CN116863388A - Sperm motility determining method and system based on neural network - Google Patents
Sperm motility determining method and system based on neural network Download PDFInfo
- Publication number
- CN116863388A CN116863388A CN202311136028.3A CN202311136028A CN116863388A CN 116863388 A CN116863388 A CN 116863388A CN 202311136028 A CN202311136028 A CN 202311136028A CN 116863388 A CN116863388 A CN 116863388A
- Authority
- CN
- China
- Prior art keywords
- sperm
- frame
- sperms
- area
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 230000019100 sperm motility Effects 0.000 title claims abstract description 51
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 19
- 230000033001 locomotion Effects 0.000 claims abstract description 102
- 230000000877 morphologic effect Effects 0.000 claims abstract description 27
- 238000003062 neural network model Methods 0.000 claims abstract description 26
- 210000000582 semen Anatomy 0.000 claims abstract description 13
- 238000011176 pooling Methods 0.000 claims abstract description 7
- 230000008859 change Effects 0.000 claims description 48
- 230000009182 swimming Effects 0.000 claims description 25
- 230000000694 effects Effects 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 2
- 230000004899 motility Effects 0.000 abstract description 16
- 238000004590 computer program Methods 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 102000002322 Egg Proteins Human genes 0.000 description 2
- 108010000912 Egg Proteins Proteins 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000035558 fertility Effects 0.000 description 2
- 230000004720 fertilization Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000000338 in vitro Methods 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 210000004681 ovum Anatomy 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000003703 image analysis method Methods 0.000 description 1
- 230000036512 infertility Effects 0.000 description 1
- 208000000509 infertility Diseases 0.000 description 1
- 231100000535 infertility Toxicity 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000036544 posture Effects 0.000 description 1
- 238000001303 quality assessment method Methods 0.000 description 1
- 230000001850 reproductive effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
- 210000004291 uterus Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- 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
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a sperm motility determining method and system based on a neural network, comprising the steps of obtaining a dynamic video of a shot semen sample, selecting a plurality of video frames as key frames in the dynamic video, identifying the position of sperms in each key frame, and calculating a sample movement value of the sperm; for the same sperm, extracting sperm images in each key frame, carrying out convolution and pooling operation on the sperm images to obtain a plurality of sperm images with the same size, extracting sperm image characteristics and inputting the sperm image characteristics into a neural network model to obtain sperm morphology scores; taking the average value of the morphological scores of all sperms in the sample as a morphological value; and taking the motion value, the morphology value and other characteristic values of the sperm as the input of the deep neural network to obtain the sperm motility value. The invention can quantify the movement information of the sperms and comprehensively evaluate the motility of the sperms by integrating other information of the sperms.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a sperm motility determining method and system based on a neural network.
Background
Sperm motility is one of the important markers for male fertility. Healthy sperm should exhibit a high level of activity and strong exercise capacity, which helps them effectively advance, quickly approaching the ovum, and thus increase the chance of successful conception. By evaluating the vitality of sperms, the fertility of men can be judged, doctors can find out the potential infertility reasons, and a basis is provided for formulating a more accurate treatment scheme. In assisted reproductive technology, in vitro fertilization is widely used. This technique combines a female's ovum with a male's sperm in a laboratory to form a fertilized egg, which is then transplanted back into the female's uterus. In the process of in vitro fertilization, it is particularly important to select healthy sperm. However, the conventional manual sperm motility test method is tedious and time-consuming, and has the influence of subjective factors of different operators, which may lead to inconsistency of results. In addition, manual detection is susceptible to human error, thereby affecting the accuracy of the detection results. To address these problems, the introduction of automated techniques to detect sperm motility has become a trend.
The introduction of automatic sperm motility detection has important significance in reducing human error, and effectively improves the detection accuracy. Compared with manual observation, the method can accurately measure and record the motion parameters of the sperms by adopting a computer technology, and even can acutely capture tiny motion differences, thereby more comprehensively observing the motion characteristics of the sperms. This provides more detailed information for sperm quality assessment, further improving the reliability and accuracy of the results. The evaluation of judging sperm motility is mainly dependent on a plurality of parameters including sperm density, sperm motility, sperm morphology, sperm amount, and the like. Although sperm density and semen volume are relatively easy to quantify, sperm motility and sperm morphology lack uniform standards and therefore cannot be quantified directly. In this context, neural networks provide a new approach to assessing sperm motility by computer analysis.
Disclosure of Invention
In order to solve the above-described problems, in a first aspect, the present invention provides a sperm motility assessment method comprising the steps of:
step 1, acquiring a dynamic video of a shot semen sample, selecting a plurality of video frames as key frames in the dynamic video, identifying the position of sperms in each key frame, and calculating a sample motion value of a sperm motion value;
step 2, for the same sperm, a sperm image is arranged in each key frame, the sperm image in each key frame is extracted, convolution and pooling operations are carried out on the sperm images to obtain a plurality of sperm images with the same size, the sperm images with the same size form an image sequence of the sperm, and the image sequence of the sperm is input into a neural network model to obtain a sperm morphology score; taking the average value of the morphological scores of all sperms in the sample as a morphological value;
and step 3, taking the sample motion value, the morphological value and other characteristic values of the sperm as the input of a deep neural network to obtain a sperm motility value.
Preferably, the selecting a plurality of video frames as key frames in the dynamic video, identifying the position of the sperm in each key frame, and calculating the motion value of the sample, specifically:
Step 11, selecting a segment from dynamic video, determining the length of the segment, identifying sperms in an image of the middle moment of the segment, dividing a moving area for each sperm according to the length and the maximum swimming speed of the sperm, judging whether the moving area of the sperm is overlapped with other sperms, if not, executing S12, otherwise, executing S13;
step 12, calculating the gray level change of the same active area in two adjacent frames of images in the segment, determining a key frame according to the gray level change and the size of the active area, identifying the position of a sperm in the key frame, and fitting a sperm motion curve according to the position by adopting a least square method to obtain a sperm motion track; wherein the key frames of different sperm are different;
step 13, dividing the overlapped active areas into at least one group, wherein the active areas in the group are overlapped with at least one other active area in the group, obtaining an area to be identified according to the position relation of the active areas in the group, determining an interval according to the distance of sperms in the area to be identified, extracting frame images to be identified from the segments according to the interval, determining the area corresponding to each group in the frame images to be identified, identifying the position of sperms in the area corresponding to each group, and fitting the movement curve of the sperms by adopting a least square method according to the position to obtain the movement track of the sperms;
S14, determining the movement value of the single sperm according to the position, the movement track and the length of the sperm, and taking the average value of the movement values of all sperm in the sample as the sample movement value.
Preferably, the method comprises the steps of dividing an activity area for each sperm according to the length and the maximum swimming speed of the sperm, specifically:
determining a radius according to the length and the maximum swimming speed of the sperms, obtaining two circle centers according to a target frame for identifying the sperms, and further determining two circles, wherein the radius of the two circles is determined according to the length and the maximum swimming speed of the sperms; and obtaining a rectangular area, wherein the rectangular area is a rectangle with the smallest area and containing the two circles, and the rectangular area is used as the sperm movement area.
Preferably, the two circle centers are obtained according to the target frame for identifying the sperm, specifically:
obtaining a target frame for identifying sperms, obtaining two widths of the target frame, and taking the center of a line segment where each width is located as a circle center.
Preferably, the calculating the gray scale change of the same active area in two adjacent frames of images in the segment determines a key frame according to the gray scale change and the size of the active area, specifically:
According to the formulaCalculating gray level change of the same active region in two adjacent frames of images, wherein N represents the number of pixel points in the same region, N is a positive integer,/or->Gray value representing j-th pixel of i-th frame,>gray value of j pixel point of i+1th frame,/and>representing the gray scale change of the same active area of the (i+1) th frame relative to the (i) th frame;
taking the segmented first frame as a first key frame, wherein the gray level change of the same active area of the adjacent frame between the next key frame and the last key frame is increased more than a threshold value, and the threshold value is proportional to the size of the active area; and the last frame of the segment is taken as the last key frame.
Preferably, the area to be identified is obtained according to the position relation of the active areas in the group, specifically:
determining a rectangular frame, wherein the rectangular frame comprises all active areas in the group, and the area of the rectangular frame is the smallest; and taking the rectangular frame as an area to be identified.
Preferably, the interval is determined according to the distance between the sperms in the area to be identified, and the frame image to be identified is extracted from the segment according to the interval, specifically:
calculating the average distance of sperms in the area to be identified, determining an interval according to a preset comparison relation table of the average distance and the interval, and extracting a frame image to be identified from the segment according to the determined interval on the basis of the first frame of the segment;
Or alternatively, the process may be performed,
calculating the shortest distance of sperms in the area to be identified, calculating the ratio of the shortest distance to the standard distance, taking the integer part of the product of the ratio and the standard interval as an interval, and extracting the frame image to be identified from the section according to the determined interval based on the first frame of the section.
Preferably, the determining the motion value of the single sperm according to the position, the motion track and the length of the sperm specifically comprises:
determining a moving distance according to the movement track of the sperm, and calculating the ratio of the moving distance to the length to obtain a first parameter;
fitting according to the position of the sperm to obtain a straight line, calculating according to the straight line and the position to obtain an r-square coefficient, and taking the r-square coefficient as a second parameter;
normalizing the first parameter to between 0 and 1 according to the formulaCalculating to obtain sperm movement value, wherein +.>、/>Represents the weight, and->,/>Representing the normalized first parameter, ++>Representing a second parameter.
Preferably, the image sequence of the sperm is input into a neural network model, specifically:
extracting characteristics of each image of the sperm to obtain a sequence, wherein one sperm image corresponds to one sequence, the sequence is used as input of a neural network model, and the number of input parameters of the neural network model is the same as the number of the sequences.
In addition, the invention also provides a sperm motility assessment system, which comprises the following modules:
the motion value calculation module is used for acquiring a dynamic video of the shot semen sample, selecting a plurality of video frames from the dynamic video as key frames, identifying the position of sperms in each key frame, and calculating a sample motion value of the sperm;
the morphological value calculation module is used for the same sperm, the sperm has a sperm image in each key frame, extracting the sperm image in each key frame, carrying out convolution and pooling operation on the sperm images to obtain a plurality of sperm images with the same size, forming an image sequence of the sperm by the sperm images with the same size, and inputting the image sequence of the sperm into the neural network model to obtain a sperm morphological score; taking the average value of the morphological scores of all sperms in the sample as a morphological value;
and the vitality assessment module is used for taking the sample movement value, the morphological value and other characteristic values of the sperm as the input of the deep neural network to obtain a sperm vitality value.
Preferably, the selecting a plurality of video frames as key frames in the dynamic video, identifying the position of the sperm in each key frame, and calculating the motion value of the sample, specifically:
Step 11, selecting a segment from dynamic video, determining the length of the segment, identifying sperms in an image of the middle moment of the segment, dividing a moving area for each sperm according to the length and the maximum swimming speed of the sperm, judging whether the moving area of the sperm is overlapped with other sperms, if not, executing S12, otherwise, executing S13;
step 12, calculating the gray level change of the same active area in two adjacent frames of images in the segment, determining a key frame according to the gray level change and the size of the active area, identifying the position of a sperm in the key frame, and fitting a sperm motion curve according to the position by adopting a least square method to obtain a sperm motion track;
step 13, dividing the overlapped active areas into at least one group, wherein the active areas in the group are overlapped with at least one other active area in the group, obtaining an area to be identified according to the position relation of the active areas in the group, determining an interval according to the distance of sperms in the area to be identified, extracting frame images to be identified from the segments according to the interval, determining the area corresponding to each group in the frame images to be identified, identifying the position of sperms in the area corresponding to each group, and fitting the movement curve of the sperms by adopting a least square method according to the position to obtain the movement track of the sperms;
S14, determining the movement value of the single sperm according to the position, the movement track and the length of the sperm, and taking the average value of the movement values of all sperm in the sample as the sample movement value.
Preferably, the method comprises the steps of dividing an activity area for each sperm according to the length and the maximum swimming speed of the sperm, specifically:
determining a radius according to the length and the maximum swimming speed of the sperms, obtaining two circle centers according to a target frame for identifying the sperms, and further determining two circles, wherein the radius of the two circles is determined according to the length and the maximum swimming speed of the sperms; and obtaining a rectangular area, wherein the rectangular area is a rectangle with the smallest area and containing the two circles, and the rectangular area is used as the sperm movement area.
Preferably, the two circle centers are obtained according to the target frame for identifying the sperm, specifically:
obtaining a target frame for identifying sperms, obtaining two widths of the target frame, and taking the center of a line segment where each width is located as a circle center.
Preferably, the calculating the gray scale change of the same active area in two adjacent frames of images in the segment determines a key frame according to the gray scale change and the size of the active area, specifically:
According to the formulaCalculating gray level change of the same active region in two adjacent frames of images, wherein N represents the number of pixel points in the same region, N is a positive integer,/or->Gray value representing j-th pixel of i-th frame,>gray value of j pixel point of i+1th frame,/and>representing the gray scale change of the same active area of the (i+1) th frame relative to the (i) th frame;
taking the segmented first frame as a first key frame, wherein the gray level change of the same active area of the adjacent frame between the next key frame and the last key frame is increased more than a threshold value, and the threshold value is proportional to the size of the active area; and the last frame of the segment is taken as the last key frame.
Preferably, the area to be identified is obtained according to the position relation of the active areas in the group, specifically:
determining a rectangular frame, wherein the rectangular frame comprises all active areas in the group, and the area of the rectangular frame is the smallest; and taking the rectangular frame as an area to be identified.
Preferably, the interval is determined according to the distance between the sperms in the area to be identified, and the frame image to be identified is extracted from the segment according to the interval, specifically:
calculating the average distance of sperms in the area to be identified, determining an interval according to a preset comparison relation table of the average distance and the interval, and extracting a frame image to be identified from the segment according to the determined interval on the basis of the first frame of the segment;
Or alternatively, the process may be performed,
calculating the shortest distance of sperms in the area to be identified, calculating the ratio of the shortest distance to the standard distance, taking the integer part of the product of the ratio and the standard interval as an interval, and extracting the frame image to be identified from the section according to the determined interval based on the first frame of the section.
Preferably, the determining the motion value of the single sperm according to the position, the motion track and the length of the sperm specifically comprises:
determining a moving distance according to the movement track of the sperm, and calculating the ratio of the moving distance to the length to obtain a first parameter;
fitting according to the position of the sperm to obtain a straight line, calculating according to the straight line and the position to obtain an r-square coefficient, and taking the r-square coefficient as a second parameter;
normalizing the first parameter to between 0 and 1 according to the formulaCalculating to obtain sperm movement value, wherein +.>、/>Represents the weight, and->,/>Representing the normalized first parameter, ++>Representing a second parameter.
Preferably, the image sequence of the sperm is input into a neural network model, specifically:
extracting characteristics of each image of the sperm to obtain a sequence, wherein one sperm image corresponds to one sequence, the sequence is used as input of a neural network model, and the number of input parameters of the neural network model is the same as the number of the sequences.
Finally, the present invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method as described above.
Aiming at the problems that in the prior art, sperm motility depends on the experience of medical staff to cause great subjectivity and sperm motility and morphology are difficult to quantify, the invention obtains sperm morphology scores according to a sample motility value calculated by the motility value of the sperm and by utilizing a neural network, takes the sample motility value, the morphology value and other characteristic values of the sperm as the input of a deep neural network to obtain sperm motility values; in addition, the invention also considers the relation between the sperm movement track and the straight line when calculating the sperm movement value, thereby achieving the effects of quantifying the sperm movement and comprehensively calculating the sperm motility by combining other information, and reducing the influence of the subjectivity of traditional Chinese medical personnel on the result in sperm motility evaluation.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a first embodiment
FIG. 2 is a flow chart of one embodiment of the first embodiment;
FIG. 3 is a schematic view of the center and radius of a sperm motility zone;
fig. 4 is a schematic view of a sperm motility zone.
Detailed Description
In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
The invention provides a sperm motility assessment method, as shown in figure 1, comprising the following steps:
step 1, acquiring a dynamic video of a shot semen sample, selecting a plurality of video frames as key frames in the dynamic video, identifying the position of sperms in each key frame, and calculating a sample motion value of a sperm motion value;
in the evaluation of sperm motility in semen, a dynamic video of a semen sample needs to be acquired, the dynamic video can be shot by a camera on a microscope, and the movement condition of the sperm and the morphology of the sperm can be observed by the dynamic video. Because the frame rate of the video is higher, namely, the video contains a plurality of frames of video images in one second, if each frame is processed, the calculated amount is larger, the moving speed of sperms is slower, and the change of sperms in two adjacent frames is not large, based on the method, a plurality of video frames are selected from the dynamic video as key frames. In one embodiment, key frames are determined by selecting a video frame at intervals of a certain number of frames. The above method is too fixed and the different sperm activation conditions are different, in another embodiment, a segment of the video is selected, and then a key frame is determined according to the sperm activation region, which will be described in detail below.
It should be noted that, the determination of the key frame is two ways, wherein the first way is to determine the key frame by a certain number of frames, that is, two key frames are separated by a certain number of video frames, in this way, in the step 1 and the step 2, for different sperms, the determination of the position of the sperm and the extraction of the sperm image are both in the same key frame; in addition, there is a way in which the keyframes of the different sperm are different, and in this way, for the different sperm in step 1 and step 2, the position of the sperm and the extracted sperm image are both determined to be in the keyframe corresponding to the sperm.
Step 2, for the same sperm, a sperm image is arranged in each key frame, the sperm image in each key frame is extracted, convolution and pooling operations are carried out on the sperm images to obtain a plurality of sperm images with the same size, the sperm images with the same size form an image sequence of the sperm, and the image sequence of the sperm is input into a neural network model to obtain a sperm morphology score; taking the average value of the morphological scores of all sperms in the sample as a morphological value;
the morphological information of the sperms is mainly judged according to the appearance of the sperms, the swimming postures of the sperms at different moments are different, sperm images are extracted from key frames, and as the number of the key frames is multiple, one sperm has multiple sperm images, the multiple sperm images are input into a trained neural network model, and the grading of the sperms is obtained. The neural network model preferably uses a convolutional neural network model, and a ViT network model can also be used. The sperm images in each key frame are preferably feature extracted using a ViT network model to obtain a sequence, which is then input into a ViT network model.
And step 3, taking the sample motion value, the morphological value and other characteristic values of the sperm as the input of a deep neural network to obtain a sperm motility value.
The motility value, the morphology value and other characteristic values of the sperm are input into a deep neural network, so that the motility value of the sperm or the semen can be obtained. Wherein the other characteristic value of the sperm is one or more of pH value, density and sperm volume, and of course, the other characteristic value can be other characteristic values besides pH value, density and sperm volume. In an alternative, the deep neural network is replaced with an MLP.
Preferably, the selecting a plurality of video frames as key frames in the dynamic video, identifying the position of the sperm in each key frame, and calculating the motion value of the sperm, as shown in fig. 2, specifically:
step 11, selecting a segment from dynamic video, determining the length of the segment, identifying sperms in an image of the middle moment of the segment, dividing a moving area for each sperm according to the length and the maximum swimming speed of the sperm, judging whether the moving area of the sperm is overlapped with other sperms, if not, executing S12, otherwise, executing S13;
After the sperm sample is produced, the sperm can be photographed by a camera of a computer connected with a microscope, and the photographing time can be determined by a user. After shooting, a user can select one or more sections of clear sperm videos, the morphology of the sperm can be observed, the movement condition of the sperm can be observed in the sperm videos, and according to the standard of world health organization, about 60% of the sperm with normal motility are in a movement state, and the sperm with good state move forwards along a straight line.
The method comprises the steps of selecting a middle time image as a basis after obtaining a segmented segment, and determining a swimming area or a moving area of the sperm according to the length and the maximum swimming speed of the sperm. The length described herein may be either a time length or a frame length.
After the activity area of each sperm is obtained, judging whether the activity area of one sperm overlaps with the activity areas of other sperm, if not, only extracting the area of each image frame of the sperm in the segment, and then analyzing the activity area of the sperm; if there is overlap in this way, it is necessary to eliminate the influence of overlap on sperm trajectory analysis.
In the present invention, each sperm corresponds to an active region, and in each frame of the segmented image, the active region is unchanged, and after determining the active region of the sperm, the same region may be directly truncated from the other frames. This avoids repeated region identification for each frame, e.g., (0, 0), (0, 5), (5, 0), (5, 5) for one sperm, so that the same region for other frames can be directly truncated.
Step 12, calculating the gray level change of the same active area in two adjacent frames of images in the segment, determining a key frame according to the gray level change and the size of the active area, identifying the position of a sperm in the key frame, and fitting a sperm motion curve according to the position by adopting a least square method to obtain a sperm motion track; wherein the key frames of different sperm are different;
if the operation of identifying the sperm position is performed for each frame, the calculated amount is relatively large, and still taking 3s and 50fps as examples, 150 pictures need to be processed, and the operation is performed on only one sperm, if 150 sperm exist, the data amount is larger, and on the basis of the calculated amount, the key frames in the segments are determined according to the gray level change and the size of the active area for the sperm without overlapping the active area. The gray level change refers to the gray level change of the active area, the active areas of different sperms are different, and the corresponding gray level changes are also different. The key frames for each sperm are also different, e.g., 1, 51, 101 for sperm 1 and 1, 41, 81, 121 for sperm 2, based on gray scale change and size of the active area. And then determining the position of the sperm in the key frame according to the activity area, and after identifying the track of the activity area of the sperm in all the key frames of the sperm, fitting the motion curve of the sperm by adopting a least square method to obtain the motion track of the sperm. In an alternative embodiment, the fitting of the sperm motility curve uses polynomial fitting or gradient descent.
Taking the above sperm 1 and sperm 2 as examples, in the key frames 1, 51 and 101 of the sperm 1, the positions of the sperm are (2.5), (2.5,3) and (2.5 and 4), and the motion track of the sperm 1 is obtained by fitting. Likewise, the motion profile of sperm 2 may be fitted based on the position of sperm 2 at its corresponding keyframe.
There are various ways to identify the location of sperm, including but not limited to YOLOv4, YOLOv5, or R-CNN, and conventional image analysis methods can be used, as the invention is not limited in detail. Sperm has a certain length, and the position of the sperm head is used as the position of the sperm in the process of carrying out the position identification and the motion track fitting of the sperm unless otherwise specified in the invention.
Step 13, dividing the overlapped active areas into at least one group, wherein the active areas in the group are overlapped with at least one other active area in the group, obtaining an area to be identified according to the position relation of the active areas in the group, determining an interval according to the distance of sperms in the area to be identified, extracting frame images to be identified from the segments according to the interval, determining the area corresponding to each group in the frame images to be identified, identifying the position of sperms in the area corresponding to each group, and fitting the movement curve of the sperms by adopting a least square method according to the position to obtain the movement track of the sperms;
If the active areas of different sperms are overlapped, the positions to be identified are more accurate, so that the mutual influence of other sperms in the overlapped areas can be prevented.
One group corresponds to one region to be identified and one region to be identified corresponds to one interval, for example, sperm 3, 4, 5 are located in group 1, and the region to be identified can be obtained according to the activity region of sperm 3, 4, 5, and likewise, the positions of all frames in the segment of the region to be identified are the same. The presence of an active region in a group overlapping at least one other active region in the group means that each active region in the group overlaps at least one of the remaining active regions, e.g., sperm 3, 4, 5 are in group 1, and then the active region of sperm 3 overlaps the active region of at least one of sperm 4 and 5.
And then, recognizing the positions of all sperms in the group, and fitting the track of each sperm according to the position of each sperm to obtain the motion track of each sperm in the group. In an alternative embodiment, the fitting of the sperm motility curve uses polynomial fitting or gradient descent.
In step S3, a region to be identified has a plurality of sperms, and the positions of the plurality of sperms are identified by adopting an example segmentation mode, or Mask R-CNN, yolact++, SOLOv2, or the like, which is not particularly limited in the present invention.
S14, determining the movement value of the single sperm according to the position, the movement track and the length of the sperm, and taking the average value of the movement values of all sperm in the sample as the sample movement value.
The movement speed of the sperms can be obtained through the movement track of the sperms, in addition, the movement track of the sperms reflects the movement direction of the sperms, most of sperms with better quality move along one direction, the movement direction of the sperms can be determined through the positions of the sperms, and if the movement track of the sperms is relatively close to a straight line, the quality of the sperms is higher.
The maximum motility area of sperm in a segment is related to the motility speed of sperm and the segment duration, while sperm also has a length, in one embodiment, the method comprises dividing a motility area for each sperm according to the length and the maximum motility speed of sperm, specifically:
Determining a radius according to the length and the maximum swimming speed of the sperms, obtaining two circle centers according to a target frame for identifying the sperms, and further determining two circles, wherein the radius of the two circles is determined according to the length and the maximum swimming speed of the sperms; and obtaining a rectangular area, wherein the rectangular area is a rectangle with the smallest area and containing the two circles, and the rectangular area is used as the sperm movement area.
If the length is time, the radius can be obtained by the product of the length and the maximum swimming speed, if the time is frame number, the time is calculated according to the frame number, and then the radius can be obtained by the product of the calculated time and the maximum swimming speed. The maximum swimming speed of the sperms is set in advance, and the normal speed of the sperms is 25 microns/second according to the standard of world health organization.
The circle center can be determined by identifying the sperm, and in a simple embodiment, the two circle centers are obtained according to the target frame for identifying the sperm, specifically:
obtaining a target frame for identifying sperms, obtaining two widths of the target frame, and taking the center of a line segment where each width is located as a circle center. Figure 3 shows two circle centers and radius of one sperm and figure 4 shows sperm motility zone.
When the sperm moves in the active area, the gray level of the active area is changed, and the activity condition of the sperm can be obtained according to the gray level change, in a specific embodiment, the gray level change of the same active area in two adjacent frames of images in each segment is calculated, and a key frame is determined according to the gray level change and the size of the active area, specifically:
according to the formulaCalculating gray level change of the same active region in two adjacent frames of images, wherein N represents the number of pixel points in the same region, N is a positive integer,/or->Gray value representing j-th pixel of i-th frame,>gray value of j pixel point of i+1th frame,/and>representing the gray scale change of the same active area of the (i+1) th frame relative to the (i) th frame;
taking the first frame of each segment as a first key frame, wherein the gray level change of the same active area of the adjacent frame between the next key frame and the last key frame is increased more than a threshold value, and the threshold value is in direct proportion to the size of the active area; and the last frame of each segment is taken as the last key frame.
In the above process, the active area of each frame is fixed in the image, but the change of gray scale of the active area is caused by the movement of sperm, and the sperm movement can be determined by the change of the active areas of two adjacent frames.
For a sperm, the active area has N pixel points, namely, the active area in each frame in the segment has N pixel points, then the change of the same pixel point in two adjacent frames is calculated, and the change of the pixel values of the N pixel points is accumulated to obtain the gray level change of the active area.
After the gray level change of the same active area of two adjacent frames is calculated, the change condition of the same active area of a plurality of adjacent frames is accumulated to obtain an accumulated result, if the accumulated result is larger than a threshold value, the movement of sperms reaches a certain degree, and at the moment, the sperms are identified. By the method, the data quantity for identifying the image can be effectively reduced, and the identification speed is improved.
FIG. 3 shows a schematic view of the area to be identified, i.e. the active area of a sperm, wherein the identification process is similar when there are a plurality of sperm, in particular a rectangular box is determined, said rectangular box comprising all active areas of the group, and the area of the rectangular box is minimal; and taking the rectangular frame as an area to be identified.
The denser the sperm is, the more the images need to be identified are, so that the motion track of the sperm in a dense state can be better obtained, in one embodiment, the interval is determined according to the distance of the sperm in the area to be identified, and the frame image to be identified is extracted from the segment according to the interval, specifically:
Calculating the average distance of sperms in the area to be identified, determining an interval according to a preset comparison relation table of the average distance and the interval, and extracting a frame image to be identified from the segment according to the determined interval on the basis of the first frame of the segment;
or alternatively, the process may be performed,
calculating the shortest distance of sperms in the area to be identified, calculating the ratio of the shortest distance to the standard distance, taking the integer part of the product of the ratio and the standard interval as an interval, and extracting the frame image to be identified from the section according to the determined interval based on the first frame of the section.
The motility of the sperm is not only related to the movement speed, but also whether it is a near straight line, in a specific embodiment, the motility of the sperm is determined according to the position, the movement track and the length of the sperm, specifically:
determining a moving distance according to the movement track of the sperm, and calculating the ratio of the moving distance to the length to obtain a first parameter;
fitting according to the position of the sperm to obtain a straight line, calculating according to the straight line and the position to obtain an r-square coefficient, and taking the r-square coefficient as a second parameter;
normalizing the first parameter to between 0 and 1 according to the formula Calculating to obtain sperm motility, wherein ∈>、/>Represents the weight, and->,/>Representing the first parameter after normalization,representing a second parameter.
The first parameter is the moving speed of the sperm, and the second parameter represents the approaching degree of the sperm movement to the straight line. The R-square coefficient, also known as the determinant coefficient, is 1 when the sperm positions all fall on the straight line, and 0 if there is no relation to the straight line.
In another embodiment, the r-square coefficient may be replaced with an average distance from the sperm location to the line, the average distance being used as the second parameter.
For one sperm, when there are multiple segments, the average of the sperm's motility in all segments is taken as the final sperm's motility. If the motility condition of one semen sample is needed to be obtained according to the motility of a plurality of sperms, the average value of the motility of all sperms according to the segmentation is used as the movement value of the final semen sample.
Preferably, the image sequence of the sperm is input into a neural network model, specifically:
extracting characteristics of each image of the sperm to obtain a sequence, wherein one sperm image corresponds to one sequence, the sequence is used as input of a neural network model, and the number of input parameters of the neural network model is the same as the number of the sequences. For example, there are 5 key frames, and each sperm has 5 sperm images, and the 5 sperm images form an image sequence of the sperm, and each image is subjected to feature extraction to obtain a sequence, and the 5 sequences are input into the neural network model.
In an alternative embodiment, the features of each image of sperm are extracted, one sperm image corresponding to each image feature, and the image features of the plurality of images corresponding to sperm are used as inputs to the neural network. For example, there are 5 key frames, and each sperm has 5 sperm images, the 5 sperm images form an image sequence of the sperm, each image is subjected to feature extraction to obtain a feature map, and the 5 feature maps are input into the neural network model.
Example two
The invention also provides a sperm motility assessment system, which comprises the following modules:
the motion value calculation module is used for acquiring a dynamic video of the shot semen sample, selecting a plurality of video frames from the dynamic video as key frames, identifying the position of sperms in each key frame, and calculating a sample motion value of the sperm;
the morphological value calculation module is used for the same sperm, the sperm has a sperm image in each key frame, extracting the sperm image in each key frame, carrying out convolution and pooling operation on the sperm images to obtain a plurality of sperm images with the same size, forming an image sequence of the sperm by the sperm images with the same size, and inputting the image sequence of the sperm into the neural network model to obtain a sperm morphological score; taking the average value of the morphological scores of all sperms in the sample as a morphological value;
And the activity evaluation module is used for taking the motion value, the morphological value and other characteristic values of the sperm as the input of the deep neural network to obtain a sperm activity value.
Preferably, the selecting a plurality of video frames as key frames in the dynamic video, identifying the position of the sperm in each key frame, and calculating the motion value of the sample, specifically:
step 11, selecting a segment from dynamic video, determining the length of the segment, identifying sperms in an image of the middle moment of the segment, dividing a moving area for each sperm according to the length and the maximum swimming speed of the sperm, judging whether the moving area of the sperm is overlapped with other sperms, if not, executing S12, otherwise, executing S13;
step 12, calculating the gray level change of the same active area in two adjacent frames of images in the segment, determining a key frame according to the gray level change and the size of the active area, identifying the position of a sperm in the key frame, and fitting a sperm motion curve according to the position by adopting a least square method to obtain a sperm motion track; wherein the key frames of different sperm are different;
step 13, dividing the overlapped active areas into at least one group, wherein the active areas in the group are overlapped with at least one other active area in the group, obtaining an area to be identified according to the position relation of the active areas in the group, determining an interval according to the distance of sperms in the area to be identified, extracting frame images to be identified from the segments according to the interval, determining the area corresponding to each group in the frame images to be identified, identifying the position of sperms in the area corresponding to each group, and fitting the movement curve of the sperms by adopting a least square method according to the position to obtain the movement track of the sperms;
S14, determining the movement value of the single sperm according to the position, the movement track and the length of the sperm, and taking the average value of the movement values of all sperm in the sample as the sample movement value.
Preferably, the method comprises the steps of dividing an activity area for each sperm according to the length and the maximum swimming speed of the sperm, specifically:
determining a radius according to the length and the maximum swimming speed of the sperms, obtaining two circle centers according to a target frame for identifying the sperms, and further determining two circles, wherein the radius of the two circles is determined according to the length and the maximum swimming speed of the sperms; and obtaining a rectangular area, wherein the rectangular area is a rectangle with the smallest area and containing the two circles, and the rectangular area is used as the sperm movement area.
Preferably, the two circle centers are obtained according to the target frame for identifying the sperm, specifically:
obtaining a target frame for identifying sperms, obtaining two widths of the target frame, and taking the center of a line segment where each width is located as a circle center.
Preferably, the calculating the gray scale change of the same active area in two adjacent frames of images in the segment determines a key frame according to the gray scale change and the size of the active area, specifically:
According to the formulaCalculating gray level change of the same active region in two adjacent frames of images, wherein N represents the number of pixel points in the same region, N is a positive integer,/or->Gray value representing j-th pixel of i-th frame,>gray value of j pixel point of i+1th frame,/and>representing the gray scale change of the same active area of the (i+1) th frame relative to the (i) th frame;
taking the segmented first frame as a first key frame, wherein the gray level change of the same active area of the adjacent frame between the next key frame and the last key frame is increased more than a threshold value, and the threshold value is proportional to the size of the active area; and the last frame of the segment is taken as the last key frame.
Preferably, the area to be identified is obtained according to the position relation of the active areas in the group, specifically:
determining a rectangular frame, wherein the rectangular frame comprises all active areas in the group, and the area of the rectangular frame is the smallest; and taking the rectangular frame as an area to be identified.
Preferably, the interval is determined according to the distance between the sperms in the area to be identified, and the frame image to be identified is extracted from the segment according to the interval, specifically:
calculating the average distance of sperms in the area to be identified, determining an interval according to a preset comparison relation table of the average distance and the interval, and extracting a frame image to be identified from the segment according to the determined interval on the basis of the first frame of the segment;
Or alternatively, the process may be performed,
calculating the shortest distance of sperms in the area to be identified, calculating the ratio of the shortest distance to the standard distance, taking the integer part of the product of the ratio and the standard interval as an interval, and extracting the frame image to be identified from the section according to the determined interval based on the first frame of the section.
Preferably, the determining the motion value of the single sperm according to the position, the motion track and the length of the sperm specifically comprises:
determining a moving distance according to the movement track of the sperm, and calculating the ratio of the moving distance to the length to obtain a first parameter;
fitting according to the position of the sperm to obtain a straight line, calculating according to the straight line and the position to obtain an r-square coefficient, and taking the r-square coefficient as a second parameter;
normalizing the first parameter to between 0 and 1 according to the formulaCalculating to obtain sperm movement value, wherein +.>、/>Represents the weight, and->,/>Representing the normalized first parameter, ++>Representing a second parameter.
Preferably, the sperm image features are extracted and input into a neural network model, specifically:
extracting the characteristics of each sperm image to obtain a sequence, wherein one sperm image corresponds to one sequence, the sequence is used as the input of a neural network model, and the number of input parameters of the neural network model is the same as the number of sequences.
Example III
The present invention provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a method as described in the first embodiment.
Example IV
The invention also provides a computer device comprising a processor, a memory, the memory having stored thereon a computer program which, when executed by the processor, implements a method as described in the first embodiment.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by adding necessary general purpose hardware platforms, or may be implemented by a combination of hardware and software. Based on such understanding, the foregoing aspects, in essence and portions contributing to the art, may be embodied in the form of a computer program product, which may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A neural network-based sperm motility determining method, comprising the steps of:
step 1, acquiring a dynamic video of a shot semen sample, selecting a plurality of video frames as key frames in the dynamic video, identifying the position of sperms in each key frame, and calculating a sample motion value of a sperm motion value;
step 2, for the same sperm, a sperm image is arranged in each key frame, the sperm image in each key frame is extracted, convolution and pooling operations are carried out on the sperm images to obtain a plurality of sperm images with the same size, the sperm images with the same size form an image sequence of the sperm, and the image sequence of the sperm is input into a neural network model to obtain a sperm morphology score; taking the average value of the morphological scores of all sperms in the sample as a morphological value;
And step 3, taking the sample motion value, the morphological value and other characteristic values of the sperm as the input of a deep neural network to obtain a sperm motility value.
2. The method of claim 1, wherein the selecting a plurality of video frames in the dynamic video as key frames, identifying the location of sperm in each key frame, and calculating a sample motion value of sperm, in particular:
step 11, selecting a segment from dynamic video, determining the length of the segment, identifying sperms in an image of the middle moment of the segment, dividing a moving area for each sperm according to the length and the maximum swimming speed of the sperm, judging whether the moving area of the sperm is overlapped with other sperms, if not, executing S12, otherwise, executing S13;
step 12, calculating the gray level change of the same active area in two adjacent frames of images in the segment, determining a key frame according to the gray level change and the size of the active area, identifying the position of a sperm in the key frame, and fitting a sperm motion curve according to the position by adopting a least square method to obtain a sperm motion track; wherein the key frames of different sperm are different;
Step 13, dividing the overlapped active areas into at least one group, wherein the active areas in the group are overlapped with at least one other active area in the group, obtaining an area to be identified according to the position relation of the active areas in the group, determining an interval according to the distance of sperms in the area to be identified, extracting frame images to be identified from the segments according to the interval, determining the area corresponding to each group in the frame images to be identified, identifying the position of sperms in the area corresponding to each group, and fitting the movement curve of the sperms by adopting a least square method according to the position to obtain the movement track of the sperms;
s14, determining the movement value of the single sperm according to the position, the movement track and the length of the sperm, and taking the average value of the movement values of all sperm in the sample as the sample movement value.
3. The method according to claim 2, wherein said dividing an activity area for each sperm based on said length and maximum swimming speed of the sperm, in particular:
determining a radius according to the length and the maximum swimming speed of the sperms, obtaining two circle centers according to a target frame for identifying the sperms, and further determining two circles, wherein the radius of the two circles is determined according to the length and the maximum swimming speed of the sperms; and obtaining a rectangular area, wherein the rectangular area is a rectangle with the smallest area and containing the two circles, and the rectangular area is used as the sperm movement area.
4. A method according to claim 3, wherein the two centers are obtained according to the target frame for identifying sperm, specifically:
obtaining a target frame for identifying sperms, obtaining two widths of the target frame, and taking the center of a line segment where each width is located as a circle center.
5. The method according to claim 2, wherein the calculating the gray scale change of the same active area in two adjacent frames of images in the segment determines a key frame according to the gray scale change and the size of the active area, specifically:
according to the formulaCalculating gray level change of the same active region in two adjacent frames of images, wherein N represents the number of pixel points in the same region, N is a positive integer,/or->Represents the gray value of the j-th pixel point of the i-th frame,gray value of j pixel point of i+1th frame,/and>representing the gray scale change of the same active area of the (i+1) th frame relative to the (i) th frame;
taking the segmented first frame as a first key frame, wherein the gray level change of the same active area of the adjacent frame between the next key frame and the last key frame is increased more than a threshold value, and the threshold value is proportional to the size of the active area; and the last frame of the segment is taken as the last key frame.
6. The method according to claim 2, wherein the obtaining the area to be identified according to the position relation of the active areas in the group is specifically:
determining a rectangular frame, wherein the rectangular frame comprises all active areas in the group, and the area of the rectangular frame is the smallest; and taking the rectangular frame as an area to be identified.
7. The method according to claim 2, wherein the interval is determined according to the distance of the sperm in the area to be identified, and the frame image to be identified is extracted from the segment according to the interval, in particular:
calculating the average distance of sperms in the area to be identified, determining an interval according to a preset comparison relation table of the average distance and the interval, and extracting a frame image to be identified from the segment according to the determined interval on the basis of the first frame of the segment;
or alternatively, the process may be performed,
calculating the shortest distance of sperms in the area to be identified, calculating the ratio of the shortest distance to the standard distance, taking the integer part of the product of the ratio and the standard interval as an interval, and extracting the frame image to be identified from the section according to the determined interval based on the first frame of the section.
8. The method according to claim 2, wherein the determining of the individual sperm movement value is based on the sperm position, movement trajectory and length, in particular:
Determining a moving distance according to the movement track of the sperm, and calculating the ratio of the moving distance to the length to obtain a first parameter;
fitting according to the position of the sperm to obtain a straight line, calculating according to the straight line and the position to obtain an r-square coefficient, and taking the r-square coefficient as a second parameter;
normalizing the first parameter to between 0 and 1 according to the formulaCalculating to obtain sperm movement value, wherein +.>、/>Represents the weight, and->,/>Representing the first parameter after normalization,representing a second parameter.
9. The method according to claim 1, wherein the inputting of the image sequence of sperm into the neural network model is in particular:
extracting characteristics of each image of the sperm to obtain a sequence, wherein one sperm image corresponds to one sequence, the sequence is used as input of a neural network model, and the number of input parameters of the neural network model is the same as the number of the sequences.
10. A neural network-based sperm motility determination system, the system comprising the following modules:
the motion value calculation module is used for acquiring a dynamic video of the shot semen sample, selecting a plurality of video frames from the dynamic video as key frames, identifying the position of sperms in each key frame, and calculating a sample motion value of the sperm;
The morphological value calculation module is used for the same sperm, the sperm has a sperm image in each key frame, extracting the sperm image in each key frame, carrying out convolution and pooling operation on the sperm images to obtain a plurality of sperm images with the same size, forming an image sequence of the sperm by the sperm images with the same size, and inputting the image sequence of the sperm into the neural network model to obtain a sperm morphological score; taking the average value of the morphological scores of all sperms in the sample as a morphological value;
and the vitality assessment module is used for taking the sample movement value, the morphological value and other characteristic values of the sperm as the input of the deep neural network to obtain a sperm vitality value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311136028.3A CN116863388A (en) | 2023-09-05 | 2023-09-05 | Sperm motility determining method and system based on neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311136028.3A CN116863388A (en) | 2023-09-05 | 2023-09-05 | Sperm motility determining method and system based on neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116863388A true CN116863388A (en) | 2023-10-10 |
Family
ID=88232657
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311136028.3A Pending CN116863388A (en) | 2023-09-05 | 2023-09-05 | Sperm motility determining method and system based on neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116863388A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117114749A (en) * | 2023-10-16 | 2023-11-24 | 吉林省农业科学院(中国农业科技东北创新中心) | Intelligent pig frozen semen management method and system |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20110049606A (en) * | 2009-11-05 | 2011-05-12 | 주식회사 메디칼써프라이 | A method for analyzing the morphology and motility of sperm using histogram |
CN104237121A (en) * | 2014-09-18 | 2014-12-24 | 山东大学 | Sperm motility performance detection method |
CN108875931A (en) * | 2017-12-06 | 2018-11-23 | 北京旷视科技有限公司 | Neural metwork training and image processing method, device, system |
US20190197294A1 (en) * | 2017-12-22 | 2019-06-27 | Hillel Llc | Imaging device for measuring sperm motility |
CN111080624A (en) * | 2019-12-17 | 2020-04-28 | 北京推想科技有限公司 | Sperm movement state classification method, device, medium and electronic equipment |
CN112150415A (en) * | 2020-09-04 | 2020-12-29 | 清华大学 | Multi-target sperm real-time monitoring method based on deep learning |
CN113706481A (en) * | 2021-08-13 | 2021-11-26 | 苏州贝康智能制造有限公司 | Sperm quality detection method, sperm quality detection device, computer equipment and storage medium |
CN113780145A (en) * | 2021-09-06 | 2021-12-10 | 苏州贝康智能制造有限公司 | Sperm morphology detection method, sperm morphology detection device, computer equipment and storage medium |
CN114241006A (en) * | 2021-12-20 | 2022-03-25 | 电子科技大学长三角研究院(衢州) | Sperm tracking and automatic extraction method based on machine learning |
US20220156922A1 (en) * | 2020-11-17 | 2022-05-19 | Alejandro Chávez Badiola | System for real-time automatic quantitative evaluation, assessment and/or ranking of individual sperm, aimed for intracytoplasmic sperm injection (icsi), and other fertilization procedures, allowing the selection of a single sperm |
CN114518362A (en) * | 2022-02-24 | 2022-05-20 | 上海北昂医药科技股份有限公司 | Sperm quality analysis device, system, method and readable storage medium |
-
2023
- 2023-09-05 CN CN202311136028.3A patent/CN116863388A/en active Pending
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20110049606A (en) * | 2009-11-05 | 2011-05-12 | 주식회사 메디칼써프라이 | A method for analyzing the morphology and motility of sperm using histogram |
CN104237121A (en) * | 2014-09-18 | 2014-12-24 | 山东大学 | Sperm motility performance detection method |
CN108875931A (en) * | 2017-12-06 | 2018-11-23 | 北京旷视科技有限公司 | Neural metwork training and image processing method, device, system |
US20190197294A1 (en) * | 2017-12-22 | 2019-06-27 | Hillel Llc | Imaging device for measuring sperm motility |
CN111080624A (en) * | 2019-12-17 | 2020-04-28 | 北京推想科技有限公司 | Sperm movement state classification method, device, medium and electronic equipment |
CN112150415A (en) * | 2020-09-04 | 2020-12-29 | 清华大学 | Multi-target sperm real-time monitoring method based on deep learning |
US20220156922A1 (en) * | 2020-11-17 | 2022-05-19 | Alejandro Chávez Badiola | System for real-time automatic quantitative evaluation, assessment and/or ranking of individual sperm, aimed for intracytoplasmic sperm injection (icsi), and other fertilization procedures, allowing the selection of a single sperm |
CN113706481A (en) * | 2021-08-13 | 2021-11-26 | 苏州贝康智能制造有限公司 | Sperm quality detection method, sperm quality detection device, computer equipment and storage medium |
CN113780145A (en) * | 2021-09-06 | 2021-12-10 | 苏州贝康智能制造有限公司 | Sperm morphology detection method, sperm morphology detection device, computer equipment and storage medium |
CN114241006A (en) * | 2021-12-20 | 2022-03-25 | 电子科技大学长三角研究院(衢州) | Sperm tracking and automatic extraction method based on machine learning |
CN114518362A (en) * | 2022-02-24 | 2022-05-20 | 上海北昂医药科技股份有限公司 | Sperm quality analysis device, system, method and readable storage medium |
Non-Patent Citations (4)
Title |
---|
MECIT YÜZKAT等: "Morphological Classification of Low Quality Sperm Images Using Deep Learning Networks", 2020 MEDICAL TECHNOLOGIES CONGRESS (TIPTEKNO) * |
孙丽;黄桂;龚湘君;欧建平;: "基于数字全息显微镜的人类精子的三维动态行为监测技术", 中国临床解剖学杂志, no. 04 * |
庞增拴: "精子运动能力检测系统中图像分割方法的选取与算法实现", 河北省科学院学报, no. 03 * |
张时民;: "医学检验领域人工智能技术应用与展望", 国际检验医学杂志, no. 05 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117114749A (en) * | 2023-10-16 | 2023-11-24 | 吉林省农业科学院(中国农业科技东北创新中心) | Intelligent pig frozen semen management method and system |
CN117114749B (en) * | 2023-10-16 | 2024-01-12 | 吉林省农业科学院(中国农业科技东北创新中心) | Intelligent pig frozen semen management method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111178197B (en) | Mass R-CNN and Soft-NMS fusion based group-fed adherent pig example segmentation method | |
WO2021139258A1 (en) | Image recognition based cell recognition and counting method and apparatus, and computer device | |
US20170109879A1 (en) | Computer-implemented methods, computer-readable media, and systems for tracking a plurality of spermatozoa | |
CN110245657B (en) | Pathological image similarity detection method and detection device | |
CN109145708B (en) | Pedestrian flow statistical method based on RGB and D information fusion | |
CN109544512A (en) | It is a kind of based on multi-modal embryo's pregnancy outcome prediction meanss | |
US20210327051A1 (en) | Method, computer device and storage medium of fundus oculi image analysis | |
CN108052886A (en) | A kind of puccinia striiformis uredospore programming count method of counting | |
CN116863388A (en) | Sperm motility determining method and system based on neural network | |
EP3853763A1 (en) | Image feature detection | |
CN110772286A (en) | System for discernment liver focal lesion based on ultrasonic contrast | |
US20220237789A1 (en) | Weakly supervised multi-task learning for cell detection and segmentation | |
CN113435355A (en) | Multi-target cow identity identification method and system | |
CN110197113A (en) | A kind of method for detecting human face of high-precision anchor point matching strategy | |
CN113393454A (en) | Method and device for segmenting pathological target examples in biopsy tissues | |
CN112528823A (en) | Striped shark movement behavior analysis method and system based on key frame detection and semantic component segmentation | |
Silva-Rodríguez et al. | Predicting the success of blastocyst implantation from morphokinetic parameters estimated through CNNs and sum of absolute differences | |
Aristoteles et al. | Identification of human sperm based on morphology using the you only look once version 4 algorithm | |
CN111612749B (en) | Focus detection method and device based on lung image | |
CN113780145A (en) | Sperm morphology detection method, sperm morphology detection device, computer equipment and storage medium | |
CN108154199A (en) | A kind of quick single class object detection method of high-precision based on deep learning | |
KR20210033902A (en) | Method, apparatus and software program for cervical cancer diagnosis using image analysis of artificial intelligence based technology | |
CN116311234A (en) | Semen quality assessment method based on deep learning | |
CN113627255B (en) | Method, device and equipment for quantitatively analyzing mouse behaviors and readable storage medium | |
CN114947751A (en) | Mobile terminal intelligent tongue diagnosis method based on deep learning |
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 |