A system and method for automated detection of semen quality

Abstract

This invention relates to the field of beauty and skincare, and discloses an automated system and method for detecting semen quality, comprising: a morphological defect rate calculation module for calculating the morphological defect rate corresponding to a target semen sample; a dysfunctional sperm identification module for identifying dysfunctional sperm in the target semen sample; a reproductive function scoring module for evaluating the reproductive function score corresponding to the target semen sample; a dynamic motility potential scoring module for configuring a suitable culture medium for the target semen sample, culturing the target semen sample in the suitable culture medium, recording the sperm viability baseline parameters during the culture process, calculating the sperm survival rate of the target semen sample, and evaluating the dynamic motility potential score corresponding to the target semen sample; and a semen quality analysis module for generating a semen quality analysis report corresponding to the target semen sample. This invention can improve the accuracy of semen quality detection.

Description

Technical Field

[0001] This invention relates to the field of beauty and skincare technology, and in particular to a system and method for automatically detecting semen quality. Background Technology

[0002] In the current field of medicine and reproductive health, semen quality testing plays a crucial role in assessing male fertility, diagnosing reproductive system diseases, and implementing assisted reproductive technologies. Therefore, the requirements for the accuracy, efficiency, and comprehensiveness of semen quality testing are constantly increasing.

[0003] Currently, semen quality testing mainly employs three methods: traditional manual testing, semi-automatic testing, and fully automated testing. Traditional manual testing heavily relies on the professional skills and experience of testing personnel, who need to manually observe sperm morphology, count sperm count, and assess sperm motility under a microscope. While this method can provide relatively detailed results to some extent, the testing process is time-consuming and labor-intensive, and significant subjective differences exist between different testing personnel, resulting in poor accuracy and repeatability of the test results. Semi-automatic testing utilizes some basic automated equipment, such as sperm counters, which improves testing efficiency to some extent and reduces the manual burden. However, it still requires significant manual intervention in the complex analysis of sperm motility and morphological characteristics, as well as the comprehensive evaluation of multi-dimensional data, making it difficult to achieve comprehensive and accurate semen quality testing. Although fully automated testing utilizes advanced image recognition and data analysis algorithms to quickly acquire large amounts of semen data, it has shortcomings in the accurate identification of sperm micromorphology, accurate analysis of complex motility patterns, and in-depth exploration of the uniqueness of individual semen samples, thus leading to low accuracy in semen quality testing. Summary of the Invention

[0004] This invention provides a system and method for automatically detecting semen quality, the main purpose of which is to improve the accuracy of semen quality detection.

[0005] To achieve the above objectives, the present invention provides an automated system for detecting semen quality, comprising: a morphological defect rate calculation module, a dysfunctional sperm identification module, a reproductive function scoring module, a dynamic motility potential scoring module, and a semen quality analysis module;

[0006] The morphological defect rate calculation module is used to acquire the target semen sample to be tested, collect the optical microscopic image of the target semen sample and semen dynamic data, perform image segmentation processing on the optical microscopic image of the sample to obtain a sperm segmentation image, and calculate the morphological defect rate corresponding to the target semen sample based on the sperm segmentation image.

[0007] The dysfunctional sperm identification module is used to extract trajectory dynamic data and oscillation dynamic data from the semen dynamic data, perform clustering processing on the trajectory dynamic data to obtain a clustered motion trajectory set, analyze the trajectory behavior pattern corresponding to the clustered motion trajectory set, and identify dysfunctional sperm in the target semen sample based on the trajectory behavior pattern.

[0008] The reproductive function scoring and evaluation module is used to extract the swing dynamic features from the swing dynamic data, evaluate the sperm motility intensity in the target semen sample based on the swing dynamic features, and evaluate the reproductive function score corresponding to the target semen sample by combining the morphological defect rate, the dysfunctional sperm and the sperm motility intensity.

[0009] The dynamic motility potential scoring module is used to configure the appropriate culture medium for the target semen sample, place the target semen sample in the appropriate culture medium for culture treatment, record the sperm motility baseline parameters during the culture treatment process, calculate the sperm survival rate of the target semen sample based on the sperm survival rate, and evaluate the dynamic motility potential score corresponding to the target semen sample based on the sperm survival rate.

[0010] The semen quality analysis module is used to combine the reproductive function score and the dynamic vitality potential score to generate a semen quality analysis report corresponding to the target semen sample.

[0011] A method for automatically detecting semen quality, characterized in that the method includes:

[0012] A target semen sample to be tested is obtained, and optical microscopic images and semen dynamic data of the target semen sample are collected. The optical microscopic images of the sample are processed to obtain sperm segmentation images. Based on the sperm segmentation images, the morphological defect rate corresponding to the target semen sample is calculated.

[0013] Trajectory dynamic data and oscillation dynamic data are extracted from the semen dynamic data. The trajectory dynamic data is clustered to obtain a set of clustered motion trajectories. The trajectory behavior patterns corresponding to the set of clustered motion trajectories are analyzed. Based on the trajectory behavior patterns, dysfunctional sperm in the target semen sample are identified.

[0014] Extract the swing dynamic features from the swing dynamic data, evaluate the sperm motility intensity in the target semen sample based on the swing dynamic features, and evaluate the reproductive function score corresponding to the target semen sample by combining the morphological defect rate, the dysfunctional sperm and the sperm motility intensity.

[0015] Configure the appropriate culture medium for the target semen sample, place the target semen sample in the appropriate culture medium for culture treatment, and record the sperm motility baseline parameters during the culture treatment process. Based on the motility baseline parameters, calculate the sperm survival rate of the target semen sample, and evaluate the dynamic motility potential score corresponding to the target semen sample based on the sperm survival rate.

[0016] By combining the reproductive function score and the dynamic vitality potential score, a semen quality analysis report corresponding to the target semen sample is generated.

[0017] This invention, through image segmentation processing of the optical microscopic images of the sample, can obtain images that clearly distinguish sperm from other impurities and accurately identify sperm morphology. This not only visually displays sperm morphological details but also helps in accurately calculating sperm morphological defect rates. Optionally, this invention, by clustering the trajectory dynamic data, obtains a set of clustered motion trajectories, which can integrate sperm trajectory information with similar motion characteristics and quickly locate groups of different motion patterns. This invention extracts the oscillation dynamic features from the oscillation dynamic data, which can accurately capture subtle characteristics of sperm movement, providing an intuitive and reliable quantitative basis for evaluating the reproductive function score corresponding to the target semen sample. This invention, by configuring the appropriate culture medium for the target semen sample, facilitates refined and targeted culture of the target semen sample. The target semen sample is placed in the appropriate culture medium for culture treatment, and the sperm viability baseline parameters are recorded during the culture treatment process, thus facilitating subsequent accurate calculation of sperm viability. This invention, by combining the reproductive function score and the dynamic motility potential score, generates a semen quality analysis report corresponding to the target semen sample, resulting in high-quality semen quality analysis results. Therefore, the accuracy of semen quality testing has been improved. Attached Figure Description

[0018] Figure 1 A functional block diagram of an automated semen quality detection system provided in an embodiment of the present invention;

[0019] Figure 2 This is a flowchart illustrating an automated method for detecting semen quality according to an embodiment of the present invention.

[0020] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.

[0023] In practice, the server-side equipment deployed in an automated semen quality detection system may consist of one or more devices. This automated semen quality detection system can be implemented as a business instance, a virtual machine, or hardware devices. For example, the system can be implemented as a business instance deployed on one or more devices in a cloud node. Simply put, this automated semen quality detection system can be understood as software deployed on a cloud node, used to provide automated semen quality detection services to various user terminals. Alternatively, the system can also be implemented as a virtual machine deployed on one or more devices in a cloud node. This virtual machine contains application software for managing various user terminals. Or, the system can also be implemented as a server composed of numerous identical or different types of hardware devices, with one or more hardware devices configured to provide automated semen quality detection services to various user terminals.

[0024] In terms of implementation, the automated semen quality detection system and the user terminal are mutually compatible. That is, if the automated semen quality detection system is implemented as an application installed on a cloud service platform, then the user terminal is implemented as a client that establishes a communication connection with the application; or if the automated semen quality detection system is implemented as a website, then the user terminal is implemented as a webpage; or if the automated semen quality detection system is implemented as a cloud service platform, then the user terminal is implemented as a mini-program in an instant messaging application.

[0025] Reference Figure 1 The diagram shown is a functional block diagram of an automated semen quality detection system provided in an embodiment of the present invention.

[0026] The automated semen quality detection system 100 of this invention can be set up in a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed in the cloud (e.g., a server for automated semen quality detection, a server cluster, etc.), or it can be developed as a website. Depending on the functions implemented, the automated semen quality detection system 100 includes a morphological defect rate calculation module 101, a sperm dysfunction identification module 102, a reproductive function scoring and evaluation module 103, a dynamic motility potential scoring and evaluation module 104, and a semen quality analysis module 105.

[0027] In this embodiment of the invention, in the automated semen quality tracking, each of the above modules can be implemented independently and called upon other modules. This "calling" can be understood as one module connecting to multiple modules of another type and providing corresponding services to those connected modules. In the automated semen quality detection system provided by this embodiment, the applicability of the automated semen quality detection architecture can be adjusted by adding modules and directly calling them without modifying the program code, achieving cluster-based horizontal expansion to quickly and flexibly expand the system. In practical applications, the above modules can be set in the same device or different devices, or they can be set in a virtual device, such as a service instance in a cloud server.

[0028] The following describes the components and specific workflow of an automated semen quality detection system, using specific embodiments as examples.

[0029] The morphological defect rate calculation module 101 is used to acquire the target semen sample to be tested, collect the optical microscopic image of the target semen sample and semen dynamic data, perform image segmentation processing on the optical microscopic image of the sample to obtain a sperm segmentation image, and calculate the morphological defect rate corresponding to the target semen sample based on the sperm segmentation image.

[0030] This invention, through image segmentation processing of the optical microscopic image of the sample, can obtain images that clearly distinguish sperm from other impurities and accurately identify sperm morphology. This not only visually displays sperm morphological details but also helps in accurately calculating the sperm morphological defect rate. The target semen sample is the semen entity to be tested and evaluated; the sample optical microscopic image is an image obtained by taking a picture of the target semen sample using an optical microscope; the semen dynamic data is data generated during sperm movement in the target semen sample; and the sperm segmentation image is an image formed by separating and extracting sperm from the background and other impurities from the sample optical microscopic image through image segmentation processing. Furthermore, the acquisition of the sample optical microscopic image and semen dynamic data of the target semen sample can be achieved using a high-resolution microscope and a sensor with motion trajectory tracking capabilities.

[0031] As an embodiment of the present invention, the step of performing image segmentation processing on the optical microscopic image of the sample to obtain a sperm segmentation image includes:

[0032] The optical microscopic image of the sample is subjected to image denoising processing to obtain a denoised optical microscopic image;

[0033] The denoised optical micrograph is processed to obtain a grayscale optical micrograph.

[0034] The grayscale optical microscopic image is divided into regions to obtain an optical microscopic region image, and the grayscale distribution features corresponding to the optical microscopic region image are extracted.

[0035] Based on the grayscale distribution characteristics, the segmentation threshold corresponding to the grayscale optical microscopy image is determined;

[0036] Based on the segmentation threshold, the grayscale optical microscopy image is segmented to obtain an initial sperm segmentation image;

[0037] The initial sperm segmentation image is subjected to morphological optimization processing to obtain a sperm segmentation image.

[0038] Wherein, the denoised optical microscopic image is an image obtained by denoising the sample optical microscopic image; the grayscale optical microscopic image is an image obtained by grayscale processing of the denoised optical microscopic image; the optical microscopic region image is a selected local region image divided from the grayscale optical microscopic image; the grayscale distribution feature is the distribution characteristics of pixels with different grayscale values ​​corresponding to the optical microscopic region image; and the initial sperm segmentation image is an image obtained by thresholding the grayscale optical microscopic image.

[0039] Optionally, the sample optical microscopic image can be denoised using a Gaussian filtering algorithm to obtain a denoised optical microscopic image; the denoised optical microscopic image can be converted to grayscale values ​​using a specific algorithm to obtain a grayscale optical microscopic image; the grayscale optical microscopic image can be divided into regions using an adaptive region growing algorithm to obtain an optical microscopic region image; the grayscale distribution features corresponding to the optical microscopic region image can be extracted by statistically analyzing the number and proportion of pixels with different grayscale values; based on the grayscale distribution features, the Otsu algorithm is used to determine the segmentation threshold corresponding to the grayscale optical microscopic image; based on the segmentation threshold, the grayscale optical microscopic image can be segmented using a threshold segmentation algorithm to obtain an initial sperm segmentation image; the initial sperm segmentation image can be morphologically optimized using morphological operations such as erosion, dilation, and opening / closing operations to obtain a sperm segmentation image.

[0040] This invention calculates the morphological defect rate of the target semen sample based on the sperm segmentation image. The morphological defect rate can be used to understand the morphological health of the sperm in the target semen sample, laying an important basis for the subsequent evaluation of the reproductive function score of the target semen sample. The morphological defect rate represents the ratio of morphological defects in the target semen sample.

[0041] As an embodiment of the present invention, the step of calculating the morphological defect rate corresponding to the target semen sample based on the sperm segmentation image includes:

[0042] Morphological parameters are extracted from the sperm segmentation image, and abnormal sperm in the sperm segmentation image are identified based on the morphological parameters.

[0043] Based on the morphological parameters, the sperm morphological features in the sperm segmentation images are statistically analyzed;

[0044] Based on the sperm morphological characteristics, the severity of the abnormality corresponding to the morphologically abnormal sperm is determined;

[0045] Assign abnormal weights to the morphologically abnormal sperm, and calculate the abnormal defect value corresponding to the morphologically defective sperm by combining the abnormality severity and the abnormal weights.

[0046] Based on the abnormal defect value, sperm with morphological defects are screened from the sperm with abnormal morphology;

[0047] Based on the morphologically defective sperm, the morphological defect rate corresponding to the target semen sample is calculated.

[0048] The morphological parameters are various quantitative indicators used to describe sperm morphology in the sperm segmentation image, such as length, width, area, and circumference. Abnormal sperm morphology refers to sperm in the sperm segmentation image whose morphology does not conform to normal standards. Sperm morphological characteristics are the appearance characteristics of the head, midpiece, tail, and other parts of the sperm in the sperm segmentation image, including shape, size, and proportion. Abnormality severity measures the degree to which abnormal sperm morphology may affect sperm function and fertility. Abnormality weight is the relative importance value assigned to abnormal sperm morphology based on its impact on fertility. Abnormality defect value represents a quantitative value that comprehensively reflects the degree of morphological abnormality of the morphologically defective sperm. Morphologically defective sperm are sperm among the abnormal sperm morphology that, after further evaluation, are determined to have morphological defects that affect fertility.

[0049] Optionally, morphological parameters can be extracted from the sperm segmentation image using built-in functions in an image analysis library. Based on these morphological parameters, and by comparing them with pre-defined normal morphological standards and thresholds, abnormal sperm morphology in the sperm segmentation image can be identified. Based on these morphological parameters, the morphological characteristics of the sperm in the sperm segmentation image can be statistically analyzed, such as the aspect ratio and ellipticity of the sperm head, the thickness and length of the midsegment, the curvature of the tail, and the area and perimeter of each part. Based on this, the frequency of abnormal features such as acrosome integrity, presence of head malformation, and tail curling or breakage can be quantitatively analyzed. Based on these sperm morphological characteristics, and referring to established grading standards in the medical field, the severity of the abnormality corresponding to the abnormal sperm morphology can be determined. For example, if the sperm head shows more than 50% abnormal morphological changes, or if there is significant swelling or twisting in the midsegment, or if the tail shows multiple folds or a short tail, indicating a serious deviation from the normal morphology... If the sperm exhibits normal morphological characteristics, it is classified as severely abnormal; if there is only slight irregularity in head shape and slight curvature of tail, it is classified as mildly abnormal; and if it falls between these two, it is classified as moderately abnormal. The abnormal weights corresponding to the morphologically abnormal sperm can be assigned through a combination of expert experience and data training. Combining the severity of the abnormality and the abnormal weights, a weighted calculation formula is used to calculate the abnormal defect value corresponding to the morphologically defective sperm. Based on the abnormal defect value, a reasonable screening threshold is set to screen out morphologically defective sperm from the morphologically abnormal sperm. For example, a critical value for abnormal defect values ​​determined in advance through extensive clinical sample testing and data analysis can be used as the screening threshold; morphologically abnormal sperm exceeding this threshold are judged as morphologically defective sperm. Based on the morphologically defective sperm, the morphological defect rate corresponding to the target semen sample is calculated by statistically analyzing the proportion of morphologically defective sperm in the total sperm count.

[0050] The dysfunctional sperm identification module 102 is used to extract trajectory dynamic data and oscillation dynamic data from the semen dynamic data, perform clustering processing on the trajectory dynamic data to obtain a clustered motion trajectory set, analyze the trajectory behavior pattern corresponding to the clustered motion trajectory set, and identify dysfunctional sperm in the target semen sample based on the trajectory behavior pattern.

[0051] This invention clusters the dynamic trajectory data to obtain a set of clustered motion trajectories, which can integrate sperm trajectory information with similar motion characteristics and quickly locate groups of different motion patterns. The dynamic trajectory data is the portion of the semen dynamic data that records the sperm movement path, and the oscillation dynamic data is the portion of the semen dynamic data that describes the amplitude and frequency of sperm tail oscillation. The set of clustered motion trajectories is a collection formed by grouping data with similar motion trajectories together. Furthermore, the extraction of the dynamic trajectory data and oscillation dynamic data from the semen dynamic data can be achieved through a specific extraction function, which is compiled by a programming language.

[0052] As an embodiment of the present invention, the step of clustering the trajectory dynamic data to obtain a clustered motion trajectory set includes:

[0053] The dynamic trajectory data is subjected to trajectory repair processing to obtain purified trajectory data;

[0054] Analyze the motion trajectory corresponding to the purification trajectory data, perform trajectory segmentation on the motion trajectory to obtain segmented trajectories, and extract multidimensional motion features from the segmented trajectories;

[0055] Based on the multidimensional motion features, a trajectory feature matrix corresponding to the purification trajectory data is constructed;

[0056] Calculate the matrix similarity between the trajectory feature matrices. When the matrix similarity is greater than a preset similarity, perform clustering processing on the motion trajectory to obtain a clustered motion trajectory set.

[0057] Wherein, the motion trajectory is the actual motion path corresponding to the purification trajectory data; the segmented trajectory is a fragment of the motion trajectory divided according to certain rules; the multidimensional motion features are parameters in the segmented trajectory that can reflect multiple dimensions of motion characteristics; the trajectory feature matrix is ​​a matrix constructed with multidimensional motion features as elements corresponding to the purification trajectory data; the matrix similarity represents the quantitative value of the similarity between the trajectory feature matrices; the preset similarity is a pre-set threshold for judging whether matrices are similar or not, which can be set to 0.8 or can be set according to the actual application scenario.

[0058] Furthermore, the trajectory dynamic data can be repaired using a wavelet transform algorithm to obtain purified trajectory data; the motion trajectory corresponding to the purified trajectory data can be analyzed using spatial coordinate mapping and parameter fitting algorithms; the motion trajectory can be segmented using time series segmentation or key node identification methods to obtain segmented trajectories; multidimensional motion features can be extracted from the segmented trajectories based on parameters such as velocity, acceleration, and curvature; based on the multidimensional motion features, a trajectory feature matrix corresponding to the purified trajectory data can be constructed using a data structured recombination algorithm; when the matrix similarity is greater than a preset similarity, the motion trajectory can be clustered using classic clustering algorithms such as K-Means to obtain a clustered motion trajectory set.

[0059] Furthermore, as an optional embodiment of the present invention, calculating the matrix similarity between the trajectory feature matrices includes:

[0060] Calculate the Euclidean distance between the trajectory feature matrices, and then vectorize the trajectory feature matrices to obtain a matrix vector.

[0061] Calculate the vector similarity between the matrix vectors, and determine the cosine similarity between the trajectory feature matrices based on the vector similarity;

[0062] Calculate the dynamic programming distance between the trajectory feature matrices, and combine the Euclidean distance, the cosine similarity, and the dynamic programming distance to calculate the matrix similarity between the trajectory feature matrices using the following formula.

[0063]

[0064] Where β represents the matrix similarity between trace feature matrices, d a (A,B) represents the Euclidean distance between the A-th and B-th matrices in the trajectory feature matrix, and cos(A,B) represents the cosine similarity between the A-th and B-th matrices in the trajectory feature matrix. t (A,B) represents the time series distance.

[0065] Wherein, the Euclidean distance is a metric based on the square root of the sum of the squared differences between the elements of the trajectory feature matrices; the matrix vector is a vector obtained by transforming or extracting the trajectory feature matrices in a certain way; the vector similarity represents the degree of similarity between the matrix vectors in terms of direction and length; the cosine similarity represents the similarity between the trajectory feature matrices by calculating the cosine value of the angle between the vectors; and the dynamic programming distance is the cumulative distance between the trajectory feature matrices under the optimal alignment path found by the dynamic programming algorithm.

[0066] Optionally, the Euclidean distance between the trajectory feature matrices can be calculated by taking the square root of the sum of the squares of the differences between corresponding elements of the trajectory feature matrices; the trajectory feature matrices can be vectorized by methods such as extracting elements by row or column and principal component analysis to obtain matrix vectors; the vector similarity between the matrix vectors can be calculated by calculating the inner product of the matrix vectors and combining it with the vector magnitude, based on the cosine similarity formula, and the sum of the vector similarities can be obtained to obtain the cosine similarity between the trajectory feature matrices; or the dynamic programming distance between the trajectory feature matrices can be calculated iteratively by constructing a cost matrix and using the recursive formula of the dynamic programming algorithm.

[0067] This invention analyzes the trajectory behavior patterns corresponding to the clustered movement trajectory sets to summarize sperm motility patterns, clarify the characteristics and differences of different movement patterns, and construct a precise trajectory behavior classification system. Based on the trajectory behavior patterns, it identifies dysfunctional sperm in the target semen sample, obtaining the category and proportion of dysfunctional sperm in the semen sample, providing scientific and reliable data support for clinical male fertility assessment. The trajectory behavior patterns are movement patterns with specific rules and characteristics corresponding to the clustered movement trajectory sets, and the dysfunctional sperm are sperm in the target semen sample exhibiting abnormalities in motility, morphology, or other functions. Furthermore, based on the trajectory behavior patterns, it identifies dysfunctional sperm in the target semen sample; for example, sperm exhibiting highly irregular movement trajectories or speeds significantly deviating from the normal range are classified as dysfunctional sperm.

[0068] As an embodiment of the present invention, the analysis of the trajectory behavior patterns corresponding to the clustered motion trajectory set includes:

[0069] Based on the preset basic motion states, the clustered motion trajectory set is marked with state labels to obtain the trajectory motion states;

[0070] Based on the trajectory motion state, generate a state transition log corresponding to the clustered motion trajectory set;

[0071] Based on the state transition log, construct the motion state sequence corresponding to the clustered motion trajectory set;

[0072] Calculate the state transition probability corresponding to the motion state sequence, and construct the transition matrix corresponding to each trajectory in the clustered motion trajectory set based on the state transition probability;

[0073] Based on the transition matrix, generate the pattern feature vector corresponding to each trajectory in the clustered motion trajectory set;

[0074] Based on the pattern feature vector, the trajectory behavior patterns corresponding to the clustered motion trajectory set are analyzed.

[0075] The basic motion state is a pre-defined set of states used to describe the basic motion characteristics of sperm, such as stationary, linear motion, and curvilinear motion. The trajectory motion state is the specific motion state presented after the clustered motion trajectory set is marked according to the basic motion state. The state transition log is a set of information corresponding to the clustered motion trajectory set that records the changes in the motion state of each sperm at different times. The motion state sequence is a sequence of sperm motion states arranged in chronological order corresponding to the clustered motion trajectory set. The state transition probability is a numerical value corresponding to the motion state sequence that represents the probability of transitioning from one motion state to another. The transition matrix is ​​a matrix corresponding to each trajectory in the clustered motion trajectory set that describes the transition probability between motion states. The pattern feature vector is a vector corresponding to each trajectory in the clustered motion trajectory set that reflects its motion pattern characteristics.

[0076] Optionally, based on a preset basic motion state, the clustered motion trajectory set can be marked with a state label by comparing and matching trajectories one by one to obtain the trajectory motion state; based on the trajectory motion state, a state transition log corresponding to the clustered motion trajectory set can be generated by recording the motion state changes at adjacent time points; based on the state transition log, a motion state sequence corresponding to the clustered motion trajectory set can be constructed by sorting it out in chronological order; the state transition probability corresponding to the motion state sequence can be calculated by statistically analyzing the frequency of each state transition and performing normalization; based on the state transition probability, a transition matrix corresponding to each trajectory in the clustered motion trajectory set can be constructed by filling in the corresponding state transition probability values; based on the transition matrix, a pattern feature vector corresponding to each trajectory in the clustered motion trajectory set can be generated by extracting key parameters and feature values ​​of the matrix; based on the pattern feature vector, the trajectory behavior pattern corresponding to the clustered motion trajectory set can be analyzed by using a feature induction strategy, such as sorting out the commonalities and differences of feature vectors in dimensions such as speed change, direction turning, and displacement fluctuation, and summarizing typical trajectory behavior patterns such as linear acceleration, spiral wandering, and disordered trembling.

[0077] The reproductive function scoring and evaluation module 103 is used to extract the swing dynamic features from the swing dynamic data, evaluate the sperm motility intensity in the target semen sample based on the swing dynamic features, and evaluate the reproductive function score corresponding to the target semen sample by combining the morphological defect rate, the dysfunctional sperm and the sperm motility intensity.

[0078] This invention extracts the dynamic features of the swinging motion from the swinging dynamic data, which can accurately capture the subtle characteristics of sperm movement and provide an intuitive and reliable quantitative basis for the evaluation of the reproductive function score corresponding to the target semen sample. The swinging dynamic features are a set of parameters in the swinging dynamic data that can characterize the characteristics of sperm swinging behavior, quantify the swinging behavior process, and provide key information for sperm movement status analysis and semen quality assessment, such as swinging frequency, amplitude, and path tortuosity.

[0079] As an embodiment of the present invention, the extraction of swing dynamic features from the swing dynamic data includes:

[0080] Extract the swing dynamic signal corresponding to the swing dynamic data, and perform intrinsic mode decomposition on the swing dynamic signal to obtain the signal intrinsic modes;

[0081] Calculate the modal energy ratio corresponding to the intrinsic modes of the signal, and construct the oscillation energy spectrum corresponding to the oscillation dynamic signal based on the modal energy ratio;

[0082] Peak detection is performed on the swing energy spectrum to obtain the energy spectrum peak characteristics;

[0083] Extract the time-domain and frequency-domain peak features from the energy spectrum peak features, and perform feature fusion processing on the time-domain and frequency-domain peak features to obtain the swing dynamic features in the swing dynamic data.

[0084] Wherein, the oscillating dynamic signal is the time-series signal corresponding to the oscillating dynamic data; the signal intrinsic mode is the component with different time-scale characteristics obtained by intrinsic mode decomposition of the oscillating dynamic signal; the modal energy ratio is the proportion of the energy of each component corresponding to the signal intrinsic mode in the total energy; the oscillating energy spectrum is the energy-frequency distribution spectrum constructed based on the modal energy ratio corresponding to the oscillating dynamic signal; the energy spectrum peak feature is the feature reflecting the peak characteristics of the energy distribution obtained after the oscillating energy spectrum has undergone peak detection; the time-domain peak feature and the frequency-domain peak feature are the relevant feature parameters describing the peak in the time domain and frequency domain, respectively, of the energy spectrum peak features.

[0085] Furthermore, the swing dynamic signal corresponding to the swing dynamic data can be extracted through a conversion algorithm based on sensor principles and data acquisition specifications; the intrinsic mode decomposition (EMD) algorithm can be used to process the swing dynamic signal to obtain the signal's intrinsic modes; based on the modal energy ratio, a plotting algorithm that arranges the energy ratios of each IMF according to frequency or time dimensions can be used to construct the swing energy spectrum corresponding to the swing dynamic signal; a detection algorithm based on peak search and threshold judgment can be used to detect the peaks of the swing energy spectrum to obtain energy peak features; the time-domain and frequency-domain peak features in the energy peak features can be extracted using parameter extraction methods defined according to time-domain and frequency-domain features; and the time-domain and frequency-domain peak features can be fused using fusion strategies such as splicing and weighted summation to obtain the swing dynamic features in the swing dynamic data.

[0086] Furthermore, as an optional embodiment of the present invention, the calculation of the modal energy ratio corresponding to the intrinsic modes of the signal includes:

[0087]

[0088] Where δ represents the modal energy percentage corresponding to the intrinsic modes of the signal, IMF k (t) represents the modal energy value of the k-th mode in the intrinsic modes of the signal at time t, where k represents the sequence number of the intrinsic modes of the signal, N represents the number of modes in the intrinsic modes of the signal, t represents the start time point of the intrinsic modes of the signal, and T represents the end time point of the intrinsic modes of the signal.

[0089] This invention evaluates sperm motility intensity in a target semen sample based on the aforementioned dynamic characteristics of oscillation, accurately quantifying sperm vitality during motility. By combining the morphological defect rate, the number of dysfunctional sperm, and the sperm motility intensity, a reproductive function score is assessed for the target semen sample. This allows for a comprehensive and multi-dimensional evaluation of semen quality, providing a description of the target semen sample's reproductive capabilities. Specifically, sperm motility intensity is an indicator of the vitality and strength of sperm during motility, and the reproductive function score is a quantitative value comprehensively reflecting the fertility of the target semen sample. Furthermore, based on the aforementioned dynamic characteristics of oscillation, the sperm motility intensity in the target semen sample is evaluated firstly… The dynamic characteristics of the oscillation are decomposed, and key parameters such as oscillation frequency and amplitude are extracted. These parameters are used to calculate motion data such as sperm displacement and velocity changes per unit time. Then, based on clinical experience and research data, these data are compared with standard thresholds to determine the sperm motility intensity level, thereby achieving the evaluation of sperm motility intensity in the target semen sample. Combining the morphological defect rate, the dysfunctional sperm, and the sperm motility intensity, the reproductive function score corresponding to the target semen sample is evaluated. The proportion of dysfunctional sperm is calculated based on the dysfunctional sperm. The proportion of dysfunctional sperm, the morphological defect rate, and the sperm motility intensity are normalized, and the average value of the normalized values ​​is calculated to obtain the reproductive function score corresponding to the target semen sample.

[0090] The dynamic motility potential assessment module 104 is used to configure the appropriate culture medium for the target semen sample, place the target semen sample in the appropriate culture medium for culture treatment, record the sperm viability baseline parameters during the culture treatment process, calculate the sperm survival rate of the target semen sample based on the viability baseline parameters, and assess the dynamic motility potential score corresponding to the target semen sample based on the sperm survival rate.

[0091] This invention facilitates refined and targeted culture of the target semen sample by configuring a suitable culture medium. The target semen sample is placed in the suitable culture medium for culture treatment, and sperm motility baseline parameters are recorded during the culture process. This allows for accurate subsequent calculation of sperm viability. The suitable culture medium is a liquid that simulates the survival environment of the target semen sample according to the usage scenario. The sperm motility baseline parameters are basic data reflecting sperm survival and motility during the culture process, such as the number of motile sperm per unit volume. Furthermore, the target... The steps for preparing the appropriate culture medium for the target semen sample are as follows: Query the application scenario corresponding to the target semen sample, analyze the physicochemical properties corresponding to the application scenario, determine the basic formula of the culture medium based on the physicochemical properties, adjust the ratio of nutrients and additives in the basic formula in combination with the physiological needs of sperm, and prepare the appropriate culture medium according to the adjusted formula; the recording of the sperm motility baseline parameters during the culture process can be achieved by using a microscopic observation device equipped with image analysis software. Using a phase contrast microscope or fluorescence microscope, sperm motility images are captured, and the image analysis software automatically identifies and counts the number of motile sperm per unit volume.

[0092] This invention calculates the sperm viability rate of the target semen sample based on the viability benchmark parameter, thereby understanding the sperm viability status of the target semen sample and intuitively grasping the proportion of viable sperm in the semen sample. This provides a basis for the subsequent evaluation and processing of the reproductive function score corresponding to the target semen sample. The sperm viability rate represents the proportion of viable sperm in the total number of sperm in the target semen sample.

[0093] As an embodiment of the present invention, calculating the sperm viability of the target semen sample based on the viability benchmark parameter includes:

[0094] Identify the survival evaluation indicators corresponding to the survival benchmark parameters, wherein the survival evaluation indicators include speed evaluation indicators and direction evaluation indicators;

[0095] Combining the speed evaluation index, the direction evaluation index, and the viability benchmark parameter, the sperm in the target semen sample are subjected to viability grading and labeling to obtain the labeling results;

[0096] The annotation results are divided into three categories: forward motility, non-forward motility, and inactive. The number of sperm in each category is counted to obtain the first sperm count, the second sperm count, and the third sperm count.

[0097] Combining the first sperm count, the second sperm count, and the third sperm count, the sperm viability of the target semen sample is calculated using the following formula:

[0098]

[0099] Where φ represents the sperm viability of the target semen sample, D represents the number of first sperm, E represents the number of second sperm, and F represents the number of third sperm.

[0100] The viability assessment index is an evaluation dimension corresponding to the viability benchmark parameter, thereby measuring sperm motility. The speed assessment index and the direction assessment index are key components of the viability assessment index. The former quantifies the speed of sperm movement, while the latter describes the direction of sperm movement. The forward motility category is a classification of sperm with a movement speed ≥25μm / s and a movement direction angle <45°. The non-forward motility category is a classification of sperm with a movement speed ≥5μm / s but not meeting the forward motility standard. The inactive category is a classification of sperm with a movement speed <5μm / s that has almost lost its motility.

[0101] Furthermore, the viability evaluation index corresponding to the viability benchmark parameter can be identified according to preset rules, which can be set by the World Health Organization. Combining the velocity evaluation index, the direction evaluation index, and the viability benchmark parameter, the sperm in the target semen sample can be classified and labeled using a convolutional neural network (CNN) classifier with built-in World Health Organization (WHO) classification standards to obtain the labeling results.

[0102] This invention assesses the dynamic motility potential score of a target semen sample based on the sperm survival rate. This allows for an understanding of the sperm's potential to maintain activity and achieve fertilization. The dynamic motility potential score is a quantitative measure of the sperm's continued activity and likelihood of conception during fertilization. Furthermore, the dynamic motility potential score is assessed based on the sperm survival rate. For example, a sperm survival rate ≥80% indicates that the vast majority of sperm in the sample are active, combined with forward motility... The percentage of motile sperm is considered. If the percentage of progressively motile sperm exceeds 60%, a high dynamic vitality potential score is given, indicating that the sample has excellent fertility potential. 60% ≤ sperm survival rate < 80%: Sperm survival rate is between 60% and 80%. If the percentage of non-progressively motile sperm is high and the percentage of normally morphological sperm is also high, a medium to high dynamic vitality potential score is given. If there are many abnormal sperm, the score will be lower accordingly. 40% ≤ sperm survival rate < 60%: Sperm survival rate is between 40% and 60%, and a medium dynamic vitality potential score is given. Survival rate < 40%: When the sperm survival rate is below 40%, it indicates poor sample quality and a lower score is given.

[0103] The semen quality analysis module 105 is used to combine the reproductive function score and the dynamic vitality potential score to generate a semen quality analysis report corresponding to the target semen sample.

[0104] This invention generates a semen quality analysis report corresponding to the target semen sample by combining the reproductive function score and the dynamic vitality potential score, thereby obtaining high-quality semen quality analysis results. The semen quality analysis report is a comprehensive quality analysis result corresponding to the target semen sample. Furthermore, by combining the reproductive function score and the dynamic vitality potential score, the semen quality analysis report corresponding to the target semen sample is generated, such as presenting each score in tabular form and combining textual analysis of the semen quality.

[0105] This invention, through image segmentation processing of the optical microscopic images of the sample, can obtain images that clearly distinguish sperm from other impurities and accurately identify sperm morphology. This not only visually displays sperm morphological details but also helps in accurately calculating sperm morphological defect rates. Optionally, this invention, by clustering the trajectory dynamic data, obtains a set of clustered motion trajectories, which can integrate sperm trajectory information with similar motion characteristics and quickly locate groups of different motion patterns. This invention extracts the oscillation dynamic features from the oscillation dynamic data, which can accurately capture subtle characteristics of sperm movement, providing an intuitive and reliable quantitative basis for evaluating the reproductive function score corresponding to the target semen sample. This invention, by configuring the appropriate culture medium for the target semen sample, facilitates refined and targeted culture of the target semen sample. The target semen sample is placed in the appropriate culture medium for culture treatment, and the sperm viability baseline parameters are recorded during the culture treatment process, thus facilitating subsequent accurate calculation of sperm viability. This invention, by combining the reproductive function score and the dynamic motility potential score, generates a semen quality analysis report corresponding to the target semen sample, resulting in high-quality semen quality analysis results. Therefore, the accuracy of semen quality testing has been improved.

[0106] like Figure 2 The diagram shown is a flowchart illustrating an automated method for detecting semen quality according to an embodiment of the present invention. In this embodiment, the automated method for detecting semen quality includes:

[0107] A target semen sample to be tested is obtained, and optical microscopic images and semen dynamic data of the target semen sample are collected. The optical microscopic images of the sample are processed to obtain sperm segmentation images. Based on the sperm segmentation images, the morphological defect rate corresponding to the target semen sample is calculated.

[0108] Trajectory dynamic data and oscillation dynamic data are extracted from the semen dynamic data. The trajectory dynamic data is clustered to obtain a set of clustered motion trajectories. The trajectory behavior patterns corresponding to the set of clustered motion trajectories are analyzed. Based on the trajectory behavior patterns, dysfunctional sperm in the target semen sample are identified.

[0109] Extract the swing dynamic features from the swing dynamic data, evaluate the sperm motility intensity in the target semen sample based on the swing dynamic features, and evaluate the reproductive function score corresponding to the target semen sample by combining the morphological defect rate, the dysfunctional sperm and the sperm motility intensity.

[0110] Configure the appropriate culture medium for the target semen sample, place the target semen sample in the appropriate culture medium for culture treatment, and record the sperm motility baseline parameters during the culture treatment process. Based on the motility baseline parameters, calculate the sperm survival rate of the target semen sample, and evaluate the dynamic motility potential score corresponding to the target semen sample based on the sperm survival rate.

[0111] By combining the reproductive function score and the dynamic vitality potential score, a semen quality analysis report corresponding to the target semen sample is generated.

[0112] In the several embodiments provided by this invention, it should be understood that the provided systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0113] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A system for automatically detecting semen quality, characterized in that, The automated semen quality detection system includes: a morphological defect rate calculation module, a sperm dysfunction identification module, a reproductive function scoring and evaluation module, a dynamic motility potential scoring and evaluation module, and a semen quality analysis module. The morphological defect rate calculation module is used to acquire the target semen sample to be tested, collect the optical microscopic image of the target semen sample and semen dynamic data, perform image segmentation processing on the optical microscopic image of the sample to obtain a sperm segmentation image, and calculate the morphological defect rate corresponding to the target semen sample based on the sperm segmentation image. The dysfunctional sperm identification module is used to extract trajectory dynamic data and oscillation dynamic data from the semen dynamic data, perform clustering processing on the trajectory dynamic data to obtain a clustered motion trajectory set, analyze the trajectory behavior pattern corresponding to the clustered motion trajectory set, and identify dysfunctional sperm in the target semen sample based on the trajectory behavior pattern. The reproductive function scoring and evaluation module is used to extract the swing dynamic features from the swing dynamic data, evaluate the sperm motility intensity in the target semen sample based on the swing dynamic features, and evaluate the reproductive function score corresponding to the target semen sample by combining the morphological defect rate, the dysfunctional sperm and the sperm motility intensity. The dynamic motility potential scoring module is used to configure the appropriate culture medium for the target semen sample, place the target semen sample in the appropriate culture medium for culture treatment, record the sperm motility baseline parameters during the culture treatment process, calculate the sperm survival rate of the target semen sample based on the sperm survival rate, and evaluate the dynamic motility potential score corresponding to the target semen sample based on the sperm survival rate. The semen quality analysis module is used to combine the reproductive function score and the dynamic vitality potential score to generate a semen quality analysis report corresponding to the target semen sample. The process of clustering the dynamic trajectory data to obtain a set of clustered motion trajectories includes: The dynamic trajectory data is subjected to trajectory repair processing to obtain purified trajectory data; Analyze the motion trajectory corresponding to the purification trajectory data, perform trajectory segmentation on the motion trajectory to obtain segmented trajectories, and extract multidimensional motion features from the segmented trajectories; Based on the multidimensional motion features, a trajectory feature matrix corresponding to the purification trajectory data is constructed; Calculate the matrix similarity between the trajectory feature matrices. When the matrix similarity is greater than a preset similarity, perform clustering processing on the motion trajectory to obtain a clustered motion trajectory set.

2. The automated semen quality detection system as described in claim 1, characterized in that, The step of performing image segmentation processing on the optical microscopic image of the sample to obtain a sperm segmentation image includes: The optical microscopic image of the sample is subjected to image denoising processing to obtain a denoised optical microscopic image; The denoised optical micrograph is processed to obtain a grayscale optical micrograph. The grayscale optical microscopic image is divided into regions to obtain an optical microscopic region image, and the grayscale distribution features corresponding to the optical microscopic region image are extracted. Based on the grayscale distribution characteristics, the segmentation threshold corresponding to the grayscale optical microscopy image is determined; Based on the segmentation threshold, the grayscale optical microscopy image is segmented to obtain an initial sperm segmentation image; The initial sperm segmentation image is subjected to morphological optimization processing to obtain a sperm segmentation image.

3. The automated semen quality detection system as described in claim 1, characterized in that, The step of calculating the morphological defect rate of the target semen sample based on the sperm segmentation image includes: Morphological parameters are extracted from the sperm segmentation image, and abnormal sperm in the sperm segmentation image are identified based on the morphological parameters. Based on the morphological parameters, the sperm morphological features in the sperm segmentation images are statistically analyzed; Based on the sperm morphological characteristics, the severity of the abnormality corresponding to the morphologically abnormal sperm is determined; Assign abnormal weights to the morphologically abnormal sperm, and calculate the abnormal defect value corresponding to the morphologically abnormal sperm by combining the abnormality severity and the abnormal weights. Based on the abnormal defect value, sperm with morphological defects are screened from the sperm with abnormal morphology; Based on the morphologically defective sperm, the morphological defect rate corresponding to the target semen sample is calculated.

4. The automated semen quality detection system as described in claim 1, characterized in that, The calculation of the matrix similarity between the trajectory feature matrices includes: Calculate the Euclidean distance between the trajectory feature matrices, and then vectorize the trajectory feature matrices to obtain a matrix vector. Calculate the vector similarity between the matrix vectors, and determine the cosine similarity between the trajectory feature matrices based on the vector similarity; Calculate the dynamic programming distance between the trajectory feature matrices, and combine the Euclidean distance, the cosine similarity, and the dynamic programming distance to calculate the matrix similarity between the trajectory feature matrices.

5. The automated semen quality detection system as described in claim 1, characterized in that, The analysis of the trajectory behavior patterns corresponding to the clustered motion trajectory set includes: Based on the preset basic motion states, the clustered motion trajectory set is marked with state labels to obtain the trajectory motion states; Based on the trajectory motion state, generate a state transition log corresponding to the clustered motion trajectory set; Based on the state transition log, construct the motion state sequence corresponding to the clustered motion trajectory set; Calculate the state transition probability corresponding to the motion state sequence, and construct the transition matrix corresponding to each trajectory in the clustered motion trajectory set based on the state transition probability; Based on the transition matrix, generate the pattern feature vector corresponding to each trajectory in the clustered motion trajectory set; Based on the pattern feature vector, the trajectory behavior patterns corresponding to the clustered motion trajectory set are analyzed.

6. The automated semen quality detection system as described in claim 1, characterized in that, The extraction of swing dynamic features from the swing dynamic data includes: Extract the swing dynamic signal corresponding to the swing dynamic data, and perform intrinsic mode decomposition on the swing dynamic signal to obtain the signal intrinsic modes; Calculate the modal energy ratio corresponding to the intrinsic modes of the signal, and construct the oscillation energy spectrum corresponding to the oscillation dynamic signal based on the modal energy ratio; Peak detection is performed on the swing energy spectrum to obtain the energy spectrum peak characteristics; Extract the time-domain and frequency-domain peak features from the energy spectrum peak features, and perform feature fusion processing on the time-domain and frequency-domain peak features to obtain the swing dynamic features in the swing dynamic data.

7. The automated semen quality detection system as described in claim 1, characterized in that, The calculation of sperm viability of the target semen sample based on the viability benchmark parameter includes: Identify the survival evaluation indicators corresponding to the survival benchmark parameters, wherein the survival evaluation indicators include speed evaluation indicators and direction evaluation indicators; Combining the speed evaluation index, the direction evaluation index, and the viability benchmark parameter, the sperm in the target semen sample are subjected to viability grading and labeling to obtain the labeling results; The annotation results are divided into three categories: forward motility, non-forward motility, and inactive. The number of sperm in each category is counted to obtain the first sperm count, the second sperm count, and the third sperm count. The sperm viability of the target semen sample is calculated by combining the first sperm count, the second sperm count, and the third sperm count.

8. An automated method for detecting semen quality for non-disease diagnosis and treatment purposes, the method being implemented based on the system described in claim 1, characterized in that, The method includes: A target semen sample to be tested is obtained, and optical microscopic images and semen dynamic data of the target semen sample are collected. The optical microscopic images of the sample are processed to obtain sperm segmentation images. Based on the sperm segmentation images, the morphological defect rate corresponding to the target semen sample is calculated. Trajectory dynamic data and oscillation dynamic data are extracted from the semen dynamic data. The trajectory dynamic data is clustered to obtain a set of clustered motion trajectories. The trajectory behavior patterns corresponding to the set of clustered motion trajectories are analyzed. Based on the trajectory behavior patterns, dysfunctional sperm in the target semen sample are identified. Extract the swing dynamic features from the swing dynamic data, evaluate the sperm motility intensity in the target semen sample based on the swing dynamic features, and evaluate the reproductive function score corresponding to the target semen sample by combining the morphological defect rate, the dysfunctional sperm and the sperm motility intensity. Configure the appropriate culture medium for the target semen sample, place the target semen sample in the appropriate culture medium for culture treatment, and record the sperm motility baseline parameters during the culture treatment process. Based on the motility baseline parameters, calculate the sperm survival rate of the target semen sample, and evaluate the dynamic motility potential score corresponding to the target semen sample based on the sperm survival rate. By combining the reproductive function score and the dynamic vitality potential score, a semen quality analysis report corresponding to the target semen sample is generated.

Images