A method, system and storage medium for sexing copepods

Abstract

This invention relates to the field of aquaculture and intelligent plankton identification technology, and particularly to a method, system, and storage medium for sexing copepods. The method includes the following steps: acquiring a continuous motion image sequence containing at least one live copepod individual; performing target detection and tracking on the continuous motion image sequence to obtain a trajectory point sequence of the individual; extracting motion behavior feature parameters based on the trajectory point sequence of the individual; setting a preset threshold set corresponding to the motion behavior feature parameters, and constructing multiple discrimination conditions based on the preset threshold set; wherein, when at least two of the multiple discrimination conditions are satisfied, the individual is determined to be of the first sex; otherwise, it is determined to be of the second sex. The trajectory feature extraction and combination discrimination and separation purification scheme provided by this invention into a standardized two-level scheme has significant application value.

Description

Technical Field

[0001] This invention relates to the field of aquaculture and intelligent identification technology for plankton, and particularly to a method, system, and storage medium for distinguishing male and female copepods. Background Technology

[0002] Copepods belong to the phylum Arthropoda, class Crustacea, subclass Copepoda, and are important planktonic crustaceans in marine ecosystems. Small copepods are characterized by their small size and high nutritional value, being rich in polyunsaturated fatty acids (PUFAs) as well as highly unsaturated fatty acids such as DHA and EPA. They also contain astaxanthin and various vitamins, and their nutritional value is generally higher than that of rotifers and artichokes. Copepod eggs and early larvae are particularly suitable as food for the first feeding stage of marine fish larvae, and have important applications in aquaculture seedling production and large-scale live bait production.

[0003] For the purebred isolation and propagation of free-spawning small copepods, it is usually necessary to prioritize the selection of mature females with prominent genital segments to establish a stable breeding population. While existing isolation and purification techniques are relatively mature, in practice, "how to quickly and accurately obtain mature females with prominent genital segments" remains a key challenge affecting efficiency. This is especially true for small copepods (Dichroa spp.) with a body length of less than 1 mm, where males are easily aspirated into the wells or culture plates during the isolation and purification process. This can lead to the inability of subsequent individuals in the wells to proliferate or even death, significantly increasing the number of isolation rounds, time costs, and overall isolation efficiency. Therefore, achieving rapid, accurate, and reproducible identification of mature males and females in the initial stages of isolation and purification is one of the key technical issues for improving the purebred isolation efficiency of free-spawning small copepods.

[0004] On the other hand, copepod males and females often differ in swimming behavior and kinematic characteristics (such as range of motion, speed, turning / curvature, pause-burst patterns, etc.), but there is a lack of an automated, standardized, reproducible, and batch-output trajectory feature extraction and discrimination scheme, which makes it difficult to support rapid targeted sorting and large-scale application in the separation and purification process. Therefore, a male-female discrimination technique combining high-speed microscopic imaging and motion trajectory features is needed to improve the efficiency of obtaining mature females and the success rate of separation and purification. Summary of the Invention

[0005] To address the shortcomings of the prior art, the present invention provides a method, system, and storage medium for distinguishing male and female copepods.

[0006] The first aspect of the present invention provides a method for distinguishing male and female copepods, comprising the following steps: Obtain a sequence of continuous motion images containing at least one live copepod individual; Target detection and target tracking are performed on the continuous motion image sequence to obtain the trajectory point sequence of the individual; Motion behavior feature parameters are extracted based on the trajectory point sequence of the individual; Set a preset threshold set corresponding to the motion behavior feature parameters, and construct multiple discrimination conditions based on the preset threshold set; Specifically, if at least two of the multiple discrimination conditions are met, the individual is determined to be a first-gender individual; otherwise, it is determined to be a second-gender individual.

[0007] In one embodiment of the present invention, the motion behavior characteristic parameters include the range of motion amplitude. Trajectory dispersion radius Trajectory curvature statistics Frequency of directional change Average speed Instantaneous peak speed , percentage of outbreaks Frequency of acceleration change At least three of the following categories: pause / burst ratio, pause percentage to burst percentage ratio.

[0008] In one embodiment of the present invention, at least one of the following steps is further included: Before extracting the motion behavior feature parameters, the trajectory point sequence is subjected to quality screening; preferably, the quality screening includes at least one of trajectory duration screening, trajectory missing ratio screening, and occlusion and cross processing. Before extracting the motion behavior feature parameters, the continuous motion image sequence is scaled to obtain a pixel-to-length conversion factor, and the coordinates and displacement of the trajectory point sequence are converted into actual physical units based on the conversion factor. Before extracting the motion behavior feature parameters, the trajectory point sequence is preprocessed; preferably, the preprocessing includes at least one of smoothing filtering, missing point interpolation, outlier removal, and time step uniform resampling.

[0009] In one embodiment of the present invention, the trajectory duration filtering includes a trajectory duration of not less than 2 seconds or a trajectory frame count of not less than 400 frames; and / or The trajectory missing ratio screening includes trajectory missing ratios not exceeding 20%; and / or The occlusion and crossover processing includes truncating, renumbering, or removing the corresponding trajectory when individual crossover or occlusion leads to uncertainty of identity.

[0010] In one embodiment of the present invention, the continuous motion image sequence is acquired by high-speed imaging under microscopic imaging conditions, wherein the frame rate of the high-speed imaging is not less than 100 frames / second, and infrared illumination or low-interference illumination is used during the imaging process; and / or The preset threshold set is determined by any of the following methods: Thresholds are determined based on statistics from labeled male and female samples, including the midpoint method of male and female means or the quantile threshold method; or, Search for candidate threshold combinations on the training set, and select the threshold combination that optimizes the classification performance as the preset threshold set; The preset threshold set can be calibrated and adjusted according to species, imaging magnification, frame rate and culture conditions.

[0011] In one embodiment of the present invention, the combined discrimination further includes outputting a discrimination score, which is calculated based on the ratio of the number of satisfied discrimination conditions to the total number of discrimination conditions; When the discrimination score falls into the preset gray area or the confidence level is lower than the preset threshold, a review process is triggered to obtain the final gender discrimination result. The review process includes extending the acquisition of continuous motion image sequences, adjusting the imaging acquisition conditions before acquisition, or performing manual microscopic examination and labeling verification on the individuals to be discriminated.

[0012] In one embodiment of the present invention, a directional coarse sorting is performed based on the external features of the target individual to obtain candidate individuals; The candidate individuals are subjected to the acquisition of the continuous motion image sequence and subsequent steps to output the gender determination result.

[0013] A second aspect of the present invention provides a system for distinguishing male and female copepods, comprising: An imaging acquisition module is used to acquire a sequence of continuous motion images containing at least one live copepod individual; The trajectory construction module is used to perform target detection and target tracking on the continuous motion image sequence to obtain the trajectory point sequence of the individual. The feature extraction module is used to extract motion behavior feature parameters based on the trajectory point sequence of the individual; the motion behavior feature parameters include the range of motion. Trajectory dispersion radius Trajectory curvature statistics Frequency of directional change Average speed Instantaneous peak speed , percentage of outbreaks Frequency of acceleration change At least three of the following categories: pause / burst ratio; A gender discrimination module is used to set a preset threshold set corresponding to the motion behavior feature parameters, and to construct multiple discrimination conditions based on the preset threshold set; Specifically, if at least two of the multiple discrimination conditions are met, the individual is determined to be a first-gender individual; otherwise, it is determined to be a second-gender individual.

[0014] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the processor to perform the method for distinguishing male and female copepods as described in the first aspect of the present invention.

[0015] A fourth aspect of the present invention provides a computer program product comprising a computer program that, when executed by a processor, implements the steps of the method described in the first aspect of the present invention.

[0016] Based on the above, compared with existing technologies, this invention achieves automatic and rapid identification of male and female individuals by acquiring continuous motion image sequences of copepod individuals, constructing trajectory point sequences through target detection and tracking, extracting motion behavior feature parameters, and combining multiple discrimination conditions constructed with preset thresholds for combined discrimination. This scheme significantly reduces reliance on operator experience and solves the key problem of accidental aspiration of males during the isolation and purification process of small copepods with a body length of less than 1 mm, leading to culture failure. It effectively improves the efficiency and accuracy of obtaining mature female individuals. Furthermore, based on reproducible and calibrated trajectory feature extraction and standardized discrimination rules, this method provides an automated, batch-output quality control method for the purebred isolation and large-scale cultivation of freely spawning small copepods, thereby improving the success rate of isolation and purification and overall production efficiency.

[0017] Other features and beneficial effects of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other beneficial effects of the invention can be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Unless otherwise specified, the positional relationships shown in the drawings in the following description are based on the direction in which the components are drawn in the figure.

[0019] Figure 1This is a schematic diagram of the module of the copepod sex differentiation system of the present invention; Figure 2 This is a flowchart of trajectory data acquisition for the present invention, and a schematic diagram of multi-target trajectory acquisition process. It shows the steps of target detection, cross-frame association and trajectory point sequence construction, and may include trajectory quality screening and occlusion and cross processing. Figure 3 This is a flowchart of the feature extraction and gender discrimination process of the present invention, showing the process of feature parameter extraction and threshold / combination rule discrimination; Figure 4 Example of a female individual's movement trajectory; the trajectory is a 3D visualized trajectory; the 3D visualization dimensions can include time, frame number, or other auxiliary dimensions; Figure 5 This is a schematic diagram of the movement trajectory of a male individual (example); the trajectory is a 3D visualized trajectory; the 3D visualization dimensions can include time, frame number, or other auxiliary dimensions; Figure 6 This is a schematic diagram of the imaging acquisition device arrangement provided in an embodiment of the present invention, showing the relative arrangement of the high-speed microscopic camera, the lens assembly, and the infrared illumination assembly; Figure 7 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present invention.

[0020] Figure label: 1. Light-shielding module; 2. Illumination module; 3. Sample container; 4. Sample carrying module; 5. Imaging acquisition module; 6. Data processing module; 7. Light-shielding module; 8. Illumination module; 9. Sample container; 10. Sample carrying module; 11. Imaging acquisition module; 12. Data processing module; 21. Processing device; 22. Read-only memory; 23. Random access memory; 24. Bus; 25. I / O interface; 26. Input device; 27. Output device; 28. Storage device; 29. ​​Communication device. 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. The technical features designed in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] In the description of this invention, it should be noted that all terms used in this invention (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains, and should not be construed as limiting the invention; it should be further understood that the terms used in this invention should be understood to have the same meaning as those in the context of this specification and in the relevant field, and should not be understood in an idealized or overly formal sense, except as expressly defined in this invention.

[0023] This invention aims to provide a method, system, and storage medium for sexing copepods. It acquires motion image sequences through microscopic imaging, constructs trajectory point sequences, and extracts kinematic / geometric feature parameters. Based on preset thresholds and combined discrimination rules, it automatically outputs sex identification results, reducing reliance on human experience and improving efficiency. Furthermore, as a supplement to the purebred isolation and culture route for freely ovipositing copepods (e.g., *Dichthyophthiriidae*), this invention further focuses on the crucial operational step of "selecting mature females with prominent genital segments," providing a more easily implemented and standardized sexing and directional sorting scheme. This reduces the probability of separation failure due to missorting and improves the efficiency of obtaining mature females and the success rate of separation and purification.

[0024] An embodiment of the first aspect of the present invention provides a method for distinguishing male and female copepods, comprising the following steps: Step 1: Obtain a sequence of continuous motion images containing at least one live copepod individual; The continuous motion image sequence is acquired by high-speed imaging under microscopic imaging conditions. The frame rate of the high-speed imaging is not less than 100 frames / second, and infrared illumination or low-interference illumination is used during the imaging process.

[0025] Step 2: Perform target detection and target tracking on the continuous motion image sequence to obtain the trajectory point sequence of the individual; This trajectory point sequence

[0026] in, For time or frame number, , Spatial coordinates; Optionally, before extracting the motion behavior feature parameters, the continuous motion image sequence is scaled to obtain a pixel-to-length conversion coefficient, and the coordinates and displacement of the trajectory point sequence are converted into actual physical units based on the conversion coefficient, thereby ensuring that the feature parameters are comparable and reproducible under different imaging magnifications and different acquisition conditions.

[0027] Optionally, before extracting the motion behavior feature parameters, the trajectory point sequence is subjected to quality screening; preferably, the quality screening includes at least one of trajectory duration screening, trajectory missing ratio screening, and occlusion and cross processing; and / or the trajectory point sequence is preprocessed before extracting the motion behavior feature parameters; preferably, the preprocessing includes at least one of smoothing filtering, missing point interpolation, outlier removal, and time step uniform resampling. To reduce the impact of imaging noise, tracking jitter, missing points, and cross-occlusion on subsequent feature parameters, and to reduce the impact of noise and occlusion through imaging acquisition, scale calibration, and trajectory quality screening, thereby improving the feasibility and stability of separation and purification on-site.

[0028] Step 3: Extract motion behavior feature parameters based on the trajectory point sequence of the individual; Preferably, the motion behavior characteristic parameters include at least one of the following or a combination thereof: range of motion. Trajectory dispersion radius Trajectory curvature statistics Frequency of directional change Average speed Instantaneous peak speed , percentage of outbreaks Frequency of acceleration change The ratio of pauses to bursts, and the ratio of pause percentage to burst percentage; In a preferred embodiment of the present invention, the motion behavior feature parameters are selected from at least three of the above-mentioned motion behavior feature parameters.

[0029] Step 4: Set a preset threshold set corresponding to the motion behavior feature parameters, and construct multiple discrimination conditions based on the preset threshold set; Specifically, if at least two of the multiple discrimination conditions are met, the individual is determined to be a first-gender individual; otherwise, it is determined to be a second-gender individual.

[0030] In one embodiment of the present invention, a preset threshold set corresponding to the feature parameters is set, and a combined discrimination rule is used to output the gender discrimination result; preferably, a combined discrimination strategy of "being judged as an individual of the first gender if at least two discrimination conditions are met, otherwise being judged as an individual of the second gender" is adopted. The preset threshold can be determined based on the statistics of labeled samples or training, and can be calibrated and adjusted according to species, imaging magnification, frame rate and culture conditions, which does not constitute a limitation of the present invention.

[0031] Preferably, the continuous motion image sequence may contain multiple copepod individuals. The data processing module establishes trajectory point sequences for each individual and extracts feature parameters for each individual, and then outputs gender discrimination results for each individual to achieve multi-target batch recognition.

[0032] It should be noted that this invention takes the continuous motion image sequence and its trajectory data obtained by microscopic imaging as input, and outputs the gender discrimination result under the constraints of scale calibration, trajectory quality control and (optional) verification mechanism. It belongs to the technical solution based on the image acquisition and data processing link, which can produce verifiable and repeatable technical effects in the separation and purification scenario, rather than just abstract rules or pure mathematical / statistical methods.

[0033] The trajectory duration filtering includes a trajectory duration of not less than 2 seconds or a trajectory frame count of not less than 400 frames. The trajectory missing ratio screening includes trajectories with a missing ratio not exceeding 20%. The occlusion and crossover processing includes truncating, renumbering, or removing the corresponding trajectory when individual crossover or occlusion leads to uncertainty of identity.

[0034] The preset threshold set is determined in any of the following ways: Thresholds are determined based on statistics from labeled male and female samples, including the midpoint method of male and female means or the quantile threshold method; or, Search for candidate threshold combinations on the training set, and select the threshold combination that optimizes the classification performance as the preset threshold set; The preset threshold set can be calibrated and adjusted according to species, imaging magnification, frame rate and culture conditions.

[0035] The combined discrimination also includes outputting a discrimination score, which is calculated based on the ratio of the number of satisfied discrimination conditions to the total number of discrimination conditions. When the discrimination score falls into the preset gray area or the confidence level is lower than the preset threshold, a review process is triggered to obtain the final gender discrimination result. The review process includes extending the acquisition of continuous motion image sequences, adjusting the imaging acquisition conditions before acquisition, or performing manual microscopic examination and labeling verification on the individuals to be discriminated.

[0036] In one embodiment of the present invention, a directional coarse sorting is performed based on the external features of the target individual to obtain candidate individuals; The candidate individuals are subjected to the acquisition of the continuous motion image sequence and subsequent steps to output the gender determination result.

[0037] A second aspect of the present invention provides a system for distinguishing male and female copepods, comprising: An imaging acquisition module is used to acquire a sequence of continuous motion images containing at least one live copepod individual. This imaging acquisition module may include a microscopic imaging component and a high-speed camera component, used to acquire the sequence of continuous motion images containing at least one live copepod individual at a preset imaging magnification and frame rate. The frame rate of the high-speed camera component is preferably not less than 100 fps, more preferably not less than 200 fps. In a preferred embodiment of the invention, the system further includes an illumination module for providing stable illumination to improve imaging contrast and reduce environmental fluctuations; preferably, infrared illumination or low-interference illumination is used to reduce the interference of visible light on copepod behavior and improve trajectory stability. The system may optionally incorporate a light-shielding structure, a dark-field environment, or a background suppression structure to reduce background noise.

[0038] The trajectory construction module is used to perform target detection and target tracking on the continuous motion image sequence to obtain the trajectory point sequence of the individual. In a preferred embodiment of the invention, the trajectory construction module includes a transparent sample container and its fixing structure for holding the culture medium and copepod individuals and defining the imaging field of view. The volume and geometry of the sample container can be adjusted according to the species, density, and imaging field of view; in a preferred embodiment, the container volume is approximately 40 mL, and the number of individuals is approximately 20–50. To maintain feasibility under different container sizes and release densities, the system preferably incorporates a scale calibration and threshold recalibration mechanism.

[0039] The feature extraction module is used to extract motion behavior feature parameters based on the trajectory point sequence of the individual; the motion behavior feature parameters include the range of motion. Trajectory dispersion radius Trajectory curvature statistics Frequency of directional change Average speed Instantaneous peak speed , percentage of outbreaks Frequency of acceleration change At least three of the following categories: pause / burst ratio; In a preferred embodiment of the present invention, the feature extraction module includes a processor and a memory, used for target detection and target tracking of the continuous motion image sequence to construct a trajectory point sequence.

[0040] in, For time or frame number, The spatial coordinates are used; and based on the trajectory point sequence, scale calibration, trajectory preprocessing, and / or trajectory quality screening are performed to extract feature parameters for gender determination. These feature parameters include, but are not limited to, the range of motion. (or trajectory dispersion radius) ), Trajectory curvature statistics Frequency of directional change Average speed Instantaneous peak speed , percentage of outbreaks Frequency of acceleration change And the pause / burst ratio.

[0041] A gender discrimination module is used to set a preset threshold set corresponding to the motion behavior feature parameters, and to construct multiple discrimination conditions based on the preset threshold set; Specifically, if at least two of the multiple discrimination conditions are met, the individual is determined to be a first-gender individual; otherwise, it is determined to be a second-gender individual.

[0042] In a preferred embodiment of the present invention, the gender discrimination module is used to call a preset threshold set and a combined discrimination rule to output the gender discrimination result. Preferably, a combined discrimination strategy of "being judged as an individual of the first gender if at least two discrimination conditions are met, otherwise being judged as an individual of the second gender" is adopted; the thresholds can be determined based on the statistics of labeled samples or training, and can be calibrated and adjusted according to species, imaging magnification, frame rate and culture conditions, which does not constitute a limitation of the present invention.

[0043] Furthermore, the copepod sex differentiation system provided by this invention may also include an output and verification module for outputting the sex determination result, and optionally outputting the determination score and / or confidence level; when the determination score falls within a preset gray area or the confidence level is lower than a preset threshold, a verification process is triggered to obtain the final determination result. The verification process includes one or a combination of the following: extending the acquisition time to reacquire the continuous motion image sequence, adjusting the imaging acquisition conditions to reacquire (including frame rate, magnification, illumination mode or exposure parameters), or performing manual microscopic examination and annotation verification on the individuals to be determined.

[0044] Preferably, the system supports multi-target acquisition and batch output: when the continuous motion image sequence contains multiple individuals, the data processing module constructs multiple individual trajectories respectively, and outputs qualified trajectories by combining occlusion / crossing processing and trajectory quality screening rules. Then, the gender discrimination module outputs the gender discrimination results of each individual respectively, so as to realize multi-target batch recognition and group gender structure evaluation.

[0045] A preferred embodiment of the present invention provides the following method for directional sorting of copepods, including: One method is coarse sorting based on external morphological characteristics. For example, classification rules can be established based on the body shape and appearance of the copepods to be separated, allowing for preliminary directional sorting of individuals. For instance, females can be categorized as "short and stout" or "spindle-shaped," while males, due to their more prominent outward extension of appendages during swimming, appear more like a "rhomboid" shape under low magnification. These classification rules are used to improve initial screening efficiency and reduce the probability of accidental aspiration (this description is illustrative and not restrictive). Secondly, there is re-sorting based on behavioral characteristics. For example, in some small Papilionidae species, mature females may exhibit relatively slow circling / rotating movements (which can be "flipping" circling), and short-term jumping acceleration when stimulated; mature males may tend to exhibit a "trembling propulsion" movement. Specific behaviors can be adjusted according to species and imaging conditions.

[0046] Thirdly, there is assisted identification based on kinetic characteristics, such as high-speed microscopic imaging of candidate individuals after initial screening, importing the image sequence into trajectory analysis software (such as Tracker) for tracking, obtaining trajectory point sequences and forming male and female movement feature templates for rapid on-site comparison and judgment.

[0047] Finally, behavioral trajectory verification is performed, such as high-speed microscopic imaging and trajectory analysis of candidate individuals, establishing male and female movement feature templates and comparing / discriminating them to quickly obtain mature female individuals and support male-female pairing experiments.

[0048] Referring below to 7, a schematic diagram of the structure of an electronic device (e.g., a computing device) suitable for implementing some embodiments of the present invention is shown. Figure 7 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0049] like Figure 7 As shown, the electronic device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 21, which can perform various appropriate actions and processes according to a program stored in read-only memory 22 or a program loaded from storage device 28 into random access memory 23. The random access memory 23 also stores various programs and data required for the operation of the electronic device 20. The processing unit 21, read-only memory 22, and random access memory 23 are interconnected via bus 24. Input / output interface 25 is also connected to bus 24.

[0050] Typically, the following devices can be connected to I / O interface 25: input devices 26 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 27 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 28 including, for example, magnetic tapes, hard disks, etc.; and communication devices 29. Communication device 29 allows electronic device 20 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 7 An electronic device 20 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 7 Each box shown can represent a device or multiple devices as needed.

[0051] In particular, according to some embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 29, or installed from storage device 28, or installed from read-only memory 22. When the computer program is executed by processing device 21, it performs the functions defined in the methods of some embodiments of the present invention.

[0052] It should be noted that, in some embodiments of the present invention, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of the present invention, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0053] In some embodiments, the client and server may communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and may interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0054] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire a continuous motion image sequence containing at least one live copepod individual; perform target detection and target tracking on the continuous motion image sequence to obtain a trajectory point sequence of the individual; extract motion behavior feature parameters based on the trajectory point sequence of the individual; set a preset threshold set corresponding to the motion behavior feature parameters, and construct multiple discrimination conditions based on the preset threshold set; wherein, when at least two of the multiple discrimination conditions are satisfied, the individual is determined to be of the first sex; otherwise, it is determined to be of the second sex.

[0055] Computer program code for performing operations of some embodiments of the present invention can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—and conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0056] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0057] The units described in some embodiments of the present invention can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including an imaging acquisition module, a trajectory construction module, a feature extraction module, and a gender discrimination module. The names of these units do not, in certain circumstances, constitute a limitation on the unit itself.

[0058] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0059] This invention provides a system for distinguishing males and females in copepods, including an imaging acquisition module, a trajectory construction module, a feature extraction module, and a sex determination module. Please refer to... Figure 1 The system specifically includes: a light-shielding module 1 for reducing stray ambient light and stabilizing the imaging background; an illumination module 2 for illuminating the sample, preferably using an infrared light source to reduce visible light stimulation; a sample container 3 for containing copepod individuals and culture medium / seawater medium; a sample carrying module 4 for carrying the sample container 3 and maintaining its positional stability; an imaging acquisition module 5 for acquiring continuous motion image sequences; and a data processing module 6 for trajectory construction, parameter extraction, threshold discrimination, and result output, wherein the data processing module 6 includes the aforementioned trajectory construction module, feature extraction module, and gender discrimination module.

[0060] The imaging acquisition module includes a high-speed microscopic camera, a lens assembly, and an infrared illumination assembly. The high-speed microscopic camera can be a VRI-VEO-410L-72G-M or a similar high-speed imaging device, with a frame rate of at least 60 fps, preferably 100–500 fps, and more preferably approximately 200 fps. The lens assembly can be a macro lens or a combination of microscope objectives / tubes, used for imaging copepod individuals. The magnification or equivalent imaging magnification can be set according to the individual size and field of view requirements, preferably within the range of 0.25×–2× (or equivalent reproduction magnification 1:4–2:1) to ensure that the individual's outline edges are identifiable and support subsequent trajectory tracking. Infrared illumination is preferred to reduce the impact of visible light stimulation on behavior and improve imaging stability. Adjustments based on species length, container size, density, and lens working distance are not intended to limit the invention.

[0061] Furthermore, continuous imaging of live copepod individuals within the container is performed within a preset time period, wherein the imaging field of view may contain at least one copepod individual; preferably, it contains multiple individuals. The data processing module performs target detection on each frame of the image, obtains the position information of multiple targets, and establishes a trajectory point sequence for each target through cross-frame correlation, resulting in a time-sorted trajectory set:

[0062] in Indicates the first Individual being tracked. Coordinate units can be mm or μm, converted using a scaling factor if necessary.

[0063] To ensure the reliability of the trajectory, quality screening can be performed: when the trajectory length is less than a preset threshold, the proportion of missing trajectories exceeds a preset threshold, or the identity of an individual is uncertain due to severe occlusion / intersection, the corresponding trajectory will be removed or truncated; parameters will be extracted and gender will be determined for the selected trajectories.

[0064] It should be noted that the extraction and discrimination rules for motion behavior feature parameters can be performed in the following ways: 1. Trajectory point sequence

[0065] Preferred by frame rate Obtain the sampling interval When there are dropped frames or uneven timestamps, It can be calculated from the difference between adjacent timestamps. Valid duration of the trajectory. Ideally, scale calibration should be performed before feature calculation to obtain pixel-to-length conversion factors. (mm / px); Scale calibration can be obtained by photographing a target with a graduated scale or a known size under the same imaging conditions, and the trajectory coordinates and displacement can be converted into actual length units accordingly.

[0066] 2. Range of activity

[0067]

[0068]

[0069] Optional robust scalars:

[0070] Ideally, after scaling, the coordinates should be converted from pixel units to actual length units (mm), thereby... and The unit is mm; the pixel-to-length conversion factor can be obtained through a scale or a known scale. (mm / px). Calculations can be performed in pixel units even without scale calibration, but threshold calibration and discrimination must maintain consistent unit diameter.

[0071] 3. Speed

[0072]

[0073]

[0074] The preferred unit for velocity is mm / s (calculated after dimensional calibration); where Determined by frame rate or timestamp. If the speed is calculated in pixels, the threshold should also be calibrated in pixels per second, keeping them consistent.

[0075] 4. Frequency of directional change

[0076]

[0077] Frequency of directional change The unit is 1 / s, where the turning threshold is... The threshold is configurable and can be determined based on statistics from labeled samples or by searching the training set. Direction vector. It can be obtained by normalizing adjacent displacement vectors, and the specific discrete implementation method is not limited.

[0078] 5. Curvature statistics

[0079]

[0080] curvature This can be achieved using discrete curvature or sliding window fitting curvature methods; the output can be the mean. or median As a statistical measure, stable discrimination can be achieved as long as the curvature calculation method remains consistent within the same dataset and the threshold is calibrated. The differences in implementation mentioned above do not constitute a limitation.

[0081] 6. Frequency of acceleration change

[0082] Preferably, the acceleration is first calculated from the velocity sequence.

[0083] Next, calculate the rate of change of acceleration (jerk).

[0084] And set a preset threshold The frequency of acceleration change is defined as

[0085] in, The effective duration of the trajectory; The unit is At different frame rates or with unequal sampling intervals, Take the time difference between adjacent frames; threshold The determination was made based on imaging and culture conditions.

[0086] 7. Ratio of pauses to bursts

[0087] pause threshold With the burst threshold For calibrable thresholds: a preset fixed threshold can be used, or it can be based on the quantiles of the velocity sequence (e.g.) or The determination is made to adapt to different species, magnification, frame rate and culture conditions. , and It is dimensionless.

[0088] 8. Threshold Determination Method To ensure feasibility and transferability under different species, imaging magnifications, frame rates, and culture conditions, the threshold set used for sex determination in this invention is a preset threshold set, denoted as:

[0089] in, Corresponding activity range amplitude threshold (or dispersion radius threshold). The threshold for the corresponding trajectory curvature statistics. The threshold for frequency change in the corresponding direction. Corresponding to the average speed threshold, Corresponding to the instantaneous peak velocity threshold, Corresponding to the acceleration change frequency threshold, The corresponding outbreak percentage threshold. The above threshold can be determined using one or more of the following methods: (1) Midpoint method of male and female mean Extract the corresponding feature parameters from the gender-labeled sample trajectories, and calculate the mean of the feature for female and male samples respectively. and The threshold is taken as the midpoint of the male and female mean values:

[0090] in Can correspond to Any threshold. The midpoint of the mean method can quickly provide an usable threshold when the sample size is limited, facilitating on-site deployment and iterative calibration.

[0091] (2) Quantile threshold method When the feature parameters exhibit a skewed distribution or contain outliers, quantile thresholds can be calculated separately for male and female samples, and the threshold should be selected based on the expected false positive rate / false negative rate. For example: Use the lower quantile of the male sample distribution as the male discrimination threshold; or The high quantile of the female sample distribution was used as the female exclusion threshold.

[0092] The specific quantile values ​​can be determined based on the actual species and imaging conditions.

[0093] (3) Search for the optimal threshold on the training set The labeled samples are divided into a training set (which can be further divided into a validation set). Under the constraints of a given discrimination rule (e.g., "satisfying at least two conditions"), the threshold set is... A search is performed to optimize the discrimination metrics (such as accuracy, F1 score, or weighted cost function) on the training set, thereby obtaining the optimal combination of thresholds. This approach is suitable for scenarios with a relatively sufficient sample size where higher discrimination performance is desired.

[0094] (4) Determination of the pause / burst threshold For pause / burst related features, the pause threshold With the burst threshold It can be determined by the quantiles of the velocity sequence, for example, taking... , Based on this, the percentage of pauses and the percentage of bursts are calculated. and the pause / breakout ratio; among which and These represent the 10th and 90th quantiles of the velocity sequence, respectively. The quantile values ​​can be adjusted according to practical applications. The thresholds can also be determined through optimal threshold search on the training set and can be adjusted according to species, imaging magnification, frame rate, and culture conditions, without constituting a limitation of the invention. In some embodiments, a fixed threshold can also be used as a simplified implementation (e.g., taking...). or ).

[0095] (5) Calibration and migration instructions The above thresholds are preset thresholds and can be calibrated and adjusted according to different species, culture conditions, target density, imaging magnification, frame rate, container size, and pixel-to-length conversion factor, etc., and do not constitute a limitation on the present invention. Example values ​​of the thresholds can be given in the embodiments. For example, Example 3 is an example threshold for Parvocalanus crassirostris, and Example 4 is an example threshold for Bestiolina zeylonica.

[0096] In a feasible minimum configuration, it is preferable to obtain at least 10 microscopic annotation trajectories for each sex for initial threshold calibration, and to complete the pixel-length conversion coefficient calibration before imaging; when the imaging magnification or frame rate changes, it is preferable to recalibrate the threshold set Θ to ensure the comparability of feature parameters and the consistency of discrimination results under different conditions.

[0097] 9. Judgment Rules This invention preferably employs a combined discrimination strategy of "identifying male (♂) if any two conditions are met, and female (♀) otherwise". Let the threshold set be:

[0098] Each threshold can be determined based on the mean midpoint method, quantile threshold method, or optimal threshold search of the training set as described in 5.4.8, and can be calibrated and adjusted according to different species, imaging magnification, frame rate, and culture conditions, which does not constitute a limitation of the present invention.

[0099] In a preferred embodiment, an individual is identified as a first-sex individual if at least two of the following conditions are met; otherwise, it is identified as a second-sex individual: a) (or ); b) b) (or ); c) c) ; d) d) ; e) e) ; f) f) or .

[0100] To improve the configurability and robustness of the rules across different species and imaging / culture conditions, let the total number of discriminant conditions involved in the discrimination be . The combined discrimination strategy of the present invention can be configured to "satisfy at least..." "If the condition is met, the individual is determined to be of the primary gender; otherwise, the individual is determined to be of the secondary gender." .when This is a preferred implementation method.

[0101] In a further preferred embodiment, in addition to outputting the gender discrimination category, the present invention also outputs a discrimination score or confidence level, used for engineered control of discrimination uncertainty caused by individual behavioral fluctuations, imaging noise, missing trajectories, or boundary samples. The discrimination score can be calculated based on the number of discrimination conditions satisfied and the total number of discrimination conditions, for example, defined as:

[0102] in, To satisfy the number of discriminant conditions, The total number of discrimination conditions involved in the discrimination; or, the discrimination score can be obtained by weighting the comparison results of each feature parameter with its threshold according to a preset weight.

[0103] A preset gray area range (or confidence threshold) can be set. The gray area range can be determined according to any of the following methods: ① When Approaching the discrimination threshold (e.g.) or If the comparison result between the corresponding feature parameter and its threshold shows a boundary, it is determined to fall into the gray area; the boundary can be defined, for example, as the absolute value of the difference between at least one feature parameter value and its threshold not being greater than a preset tolerance. ( ),in These are configurable parameters; ② When the trajectory quality index falls into the preset gray area, it is determined to fall into the gray area; the trajectory quality index may include situations such as the trajectory duration being close to the preset threshold, the trajectory missing ratio being close to the preset threshold, or the effective duration being insufficient after the trajectory is truncated due to cross occlusion. ③ When judging the score Falling into the preset gray area Or the output confidence level is lower than the preset threshold. At that time, it was determined to have fallen into the gray area, among which These are configurable parameters.

[0104] When the discrimination score or confidence level falls within the gray area range, a verification process is triggered to obtain the final discrimination result. The verification process includes one or a combination of the following: extending the acquisition time to reacquire the continuous motion image sequence, adjusting the imaging acquisition conditions (including but not limited to frame rate, magnification, illumination mode, or exposure parameters) and then reacquiring, or performing manual microscopic examination and annotation on the individuals to be discriminated. Through the above-mentioned confidence level output and verification triggering mechanism, this invention can achieve reliable output under controllable errors while maintaining the efficiency of automatic or semi-automatic discrimination, improving repeatability and robustness for different groups and under different conditions.

[0105] Preferably, the first gender individual and the second gender individual correspond to male and female, respectively; however, without changing the discrimination rule structure and threshold calibration method, the correspondence between the first gender and the second gender can also be reversed, without affecting the implementation of the present invention.

[0106] The following describes embodiments of the present invention. These embodiments are exemplary and are only used to explain the present invention, and should not be construed as limiting the invention. Where techniques or conditions are not specified in the embodiments, they are performed according to the techniques or conditions described in the literature in the art or according to the product instructions. Reagents or instruments used, unless otherwise specified, are all commercially available conventional products.

[0107] Example 1 Trajectory acquisition and gender determination under multi-target conditions Copepod individuals are collected by placing them in a transparent sample container, preferably with a volume of about 40 mL and a number of individuals of about 20–50. Stable illumination (preferably infrared illumination) is provided by an illumination module in a darkened environment, and a continuous motion image sequence is acquired by an imaging acquisition module; preferably, high-speed microscopic imaging is used for acquisition. The data processing module performs target detection and multi-target tracking on each frame of the image, obtaining the two-dimensional trajectory point sequence of each individual. .

[0108] To reduce mismatches caused by cross-occlusion, it is preferable to perform occlusion / cross-over processing and trajectory quality screening: when an individual's identity is uncertain due to cross-occlusion or occlusion, the corresponding trajectory is truncated, renumbered, or removed; and a quality threshold is set for the trajectory, with unqualified trajectories being removed. The preferred trajectory quality threshold is: (1) The duration of the trajectory is ≥ 2 s or the number of trajectory frames is ≥ 400 frames (under the condition of approximately 200 fps); (2) The proportion of missing trajectories is ≤ 20%; (3) When individuals are severely occluded / intersected, the corresponding trajectory is truncated or removed.

[0109] Calculate the range of activity for each of the selected trajectories. Curvature statistics Frequency of directional change Average speed Instantaneous peak speed Frequency of acceleration change The system collects characteristic parameters such as pause / burst ratio and outputs gender discrimination results based on preset thresholds and combination rules. The qualified trajectory data obtained can be used for subsequent threshold calibration, batch gender discrimination, and group gender structure assessment.

[0110] Example 2 Two-stage discrimination process in separation and purification In the isolation and purification process of free-laying small copepods (Dichthyophthirius multifiliis), rapid coarse sorting based on morphological characteristics is first performed to obtain candidate female individuals; then, high-speed microscopic imaging and trajectory analysis are used to verify the candidate individuals. Specifically, continuous motion image sequences are acquired from the candidate individuals, trajectory point sequences are constructed, and the data are extracted. Based on the characteristic parameters, the sex determination result is output according to the combination discrimination rules described in 5.4.8.

[0111] This two-stage discrimination process significantly reduces the risk of males being mistakenly sucked into the well plates / culture wells, leading to "non-proliferation or death," and improves the efficiency of obtaining mature female individuals with genital protrusions. Compared to methods that rely solely on manual morphological observation, introducing high-speed camera trajectory verification after coarse morphological classification significantly improves the reliability of sex determination. Validated, this method achieves a discrimination accuracy of over 90% under the aforementioned acquisition conditions, thereby reducing the probability of culture failure due to missorting and improving the efficiency of obtaining mature female individuals.

[0112] Example 3 Threshold examples and discrimination results of *Dichroa hygrophora* Based on the two-level discrimination process described in Example 2, an example threshold set is established for freely ovipositing copepods of the family Parvocalanus (e.g., Parvocalanus crassirostris). The threshold can be determined using the midpoint of the mean of labeled male and female samples, or through optimal threshold search on the training set. The threshold is an example value calibrated under the conditions of imaging magnification, frame rate, infrared illumination, and pixel-length conversion factor in this example. For example, an individual is classified as male (♂) if it meets any two of the following conditions: (1) Average speed ; (2) Peak instantaneous velocity ; (3) Outbreak rate Where burst is defined as instantaneous velocity The state; (4) Frequency of acceleration change ; (5) Range of activities .

[0113] Individuals that do not meet any two of the above criteria are classified as female (♀). The thresholds mentioned are example thresholds and can be calibrated and adjusted according to species, culture conditions, and imaging parameters, and do not constitute a limitation of the present invention. Compared to methods relying solely on manual morphological observation, introducing high-speed camera trajectory verification after coarse morphological classification can significantly improve the reliability of sex determination. Verification has shown that the accuracy rate of sex determination in this embodiment can reach over 90% under the aforementioned acquisition conditions, and the determination effect can be repeatedly obtained under the same process, indicating that the present invention is engineering feasible and reproducible.

[0114] Example 4 Threshold examples and discrimination results of Ceylon dwarf water flea Based on the two-level discrimination process described in Example 2, an example threshold set is established for freely ovipositing copepods of the family Bestiolina (e.g., *Bestiolina zeylonica*). The threshold can be determined using the midpoint of the mean of labeled male and female samples, or through optimal threshold search on the training set. The threshold is an example value calibrated under the conditions of imaging magnification, frame rate, infrared illumination, and pixel-length conversion factor in this example. For example, an individual is classified as male (♂) if it meets any two of the following conditions: (1) Average speed ; (2) Peak instantaneous velocity ; (3) Outbreak rate Where burst is defined as instantaneous velocity The state, for example, is taken (Example burst threshold, calibrable); (4) Frequency of acceleration change ; (5) Range of activities .

[0115] Individuals that do not meet any two of the above criteria are classified as female (♀). The thresholds mentioned are example thresholds and can be calibrated and adjusted according to species, culture conditions, and imaging parameters, and do not constitute a limitation of the present invention. Compared to methods relying solely on manual morphological observation, introducing high-speed camera trajectory verification after coarse morphological classification can significantly improve the reliability of sex determination. Verification has shown that the accuracy rate of sex determination in this embodiment can reach over 90% under the aforementioned acquisition conditions, and the determination effect can be repeatedly obtained under the same process, indicating that the present invention is engineering feasible and reproducible.

[0116] Example 5 Establishment of male and female behavior templates Small copepod individuals of the family Physalis to be separated were selected and high-speed video was acquired under microscopic imaging conditions to obtain a continuous motion image sequence. The image sequence was then imported into trajectory analysis software (such as Tracker) for target tracking to obtain the individual trajectory point sequence. Based on the trajectory, a two-dimensional trajectory diagram can be drawn, and a three-dimensional visualization method can be optionally used to display the trajectory shape, so as to form a motion template that can be intuitively compared.

[0117] Trajectory templates were established for both female and male individuals, and typical movement patterns were recorded for rapid identification during separation and purification, personnel training, and operational standardization. For example, the female template exhibited relatively slow circling / rotating movement characteristics, while the male template exhibited trembling propulsive movement characteristics. Subsequently, during actual separation and purification, operators could compare the results with the morphological coarse classification rules and movement templates to improve the accuracy and efficiency of selecting mature female individuals with genital protrusions.

[0118] Example 6 Multi-target population data collection and sampling verification To verify that the method of this invention is not only effective for a single experimental sample, but also has reproducibility and stable technical effects under different populations and imaging / culture conditions, this embodiment conducts cross-population and cross-species verification on free-laying small copepods (preferably from the family Physalis). The verification follows the same technical process: high-speed microscopic imaging to acquire continuous motion image sequences—target detection and tracking to construct trajectory—scale calibration—feature extraction—threshold calibration—combined discrimination output, and manual microscopic annotation is used as a true value control to statistically analyze discrimination performance indicators.

[0119] (1) Validation subjects and grouping design (cross species / cross population) In one example implementation, two free-laying copepods of the family Parvocalanus were selected as verification subjects: species A is Parvocalanus crassirostris, and species B is Bestiolinazeylonica. Motion image sequences of male and female individuals were collected respectively, and a sex-labeled sample set was established. Based on this, threshold calibration and sex determination were performed to verify the applicability and transferability of the invention to different species within the same family.

[0120] (2) Data collection and trajectory construction (individual labeling + group sampling verification) The continuous motion image sequence is acquired by high-speed photography under microscopic imaging conditions, with a frame rate preferably not less than 100 fps, more preferably not less than 200 fps; the illumination method is preferably infrared illumination or low-interference illumination.

[0121] To balance authenticity and verifiability, this embodiment adopts a verification path of "individual labeling verification + group sampling verification": a) Individual labeling verification: Sex was collected and labeled under a microscope to form a sex-labeled sample set; among them, the number of female samples in species A was 35 and the number of male samples was 20; the number of female samples in species B was 40 and the number of male samples was 26.

[0122] b) Population Sampling Verification: Approximately 20–50 individuals are placed in a 40 mL container for multi-target data collection. Multiple trajectories are constructed through target detection and multi-target tracking. The trajectory results after automatic / semi-automatic discrimination are sampled for verification. In one example, 30 trajectories are selected for microscopic examination and comparison, with 27 trajectories (27 / 30) matching, corresponding to a sampling verification accuracy rate of 90%. Preferably, the sampling verification adopts the "immediate isolation sampling after discrimination - microscopic annotation" method, that is, after the discrimination results are output, the corresponding individuals are isolated and microscopic annotation is completed within a preset short time window to reduce the impact of individual migration, mixing with other individuals, or changes in state on the consistency of verification and improve the traceability of the true control value.

[0123] When cross-occlusion causes uncertainty in identity, the corresponding trajectory is truncated, renumbered, or removed; and trajectory quality is screened: trajectory duration ≥ 2 s or trajectory frame count ≥ 400 frames, trajectory missing ratio ≤ 20%, so as to output reliable trajectories for subsequent analysis.

[0124] (3) Scale calibration, feature extraction and threshold calibration (emphasizing "calibrable" and "reproducible") For each set of data, scale the data to obtain a pixel-to-length conversion factor, and calculate the range of motion. (or dispersion radius) average speed Instantaneous peak speed Curvature statistics Frequency of directional change Frequency of acceleration change , percentage of outbreaks Characteristic parameters such as pause / burst ratio.

[0125] Threshold set Based on the criteria in 5.4.8, the midpoint of the mean method is preferred for rapid calibration; when the sample size is sufficient, the optimal threshold search on the training set can be used to determine the threshold combination. For burst-related parameters, the burst threshold... A fixed threshold or velocity sequence quantile can be used (e.g.) or The calibration is determined.

[0126] (4) Combination of discriminant output and confidence / verification mechanism The combined discrimination strategy described in 5.4.9 is used to output the gender discrimination result: when at least one of the discrimination conditions is met... Item (of which) If the condition is met, the individual is classified as the first gender; otherwise, the individual is classified as the second gender.

[0127] In a preferred embodiment, a discrimination score or confidence level is output simultaneously to handle boundary samples caused by individual behavior fluctuations, imaging noise, or missing trajectories. The discrimination score can be defined as the number of conditions met / the total number of conditions, or obtained by weighting according to feature weights. When the discrimination score is in a preset gray area or the confidence level is lower than a preset threshold, a review process is triggered (e.g., extending the acquisition time to reacquire a longer time sequence, adjusting the imaging conditions to reacquire, or performing manual microscopic examination and annotation verification) to achieve controllable errors and stable output in engineering applications.

[0128] In addition, for individuals / trajectories whose discrimination results fall within the preset gray area or whose confidence level is lower than the preset threshold under group collection conditions, a verification can be performed by "extending collection - re-discrimination - manual microscopic examination and annotation verification when necessary", thereby further reducing the probability of misjudgment under group collection conditions and improving the consistency and stability of discrimination results.

[0129] (5) Verification indicators and results The discrimination results are compared with the true values ​​labeled by manual microscopic examination. At least the accuracy is calculated, and male recall (Recall♂), female recall (Recall♀), and specificity can be calculated to quantify the stability of different species / populations.

[0130] Accuracy is calculated based on the true values ​​of microscopic examination, using the following formula:

[0131] in, The number of samples that were determined to be male and examined under a microscope were male. The number of samples that were identified as female and examined under a microscope were female. This represents the number of samples that were initially identified as male but underwent microscopic examination and were female. The number of samples that were initially identified as female but underwent microscopic examination and were found to be male. .

[0132] In the verification of group sampling, to obtain the true microscopic values ​​of the sampled objects, it is preferable to establish a correspondence using the "verification after discrimination" method: for example, immediately perform microscopic examination and labeling after isolating and sampling the sampled objects, or confirm the targets one by one according to trajectory number within the same container / same field of view, and record it in the form of sampling hits / total sampling (e.g., 27 / 30). The accuracy of the control scheme "morphological coarse classification only" is also calculated based on the true microscopic values ​​to ensure consistent comparison and verifiable results.

[0133] In an example verification: — For species A (Parvocalanus crassirostris), the sex determination accuracy was approximately 95% on a labeled sample set of 35 females and 20 males; — For species B (Bestiolina zeylonica), the sex determination accuracy was approximately 96% on a labeled sample set of 40 females and 26 males; — In the multi-target population sampling review, 27 out of 30 trajectories were matched by microscopic examination (27 / 30), and the sampling review accuracy rate was 90%.

[0134] Compared to a coarse morphology-only control scheme (for example, the accuracy of coarse morphology classification is about 70%–85%), the accuracy of sex determination can be improved to about 95%–96% after introducing trajectory verification by high-speed microscopic imaging. Moreover, it can still achieve a verification hit rate of about 90% in multi-target population sampling verification, thus obtaining a more stable and repeatable discrimination effect in real separation and purification scenarios.

[0135] The above results show that the method of the present invention can still achieve stable and repeatable discrimination results in the collection scenarios of different species and groups within the same family; the threshold set can be calibrated and adjusted according to the threshold determination method in section 8, which does not constitute a limitation.

[0136] Table 1. Comparison of discrimination performance in cross-species and group scenarios (Example data / sampling range)

[0137] Note: The accuracy rates in the table are example data or sampling statistical ranges, and the statistical caliber is based on manual microscopic examination and labeling as the true value. Specific values ​​may be recalibrated and adjusted according to species, batch, and imaging / culture conditions, and do not constitute a limitation of the present invention.

[0138] Example 7 Cross-condition calibration and transfer validation To verify the feasibility and transferability of the method under different imaging and culture conditions, this embodiment uses free-laying small copepods (preferably from the family Physalis) as the object to conduct cross-condition verification. While maintaining the overall workflow of the invention (high-speed microscopic imaging to acquire image sequences—target detection and tracking to construct trajectory—scale calibration—feature calculation—threshold calibration / recalibration—combined discriminant output), key imaging and culture parameters (including but not limited to imaging magnification, frame rate, illumination mode / intensity, container size, and deployment density) are changed. The robustness and consistency of the method are evaluated through threshold recalibration and rule-based parameterization.

[0139] (1) Data acquisition, scale calibration and trajectory acquisition Under different combinations of conditions, the individuals to be tested are placed in a transparent sample container to acquire motion image sequences. Under group acquisition conditions, the preferred container volume is about 40 mL and the number of individuals is about 20–50 per batch; under light-shielding conditions, stable illumination is provided by an illumination module (preferably infrared illumination), and continuous motion image sequences are acquired using a high-speed microscopic camera (the example frame rate is 200 fps, and the frame rate and magnification can be adjusted according to the combination of conditions).

[0140] The trajectory point sequence of each object is obtained through object detection and multi-object tracking. When cross-occlusion occurs, leading to uncertainty about the identity, the corresponding trajectory is truncated, renumbered, or removed. A trajectory quality screening threshold is adopted: trajectory duration ≥ 2 seconds or trajectory frame count ≥ 400 frames, and trajectory missing ratio ≤ 20%, in order to obtain qualified trajectories for subsequent analysis.

[0141] Furthermore, pixel-to-length conversion is performed using a scale or known scale to obtain pixel-to-length conversion coefficients, and displacement, velocity, and acceleration-related parameters are uniformly converted into actual physical units to ensure the comparability of feature parameters under different magnification / frame rate conditions.

[0142] (2) Feature extraction and threshold recalibration Calculate the range of motion for each qualified trajectory. (or dispersion radius) average speed Instantaneous peak speed Curvature statistics Frequency of directional change Frequency of acceleration change , percentage of outbreaks Characteristic parameters, etc.

[0143] Threshold set Based on 5.4.8, the threshold is determined and can be recalibrated according to changes in magnification, frame rate, lighting method, container size, and culture conditions. The midpoint of the mean of male and female labeled samples is preferred for rapid calibration. When the sample size is sufficient, the optimal threshold search on the training set can be used to determine the threshold combination. For outbreak-related features, the outbreak threshold... A fixed threshold or velocity sequence quantile can be used (e.g.) or The calibration is determined.

[0144] (3) Combination discrimination, confidence output and verification methods For each qualified trajectory, gender is determined according to the combined discrimination strategy described in 5.4.9 (meeting at least...). If a given condition is met, the individual is classified as a primary gender; otherwise, they are classified as a secondary gender. In a preferred embodiment, a discrimination score / confidence level is output simultaneously; when the discrimination score / confidence level falls within a preset gray area, a review process is triggered (extending the acquisition time, adjusting imaging conditions for re-acquisition, or manual microscopic examination and annotation) to achieve reliable output under controllable error.

[0145] To address the issue of individual entry and exit from the field of vision and cross-occlusion during group data collection, which makes it difficult to maintain identity over a long period, this embodiment adopts a verification path of "individual annotation verification + group sampling verification": a) Individual labeling verification: Manual microscopic labeling of individuals collected under different combinations of conditions is performed to form a sex-labeled sample set for threshold calibration and accuracy statistics; b) Population sampling verification: Multiple target trajectories were acquired under a population sampling condition of approximately 40 mL. After trajectory screening, the sampled trajectories were verified by "immediate capture microscopy after discrimination" to verify the consistency and stability of the method in real separation and purification scenarios.

[0146] (4) Explanation of cross-condition applicability (echoing the chain of evidence in Examples 3 / 4 / 6) By employing scale calibration, threshold recalibration, and combined discrimination (along with gray area verification), this invention can still output stable sex determination results under different combinations of magnification, frame rate, and culture conditions. The cross-condition verification described is merely an example; the threshold and verification strategies can be calibrated and adjusted according to species differences, imaging parameters, and culture conditions, and do not constitute a limitation of this invention.

[0147] Example 8 Effect verification and control To verify the effectiveness and reproducibility of the method of the present invention in the isolation and purification of free-egg-laying small copepods (e.g., the family Phytopodidae), the method was validated on two target species, and "morphological coarse classification only" was set as a control scheme.

[0148] (1) Control scheme (rough morphological classification only): The operator performs rough sexing based on morphological characteristics under a stereomicroscope, and then verifies the results by microscopic examination and labeling. The results show that the rough morphological classification accuracy of Parvocalanus samples is about 75%–85%; the rough morphological classification accuracy of Bestiolina samples is about 70%–85%.

[0149] (2) The present invention's scheme (coarse morphological segmentation + high-speed trajectory verification): After coarse morphological segmentation, high-speed microscopic imaging is performed on candidate individuals, and the gender result is output according to the trajectory features and combination discrimination rules of the present invention. The result is then verified through microscopic examination and annotation. The results show: For the *Parvocalanus crassirostris* species sample, with 35 females and 20 males, the accuracy rate for sex determination under the stated collection conditions was approximately 95%. For the Bestiolina zeylonica species sample, with 40 females and 26 males, the sex determination accuracy was approximately 96% under the stated collection conditions.

[0150] (3) Group collection sampling verification: Under the conditions of group collection of about 40 mL container and 20–50 individuals, multi-target tracking and trajectory quality screening were used to output qualified trajectories, and 30 of the trajectories were sampled and verified by "immediate capture microscopic examination after discrimination". The microscopic examination hit 27 / 30.

[0151] In summary, compared to methods that rely solely on morphological observation, this invention significantly improves the reliability of sex determination through a two-stage discrimination process of "morphological coarse classification + high-speed trajectory verification," and maintains good performance in different species samples and population collection scenarios. This reduces the risk of mis-absorbing males, leading to non-proliferation / death within the well plate, and improves the efficiency of obtaining mature female individuals from the genital protuberance and the success rate of separation and purification.

[0152] In summary, compared with the prior art, the present invention has at least the following beneficial effects: (1) It can realize automatic or semi-automatic auxiliary identification of male and female copepods, reduce the dependence on operator experience, and improve the efficiency of separation, purification and directional sorting; (2) Supports multi-target trajectory construction and batch output, which can be used for population gender structure assessment, screening and quality control in the expansion process; (3) Introducing a two-level discrimination strategy of “morphological coarse classification + behavior / trajectory verification (or motion template comparison)” into the separation and purification process can significantly reduce the risk of accidental aspiration of males leading to non-proliferation or death in the well plate / culture well, and improve the success rate and efficiency of obtaining mature female individuals with genital protrusions. (4) The feature parameters are clearly defined, the calculation methods are consistent and reproducible; the threshold set can be calibrated and recalibrated according to different species, imaging magnification, frame rate and culture conditions, with strong adaptability and transferability. (5) The robustness of the judgment can be improved by threshold recalibration or training set optimization, and standardized operating procedures and training sample library can be formed for personnel training and process promotion. (6) Combining infrared or low-interference illumination with high-speed microscopic imaging reduces the interference of external stimuli on behavior and improves trajectory stability. Compared with relying solely on manual morphological observation, introducing trajectory verification by high-speed microscopic imaging after morphological coarse classification can significantly improve the reliability of gender discrimination.

[0153] As shown in Examples 3–4 and 6 and Table 1, the present invention achieves stable discrimination results in different species and populations of the free-laying Paecilomyces family, and significantly improves upon the morphological coarse-classification control scheme. This demonstrates that the present invention is not only statistically correlated with a single sample, but also possesses reproducible engineering technology effects. Verification shows that under the collection and processing conditions described in this invention, the method can achieve stable discrimination results: collecting samples under single-individual conditions and constructing a labeled sample set through manual microscopic annotation (examples could be 20 females and 20 males, or a larger sample size labeled set can be used), extracting the range of activity. Average speed Instantaneous peak speed Frequency of acceleration change , percentage of outbreaks Characteristic parameters such as gender are used to determine sex based on a preset threshold set and combined discrimination rules. The discrimination performance is statistically analyzed by comparing the true values ​​under a microscope. At the same time, under the condition of approximately 40 mL of population collection, traceable trajectory segments are obtained through trajectory quality screening. The extracted trajectories / individuals are sampled and verified using a "post-discrimination verification" method. The verification method includes, but is not limited to: immediate isolation and sampling after discrimination and microscopic annotation, or confirmation of each individual according to the trajectory number, to verify the consistency and stability of the method in real separation and purification scenarios.

[0154] In an example validation: for freely ovipositing copepods of the family Parvocalanus, the sex identification accuracy for species A, *Parvocalanus crassirostris*, was approximately 95% on a microscopically labeled sample set of 35 females and 20 males; for species B, *Bestiolina zeylonica*, the sex identification accuracy was approximately 96% on a microscopically labeled sample set of 40 females and 26 males. Furthermore, in the sampling verification of multi-target population collections, 30 tracks were selected for microscopic comparison, with 27 hits (27 / 30), corresponding to a sampling verification accuracy of 90%. As a control, when using only morphological coarse classification, the coarse classification accuracy for species A was approximately 75%–85%, and for species B, it was approximately 70%–85%. These results demonstrate that the present invention achieves stable and reproducible discrimination results in collection scenarios involving different species and populations within the same family.

[0155] It should be noted that the above-mentioned technical effects are not simply observational conclusions of biological phenomena, nor are they merely statistical correlations. Rather, they are based on a repeatable, engineered process: This invention obtains reproducible motion image sequences through high-speed microscopic imaging, constructs trajectories through target detection and tracking, and combines scale calibration, trajectory quality screening, and feature parameter extraction with a calibrable threshold set and "at least..." The system outputs gender results based on a combination of criteria that satisfy certain conditions. When a sample is in a boundary situation due to fluctuations in individual behavior, imaging noise, or missing trajectories, an optional discrimination score / confidence level can be output, triggering a gray area verification mechanism to control errors. Thus, this invention enables engineered, reproducible, and transferable sex determination output under different populations and imaging / culture conditions.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for distinguishing males and females of copepods, characterized in that, Includes the following steps: Obtain a sequence of continuous motion images containing at least one live copepod individual; Target detection and target tracking are performed on the continuous motion image sequence to obtain the trajectory point sequence of the individual; Motion behavior feature parameters are extracted based on the trajectory point sequence of the individual; Set a preset threshold set corresponding to the motion behavior feature parameters, and construct multiple discrimination conditions based on the preset threshold set; Specifically, if at least two of the multiple discrimination conditions are met, the individual is determined to be a first-gender individual; otherwise, it is determined to be a second-gender individual.

2. The method for distinguishing male and female copepods according to claim 1, characterized in that, The motion behavior characteristic parameters include the range of motion. Trajectory dispersion radius Trajectory curvature statistics Frequency of directional change Average speed Instantaneous peak speed , percentage of outbreaks Frequency of acceleration change At least three of the following categories: pause / burst ratio, pause percentage to burst percentage ratio.

3. The method for distinguishing male and female copepods according to claim 1, characterized in that, It also includes at least one of the following steps: Before extracting the motion behavior feature parameters, the trajectory point sequence is subjected to quality screening; preferably, the quality screening includes at least one of trajectory duration screening, trajectory missing ratio screening, and occlusion and cross processing. Before extracting the motion behavior feature parameters, the continuous motion image sequence is scaled to obtain a pixel-to-length conversion factor, and the coordinates and displacement of the trajectory point sequence are converted into actual physical units based on the conversion factor. Before extracting the motion behavior feature parameters, the trajectory point sequence is preprocessed; preferably, the preprocessing includes at least one of smoothing filtering, missing point interpolation, outlier removal, and time step uniform resampling.

4. The method for distinguishing male and female copepods according to claim 3, characterized in that, The trajectory duration filtering includes a trajectory duration of not less than 2 seconds or a trajectory frame count of not less than 400 frames; and / or The trajectory missing ratio screening includes trajectories with a missing ratio not exceeding 20%; and / or The occlusion and crossover processing includes truncating, renumbering, or removing the corresponding trajectory when individual crossover or occlusion leads to uncertainty of identity.

5. The method for distinguishing male and female copepods according to claim 1, characterized in that, The continuous motion image sequence is acquired by high-speed imaging under microscopic conditions, wherein the frame rate of the high-speed imaging is not less than 100 frames / second, and infrared illumination or low-interference illumination is used during the imaging process; and / or The preset threshold set is determined by any of the following methods: Thresholds are determined based on statistics from labeled male and female samples, including the midpoint method of male and female means or the quantile threshold method; or, Search for candidate threshold combinations on the training set, and select the threshold combination that optimizes the classification performance as the preset threshold set; The preset threshold set can be calibrated and adjusted according to species, imaging magnification, frame rate and culture conditions.

6. The method for distinguishing male and female copepods according to claim 1, characterized in that, The combined discrimination also includes outputting a discrimination score, which is calculated based on the ratio of the number of satisfied discrimination conditions to the total number of discrimination conditions; When the discrimination score falls into the preset gray area or the confidence level is lower than the preset threshold, a review process is triggered to obtain the final gender discrimination result. The review process includes extending the acquisition of continuous motion image sequences, adjusting the imaging acquisition conditions before acquisition, or performing manual microscopic examination and labeling verification on the individual to be discriminated.

7. The method for distinguishing male and female copepods according to claim 1, characterized in that, Targeted coarse sorting is performed based on the physical characteristics of the target individual to obtain candidate individuals; The candidate individuals are subjected to the acquisition of the continuous motion image sequence and subsequent steps to output the gender determination result.

8. A system for distinguishing males and females in copepods, characterized in that, include: An imaging acquisition module is used to acquire a sequence of continuous motion images containing at least one live copepod individual; The trajectory construction module is used to perform target detection and target tracking on the continuous motion image sequence to obtain the trajectory point sequence of the individual. The feature extraction module is used to extract motion behavior feature parameters based on the trajectory point sequence of the individual; the motion behavior feature parameters include the range of motion. Trajectory dispersion radius Trajectory curvature statistics Frequency of directional change Average speed Instantaneous peak speed , percentage of outbreaks Frequency of acceleration change At least three of the following categories: pause / burst ratio; A gender discrimination module is used to set a preset threshold set corresponding to the motion behavior feature parameters, and to construct multiple discrimination conditions based on the preset threshold set; Specifically, if at least two of the multiple discrimination conditions are met, the individual is determined to be a first-gender individual; otherwise, it is determined to be a second-gender individual.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it causes the processor to perform the method for distinguishing male and female copepods as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

Images