Liquid drop wrapped cell detection method and device based on PMT signal and AI, and medium

Abstract

The invention discloses a liquid drop wrapped cell detection method and device based on a PMT signal and AI and a medium. The method comprises the following steps: collecting a PMT waveform signal of a liquid drop detection area; performing peak detection on the PMT waveform signal, and intercepting a periodic waveform; performing feature extraction on the periodic waveform through an artificial intelligence neural network to obtain a waveform feature of one period; and carrying out feature classification on the waveform features, and outputting a classification result as a detection result. According to the method, waveform data generated by liquid drop wrapping cells can be effectively extracted through a waveform feature extraction algorithm for realizing dynamic parameter adjustment, automatic extraction of waveform features is performed by using an artificial intelligence neural network CNN, and the problem that a traditional PMT waveform classification algorithm is insufficient in complex waveform processing capability is effectively solved. Compared with a traditional wave crest detection algorithm, the method has the advantages that the PMT waveform characteristics of the liquid drop wrapped single cells can be captured more accurately, and a more accurate classification result of the liquid drop wrapped cells is given.

Description

Technical Field

[0001] This invention relates to the field of single-cell signal processing technology, specifically to a method, device, and medium for detecting droplet-encapsulated cells based on PMT signals and AI. Background Technology

[0002] The droplet microfluidic single-cell analysis and screening system is a high-throughput cell screening device based on droplet microfluidic technology. Due to its high throughput, low consumption and high precision, it has broad application prospects in biomedical research, drug screening and environmental monitoring.

[0003] Traditional droplet microfluidic single-cell screening systems convert optical images into electrical signals, then use waveform analysis to classify and screen microdroplets for those containing single cells. However, in actual detection processes, the diversity of cell size and morphology, as well as subtle differences in droplet formation (such as flow rate fluctuations and pressure changes), lead to significant variability and noise interference in the detected waveforms. Traditional methods struggle to adapt to these complex variations, making misjudgments and missed detections common. Summary of the Invention

[0004] This invention aims to solve the technical problems mentioned in the background section. This invention provides an adaptive feature extraction and classification method for single-cell encapsulated droplet waveforms. This method utilizes a dynamically adjusted waveform feature extraction algorithm and a learned classification model to effectively overcome the limitations of traditional classification algorithms in handling complex waveforms. Based on the classification model, a single-cell sorting method is also provided.

[0005] The first aspect of this invention provides a method for detecting droplet-encapsulated cells based on PMT signals and AI, comprising the following steps: Collect PMT waveform signals from the droplet detection area; Peak detection is performed on the PMT waveform signal, and periodic waveforms are extracted; The waveform features are obtained by extracting features from the periodic waveform using an artificial intelligence neural network. The waveform features are classified, and the classification results are output as the detection results.

[0006] Furthermore, the acquisition of the PMT waveform signal in the droplet detection area specifically includes the following steps: When the droplet-encapsulated cells flow through the PMT detection area, the fluorescence signal emitted by the droplet-encapsulated cells in the area is collected by a photomultiplier tube and converted into an electrical signal; The waveform of the electrical signal is extracted to obtain the waveform signal as the PMT waveform signal for the formation of cells encapsulated by droplets.

[0007] Furthermore, the extraction of the waveform of the electrical signal specifically includes the following steps: Create a sliding window of a preset size; The sliding window is used to slide through the time series data of the electrical signal to obtain multiple sets of window data; The window data is statistically analyzed and sorted to obtain the waveform signal of the electrical signal.

[0008] Furthermore, the step of performing peak detection on the PMT waveform signal and extracting the periodic waveform specifically includes the following steps: For each waveform data point in the PMT waveform signal x i A waveform peak is identified when all of the following conditions are met: ; in, d min Indicates the minimum peak spacing. w min Indicates the minimum peak width. p min Indicates the minimum significance; h threshold The dynamic peak height threshold is calculated using the following formula: ; in, h base It is the basic threshold. α It is a noise factor; σ noise The noise level of the waveform data points is calculated using the following formula: ; in is the waveform mean, and N is the waveform signal length; Based on the identified waveform peak values, waveform feature data from multiple cycles are extracted as periodic waveforms.

[0009] Furthermore, the artificial intelligence neural network is a convolutional neural network; the step of extracting features from the periodic waveform using the artificial intelligence neural network to obtain waveform features specifically includes the following steps: Receive periodic waveform; Local features are extracted from the periodic waveform to obtain the local features of the droplet-encapsulated cell waveform as the first feature; The first feature is subjected to dimensionality reduction processing; The first feature after dimensionality reduction is subjected to high-order feature extraction to obtain the high-order feature of the droplet-encapsulated cell waveform as the second feature. The second feature is nonlinearly mapped to obtain the waveform features of the droplet-encapsulated cell waveform.

[0010] Furthermore, the convolutional neural network is trained through the following steps: Obtain the droplet-encapsulated cell cycle waveform for training, as the training cycle waveform; The waveforms of the training cycles are manually labeled and divided into training datasets and validation datasets. The preset convolutional neural network is trained using the training dataset, and the accuracy and recall of the output of the convolutional neural network are verified using the validation dataset. When the accuracy and recall of the convolutional neural network output reach the preset requirements, it is determined that the convolutional neural network has completed training.

[0011] Furthermore, the feature classification of the waveform features specifically includes the following steps: When the waveform feature has a first activation mode, the droplet-encapsulated cell is classified as an empty package. When the waveform feature has a second activation mode, it is determined that the droplet-encapsulated cell is classified as a single package. When the waveform feature has a third activation mode, it is determined that the droplet-encapsulated cell is classified as a multi-packet type.

[0012] A second aspect of the present invention discloses an electronic device, including a processor and a memory; The memory is used to store programs; The processor executes the program to implement a method for detecting droplet-encapsulated cells based on PMT signals and AI as described in the first aspect.

[0013] A third aspect of the present invention discloses a computer-readable storage medium storing a program that is executed by a processor to implement a method for detecting droplet-encapsulated cells based on PMT signals and AI as described in the first aspect.

[0014] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned method.

[0015] The embodiments of the present invention have the following beneficial effects: The present invention provides a method, device, and medium for detecting droplet-encapsulated cells based on PMT signals and AI, which fully considers the actual parameters during the droplet generation process and realizes dynamic adjustment of peak detection parameters. The present invention, through a waveform feature extraction algorithm that achieves dynamic parameter adjustment, can effectively extract waveform data generated by droplet-encapsulated cells, and uses an artificial intelligence neural network for automatic waveform feature extraction, effectively overcoming the problem of insufficient processing capability of traditional classification algorithms for complex waveforms. Compared with traditional peak detection algorithms, the present invention can more accurately capture the waveform features of single-cell-encapsulated droplet waveforms, providing more accurate droplet-encapsulated cell classification results, and improving the detection accuracy by more than 25%.

[0016] Additional aspects and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description or may be learned by practice of the invention. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of the basic steps of a droplet-encapsulated cell detection method based on PMT signal and AI according to the present invention; Figure 2 This is a schematic diagram of the steps for detecting the cell peak encapsulation in the present invention; Figure 3 This is a schematic diagram of the structure of the convolutional neural network of the present invention; Figure 4 This is a schematic diagram of the convolutional neural network training process of the present invention; Figure 5 This is an illustration of the effect of droplets encapsulating cells in an empty package embodiment of the present invention; Figure 6 This is a waveform diagram of cell cycle encapsulated by droplets in an empty package embodiment of the present invention; Figure 7 This is a diagram illustrating the effect of droplets encapsulating cells in a single-package embodiment of the present invention; Figure 8 This is a waveform diagram of cell cycle encapsulated by droplets in a single-package embodiment of the present invention; Figure 9 These are illustrations of the effect of droplets encapsulating cells in multiple embodiments of the present invention; Figure 10 These are droplet-encapsulated cell cycle waveforms from multiple embodiments of the present invention; Figure 11This is a schematic diagram of the structure of an electronic device according to the present invention; Figure 12 This is a schematic diagram of a computer-readable storage medium structure according to the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] Traditional droplet microfluidic single-cell screening systems use optical images to convert them into electrical signals, and then use the waveform analysis of these electrical signals to classify and screen microdroplets that contain single cells.

[0021] In waveform analysis, methods such as simple threshold comparison of certain waveform parameters and fixed template matching have limited processing capabilities for detecting droplets containing single cells. In actual detection processes, the diversity of cell size and morphology, as well as subtle differences in droplet formation (such as flow rate fluctuations and pressure changes), lead to significant variability and noise interference in the detected waveforms. Traditional methods struggle to adapt to these complex variations, making misclassification and missed detections common. For example, threshold comparison methods may misclassify droplets containing multiple cells or even those without cells as single-cell-containing droplets, severely impacting the accuracy and reliability of experimental results.

[0022] Regarding waveform classification, current methods largely rely on manual feature extraction, such as manually extracting simple features like peak value, width, and area, and then using traditional classification algorithms to classify the waveforms. This approach not only consumes a significant amount of manpower and time, but the extracted features often fail to fully reflect the essential characteristics of the waveform, resulting in insufficient generalization ability of the classification model. When faced with different waveform features generated under different experimental conditions (such as different cell types and different droplet generation devices), the accuracy of traditional classification algorithms drops significantly, failing to meet the high-precision and high-efficiency requirements of single-cell high-throughput screening and biopharmaceutical droplet detection.

[0023] Furthermore, current droplet detection systems typically operate with waveform acquisition, feature extraction, and classification processes operating independently, lacking effective collaborative optimization. Redundancy in data transmission and processing between these stages leads to inefficiency throughout the detection process, hindering real-time and rapid droplet detection. Moreover, this fragmented system architecture hinders system maintenance and upgrades; traditional methods of waveform analysis and classification to determine whether single-cell microdroplets are encapsulated are no longer sufficient to meet the demands of current life science research and biopharmaceutical technologies.

[0024] In summary, most existing waveform classification schemes for droplet-encapsulated cells rely on manual feature extraction, such as manually extracting simple features like peak value, width, and area, before classifying the waveform using traditional algorithms. This approach is not only labor-intensive and time-consuming, but the extracted features often fail to fully reflect the essential characteristics of the waveform, resulting in insufficient generalization ability of the classification model. When faced with different waveform characteristics generated under varying experimental conditions (such as different cell types and different droplet generation devices), the accuracy of traditional classification algorithms drops significantly, failing to meet the high-precision and high-efficiency requirements of single-cell high-throughput screening and biopharmaceutical applications for droplet detection.

[0025] In view of this, such as Figure 1 As shown, the first embodiment of the present invention provides a method for detecting droplet-encapsulated cells based on PMT signals and AI, including the following steps: S1. Acquire the PMT waveform signal in the droplet detection area; S2. Perform peak detection on the PMT waveform signal and extract the periodic waveform; S3. Extract features from the periodic waveform using an artificial intelligence neural network to obtain waveform features; S4. Perform feature classification on the waveform features and output the classification results as the detection results.

[0026] This invention utilizes a two-level processing architecture of waveform feature extraction and artificial intelligence network classification to automate the entire process from raw waveform acquisition to determining the state of droplets containing target cells, for both droplets without target cells and droplets containing target cells.

[0027] The detection method for single cells encapsulated in droplets according to embodiments of the present invention is applicable to fluorescently labeled cells, which can be cells that express their own fluorescent labels or cells labeled using fluorescent labeling technology. Fluorescent labeling is used to label and index the contents of the droplet (cells in this embodiment of the invention). Applying a fluorescent reagent to the dispersed phase solution can apply fluorescent labeling to the cells. Specific fluorescent labels include nucleic acid dyes, fluorescently tagged antibodies, etc. The specific type of fluorescent label is determined according to the cell type, and this embodiment of the invention does not limit this.

[0028] Cell-containing aqueous and oil phases generate water-in-oil cell-containing droplets (samples to be tested) in a microfluidic channel. The microfluidic channel can be a microfluidic microdroplet chip. The immiscible liquid phases mix rapidly within the micrometer-scale microfluidic channel, causing the fluids to interact under pressure and flow conditions, forming continuous but discontinuous droplets that pass through the microfluidic channel one by one, encapsulating the cells. The droplets are classified into two types: water-in-oil and oil-in-water, depending on the wettability of the channel wall and the difference between the dispersed and continuous phases. This embodiment of the invention uses a water-in-oil system to form droplets as the sample to be tested.

[0029] The implementation process of each step of this invention is described in detail below: S1. Collect the PMT waveform signal in the droplet detection area.

[0030] Droplet microfluidics refers to the rapid mixing of immiscible liquid phases within micrometer-scale channels. This allows the fluids to interact under pressure and flow conditions, forming continuous droplets ranging in volume from nanoliters to picoliters, which can then encapsulate single cells. Secretions from the cells encapsulated within the droplets accumulate, reaching a concentration within the detection range, facilitating further single-cell analysis.

[0031] In this embodiment of the invention, the acquisition of the PMT waveform signal in the droplet detection region specifically includes the following steps: S1-1. When droplets encapsulate cells and flow through the PMT detection area, the fluorescence signal emitted by the droplets encapsulating cells in the area is collected by a photomultiplier tube and converted into an electrical signal.

[0032] In this embodiment of the invention, when a droplet carrying fluorescently labeled cells flows through the detection area of ​​the chip microchannel, an external excitation light source (such as a laser of a specific wavelength) precisely irradiates the droplet, exciting the fluorescent label within the cell to emit fluorescent photons of a characteristic wavelength. After the optical signal enters the PMT detection path, it is converted into an electrical signal inside the PMT through the photoelectric effect.

[0033] S1-2. Extract the waveform of the electrical signal to obtain the waveform signal as the PMT waveform signal for the formation of cells encapsulated by droplets.

[0034] The sliding window smoothing algorithm used in this embodiment of the invention refers to smoothing the signal by sliding a fixed-size window across the original signal and calculating the statistics (such as the average, median, or weighted average) of the data within the window to replace the current point. Since noise typically exhibits high-frequency fluctuations, smoothing the signal can suppress these interferences.

[0035] In a preferred embodiment, extracting the waveform of the electrical signal specifically includes the following steps: S1-2-1. Create a sliding window of a preset size; S1-2-2. Use a sliding window to slide through the time series data of the electrical signal to obtain multiple sets of window data; S1-2-3. Perform statistics and sorting on the window data to obtain the waveform signal of the electrical signal.

[0036] S2. Perform peak detection on the PMT waveform signal and extract the periodic waveform.

[0037] like Figure 2As shown in the embodiment of the present invention, peak detection and periodic waveform extraction of the PMT waveform signal are performed, specifically including the following steps: S2-1. For each waveform data point in the PMT waveform signal x i A waveform peak is identified when all of the following conditions are met: ; in, d min Indicates the minimum peak spacing. w min Indicates the minimum peak width. p min Indicates the minimum significance; h threshold The dynamic peak height threshold is calculated using the following formula: ; in, h base It is the basic threshold. α It is a noise factor; σ noise The noise level of the waveform data points is calculated using the following formula: ; in is the waveform mean, and N is the waveform signal length.

[0038] In this embodiment of the invention, the minimum peak spacing, minimum peak width, minimum significance, basic threshold, and noise factor are preset parameters.

[0039] S2-2. Based on the identified waveform peak values, extract waveform feature data for multiple cycles as periodic waveforms.

[0040] In this embodiment of the invention, peak values ​​are identified by combining parameters such as minimum peak spacing, minimum peak width, and minimum peak saliency. The waveform feature data of a periodic wave is extracted by adaptively periodically capturing the waveform signal and used as the periodic waveform.

[0041] S3. Extract features from the periodic waveform using an artificial intelligence neural network to obtain waveform features.

[0042] The artificial intelligence neural network used in this embodiment of the invention is a convolutional neural network (CNN). A CNN is a neural network that simulates human visual processing. It extracts local features through convolutional layers, reduces spatial dimensionality using pooling layers, and combines fully connected layers for classification. The convolutional operation utilizes a filter to slide and scan the input, the activation function introduces non-linearity, pooling (such as max pooling) reduces parameters, and the fully connected layer integrates features for output prediction.

[0043] The convolutional neural network structure constructed in this embodiment of the invention is as follows: Figure 3 As shown. In this embodiment of the invention, feature extraction of the periodic waveform is performed using an artificial intelligence neural network to obtain waveform features, specifically including the following steps: S3-1. Received periodic waveform.

[0044] S3-2. Local feature extraction is performed on the periodic waveform to obtain the local features of the droplet-encapsulated cell waveform as the first feature; S3-3. Perform dimensionality reduction on the first feature; S3-4. Extract higher-order features from the first feature after dimensionality reduction to obtain the higher-order features of the droplet-enclosed cell waveform as the second feature; S3-5. Perform nonlinear mapping on the second feature to obtain the waveform features of the droplet-encapsulated cell waveform.

[0045] In this embodiment of the invention, the first step is to capture local features of the input periodic waveform using the first convolutional layer of a convolutional neural network. This step focuses on the local structural characteristics of the waveform. Local features that a periodic waveform may possess include, but are not limited to, pulse rise / fall slope, peak width, amplitude of local extrema, and gradient changes between adjacent sampling points. The first feature is a set of low-order features reflecting the microscopic morphology of the waveform. Next, dimensionality reduction is performed on the first feature to eliminate redundant information and improve computational efficiency. Specifically, methods such as Principal Component Analysis (PCA) can be used to map high-dimensional local features to a low-dimensional latent space while retaining key discriminative information. Based on the dimensionality-reduced features, this embodiment of the invention uses a second convolutional layer to perform high-order feature abstraction to uncover global correlations and complex patterns between local features, such as the amplitude correlation of multiple pulse periods, the statistical regularity of waveform distortion patterns (e.g., bimodal phenomena), and signal stability characteristics under noisy backgrounds. Finally, a nonlinear transformation function (the Sigmoid function in this embodiment of the invention) is applied to the second feature to map the higher-order feature to the optimized feature space, and the final output waveform feature is used as the input vector for the classification or quantitative analysis of droplet-encapsulated cells.

[0046] In some embodiments, before the feature extraction step of the periodic waveform using an artificial intelligence neural network, the following steps are also included: S3-0. Interpolate the periodic waveform to give each periodic waveform a standardized length.

[0047] By standardizing the periodic waveforms, each periodic waveform can meet the input requirements of an artificial intelligence neural network.

[0048] Preferably, such as Figure 4 As shown, the convolutional neural network used in this embodiment of the invention is trained through the following steps: Obtain the droplet-encapsulated cell cycle waveform for training, as the training cycle waveform; The waveforms of the training cycle are manually labeled and divided into training dataset and validation dataset; The pre-defined convolutional neural network is trained using a training dataset, and the accuracy and recall of the output of the convolutional neural network are validated using a validation dataset. When the accuracy and recall of the convolutional neural network output reach the preset requirements, the convolutional neural network is considered to have completed training.

[0049] In this embodiment of the invention, the accuracy threshold is set to 0.98 and the recall threshold is set to 0.95. When the accuracy and recall output by the convolutional neural network reach the preset requirements, it is determined that the convolutional neural network has completed training.

[0050] S4. Perform feature classification on the waveform features and output the classification results as the detection results.

[0051] In this embodiment of the invention, feature classification of waveform features specifically includes the following steps: S4-a. When the waveform features have the first activation mode, the droplet-encapsulated cell is classified as an empty package; an empty package refers to a droplet that does not encapsulate the target cell (a cell with a specific fluorescent label that can emit fluorescence at a specific wavelength).

[0052] S4-b. When the waveform features have a second activation mode, the droplet-encapsulated cell is classified as a single-cell encapsulation; a single-cell encapsulation means that the droplet encapsulates only one target cell (droplet single-cell encapsulation).

[0053] S4-c. When the waveform features have a third activation mode, the droplet-encapsulated cells are classified as multiple-encapsulation cells; multiple-encapsulation cells refer to droplets encapsulating two or more target cells.

[0054] In step S3, the convolutional neural network obtains waveform features with multiple activation modes through hierarchical feature extraction, which can effectively classify droplet-encapsulated cells as empty, single, or multiple envelopes.

[0055] Specifically, for empty-packed droplet cells, such as Figure 5 , 6 As shown, because the cells do not contain fluorescent labels, their fluorescence emission is very weak, and the electrical signal waveform acquired and converted by PMT is stable with low peaks. An artificial intelligence neural network extracts its waveform features, and then performs classification and judgment.

[0056] For single-cell droplets, such as Figure 7 , 8 As shown, because it contains a single cell, when a droplet flows through the detection area, the fluorescence emitted by the cell within the droplet at a specific wavelength is detected by the PMT. The electrical signal waveform produces a typical single-peak pulse, with the steep rising and falling edges of the pulse forming isolated high-amplitude spikes on the waveform. A specific convolutional kernel of the convolutional neural network precisely responds to this local feature, forming a strong red, spike-like high-activation region within a range of several sampling points around the pulse center. This highly concentrated activation pattern is essentially the convolutional kernel's characteristic encoding of the transient light intensity changes caused by the cell, and its activation intensity is positively correlated with the pulse amplitude.

[0057] Finally, for multi-celled droplet cells, such as Figure 9 , 10 As shown, since there are multiple cells inside the multi-cell droplet, these cells are randomly distributed within the droplet. When they pass through the detection area one after another or simultaneously, they will generate superimposed wide pulses or multiple separate secondary peaks, accompanied by baseline rise. This complex scattering pattern is deconstructed into multiple feature responses by the convolutional neural network: on the one hand, in terms of time, multiple convolutional kernels will capture secondary peaks at different positions respectively.

[0058] Overall, the detection method provided in this invention effectively solves the problem of insufficient processing capability of traditional classification algorithms for complex waveforms by effectively extracting waveform data generated by droplets encapsulating cells and using artificial intelligence neural networks for automatic extraction of waveform features. Compared with traditional peak detection algorithms, this invention can more accurately capture the waveform features of single-cell encapsulated droplets, providing more accurate droplet cell classification results, and improving the detection accuracy by more than 25%.

[0059] Figure 11This is a schematic diagram of the electronic device proposed in the second embodiment of the present invention. The memory in this embodiment stores program instructions for implementing the droplet-encapsulated cell detection method based on PMT signals and AI in any of the above embodiments. The processor executes the program instructions stored in the memory to perform single-cell dynamic feature detection based on PMT and AI. The processor can also be referred to as a CPU (Central Processing Unit). The processor may be an integrated circuit chip with signal processing capabilities. The processor can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0060] The methods described in the first embodiment of the present invention are applicable to the embodiments of the present electronic device. The specific functions implemented by the embodiments of the present electronic device are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above methods.

[0061] Figure 12 This is a schematic diagram of the structure of a computer-readable storage medium according to the third embodiment of the present invention. The computer-readable storage medium of the fourth embodiment of the present invention stores program instructions capable of implementing the above-described method for detecting droplet-encapsulated cells based on PMT signals and AI. These program instructions can be stored in the storage medium in the form of a software product, including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of the present invention. The aforementioned computer-readable storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.

[0062] The methods described in the first embodiment of the present invention are applicable to the computer-readable storage medium embodiment. The specific functions implemented by the computer-readable storage medium embodiment are the same as those in the above method embodiment, and the beneficial effects achieved are also the same as those achieved by the above method.

[0063] This embodiment also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned related steps to realize the droplet-encapsulated cell detection method based on PMT signals and AI provided in the above embodiment.

[0064] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present invention are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0065] Those skilled in the art will understand that modules in the device of the embodiments of the present invention can be adaptively modified and placed in one or more devices different from those embodiments. Modules, units, or components in the embodiments of the present invention can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the corresponding claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the corresponding claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0066] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0067] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0068] Furthermore, the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. In particular, for embodiments such as apparatus and devices, since they are basically similar to the method embodiments, the relevant parts can be referred to the description of the method embodiments. The apparatus, devices, and other embodiments described above are merely illustrative, and the modules, units, etc., described as separate components may or may not be physically separate, that is, they may be located in one place or distributed in multiple places, such as nodes in a system network. Specifically, some or all of the modules and units can be selected according to actual needs to achieve the purpose of the above-described embodiment solutions. Those skilled in the art can understand and implement this without creative effort.

[0069] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0070] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0071] Furthermore, the terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying relative importance, or implicitly specifying the number of technical features indicated in this embodiment. Therefore, features defined with terms such as "first" and "second" in the embodiments of this invention can explicitly or implicitly indicate that the embodiment includes at least one of those features. In the description of this invention, the word "multiple" means at least two or more, such as two, three, four, etc., unless otherwise explicitly specified in the embodiments.

[0072] In embodiments of the present invention, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, components, features, and elements with the same names in different embodiments of the present invention may have the same meaning or different meanings, the specific meaning of which must be determined by its interpretation in that specific embodiment or further in conjunction with the context of that specific embodiment.

[0073] Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention. Other embodiments of the present invention will readily conceive of by considering the specification and practicing the invention. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

Claims

1. A method for detecting droplet-encapsulated cells based on PMT signal and AI, characterized in that, The method comprises the following steps: Collecting a PMT waveform signal of a droplet detection area; Performing peak detection on the PMT waveform signal to intercept a periodic waveform; Extracting features of the periodic waveform through an artificial intelligence neural network to obtain a waveform feature; Performing feature classification on the waveform feature, and outputting a classification result as a detection result.

2. The method according to claim 1, wherein the method is based on PMT signal and AI for detecting the droplet-encapsulated cells. The collecting of the PMT waveform signal of the droplet detection area specifically comprises the following steps: When the droplet-encapsulated cell flows through the PMT detection area, collecting a fluorescence signal emitted by the droplet-encapsulated cell in the area through a photomultiplier tube and converting the fluorescence signal into an electrical signal; Extracting a waveform of the electrical signal to obtain a waveform signal as a PMT waveform signal formed by the droplet-encapsulated cell.

3. The method according to claim 2, wherein the PMT signal and the AI are used to detect the droplet-encapsulated cells. The extracting of the waveform of the electrical signal specifically comprises the following steps: Establishing a sliding window of a preset size; Sliding the sliding window in time series data of the electrical signal to obtain a plurality of window data; Performing statistics on the window data and sorting to obtain a waveform signal of the electrical signal.

4. The method according to claim 1, wherein the method is based on PMT signal and AI for detecting the droplet-encapsulated cells. The performing of the peak detection on the PMT waveform signal to intercept the periodic waveform specifically comprises the following steps: For each waveform data point in the PMT waveform signal x i is identified as a waveform peak when all of the following conditions are met: ； wherein, d min denotes the minimum peak distance, w min denotes the minimum peak width, p min denotes the minimum saliency; h threshold represents the dynamic peak height threshold, which is calculated by the following equation: ； wherein h base is a base threshold value, α is a noise factor; According to the identified waveform peak, intercepting waveform feature data of multiple periods as the periodic waveform. noise is the noise level of the wave form data points, calculated by the following equation: ； wherein is the waveform mean, N is the waveform signal length; The artificial intelligence neural network is a convolutional neural network; the extracting of the features of the periodic waveform through the artificial intelligence neural network to obtain the waveform feature specifically comprises the following steps:

5. The method according to claim 1, wherein, Receiving a periodic waveform; Performing local feature extraction on the periodic waveform to obtain local features of a droplet-encapsulated cell waveform as first features; Performing dimension reduction processing on the first features; Performing high-order feature extraction on the first features after the dimension reduction processing to obtain high-order features of the droplet-encapsulated cell waveform as second features; Performing nonlinear mapping on the second features to obtain waveform features of the droplet-encapsulated cell waveform. The convolutional neural network is obtained through the following steps of training: Obtaining periodic waveforms of droplet-encapsulated cells for training as training periodic waveforms; 6. The method of claim 1, wherein the method is based on PMT signal and AI for detecting droplet-encapsulated cells. Performing manual annotation on the training periodic waveforms to divide them into a training data set and a verification data set; Training a preset convolutional neural network using the training data set, and verifying the accuracy and recall rate of the output of the convolutional neural network using the verification data set; When the accuracy and recall rate of the output of the convolutional neural network meet preset requirements, determining that the convolutional neural network is trained. The performing of the feature classification on the waveform feature specifically comprises the following steps: When the waveform feature has a first activation mode, determining that the droplet-encapsulated cell is empty package classification; 7. The method of claim 1, wherein the PMT signal and the AI are based on the droplet-encapsulated cells. When the waveform feature has a second activation mode, determining that the droplet-encapsulated cell is single package classification; When the waveform feature has a third activation mode, determining that the droplet-encapsulated cell is multiple package classification. A processor and a memory are included; The memory is used to store a program; 8. An electronic device, comprising: The processor executes the program to implement the method for detecting a droplet-encapsulated cell based on a PMT signal and AI according to any one of claims 1-7. ​ ​ 9. A computer-readable storage medium, characterized in that, The storage medium stores a program, and the program is executed by the processor to realize the PMT signal and AI-based droplet cell detection method in any one of claims 1-7.

Images