CN112288760B - Adherent cell image screening method and system and cell image analysis method - Google Patents
Adherent cell image screening method and system and cell image analysis method Download PDFInfo
- Publication number
- CN112288760B CN112288760B CN202011185767.8A CN202011185767A CN112288760B CN 112288760 B CN112288760 B CN 112288760B CN 202011185767 A CN202011185767 A CN 202011185767A CN 112288760 B CN112288760 B CN 112288760B
- Authority
- CN
- China
- Prior art keywords
- cell
- single cell
- image
- cell image
- qualified
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
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
- G06T7/13—Edge detection
-
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10056—Microscopic image
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention discloses a method and a system for screening adhesion cell images and a cell image analysis method. The adherent cell image screening method comprises (1) detecting the cell edge of a single cell image obtained after image segmentation processing to obtain the morphological characteristics of the single cell; (2) judging whether the single cell image is qualified or not according to morphological characteristics; and if the single cell image is judged to be qualified, keeping the single cell image, otherwise, screening out the single cell image. The system comprises a morphological characteristic acquisition module and a single cell screening module. The cell image analysis method is used for preprocessing the cell image to be analyzed according to the adherent cell image screening method provided by the invention. The method for screening the adhesion cell image fundamentally avoids cell image analysis noise caused by the adhesion state of the cell image during cell image acquisition in cell image analysis, and remarkably reduces the interference of inhomogeneous cells on subsequent result calculation.
Description
Technical Field
The invention belongs to the technical field of biological image processing, and particularly relates to a method and a system for screening out an adhesion cell image and a cell image analysis method.
Background
The demand for analyzing and processing cell images by a computer based on various algorithms is increasing, and the method firstly divides cell images obtained by experiments to obtain images of single cells, and then inputs different models for training and predicting according to analyzed targets.
However, the cells have different shapes and different densities, and the cells may stick together in cell culture. The cell image segmentation method, whether the cell segmentation of the traditional algorithm such as a binarization method and a watershed algorithm, or an FCN full convolution neural network and a U-Net network based on deep learning, cannot well detect and segment the edge of each cell in the adherent cells, and the cut cells are possibly incomplete. The cell images with incomplete edges as data belong to noise data, and the data are added into subsequent model training or analysis, so that the performance of the model, such as accuracy and the like, can be seriously interfered, and the analysis effect is poor. The more the noise data, the more the interference to the accuracy and other indexes of the model.
At present, in order to solve the problem of data interference caused by incomplete image segmentation due to cell adhesion, an optimization method is proposed in a 'Caps-Unet-based adhesion cell nucleus edge detection and segmentation' paper in 2020 of Li Xingwei university of Shanxi, but the method cannot significantly reduce the interference of non-uniform cells on subsequent result operation. However, since two or more cells are overlapped with each other on a physical plane, boundaries thereof cross each other, or cells are closely arranged and pressed against each other, deformation occurs, so that information on cell boundary membranes is lost during image acquisition. Therefore, none of the prior art, including the above-described methods, is able to separate single cell images without pixel loss. Even if the cell edge detection and segmentation algorithm is optimized, the obtained single cell image still has pixel loss or interferes with the subsequent image analysis operation.
Disclosure of Invention
Aiming at the defects or improvement requirements of the prior art, the invention provides a method and a system for screening an adhesion cell image and a cell image analysis method, aiming at automatically screening the cell image to be analyzed, filtering the adhesion cell image and keeping a single cell image with complete form, so that the noise influencing the cell image analysis is fundamentally eliminated, and the accuracy of the cell image analysis result is remarkably improved, thereby solving the technical problems of large noise and inaccurate result of the cell image analysis technology caused by the fact that the adhesion cell or the cell image with incomplete form can not be avoided when the cell image is acquired and the edge is detected in the prior art.
To achieve the above object, according to one aspect of the present invention, there is provided a method for screening an image of adherent cells, comprising the steps of:
(1) detecting the cell edge of a single cell image obtained after image segmentation processing to obtain the morphological characteristics of the single cell;
(2) judging whether the single cell image is qualified or not according to the morphological characteristics of the single cell image obtained in the step (1); and if the single cell image is judged to be qualified, keeping the single cell image, and otherwise, screening out the single cell image.
Preferably, the adherent cell image screening method comprises the following morphological characteristics in step (1): cell area fraction, aspect ratio, convexity, and/or circularity.
Preferably, the adherent cell image screening method comprises the steps of (1) detecting the cell edge of the adherent cell image, and preferably performing a cell edge detection algorithm by using Unet;
the cell area ratio is the ratio of pixels in the cells to total pixels of the image;
the aspect ratio is a proportional relation between the width and the height of the single cell image;
the convexity is a physical quantity that characterizes whether the cell contains a convex hull;
the roundness e is calculated by the following method: e ═ S (4 pi ═ S) 2 )/L 2 Wherein S is the cell area and L is the cell perimeter.
Preferably, in the adherent cell image screening method, the step (2) of judging whether the single cell image is qualified or not adopts a threshold method, a machine learning-based classification algorithm and a clustering algorithm; preferably, a threshold method is adopted, and specifically, the method comprises the following steps:
when the morphological characteristics of the single cell image all fall within a preset qualified threshold range, judging that the single cell is qualified; otherwise, judging that the single cell is unqualified;
preferably, the qualified threshold range of the cell area ratio in the step (2) of the adherent cell image screening method is 65-80%; the qualified threshold range of the aspect ratio is that the aspect ratio is 0.8-1.4; if the qualified threshold range of the convexity is not, the convex hull is not available; the roundness qualified threshold range is more than 0.85.
According to another aspect of the present invention, there is provided an adherent cell image screening system comprising: the device comprises a morphological characteristic acquisition module and a single cell screening module;
the morphological characteristic acquisition module is used for detecting the cell edge of a single cell image obtained after image segmentation processing, acquiring the morphological characteristic of the single cell and submitting the morphological characteristic to the single cell screening module;
the single cell screening module is used for judging whether the single cell image is qualified or not according to the morphological characteristics of the single cell image and reserving the single cell image.
Preferably, the adherent cell image screening system comprises the following morphological characteristics: cell area ratio, aspect ratio, convexity, and/or circularity;
the cell area ratio is the ratio of pixels in the cells to the total pixels of the image;
the aspect ratio is a proportional relation between the width and the height of the single cell image;
the convexity is a physical quantity that characterizes whether the cell contains a convex hull;
the roundness e is calculated by the following method: e ═ S (4 pi ═ S) 2 )/L 2 Wherein S is the cell area and L is the cell perimeter.
Preferably, the adherent cell image screening system detects the cell edge thereof, and preferably adopts a cell edge detection algorithm of Unet; and judging whether the single cell image is qualified or not by adopting a threshold value method.
Preferably, the adherent cell image screening system is used for judging that the single cell is qualified when the morphological characteristics of the single cell image all fall within a preset qualified threshold range; otherwise, judging that the single cell is unqualified;
wherein, the optimal qualified threshold range of the cell area ratio is 65-80 percent; the preferable qualified threshold range of the aspect ratio is 0.8-1.4; if the qualified threshold range of the convexity is not, the convex hull is not available; the roundness of the steel is preferably in a qualified threshold range of more than 0.85.
According to another aspect of the present invention, there is provided a cell image analysis method including the steps of: according to the adherent cell image screening method provided by the invention, a cell image to be analyzed is preprocessed; and performing cell image analysis by using a cell image set formed by the reserved single cell images as an analysis object.
In general, the above technical solutions contemplated by the present invention can achieve the following advantageous effects compared to the prior art.
The screening method of the adhesion cell image provides a new preprocessing step for cell image analysis and calculation, and the cell image is screened after the cell image is divided, so that cell information which does not meet the standard is directly excluded from a cell image analysis algorithm, cell image analysis noise caused by the adhesion state of the cell image during cell image acquisition in cell image analysis is fundamentally avoided, the consistency of input data of the cell image analysis is improved, and the interference of inhomogeneous cells on subsequent result calculation is remarkably reduced.
Drawings
FIG. 1 is a schematic flow chart of a method for screening out adherent cell images according to the present invention;
FIG. 2 is a set of images of cells to be processed according to the adherent cell image screening method provided in the embodiment of the present invention;
FIG. 3 is a set of single cell images obtained after the adherent cell image screening method provided by the embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The invention provides an adherent cell image screening method, which comprises the following steps:
(1) detecting the cell edge of a single cell image obtained after image segmentation processing to obtain the morphological characteristics of the single cell; the morphological characteristics include: cell area fraction, aspect ratio, convexity, and/or circularity.
Detecting the cell edge of the cell, preferably adopting a cell edge detection algorithm by using Unet;
the cell area ratio is the ratio of pixels in the cells to the total pixels of the image;
the aspect ratio is a proportional relation between the width and the height of the single cell image;
the convexity is a physical quantity that characterizes whether the cell contains a convex hull;
the roundness e is calculated by the following method: e ═ S (4 pi ═ S) 2 )/L 2 Wherein S is the cell area and L is the cell perimeter.
(2) Judging whether the single cell image is qualified or not according to the morphological characteristics of the single cell image obtained in the step (1); and if the single cell image is judged to be qualified, keeping the single cell image, and otherwise, screening out the single cell image.
Judging whether the single cell image is qualified or not, wherein a threshold method, a machine learning-based classification algorithm and a clustering algorithm can be adopted; the threshold method does not need prior modeling, does not need a large amount of storage and operation resources, has the fastest operation speed, and has the classification effect equivalent to that of an intelligent learning algorithm when cell images of a cell edge detection algorithm are adopted for carrying out threshold method judgment through testing; the threshold method is preferred in the present invention because of the large number of cells.
Adopting a threshold method, namely judging that the single cell is qualified when the morphological characteristics of the single cell image fall within a preset qualified threshold range; otherwise, judging that the single cell is unqualified. Wherein, the optimal qualified threshold range of the cell area ratio is 65-80 percent; the preferable qualified threshold range of the aspect ratio is that the aspect ratio is 0.8-1.4; if the qualified threshold range of the convexity is not, the convex hull is not available; the roundness of the steel is preferably in a qualified threshold range of more than 0.85.
The invention provides an adherent cell image screening system, which comprises: the device comprises a morphological characteristic acquisition module and a single cell screening module;
the morphological characteristic acquisition module is used for detecting the cell edge of a single cell image obtained after image segmentation processing, acquiring the morphological characteristic of the single cell and submitting the morphological characteristic to the single cell screening module; the morphological characteristics include: cell area fraction, aspect ratio, convexity, and/or circularity.
Detecting the cell edge of the cell, preferably performing a cell edge detection algorithm by using Unet;
the cell area ratio is the ratio of pixels in the cells to the total pixels of the image;
the aspect ratio is a proportional relation between the width and the height of the single cell image;
the convexity is a physical quantity that characterizes whether the cell contains a convex hull;
the roundness e is calculated by the following method: e ═ S (4 pi ═ S) 2 )/L 2 Wherein S is the cell area and L is the cell perimeter.
The single cell screening module is used for judging whether the single cell image is qualified or not according to the morphological characteristics of the single cell image and reserving the single cell image.
Judging whether the single cell image is qualified or not, and adopting a threshold value method, a machine learning-based classification algorithm and a clustering algorithm; preferably, a threshold method is adopted, specifically:
when the morphological characteristics of the single cell image all fall within a preset qualified threshold range, judging that the single cell is qualified; otherwise, judging that the single cell is unqualified;
wherein, the optimal qualified threshold range of the cell area ratio is 65-80%; the preferable qualified threshold range of the aspect ratio is that the aspect ratio is 0.8-1.4; if the qualified threshold range of the convexity is not, the convex hull is not available; the roundness of the steel pipe is preferably in a qualified threshold range of more than 0.85.
The cell image analysis method provided by the invention comprises the following steps:
according to the adherent cell image screening method provided by the invention, a cell image to be analyzed is preprocessed; and performing cell image analysis by taking a cell image set formed by the reserved single cell images as an analysis object.
The following are examples:
an adherent cell image screening method comprises the following steps:
(1) detecting the cell edge of a single cell image obtained after image segmentation processing to obtain the morphological characteristics of the single cell; the morphological characteristics include: cell area fraction, aspect ratio, and convexity.
Detecting the cell edge of the cell, and performing a cell edge detection algorithm by using Unet; the method comprises the following specific steps:
A) pretreatment: preprocessing a cell image to generate a mask image and inputting data;
taking the original cell picture (shown in figure 2) in the folder A as the input of the u-net model, obtaining a mask map corresponding to the cell image, and storing the mask map in the folder B;
B) optimizing the edge: extracting a cell contour, carrying out binarization coding on the cell contour, and optimizing the cell edge;
binarizing a cell photo mask map in a B folder, wherein the inside of a cell is white, and the outside of the cell is black;
filtering the binarized picture by using a median filtering media blank (the linear size parameter of the aperture is set to be 5) of OpenCV (open cell container), so that the edge part of the cell is smoother;
C) morphological feature extraction: various characteristic information of the cells is acquired through the mask image as follows:
the cell area ratio cell _ area _ ratio is a ratio of intra-cell pixels to total pixels of the image, i.e., inside/(inside + outside), and is used as a cell area ratio;
the width/height ratio is the proportional relation width/height between the width and the height of the single cell image;
the convex is a physical quantity for characterizing whether the cell contains a convex hull or not;
traversing the cell mask graph processed in the step B), calculating the number of white pixel points and recording the number of the white pixel points as inside, and calculating the number of black pixel points and recording the number of the black pixel points as out;
secondly, obtaining the width and height of the image by using a shape function of OpenCV; thirdly, carrying out convexity detection on the cell mask outline by using an iscontourConvex function of OpenCV, and recording the result as convex;
(2) judging whether the single cell image is qualified or not according to the morphological characteristics of the single cell image obtained in the step (1); and if the single cell image is judged to be qualified, keeping the single cell image, and otherwise, screening out the single cell image.
In this embodiment, a threshold method is used to determine whether the single cell image is qualified, specifically:
if the cell _ area _ ratio is 65-80%, the width _ height _ ratio is within 0.8-1.4, the context is FALSE, the cell is qualified, otherwise, the cell is unqualified; the rejected cell photographs were sorted into folder C, and the remaining photographs remained unchanged in the original folder, as shown in fig. 3.
When the images are not screened, all the images are regarded as qualified images, and data interference is large; after screening, images of the majority of the disqualified cells were selected, as shown in the table below.
Without screening | After screening | |
Conform to | 65450 | 56920 |
Non-conforming to | 0 | 8530 |
In this 65450 batch of data, the true sample distribution is:
meet with | Non-conforming to |
56706 | 8744 |
The correct screening rate of the screening method of this example was calculated to be 97.5%; the accuracy of the method is higher than that of an intelligent classification algorithm, and the method does not need to be clustered into more than two types of cells.
It will be understood by those skilled in the art that the foregoing is only an exemplary embodiment of the present invention, and is not intended to limit the invention to the particular forms disclosed, since various modifications, substitutions and improvements within the spirit and scope of the invention are possible and within the scope of the appended claims.
Claims (5)
1. An adherent cell image screening method is characterized by comprising the following steps:
(1) detecting the cell edge of a single cell image obtained after image segmentation processing to obtain the morphological characteristics of the single cell; the morphological characteristics include: cell area ratio, aspect ratio, convexity, and/or circularity;
(2) judging whether the single cell image is qualified or not according to the morphological characteristics of the single cell image obtained in the step (1); if the single cell image is judged to be qualified, the single cell image is reserved, otherwise, the single cell image is screened out; the threshold method is adopted for judging whether the single cell image is qualified, and specifically comprises the following steps:
when the morphological characteristics of the single cell image all fall within a preset qualified threshold range, judging that the single cell is qualified; otherwise, judging that the single cell is unqualified;
the qualified threshold range of the cell area ratio is 65-80 percent; the qualified threshold range of the aspect ratio is that the aspect ratio is 0.8-1.4; if the qualified threshold range of the convexity is not, the convex hull is not available; the qualified threshold range of the roundness is more than 0.85;
the cell area ratio is the ratio of pixels in the cells to total pixels of the image;
the aspect ratio is a proportional relation between the width and the height of the single cell image;
the convexity is a physical quantity that characterizes whether the cell contains a convex hull;
the roundness e is calculated by the following method: e ═ S (4 pi ═ S) 2 )/L 2 Wherein S is the cell area and L is the cell perimeter.
2. The method for screening adhesion cell images as claimed in claim 1, wherein the step (1) of detecting the cell edges adopts a cell edge detection algorithm by using Unet.
3. An adherent cell image screening system, comprising: the device comprises a morphological characteristic acquisition module and a single cell screening module; the morphological characteristics include: cell area ratio, aspect ratio, convexity, and/or circularity;
the morphological characteristic acquisition module is used for detecting the cell edge of a single cell image obtained after image segmentation processing, acquiring the morphological characteristic of the single cell and submitting the morphological characteristic to the single cell screening module;
the single cell screening module is used for judging whether the single cell image is qualified or not according to the morphological characteristics of the single cell image; if the single cell image is judged to be qualified, retaining the single cell image; judging whether the single cell image is qualified or not by adopting a threshold value method; when the morphological characteristics of the single cell image all fall within a preset qualified threshold range, judging that the single cell is qualified; otherwise, judging that the single cell is unqualified;
wherein, the qualified threshold range of the cell area ratio is 65-80%; the qualified threshold range of the aspect ratio is that the aspect ratio is 0.8-1.4; if the qualified threshold range of the convexity is not, the convex hull is not available; the qualified threshold range of the roundness is more than 0.85;
the cell area ratio is the ratio of pixels in the cells to total pixels of the image;
the aspect ratio is a proportional relation between the width and the height of the single cell image;
the convexity is a physical quantity that characterizes whether the cell contains a convex hull;
the roundness e is calculated by the following method: e ═ S (4 pi ═ S) 2 )/L 2 Wherein S is the cell area and L is the cell perimeter.
4. The adherent cell image removal system of claim 3, wherein the detection of the cell edges is performed using a cell edge detection algorithm using Unet.
5. A method of cellular image analysis comprising the steps of: the adherent cell image screening method according to claim 1 or 2, wherein the cell image to be analyzed is preprocessed; and performing cell image analysis by using a cell image set formed by the reserved single cell images as an analysis object.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011185767.8A CN112288760B (en) | 2020-10-30 | 2020-10-30 | Adherent cell image screening method and system and cell image analysis method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011185767.8A CN112288760B (en) | 2020-10-30 | 2020-10-30 | Adherent cell image screening method and system and cell image analysis method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112288760A CN112288760A (en) | 2021-01-29 |
CN112288760B true CN112288760B (en) | 2022-08-19 |
Family
ID=74353624
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011185767.8A Active CN112288760B (en) | 2020-10-30 | 2020-10-30 | Adherent cell image screening method and system and cell image analysis method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112288760B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113361478B (en) * | 2021-07-05 | 2023-08-22 | 上海大学 | Deformation tracking method and system in cell movement process |
CN114998332B (en) * | 2022-08-02 | 2022-11-22 | 珠海横琴圣澳云智科技有限公司 | Method and device for determining karyotype abnormal cells |
CN117253227B (en) * | 2023-10-23 | 2024-06-18 | 深圳市茵冠生物科技有限公司 | Monitoring method and system of artificial intelligence cooperative automation equipment |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104933701A (en) * | 2015-05-18 | 2015-09-23 | 重庆大学 | Mammary glandular cell segmentation method based on multi-scale growth and double-strategy adhesion-removing model |
CN108320289A (en) * | 2018-02-12 | 2018-07-24 | 华南理工大学 | A kind of bone marrow cell dividing method based on rarefaction representation and morphological operation |
CN110110709A (en) * | 2019-05-27 | 2019-08-09 | 湖南开启时代智能科技有限公司 | A kind of red white corpuscle differential counting method, system and equipment based on image procossing |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI496112B (en) * | 2013-09-13 | 2015-08-11 | Univ Nat Cheng Kung | Cell image segmentation method and a nuclear-to-cytoplasmic ratio evaluation method using the same |
-
2020
- 2020-10-30 CN CN202011185767.8A patent/CN112288760B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104933701A (en) * | 2015-05-18 | 2015-09-23 | 重庆大学 | Mammary glandular cell segmentation method based on multi-scale growth and double-strategy adhesion-removing model |
CN108320289A (en) * | 2018-02-12 | 2018-07-24 | 华南理工大学 | A kind of bone marrow cell dividing method based on rarefaction representation and morphological operation |
CN110110709A (en) * | 2019-05-27 | 2019-08-09 | 湖南开启时代智能科技有限公司 | A kind of red white corpuscle differential counting method, system and equipment based on image procossing |
Also Published As
Publication number | Publication date |
---|---|
CN112288760A (en) | 2021-01-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112288760B (en) | Adherent cell image screening method and system and cell image analysis method | |
CN113989279B (en) | Plastic film quality detection method based on artificial intelligence and image processing | |
CN111292305B (en) | Improved YOLO-V3 metal processing surface defect detection method | |
CN115457041B (en) | Road quality identification and detection method | |
CN106056118B (en) | A kind of identification method of counting for cell | |
CN114749342B (en) | Lithium battery pole piece coating defect identification method, device and medium | |
CN114140669B (en) | Welding defect recognition model training method and device and computer terminal | |
CN107622277A (en) | A kind of complex-curved defect classification method based on Bayes classifier | |
CN114782330B (en) | Grate abnormity detection method and system based on artificial intelligence | |
CN114820625B (en) | Automobile top block defect detection method | |
CN114782475A (en) | Corrugated carton line pressing defect detection optimization method based on artificial intelligence system | |
CN110969620A (en) | Method and device for detecting magnetic shoe ripple defects | |
CN114495098B (en) | Diaxing algae cell statistical method and system based on microscope image | |
CN114926441A (en) | Defect detection method and system for machining and molding injection molding part | |
CN115294377A (en) | System and method for identifying road cracks | |
CN113435460A (en) | Method for identifying brilliant particle limestone image | |
CN115797813B (en) | Water environment pollution detection method based on aerial image | |
CN111242899A (en) | Image-based flaw detection method and computer-readable storage medium | |
CN113240623A (en) | Pavement disease detection method and device | |
CN117635615B (en) | Defect detection method and system for realizing punching die based on deep learning | |
CN116309577A (en) | Intelligent detection method and system for high-strength conveyor belt materials | |
CN108182676A (en) | A kind of sperm fragment rate detection method, device, equipment and readable storage medium storing program for executing | |
CN110889418A (en) | Gas contour identification method | |
CN115496716A (en) | Single and double micronucleus cell image detection method based on deep learning and related equipment | |
CN110619273B (en) | Efficient iris recognition method and recognition device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20220601 Address after: 518048 No. 323-m, third floor, comprehensive Xinxing phase I, No. 1, Haihong Road, Fubao community, Fubao street, Futian District, Shenzhen, Guangdong Province Applicant after: Shenzhen Taili Biotechnology Co.,Ltd. Address before: 523576 building 3 and 4, gaobao green technology city, Tutang Industrial Zone, Changping Town, Dongguan City, Guangdong Province Applicant before: DONGGUAN TAILI BIOLOGICAL ENGINEERING CO.,LTD. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |