub-sampling and computation
CN114463745A (zh) 目标蛋白快速识别及定量方法
Su et al. ROI-BMC-DNNet: An efficient automatic analysis model of whole-slide scanned bone marrow aspirate images for the diagnosis of hematological disorders
Saeed-Vafa et al. Practical applications of digital pathology
Ganapathy Cancer detection using deep neural network differentiation of squamous carcinoma cells in oral pathology
WO2023240819A1 (zh) 一种甲状腺疾病病理分析模块
US20240062365A1 (en) Imaging-based microorganosphere drug assays
Hülber et al. Influence of sample preparation optimization on the accuracy of dose assessment of an automatic non-fluorescent MN scoring system

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: 21967552

Country of ref document: EP

Kind code of ref document: A1