Method for judging activity of takifugu obscurus sperm cryopreservation liquid based on computer vision

Abstract

The invention discloses a method for judging the activity of a takifugu obscurus sperm cryopreservation solution based on computer vision, which comprises the following steps of: performing long-term ultralow-temperature preservation by using the takifugu obscurus sperm cryopreservation solution, and observing and recording a sperm activity video under a microscope after unfreezing and activating; carrying out data enhancement on the video by using a binarization method; based on two deep learning algorithms of target detection and target tracking, tracking and capturing position changes of a plurality of Takifugu obscurus sperm targets between continuous frames, analyzing and identifying key motion and tracks of sperms, and performing simulation training on the key motion and tracks; the movement and survival states of the sperms are judged by counting the displacement of the sperms among different frames; and finally generating a visual sperm survival proportion chart. According to the method provided by the invention, the takifugu obscurus sperms with high activity can be preserved at ultra-low temperature for a long time, technical support is provided for annual breeding and germplasm resource protection, the takifugu obscurus sperm activity detection precision is also improved, the detection identification process is simplified, and the cost is reduced.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence and aquaculture technology, and in particular to a method for judging the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision. Background Art

[0002] The dark-spotted pufferfish (Takifugufasciatus), commonly known as "pufferfish", is one of the three delicacies of the Yangtze River in my country and is a traditional and precious edible fish in my country. Its meat is tender and delicious, and its protein content is rich, which makes it deeply loved by consumers in my country. Therefore, the dark-spotted pufferfish has become a high-end and popular fish among today's economic aquatic fish. However, the dark-spotted pufferfish is a type of fish that lays eggs once a year and takes 2-3 years to reach sexual maturity, so breeding can only be carried out in spring. However, in today's large-scale breeding process, due to reasons such as long-term inbreeding and aging of parent fish, the germplasm resources of the dark-spotted pufferfish have gradually degraded. Therefore, it is necessary to introduce more high-quality parents to improve the breeding value of the dark-spotted pufferfish. However, in order to avoid causing great damage to the living parents during logistics transportation during the introduction of high-quality parents, which would affect the quality of the sperm and eggs collected subsequently, and the dark-spotted pufferfish eggs are not easy to preserve in vitro, we chose to collect mature sperm from male fish and store them at ultra-low temperatures for a long time in the cryopreservation solution provided by the present invention, and then combine them with mature eggs when conditions are suitable to obtain high-quality fertilized eggs. This not only reduces logistics risks and cost losses, but also provides a technical means for expanding the scope of use of improved dark-spotted pufferfish male parents and protecting their germplasm resources.

[0003] At the same time, sperm motility testing and analysis are crucial for improving the quality of Takifugu obscurus sperm resources. Sperm motility directly impacts the reproductive success rate and offspring growth and development of Takifugu obscurus. Therefore, accurate and effective sperm motility testing not only helps assess the genetic quality of Takifugu obscurus populations but also provides a scientific basis for optimizing breeding and population management. However, existing methods for detecting fish sperm motility generally rely on manual microscopic observation, evaluation, and counting of active sperm. This process not only requires high levels of professional skill and experience but is also susceptible to human error, resulting in inaccurate and poor reproducibility in counts. Furthermore, this manual operation is extremely time-consuming and labor-intensive, increasing labor costs over the long term. Another common method utilizes equipment such as computer-assisted sperm analysis systems for observation. While this method can improve detection efficiency and accuracy, the high cost of the equipment and challenges with system maintenance and upgrades have limited its widespread application. Therefore, developing a computer vision-based method for determining the activity of dark-spotted pufferfish sperm cryopreservation fluid can not only reduce labor and equipment costs, but also improve the accuracy and efficiency of sperm detection while being efficient and convenient. It is an important breakthrough in the aquaculture industry in terms of germplasm resource protection and breeding technology improvement. Summary of the Invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a method for determining the activity of Takifugu obscurus sperm cryopreservation fluid based on computer vision. The purpose is to reduce labor costs and equipment costs while ensuring that high-quality in vitro Takifugu obscurus sperm still has high activity after long-term ultra-low temperature storage and thawing, so that sperm and egg can combine to form a fertilized egg; and in the sperm sample activity detection stage, not only the convenience and speed of operation are improved, but also the precision and accuracy of detection are significantly improved.

[0005] Technical solution: The present invention provides a method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision, the method comprising the following steps:

[0006] Step 1: The testis of a male Takifugu obscurus with mature gonads is collected, ground, and then a buffer solution is added. The sperm dilution is collected by filtration with a mesh, and the sperm activity is detected under a microscope after activation with pure water, and a video of the original sperm activity is recorded;

[0007] Step 2: Threshold the sperm activity video, use binarization to adjust its grayscale value, and convert and segment the image in the video into a grayscale image;

[0008] Step 3: Load the target detection model in the integrated development environment, initialize the target tracker, input the sperm activity video processed in Step 2, draw a detection box and tracking ID for each sperm in the video, perform multi-target tracking and motion trajectory detection, and implement simulation training;

[0009] Step 4: Based on the displacement statistics of sperm activity simulation training, a displacement threshold for survival or death is set, and the current sperm position is continuously updated in real time. If the displacement exceeds the set threshold, it is judged to be alive;

[0010] Step 5: Update the total number of sperm sampled globally and the number of surviving sperm in the statistical video in real time, define the percentage formatting function, draw a pie chart of survival and death, and generate a visual survival ratio chart;

[0011] Step 6: If the effective sperm activity is greater than 95%, the sperm diluent, the dark puffer freezing preservation solution and the antifreeze are mixed to obtain a sperm freezing mixture, which is then subjected to gradient cooling and finally placed in liquid nitrogen for long-term ultra-low temperature storage; after the storage is completed, the sperm freezing mixture is removed from the liquid nitrogen, gradiently heated and thawed, and dark puffer activation solution is added for resuscitation, the effective sperm activity is observed under an electron microscope, and a sperm activity video is recorded; then steps 2 to 5 are repeated.

[0012] Furthermore, the buffer solution in step 1 is a PBS buffer solution containing 1% penicillin-streptomycin.

[0013] Furthermore, the specific steps of step 2 include: processing the dark-striped pufferfish sperm activity video using a global fixed threshold binarization method, adjusting the grayscale value of each frame image in the video, constructing a function core algorithm of a deep learning framework, combining edge detection image and threshold segmentation image technology, and segmenting the grayscale sperm video image through a hard-coded threshold, and using the weighted image grayscale conversion formula MGV=0.2989×red+0.5870×green+0.1140×blue to adjust the set predetermined grayscale threshold, thereby dividing the pixels in the image into white and black categories.

[0014] Furthermore, the specific steps of step 3 include: (1) loading a real-time target detection algorithm in the integrated development environment, setting it to CPU or GPU mode, designing a sperm target detection model using a convolutional neural network (CNN) deep learning algorithm, and initializing a simple and efficient target tracking algorithm based on the Kalman filter and the Hungarian algorithm, setting parameters to control the continuous frame threshold that the sperm target can tolerate after being lost, otherwise it will be removed, and setting the sperm target to be considered as a valid target to appear at least multiple times to ensure that the tracker will not mistakenly track some false targets that may only appear occasionally;

[0015] (2) A target detection algorithm is used to generate a bounding box represented by a rectangle for each detected sperm, marking its position and serial number. The multi-target tracking (MOT) algorithm is then used to track multiple sperm targets that appear simultaneously in the binary sperm video, recording the unique ID and motion path of each sperm target and updating its position in real time in each frame.

[0016] Furthermore, the specific steps of step 4 include: (1) based on sperm activity simulation training under deep learning, statistics of sperm activity displacement information, simulation of sperm activity through position changes in a two-dimensional space, setting the current position of each sperm to (x, y) and updating the position as the number of frames changes: the displacement of the sperm at each moment is obtained by calculating the distance between the previous and next positions, the formula is d = √[(x2-x1) 2 +(y2-y1) 2 ]; (x1, y1) is the position at the previous moment, and (x2, y2) is the position at the current moment.

[0017] (2) Using the optical flow method and background subtraction method, the statistical data of the pre-trained sperm model are compared and analyzed with the initial manual detection results of sperm activity, so as to set an appropriate displacement threshold for survival or death. When the displacement of the sperm is greater than the threshold, the sperm is judged to be alive. If the displacement is less than the threshold, the sperm may have lost its vitality and is judged to be dead. In each iteration, the current position of the sperm is updated in real time.

[0018] Furthermore, the specific steps of step 5 include: (1) updating and recording the last position of each sperm through statistical data, and storing the sperm IDs of survival, death and total sperm respectively, and calculating the number of survival and total sperm targets tracked;

[0019] (2) After each frame of video is processed, the proportion of surviving sperm in the current frame is calculated through a custom function, and a data visualization pie chart is created to plot the ratio of surviving to dead sperm. Ultimately, the number and status of sperm can be dynamically tracked during real-time video processing, and a visual sperm survival ratio chart can be generated.

[0020] Furthermore, the formula of the sperm cryopreservation solution described in step 6 is as follows: 650-655 mg of NaCl, 80-85 mg of KCl, 30-35 mg of NaHCO3, 200-210 mg of NaH2PO4·12H2O, 10-15 mg of KH2PO4·3H2O, 80-85 mg of trisodium citrate dihydrate, 350-355 mg of calf serum, 20-25 mg of vitamin C, 15-20 mg of xylitol, 2-3 mg of metformin, 200-210 ml of zinc gluconate, 10-12 ml of DMSO, 8-10 ml of methanol, and 2-5 ml of ethylene glycol are added to every 100 ml of double-distilled water; the volume ratio of the sperm diluent to the cryopreservation solution is 1:2.

[0021] Furthermore, the activation solution in step 6 contains 200-250 mg of NaHCO3 and 10-15 mg of MgCl2 per 100 ml, and is made up to 100 ml with double distilled water; the volume ratio of the sperm freezing mixture to the activation solution is 2:1.

[0022] Furthermore, the long-term ultra-low temperature storage time in step 6 includes 24 hours, 4 days, 8 days and 16 days.

[0023] Furthermore, the gradient cooling in step 6 is to balance the sperm freezing mixture on ice for 10 minutes, then fumigate it 7 cm above the liquid nitrogen surface for 10 minutes, and then balance it on the liquid nitrogen surface for 5 minutes; the gradient heating is to balance the sperm freezing mixture at the liquid nitrogen port for 5 minutes, and then thaw it in a water bath at 28°C until it is completely dissolved.

[0024] Beneficial Effects: Compared with the prior art, the present invention has the following significant advantages: (1) The computer vision-based method for determining the activity of Takifugu obscurus sperm cryopreservation fluid provided by the present invention applies artificial intelligence to the field of aquaculture. By utilizing advanced image processing and deep learning algorithms, sperm activity detection can be automatically performed on a computer by simply providing video data of sperm activity. This method no longer relies on traditional hardware equipment for cumbersome biological testing and is convenient for installation and use on various types of computers. It greatly simplifies the Takifugu obscurus sperm activity detection process. Through automated identification and analysis, it not only improves detection efficiency but also significantly reduces dependence on manual operation, thereby reducing labor costs.

[0025] (2) The best combination of freezing preservation solution and activation solution for Takifugu obscurus was screened out by the method provided by the present invention. The effective activity of Takifugu obscurus sperm after being stored in liquid nitrogen for 4 days reached 85.27%. Moreover, after being thawed and revived for 16 days in liquid nitrogen, the sperm still had more than 80% of stable activity, which fully met the time required for conventional logistics transportation and the actual production needs of sperm activity, and avoided huge damage to the living male parent during logistics transportation. This not only reduced logistics risks and cost losses, but also provided technical means for expanding the scope of use of improved male parents and protecting their germplasm resources, thereby improving the genetic diversity, reproductive efficiency and genetic quality of Takifugu obscurus. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a map of the sperm ID distribution of Takifugu obscurus in the integrated development environment.

[0027] Figure 2 This is a set of visualization results of the sperm motility of the dark-spotted pufferfish in the integrated development environment.

[0028] Figure 3 This is a group diagram of sperm activity at 24h, 4d, 8d, and 16d in Example 1 (without activation solution).

[0029] Figure 4 This is a group diagram of sperm activity after activating solution 1 was applied to Example 1 at 24h, 4d, 8d, and 16d.

[0030] Figure 5 This is a group diagram of sperm activity after activating solution 2 was applied to Example 1 at 24h, 4d, 8d, and 16d.

[0031] Figure 6 This is a group diagram of sperm activity after activating solution 3 was applied to Example 1 at 24h, 4d, 8d, and 16d.

[0032] Figure 7 This is a group diagram of sperm activity at 24h, 4d, 8d, and 16d of Example 2 (without activation solution).

[0033] Figure 8 This is a group diagram of sperm activity after activating solution 1 was applied to Example 2 at 24h, 4d, 8d, and 16d.

[0034] Figure 9 This is a group diagram of sperm activity after activating solution 2 was applied to Example 2 at 24h, 4d, 8d, and 16d.

[0035] Figure 10 This is a group diagram of sperm activity after activating solution 3 was applied to Example 2 at 24h, 4d, 8d, and 16d.

[0036] Figure 11This is a set of diagrams showing the detection of sperm activity rates of dusky puffer fish stored at ultra-low temperatures for 24 hours, 4 days, 8 days, and 16 days. DETAILED DESCRIPTION

[0037] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0038] The present invention provides two fish sperm cryopreservation solution formulas and three activation solution formulas for parallel testing, and uses a computer vision-based method for determining the activity of Takifugu obscurus sperm cryopreservation solution to perform detailed detection and analysis of sperm activity (steps as shown in FIG). Figure 1 As shown in the figure, we finally obtained sperm recovery activity test results with different effects. Taking time factors into consideration, we determined the optimal formula of sperm freezing preservation solution and activation solution. The sperm recovery activity obtained under this formula is the highest.

[0039] The following will further clearly and completely introduce the specific technical solutions and the inventive points of this invention in conjunction with the embodiments of the present invention; the specific embodiments are as follows:

[0040] Example 1:

[0041] Step 1. In November 2024, 5 sexually mature male dark-striped pufferfish were shipped back to Nanjing by Zhongyang Group Co., Ltd. and temporarily raised in the laboratory to simulate the natural day and night environment; 1 dark-striped pufferfish male with fully mature gonadal development was selected, 5 ml of Fisherman's Treasure was poured into the male fish's water body to anesthetize it for 5 minutes, and then placed on ice for dissection; the testis was removed with tweezers in a culture dish, and the blood vessels and other connective tissues attached to the testis were scraped off; then 2 times the volume of the testis was added with PBS buffer containing 1% penicillin-streptomycin for thorough grinding, and then filtered with a 40-mesh sieve, and the filtration was repeated three times. The filtrate was taken as the fresh semen diluent, and then it was pipetted into 2 50 ml test tubes and placed in a 4 ° C refrigerator for precooling for 30 minutes for use; then the fresh semen activity test was performed to see whether its sperm activity exceeded 95%, and whether it could be used for the subsequent dark-striped pufferfish sperm freezing preservation solution formula test; therefore, the precooled fresh semen diluent was taken out of the refrigerator After being placed at room temperature for 6 hours, the sperm was activated with pure water, and slices were prepared with a ratio of pure water to fresh sperm diluent = 1:2. The fresh sperm activity was observed under a microscope, and a fresh sperm activity video was recorded. The activity was efficiently detected by a computer vision-based sperm activity judgment method for dark-striped pufferfish designed by the present invention, and the final fresh sperm effective activity was higher than 95%, so it can be used for subsequent experiments. Then, 1 tube of 50 ml of sperm diluent was fully mixed with 100 ml of sperm freezing preservation solution, and 20 ml of antifreeze was added to obtain a final sperm freezing mixture. The mixture was pre-cooled in a 4°C refrigerator for 30 minutes. Then, 6 tubes were pipetted into 2 ml cryopreservation tubes, with a sample volume of 1 ml per tube. The 6 cryopreservation tube samples were subjected to gradient cooling, that is, they were placed on ice for 10 minutes, then fumigated 7 cm above the liquid nitrogen surface for 10 minutes, then balanced on the liquid nitrogen surface for 5 minutes, and finally placed in liquid nitrogen for storage for 24 hours, 4 days, 8 days and 16 days respectively.

[0042] Among them, the preparation method of sperm cryopreservation solution is as follows: take NaCl 700mg, KCl 80mg, NaHCO3 220mg, CaCl2·2H2O 10mg, MgCl2·6H2O 14mg, D-Glucose 50mg, and raffinose 200mg as the base solution, take fetal bovine serum 350mg, L-arginine monohydrochloride 30mg, maltitol 12mg, and streptomycin sulfate 10mg as the active preservation solution, take DMSO 12ml, glycerol 4ml, propylene glycol 1ml, and dimethylacetamide 0.2ml as antifreeze, add double distilled water 100ml, and add hydrochloric acid to adjust the pH to 8.0, then fully dissolve and fix the volume to obtain sperm cryopreservation solution.

[0043] Step 2: After 24 h, 4 d, 8 d, and 16 d, the cryotube samples were removed from the liquid nitrogen and subjected to gradient temperature increase: equilibrated at the liquid nitrogen port for 5 min, then thawed in a 28°C water bath until completely dissolved, and equilibrated at room temperature for 5 min; a group without activation solution and a group with activation solution were set up respectively. The activation solution group mixed the activation solution and the sperm freezing mixture in a ratio of 1:2. A small amount of the mixture was aspirated and sliced, which was placed under an electron microscope to observe sperm activity and record a video for 15 s.

[0044] Among them, the formula of activation solution 1 is: NaHCO3500mg, NaCl 700mg, add double distilled water to 100ml, add hydrochloric acid to adjust the pH value to 8.8, the formula of activation solution 2 is: NaCl 700mg, Na2CO3200mg, add double distilled water to 100ml, add hydrochloric acid to adjust the pH value to 8.8, the formula of activation solution 3 is: NaHCO3200mg, MgCl210mg, add double distilled water to 100ml, add NaOH to adjust the pH value to 8.8;

[0045] Step 3: Based on the computer vision library, the original video file of dark-striped pufferfish sperm activity is input, and a function algorithm of the deep learning framework is constructed to convert the input original sperm activity video image into a grayscale image. The grayscale image is then binarized using a global fixed thresholding method. The video is read frame by frame, and a hard-coded threshold is called to perform grayscale conversion and segmentation on the video image, and the processed binary image is displayed.

[0046] The following steps are used to input the original video file of Takifugu obscurus sperm activity based on the computer vision library and build a function algorithm for the deep learning framework:

[0047] Step 3-1. After copying the original sperm activity video from the computer, enter the integrated development environment and paste it in the root directory venv. Select "Open with" to copy the path, cut it into the designed "input_video_path=" and "output_video_path=" input functions, and run the file.

[0048] Step 3-2: Binarize the original sperm activity color file. Use the rgb_to_gray function to convert the input image from color (RGB format) to a grayscale image to obtain a data image. Then, combine the edge detection image and threshold segmentation image technology, and use the apply_threshold function in the hard-coded threshold function to perform ideal segmentation processing (binarization) on the grayscale image. The conversion formula from RGB image to grayscale image is: MGV = 0.2989 × R + 0.5870 × G + 0.1140 × B (where R, G, and B represent the values of the red, green, and blue color channels of each pixel in the image, respectively).

[0049] Step 3-3, perform color space conversion on each frame of the grayscale sperm activity video (binary_frame) after binary segmentation, and reconvert the obtained single-channel gray image to the three-channel image position of gray and white pixels, with the same value for each channel; and re-encode it into video output, finally obtaining the binary sperm activity video of data type

[0050] Step 4. In a computer integrated development environment, build a deep learning framework that supports video processing and image recognition model loading, and load a sperm activity target detection model that automatically learns high-level spatial features of images and videos. In order to make this judgment method applicable to computers equipped with different central processing units, you can choose to set the CPU or GPU mode according to the computer processing capacity. First, find the CUDA version corresponding to the computer and install cuda. Then open the system command window and enter nvcc-V to configure the environment variables. If the computer is equipped with an NVIDIA graphics card, install the Cudnn program to call the GPU to install Pytorch for sperm activity judgment processing. If the computer only supports CPU processing, after installing Pytorch, use the CPU only option, that is, "pip install torch==2.4.0torchvision==0.19.0torchaudio==2.4.0--index-url https: / / download.pytorch.org / whl / cpu" to install it in the terminal.

[0051] In order to improve the efficiency of target detection in this embodiment, the computer processing capacity is set to GPU mode, and a convolutional neural network (CNN) deep learning algorithm is used to design a model specifically for sperm target detection to perform image feature deep learning. Then, the optimal threshold of each frame image is predicted by the fully connected layer, and data association is performed in combination with the Kalman filter and the Hungarian algorithm. The sperm activity target tracking algorithm is initialized, specifically by using the SORT algorithm (Simple Online and Realtime Tracking) to match each detected target with the target of the previous frame, setting the tolerance threshold of continuous frames, setting parameters to (tracker = Sort (max_age = 10, min_hits = 3)) to control the sperm target to tolerate 10 consecutive frames after loss, otherwise it will be removed, and setting the sperm target to be considered as a valid target requires at least 3 appearances to avoid the sperm target being misjudged as disappeared or lost due to temporary occlusion or leaving the field of view. Then input the binary sperm activity video of the data type obtained in step 3, such as Figure 1 , perform neural convolution operations on the image in the field of view, detect all active sperm targets in each frame of the image, identify each active sperm target in the image, and generate a green rectangular bounding box (boundingbox) for each detected active sperm object according to the given coordinates (x1, y1) and (x2, y2). The border width is set to 2 pixels. The bounding box contains the location information of the object (such as the serial number in the upper left corner of the rectangle). Each bounding box represents the location of a sperm, and the category and confidence of the target can be returned through the algorithm. Finally, with the help of the multi-target (MOT) tracking algorithm, the motion trajectories of multiple sperm are tracked simultaneously, and unique IDs are assigned to them. The motion trajectories of these sperm are tracked in consecutive frames, and their positions are updated in real time to complete the training of the sperm activity motion trajectory model.

[0052] Step 5: Under the background of deep learning, supervised learning methods are used to train the dark-striped pufferfish active sperm model. The displacement information of all active sperm targets in multiple frames is counted, and a mathematical model is established to describe the displacement and movement characteristics of sperm in two-dimensional space. During each iteration, the current position of the sperm is updated in real time. The displacement of active sperm at different time points is calculated by calculating the distance between the previous and next two frames. The formula is d = √[(x2-x1) 2 +(y2-y1) 2]; (x1, y1) is the position at the previous moment, and (x2, y2) is the position at the current moment. Finally, the optical flow method and background subtraction method are used to compare and analyze the statistical data of the pre-trained sperm model with the initial manual detection results of sperm activity, and the survival or death threshold of active sperm is appropriately set to ensure accuracy while avoiding misjudgment of sperm due to small or weak movement. The threshold is set as follows: if displacement>2:alive_ids.add(track_id). When the displacement of the sperm is greater than 2 pixels, we determine that the sperm is still alive, indicating that it has enough activity to move effectively; when the displacement of the sperm is less than 2 pixels, it means that the movement of the sperm shows a weakening trend and may have lost its activity, so we judge it as dead.

[0053] Step 6: Use statistical data to continuously update and record the last position of all active sperm of Takifugu obscurus, distinguish and store the target IDs of living and dead sperm, and calculate the number of living and total sperm tracked; calculate the proportion of living sperm in the current frame through a custom function, such as Figure 2 , create a data visualization pie chart (Sperm Survival Ratio) that plots the ratio of live to dead sperm, showing the relationship between sperm survival and death: live sperm are represented by blue "Alive" and dead sperm are represented by red "Dead". Autopct_format is used to ensure that these percentages are only displayed to two decimal places. Finally, the number and status of sperm can be dynamically tracked during real-time video processing, and a visual sperm survival ratio chart can be generated for easy monitoring and analysis.

[0054] The activity of the frozen mixture of Takifugu obscurus sperm observed under a microscope without activation solution after ultra-low temperature storage in liquid nitrogen for 24h, 4d, 8d, and 16d is shown in the figure below. Figure 3 As shown in Figure 2, the activity of sperm frozen mixtures based on activation solution formulas 1, 2 and 3 after ultra-low temperature storage in liquid nitrogen for 24 hours, 4 days, 8 days and 16 days is shown in Figure 2. Figure 4 、 Figure 5 and Figure 6 The specific activity data of Takifugu obscurus sperm after thawing are shown in Table 1:

[0055] Table 1 Sperm activity of Takifugu obscurus after thawing (%)

[0056]

[0057] As can be seen from Table 1, in the group without activation fluid, the activity of sperm revived after 24 hours was average, and with the extension of ultra-low temperature storage time, the activity of sperm revived showed a sharp decline; in the group with activation fluid, although the activity of sperm revived after 24 hours was good, the activity of sperm revived also declined seriously after the time was extended.

[0058] Example 2:

[0059] Example 2 is basically the same as Example 1, except that:

[0060] In the step of mixing the sperm diluent with the sperm cryopreservation solution, the formula of the sperm freezing mixture is as follows: 652 mg of NaCl, 83 mg of KCl, 30 mg of NaHCO3, 200 mg of NaH2PO4·12H2O, and 10 mg of KH2PO4·3H2O are weighed as the base solution; 80 mg of trisodium citrate dihydrate, 350 mg of calf serum, 20 mg of vitamin C, 15 mg of xylitol, 2 mg of metformin, and 200 ml of zinc gluconate are taken as the active preservation solution; 10 ml of DMSO, 8 ml of methanol, and 5 ml of ethylene glycol are taken as antifreeze agents; 100 ml of double distilled water is added, and the mixture is stirred with a magnetic stirrer for 5 minutes. After sufficient dissolution, hydrochloric acid is added to adjust the pH to 8.0, and the volume is fixed to obtain the sperm cryopreservation solution.

[0061] The activity of the frozen mixture of Takifugu obscurus sperm observed under a microscope without activation solution after ultra-low temperature storage in liquid nitrogen for 24h, 4d, 8d and 16d is shown in the figure. Figure 7 As shown in Figure 2, the activity of sperm frozen mixtures based on activation solution formulas 1, 2 and 3 after ultra-low temperature storage in liquid nitrogen for 24 hours, 4 days, 8 days and 16 days is shown in Figure 2. Figure 8 、 Figure 9 and Figure 10 The specific activity data of the thawed Takifugu obscurus sperm are shown in Table 2:

[0062] Table 2 Sperm activity of Takifugu obscurus after thawing (%)

[0063]

[0064] In this embodiment, the activity of the revived sperm was high after 24 hours, and the activity of the revived sperm slowly decreased with the extension of the ultra-low temperature storage time.

[0065] Combining the above two embodiments with the results of a computer vision-based cryopreservation activity test of Takifugu obscurus sperm in the present invention, it is concluded that Figure 11The visualization results showed that in the absence of activation solution, the dark-striped pufferfish sperm in the sperm cryopreservation solution formula provided in Example 2 still had high activity after 24 hours, 4 days, 8 days and 16 days. Finally, after 16 days of ultra-low temperature freezing, the effective activity still reached more than 60%. In the screening of three different sperm recovery activation solutions, it can be concluded that after the frozen sperm was activated by activation solution 3, the sperm detection activity rate was the highest. After 16 days of ultra-low temperature freezing, the effective activity still reached more than 80%. Therefore, with the help of a computer vision-based method for determining the freezing preservation solution of dark-striped puffer sperm provided by the present invention, we screened out the most effective formula for the freezing preservation solution of dark-striped puffer sperm: weigh 650-655mg of NaCl, 80-85mg of KCl, 30-35mg of NaHCO3, 200-210mg of NaH2PO4·12H20, and 10-15mg of KH2PO4·3H20 as the base solution, 80-85mg of trisodium citrate dihydrate, 350-355mg of calf serum, 20-25mg of vitamin C, 15-20mg of xylitol, 2-3mg of metformin, and 200-210ml of zinc gluconate as the active preservation solution, and DMSO 10-12ml, 8-10ml of methanol, and 2-5ml of ethylene glycol are used as antifreeze agents. After adding 100ml of double-distilled water, stir with a magnetic stirrer for 5 minutes. After fully dissolved, add hydrochloric acid to adjust the pH to 8.0 and make up the volume to sperm freezing preservation solution. The sperm diluent and freezing preservation solution are mixed in a volume ratio of 1:2 to obtain sperm freezing mixture; the activation solution used for recovery is formulated as: NaHCO3 200-250mg, MgCl2 10-15mg, add double-distilled water to make up the volume to 100ml, and add NaOH to adjust the pH to 8.8; when detecting sperm activity under an electron microscope, the volume ratio of sperm freezing mixture to activation solution is 2:1.

[0066] Through two examples of a computer vision-based method for determining the activity of Takifugu obscurus sperm cryopreservation fluid, we determined the optimal formula and preparation method for Takifugu obscurus sperm cryopreservation fluid. This is not only of great significance for year-round artificial breeding and germplasm resource protection of Takifugu obscurus, but also greatly simplifies the sperm activity detection process and reduces labor costs.

Claims

1. A computer vision-based method for determining the activity of Takifugu obscurus sperm cryopreservation solution, characterized in that: The method comprises the following steps: Step 1: The testis of a male Takifugu obscurus with mature gonads is collected, ground, and then a buffer solution is added. The sperm dilution is collected by filtration with a mesh, and the sperm activity is detected under a microscope after activation with pure water, and a video of the original sperm activity is recorded; Step 2: Threshold the sperm activity video, use binarization to adjust its grayscale value, and convert and segment the image in the video into a grayscale image; Step 3: Load the target detection model in the integrated development environment, initialize the target tracker, input the sperm activity video processed in Step 2, draw a detection box and tracking ID for each sperm in the video, perform multi-target tracking and motion trajectory detection, and implement simulation training; Step 4: Based on the displacement statistics of sperm activity simulation training, a displacement threshold for survival or death is set, and the current sperm position is continuously updated in real time. If the displacement exceeds the set threshold, it is judged to be alive; Step 5: Update the total number of sperm sampled globally and the number of surviving sperm in the statistical video in real time, define the percentage formatting function, draw a pie chart of survival and death, and generate a visual survival ratio chart; Step 6: If the effective sperm activity is greater than 95%, the sperm diluent, the dark puffer freezing preservation solution and the antifreeze are mixed to obtain a sperm freezing mixture, which is then subjected to gradient cooling and finally placed in liquid nitrogen for long-term ultra-low temperature storage; after the storage is completed, the sperm freezing mixture is removed from the liquid nitrogen, gradiently heated and thawed, and dark puffer activation solution is added for resuscitation, the effective sperm activity is observed under an electron microscope, and a sperm activity video is recorded; then steps 2 to 5 are repeated.

2. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, wherein: The buffer described in step 1 is PBS buffer containing 1% penicillin-streptomycin.

3. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, wherein: The specific steps of step 2 include: processing the dark-striped pufferfish sperm activity video using a global fixed threshold binarization method, adjusting the grayscale value of each frame image in the video, constructing a function core algorithm of a deep learning framework, combining edge detection image and threshold segmentation image technology, and segmenting the grayscale sperm video image through a hard-coded threshold, and using the weighted image grayscale conversion formula MGV = 0.2989×red+0.5870×green+0.1140×blue to adjust the set predetermined grayscale threshold, thereby dividing the pixels in the image into white and black categories.

4. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, wherein: The specific steps of step 3 include: (1) loading a real-time target detection algorithm in the integrated development environment, setting it to CPU or GPU mode, designing a sperm target detection model using the convolutional neural network (CNN) deep learning algorithm, and initializing a simple and efficient target tracking algorithm based on the Kalman filter and the Hungarian algorithm. Setting parameters to control the continuous frame threshold that the sperm target can tolerate after being lost, otherwise it will be removed, and setting the sperm target to be considered a valid target to appear at least multiple times to ensure that the tracker will not mistakenly track some false targets that may only appear occasionally; (2) A target detection algorithm is used to generate a bounding box represented by a rectangle for each detected sperm, marking its position and serial number. The multi-target tracking (MOT) algorithm is then used to track multiple sperm targets that appear simultaneously in the binary sperm video, recording the unique ID and motion path of each sperm target and updating its position in real time in each frame.

5. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, wherein: The specific steps of step 4 include: (1) sperm activity simulation training based on deep learning, statistical sperm activity displacement information, simulating sperm activity through position changes in a two-dimensional space, setting the current position of each sperm to (x, y) and updating the position as the frame number changes: the displacement of the sperm at each moment is obtained by calculating the distance between the previous and next positions, and the formula is d = √[(x2-x1) 2 +(y2-y1) 2 ]; (x1, y1) is the position at the previous moment, and (x2, y2) is the position at the current moment; (2) Using the optical flow method and background subtraction method, the statistical data of the pre-trained sperm model are compared and analyzed with the initial manual detection results of sperm activity. A displacement threshold for survival or death is set. When the displacement of the sperm is greater than the threshold, the sperm is judged to be alive. If the displacement is less than the threshold, the sperm may have lost its vitality and is judged to be dead. In each iteration, the current position of the sperm is updated in real time.

6. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, wherein: The specific steps of step 5 include: (1) updating and recording the last position of each sperm through statistical data, and storing the sperm IDs of survival, death and total sperm respectively, and calculating the number of survival and total sperm targets tracked; (2) After processing each frame of video, the proportion of surviving sperm in the current frame is calculated through a custom function, and a data visualization pie chart is created to plot the ratio of surviving to dead sperm. Ultimately, the number and status of sperm can be dynamically tracked during real-time video processing, and a visual sperm survival ratio chart can be generated.

7. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, characterized in that: The formula of the sperm cryopreservation solution described in step 6 is NaCl 650-655 mg, KCl 80-85 mg, NaHCO 3 30-35 mg, NaH 2 PO 4 · 12H 2 O 200-210 mg, KH 2 PO 4 · 3H 2 O 10-15 mg, trisodium citrate dihydrate 80-85 mg, calf serum 350-355 mg, vitamin C 20-25 mg, xylitol 15-20 mg, metformin 2-3 mg, zinc gluconate 200-210 ml, DMSO 10-12 ml, methanol 8-10 ml, and ethylene glycol 2-5 ml. Add 100 ml of double-distilled water, mix, and fully adjust to volume. The volume ratio of the sperm diluent to the cryopreservation solution is 1:

2.

8. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, characterized in that: The activation solution in step 6 contains 200-250 mg of NaHCO3 and 10-15 mg of MgCl2 per 100 ml, and is made up to 100 ml with double distilled water; the volume ratio of the sperm freezing mixture to the activation solution is 2:

1.

9. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, characterized in that: The long-term cryopreservation time in step 6 includes 24 hours, 4 days, 8 days and 16 days.

10. The method for determining the activity of Takifugu obscurus sperm cryopreservation solution based on computer vision according to claim 1, characterized in that: The gradient cooling in step 6 is to balance the sperm freezing mixture on ice for 10 minutes, then fumigate it 7 cm above the liquid nitrogen surface for 10 minutes, and then balance it on the liquid nitrogen surface for 5 minutes; the gradient heating is to balance the sperm freezing mixture at the liquid nitrogen port for 5 minutes, and thaw it in a water bath at 28°C until it is completely dissolved.

Images