CN117218650A - Method, apparatus and storage medium for automatically extracting multiple fluorescence reference spectra - Google Patents
Method, apparatus and storage medium for automatically extracting multiple fluorescence reference spectra Download PDFInfo
- Publication number
- CN117218650A CN117218650A CN202311243184.XA CN202311243184A CN117218650A CN 117218650 A CN117218650 A CN 117218650A CN 202311243184 A CN202311243184 A CN 202311243184A CN 117218650 A CN117218650 A CN 117218650A
- Authority
- CN
- China
- Prior art keywords
- sample
- vector
- reference spectra
- multiple fluorescence
- implicit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001228 spectrum Methods 0.000 title claims abstract description 79
- 238000000034 method Methods 0.000 title claims abstract description 51
- 239000013598 vector Substances 0.000 claims abstract description 50
- 238000012549 training Methods 0.000 claims abstract description 26
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000009826 distribution Methods 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 14
- 238000005070 sampling Methods 0.000 claims description 12
- 238000012216 screening Methods 0.000 claims description 7
- 230000003595 spectral effect Effects 0.000 claims description 7
- 238000002189 fluorescence spectrum Methods 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000003384 imaging method Methods 0.000 abstract description 11
- 238000000605 extraction Methods 0.000 abstract description 6
- 238000000799 fluorescence microscopy Methods 0.000 abstract description 2
- 239000000523 sample Substances 0.000 description 58
- 239000000975 dye Substances 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 230000002093 peripheral effect Effects 0.000 description 6
- 238000000701 chemical imaging Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000000513 principal component analysis Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000004913 activation Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 210000004027 cell Anatomy 0.000 description 2
- 239000007850 fluorescent dye Substances 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012472 biological sample Substances 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012632 fluorescent imaging Methods 0.000 description 1
- 238000001215 fluorescent labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000000386 microscopy Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000000399 optical microscopy Methods 0.000 description 1
- 210000003463 organelle Anatomy 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Landscapes
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The application relates to the technical field of fluorescence microscopy imaging, in particular to a method and a system for automatically extracting multiple fluorescence reference spectra, wherein the method comprises the following steps: inputting a lambda spectrum image stack; performing effective data preprocessing and data redistribution on the lambda spectrum image stack; constructing a training variation self-encoder network and training the training variation self-encoder; generating an implicit space vector and clustering the implicit space vector; multiple fluorescence reference spectra are acquired. The application can realize the automatic extraction of multiple fluorescence reference spectrums when the dye information is unknown or the deviation between the experimental environment and the ideal environment is large.
Description
Technical Field
The application relates to the technical field of fluorescence microscopy imaging, in particular to a method, equipment and a storage medium for automatically extracting multiple fluorescence reference spectrums.
Background
Optical microscopy imaging is an important technical means of cell research, and with the continuous progress of fluorescence labeling technology and microscopy instruments, the generation and imaging of complex biological samples containing multiple fluorophores are gradually possible. In order to accurately and quantitatively analyze data simultaneously containing a plurality of fluorescence channels, reliable splitting of spectrum signals of different fluorescence channels is particularly important.
The single-channel reference spectrum curve is the basis of a multiple fluorescence spectrum splitting technology, and the existing reference spectrum acquisition mode mainly comprises the following steps: based on a dye database, single fluorescent samples were made and imaged based on lambda spectra. The dye database-based mode is to search a dye spectrum curve matched with an actual sample from an official platform and directly serve as a reference spectrum of the fluorescent dye. Based on the single fluorescent sample mode, a fluorescent sample only comprising a single cell structure needs to be manufactured for each channel, and then each sample is subjected to spectral imaging, so that a reference spectrum of the sample is obtained. The lambda spectral imaging-based approach may generate the reference spectrum by manual region-rounding, or by constructing feature spaces such as Principal Component Analysis (PCA), sine/cosine transform, etc. After the reference spectrum is obtained in the above way, the single separated fluorescent channel signal can be obtained by combining the existing spectrum splitting technology, such as linear splitting.
The dye information needs to be known in advance because of the change of the environment such as temperature, humidity and the like of an actual experiment, and certain deviation can occur between a theoretical reference spectrum and an actual spectrum curve, so that the resolution result is inaccurate. The spectral curve obtained by making a single fluorescent sample is accurate, but the making process is complex, and has high implementation difficulty for a part of subcellular organelles which are closely related in structure. In a lambda spectrum imaging mode, manual region circling is required to manually judge the pixel region of a single sample in an image, the process depends on experience of an operator and a sample structure to be detected, and when different dyes are closely related in space or have stronger background signal interference, a reference spectrum obtained by the method has larger deviation than a real curve. While conventional means such as Principal Component Analysis (PCA) and subharmonic analysis based on sine/cosine transformation are commonly used at present, the automatic extraction of a reference spectrum can be realized, but the selection of a characteristic space is depended on, and the adaptability to multiple channels is still to be improved due to the fact that the dimension of the generated characteristic vector is low.
In summary, how to automatically extract multiple fluorescence reference spectra when dye information is unknown or the deviation between experimental environment and ideal environment is large is a current problem.
Disclosure of Invention
The application provides a method and a system for automatically extracting multiple fluorescence reference spectrums, which can realize the automatic extraction of the multiple fluorescence reference spectrums when dye information is unknown or the deviation between an experimental environment and an ideal environment is large. The application provides the following technical scheme:
in a first aspect, the present application provides a method of automatically extracting multiple fluorescence reference spectra, the method comprising:
inputting a lambda spectrum image stack;
performing effective data preprocessing and data redistribution on the lambda spectrum image stack;
constructing a training variation self-encoder network and training the training variation self-encoder;
generating an implicit space vector and clustering the implicit space vector;
multiple fluorescence reference spectra are acquired.
In one specific embodiment, the performing effective data preprocessing and data redistribution on the lambda spectral image stack comprises:
calculating an average value μ of the stack of lambda spectral images m And standard deviation sigma m ；
Rearranging the data structure of the lambda spectrum image stack, wherein the newly arranged data format is L.M, L represents the total data sample number, and M is the length of the eigenvector in the lambda direction;
screening effective samples from the data samples; wherein, for each sample M, it can be expressed as a one-dimensional vector with length M, and the average value mu is calculated i Standard deviation sigma i The method comprises the steps of carrying out a first treatment on the surface of the If mu is i Greater than 0.2 mu m And sigma (sigma) i Greater than 0.2 sigma m The sample is considered to be a valid sample and is reserved, otherwise, the sample is removed; the effective data sample after screening is L e * M, wherein L e Is the number of valid data samples;
for the valid data sample L e * And M is normalized.
In a specific embodiment, the constructing a training variation self-encoder network and training the training variation self-encoder comprises:
constructing a variation self-Encoder network structure, wherein the variation self-Encoder network structure comprises an encoding Encoder module and a decoding Decoder module;
constructing a loss function loss=loss_recon+loss_dist;
inputting the valid data sample L e *M norm Training the variable self-encoder network parameters;
and if the loss function loss is smaller than the set threshold value or reaches the set training times, training is completed.
In a specific embodiment, the loss function loss consists of two parts:
representing the degree of difference between the decoding Decoder module generating vector m' and the input sample m, wherein I 2 Representing a vector binary norm;
wherein (1)>Indicating the desire z μ Variance is->N (0,I) represents a standard normal distribution of 0 and variance I, I represents an identity matrix, D KL Representing the Kullback-Leibler divergence between the two probability distributions, i.e., the degree of difference between the implicit z vector distribution and the standard normal distribution output by the encoding Encoder module.
In a specific embodiment, the construction loss function loss=loss_recon+loss_dist further includes:
sample m input code Encoder module, output implicit vector z μ And z σ ；
According to a Gaussian distributionSampling and outputting a sampling vector z s ；
Sampling vector z s The input to the decode module outputs a generated vector m'.
In a specific embodiment, the generating the implicit spatial vector includes:
taking a coding Encoder module of a trained variation self-coder;
for valid data sample L e *M norm Generating an implicit spatial vector set L e * Z; wherein for each sample m, the input variation is derived from the encoding Encoder module of the Encoder, outputting an implicit vector z μ The vector length is Z.
In a specific embodiment, the clustering of the implicit spatial vectors includes:
setting the clustering type as N for the mixed fluorescence spectrum sample with the channel number of N;
clustering the implicit space vector by a sampling clustering method;
generating a cluster of clustered samples L 1 *Z、L 2 *Z、L 3 *Z、…、L N * Z and the corresponding sample cluster centroid Z 1 、z 2 、z 3 、…、z N Wherein L is 1 +L 2 +L 3 +…+L N ＝L e 。
In a specific embodiment, the obtaining multiple fluorescence reference spectra comprises:
taking a single sample cluster L i * Z and sample cluster centroid Z i ；
For a single sample cluster L i * Z, search centroid Z i Corresponding K adjacent points form an implicit space adjacent set L knn * Z, wherein k=0.3×l;
for the implicit spatial neighbor set L knn * Z, backtracking it back in the valid data sample L e * Corresponding neighbor set L in M knn *M；
Spatial neighbor set L for data samples knn * M, calculating M i ＝(M 1 +M 2 +M 3 +…+M Lknn )/L knn ，M i The reference spectrum to be acquired is obtained;
for the clustered sample cluster L 1 *Z、L 2 *Z、L 3 *Z、…、L N * Z, for each sample cluster L in turn i * And Z (i=1 to N) performs the above operation, so that reference spectra of all fluorescence channels can be obtained.
In a second aspect, the present application provides an electronic device comprising a processor and a memory; stored in the memory is a program that is loaded and executed by the processor to implement a method of automatically extracting multiple fluorescence reference spectra as claimed in any one of claims 1 to 7.
In a third aspect, the present application provides a computer readable storage medium having stored therein a program which when executed by a processor is adapted to carry out a method of automatically extracting multiple fluorescence reference spectra as claimed in any one of claims 1 to 7.
In summary, the beneficial effects of the present application at least include:
1) The method can be used for automatically acquiring multiple fluorescence reference spectrums with unknown dye information;
2) When the deviation between the experimental environment and the ideal environment is large, more accurate results can be obtained by adopting the method;
3) The automatic acquisition of multiple fluorescence reference spectrums can be realized without the intervention of experimental personnel;
4) The variation self-coding technology adopted for generating the implicit space vector has strong interpretability based on a mixed Gaussian model, and the implicit space vector learned by the unsupervised generation model has higher robustness for acquiring multiple fluorescence (fluorescence channel number > 3) reference spectrums.
And generating a multidimensional implicit space by a variation self-coding technology, and clustering or classifying the implicit space, thereby realizing automatic extraction of multiple fluorescence reference spectrums. When the dye information is unknown or the deviation between the experimental environment and the ideal environment is large, the method can obtain ideal results.
The foregoing description is only an overview of the present application, and is intended to provide a better understanding of the present application, as it is embodied in the following description, with reference to the preferred embodiments of the present application and the accompanying drawings.
Drawings
FIG. 1 is a flow chart of a method for automatically extracting multiple fluorescence reference spectra according to an embodiment of the present application.
Fig. 2 is a schematic diagram of a variable self-encoder according to an embodiment of the present application.
FIG. 3 is a flow chart of a system for obtaining multiple fluorescence reference spectra according to an embodiment of the present application.
FIG. 4 is a block diagram of an electronic device for automatically extracting multiple fluorescence reference spectra, provided in one embodiment of the application.
Detailed Description
The following describes in further detail the embodiments of the present application with reference to the drawings and examples. The following examples are illustrative of the application and are not intended to limit the scope of the application.
Optionally, the method for automatically extracting multiple fluorescence reference spectra provided by each embodiment of the present application is described by taking an example in an electronic device, where the electronic device is a terminal or a server, and the terminal may be a mobile phone, a computer, a tablet computer, a scanner, an electronic eye, a monitoring camera, etc., and the embodiment does not limit the type of the electronic device.
Referring to fig. 1, a flow chart of a method for automatically extracting multiple fluorescence reference spectra according to an embodiment of the present application includes at least the following steps:
s100, inputting a lambda spectrum image stack.
Wherein the lambda spectral image stack is provided by a fluorescent imaging device with a spectral imaging function, such as a laser scanning confocal microscope equipped with a spectral imaging module, etc.
Specifically, the lambda spectral image stack size is x×y×m, where X is the imaging device lateral resolution size, which is CMOS device lateral resolution for an imaging device equipped with, for example, CMOS light sensing devices, and sample lateral scanning resolution for an imaging device equipped with, for example, PMT light sensing devices. Y is the size of the imaging device longitudinal resolution, which is CMOS device longitudinal resolution for imaging devices equipped with, for example, CMOS light sensing devices, and sample longitudinal scanning resolution for imaging devices equipped with, for example, PMT light sensing devices. M is the number of sampling points of the spectrum along the wavelength lambda direction, namely the length of the eigenvector of the lambda direction.
S200, data preprocessing and redistribution.
The specific process for screening and redistributing the valid data of the lambda spectrum image stack in the above step S100 includes:
s201, calculating the average value mu of the S100lambda spectrum image stack m And standard deviation sigma m 。
S202, rearranging the lambda spectrum image stack data structure of the S100, wherein the new arrangement data format is L X M, wherein L=X X Y, the total data sample number is represented, and M is the length of the eigenvector in the lambda direction.
S203, screening out effective samples from the data samples in the step S202.
In particular, the method comprises the steps of,for each sample M, it can be expressed as a one-dimensional vector, with length M, and its mean value μ is calculated i Standard deviation sigma i . If mu is i Greater than 0.2 mu m And sigma (sigma) i Greater than 0.2 sigma m And the sample is considered to be a valid sample and is reserved, otherwise, the sample is rejected. The effective data sample after screening is L e * M, wherein L e Is the number of valid data samples.
S204, regarding the valid data sample L in the above S203 e * And M is normalized.
Specifically, for each sample m, the maximum component m of the sample is searched for max Normalize the sample m norm ＝m/m max Normalized sample is le×m norm 。
S300, training the variable self-encoder.
Referring to fig. 2, a schematic diagram of a variable self-encoder according to an embodiment of the present application is shown, and the training process is specifically as follows:
s301, constructing a variable self-Encoder network structure, which mainly comprises an encoding Encoder module and a decoding Decoder module:
alternatively, the encoding Encoder module may consist of a linear layer, an activation function layer, and a linear layer;
alternatively, the decoding Decoder module may be composed of a linear layer, an activation function layer, and a linear layer.
The data forward propagation flow, that is, the flow from data input to data output, comprises:
s3011, inputting a sample m into an encoding Encoder module to output an implicit vector z μ And z σ 。
S3012, according to Gaussian distributionSampling and outputting a sampling vector z s 。
S3013, sample vector z s The input to the decode module outputs a generated vector m'.
S302, constructing a loss function loss=loss_record+loss_dist.
Wherein the loss function loss consists of two parts:
representing the degree of difference between the decoded Decoder module generated vector m' and the input samples m, wherein I II 2 Representing the vector binary norm.
Wherein->Indicating the desire z μ Variance is->N (0,I) represents a standard normal distribution of 0 and variance I, I represents an identity matrix, D KL Representing the Kullback-Leibler divergence (KL divergence for short) between two probability distributions, i.e., the degree of difference between the implicit z vector distribution output by the encoding Encoder module and the standard normal distribution.
S303, inputting the valid data sample L in the above S204 e *M norm The training variation is derived from the encoder network parameters.
S304, if the loss function loss in the S302 is smaller than the set threshold value or reaches the set training times, the training is completed.
S400, generating an implicit space vector.
The specific flow comprises the following steps:
s401, taking the coding Encoder module of the variable self-coder trained in the S300.
S402, regarding the valid data sample L in the above S204 e *M norm Generating an implicit spatial vector set L e *Z。
For each sample m, the encoded Encoder module of the variable self-Encoder in S300 is input to output an implicit vector z μ The vector length is Z.
S500, implicit space clustering.
The specific flow comprises the following steps:
s501, setting the clustering type as N for the mixed fluorescence spectrum sample with the channel number of N.
S502, clustering the implicit space vector by a sampling clustering method.
Optionally, the selected clustering method may be k-cluster clustering;
alternatively, the selected clustering method may be t-sne clustering, and the clustering method may reduce the dimension of the data before clustering.
S503, generating a clustered sample cluster L 1 *Z、L 2 *Z、L 3 *Z、…、L N * Z and the corresponding sample cluster centroid Z 1 、z 2 、z 3 、…、z N Wherein L is 1 +L 2 +L 3 +…+L N ＝L e 。
S600, acquiring multiple fluorescence reference spectrums.
Referring to fig. 3, a flow chart of acquiring multiple fluorescence reference spectra according to an embodiment of the present application is shown, specifically, for the clustered sample cluster L described in S503 1 *Z、L 2 *Z、L 3 *Z、…、L N * Z, calculating each sample cluster L in turn i * The reference spectrum corresponding to Z (i=1 to N), the procedure comprises:
s601, taking a single sample cluster L i * Z and sample cluster centroid Z i (i＝1～N)。
S602, for the single sample cluster L i * Z, search centroid Z i Corresponding K adjacent points form an implicit space adjacent set L knn * Z, where k=0.3×l i 。
S603, for the implicit spatial neighbor set L in S602 above knn * Z, backtracking it back in the valid data sample L e * Corresponding neighbor set L in M knn *M。
S604, for the data sample space adjacent set L in the S603 knn * M, calculating M i ＝(M 1 +M 2 +M 3 +…+M Lknn )/L knn ，M i The reference spectrum to be acquired is obtained;
s605, for the clustered sample cluster L in the S503 1 *Z、L 2 *Z、L 3 *Z、…、L N * Z, for each sample cluster L in turn i * And Z (i=1 to N) executing the steps S601 to S604, so that the reference spectrums of all the fluorescence channels can be obtained.
In summary, in order to overcome the defects in the prior art, the application designs a method for automatically extracting multiple fluorescence reference spectra, which generates a multidimensional implicit space through a variation self-coding technology and clusters or classifies the implicit space, thereby realizing the automatic extraction of the multiple fluorescence reference spectra. When the dye information is unknown or the deviation between the experimental environment and the ideal environment is large, the method can obtain ideal results. In addition, the method does not need the intervention of experimenters, the variation self-coding technology adopted for generating the implicit space is based on a mixed Gaussian model, any mixed spectrum distribution is regarded as the linear superposition of a plurality of independent reference spectrum distributions, the interpretation is strong, and the subsequent classification and the extraction of the reference spectrum have higher robustness.
Fig. 4 is a block diagram of an electronic device provided in one embodiment of the application. The device comprises at least a processor 401 and a memory 402.
Processor 401 may include one or more processing cores such as: 4 core processors, 8 core processors, etc. The processor 401 may be implemented in at least one hardware form of DSP (Digital Signal Processing ), FPGA (Field-Programmable Gate Array, field programmable gate array), PLA (Programmable Logic Array ). The processor 401 may also include a main processor, which is a processor for processing data in an awake state, also called a CPU (Central Processing Unit ), and a coprocessor; a coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 401 may integrate a GPU (Graphics Processing Unit, image processor) for rendering and drawing of content required to be displayed by the display screen. In some embodiments, the processor 401 may also include an AI (Artificial Intelligence ) processor for processing computing operations related to machine learning.
Memory 402 may include one or more computer-readable storage media, which may be non-transitory. Memory 402 may also include high-speed random access memory, as well as non-volatile memory, such as one or more magnetic disk storage devices, flash memory storage devices. In some embodiments, a non-transitory computer readable storage medium in memory 402 is used to store at least one instruction for execution by processor 401 to implement the method of automatically extracting multiple fluorescence reference spectra provided by the method embodiments of the present application.
In some embodiments, the electronic device may further optionally include: a peripheral interface and at least one peripheral. The processor 401, memory 402, and peripheral interfaces may be connected by buses or signal lines. The individual peripheral devices may be connected to the peripheral device interface via buses, signal lines or circuit boards. Illustratively, peripheral devices include, but are not limited to: radio frequency circuitry, touch display screens, audio circuitry, and power supplies, among others.
Of course, the electronic device may also include fewer or more components, as the present embodiment is not limited in this regard.
Optionally, the present application further provides a computer readable storage medium, in which a program is stored, and the program is loaded and executed by a processor to implement the method for automatically extracting multiple fluorescence reference spectra according to the above method embodiment.
Optionally, the present application further provides a computer product, where the computer product includes a computer readable storage medium, where a program is stored, and the program is loaded and executed by a processor to implement the method for automatically extracting multiple fluorescence reference spectra according to the above method embodiment.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.
Claims (10)
1. A method for automatically extracting multiple fluorescence reference spectra, the method comprising: inputting a lambda spectrum image stack;
performing effective data preprocessing and data redistribution on the lambda spectrum image stack;
constructing a training variation self-encoder network and training the training variation self-encoder;
generating an implicit space vector and clustering the implicit space vector;
multiple fluorescence reference spectra are acquired.
2. The method of claim 1, wherein the performing effective data preprocessing and data redistribution on the lambda spectral image stack comprises:
calculating an average value μ of the stack of lambda spectral images m And standard deviation sigma m ；
Rearranging the data structure of the lambda spectrum image stack, wherein the newly arranged data format is L.M, L represents the total data sample number, and M is the length of the eigenvector in the lambda direction;
screening effective samples from the data samples; wherein, for each sample M, it can be expressed as a one-dimensional vector with length M, and the average value mu is calculated i Standard deviation sigma i The method comprises the steps of carrying out a first treatment on the surface of the If mu is i Greater than 0.2 mu m And sigma (sigma) i Greater than 0.2 sigma m The sample is considered to be a valid sample and is reserved, otherwise, the sample is removed; the effective data sample after screening is L e * M, wherein L e Is the number of valid data samples;
for the valid data sample L e * And M is normalized.
3. The method of automatically extracting multiple fluorescence reference spectra of claim 2, wherein the constructing a training variable self-encoder network and training the training variable self-encoder comprises:
constructing a variation self-Encoder network structure, wherein the variation self-Encoder network structure comprises an encoding Encoder module and a decoding Decoder module;
constructing a loss function loss=loss_recon+loss_dist;
inputting the valid data sample L e *M norm Training the variable self-encoder network parameters;
and if the loss function loss is smaller than the set threshold value or reaches the set training times, training is completed.
4. A method of automatically extracting multiple fluorescence reference spectra according to claim 3, wherein the loss function loss consists of two parts:
representing the degree of difference between the decoding Decoder module generating vector m' and the input sample m, wherein I 2 Representing a vector binary norm;
wherein (1)>Indicating the desire z μ Variance is->N (0,I) represents a standard normal distribution of 0 and variance I, I represents an identity matrix, D KL Representing the Kullback-Leibler divergence between the two probability distributions, i.e., the degree of difference between the implicit z vector distribution and the standard normal distribution output by the encoding Encoder module.
5. The method of automatically extracting multiple fluorescence reference spectra of claim 3, wherein constructing the loss function loss = loss_record + loss_dist further comprises, prior to:
sample m input code Encoder module, output implicit vector z μ And z σ ；
According to a Gaussian distributionSampling and outputting a sampling vector z s ；
Sampling vector z s The input to the decode module outputs a generated vector m'.
6. The method of automatically extracting multiple fluorescence reference spectra of claim 3, wherein said generating an implicit spatial vector comprises:
taking a coding Encoder module of a trained variation self-coder;
for valid data sample L e *M norm Generating an implicit spatial vector set L e * Z; wherein for each sample m, the input variation is derived from the encoding Encoder module of the Encoder, outputting an implicit vector z μ The vector length is Z.
7. The method of automatically extracting multiple fluorescence reference spectra of claim 6, wherein said clustering of implicit spatial vectors comprises:
setting the clustering type as N for the mixed fluorescence spectrum sample with the channel number of N;
clustering the implicit space vector by a sampling clustering method;
generating a cluster of clustered samples L 1 *Z、L 2 *Z、L 3 *Z、…、L N * Z and the corresponding sample cluster centroid Z 1 、z 2 、z 3 、…、z N Wherein L is 1 +L 2 +L 3 +…+L N ＝L e 。
8. The method of automatically extracting multiple fluorescence reference spectra of claim 7, comprising: the acquiring multiple fluorescence reference spectra includes:
taking a single sample cluster L i * Z and sample cluster centroid Z i ；
For a single sample cluster L i * Z, search centroid Z i Corresponding K adjacent points form an implicit space adjacent set L knn * Z, wherein k=0.3×l;
for the implicit spatial neighbor set L knn * Z, backtracking it back in the valid data sample L e * Corresponding neighbor set L in M knn *M；
Spatial neighbor set L for data samples knn * M, calculating M i ＝(M 1 +M 2 +M 3 +…+M Lknn )/L knn ，M i The reference spectrum to be acquired is obtained;
for the clustered sample cluster L 1 *Z、L 2 *Z、L 3 *Z、…、L N * Z, for each sample cluster L in turn i * And Z (i=1 to N) performs the above operation, so that reference spectra of all fluorescence channels can be obtained.
9. An electronic device comprising a processor and a memory; stored in the memory is a program that is loaded and executed by the processor to implement a method of automatically extracting multiple fluorescence reference spectra as claimed in any one of claims 1 to 8.
10. A computer readable storage medium, characterized in that the storage medium has stored therein a program which, when executed by a processor, is adapted to carry out a method of automatically extracting multiple fluorescence reference spectra according to any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311243184.XA CN117218650A (en) | 2023-09-26 | 2023-09-26 | Method, apparatus and storage medium for automatically extracting multiple fluorescence reference spectra |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311243184.XA CN117218650A (en) | 2023-09-26 | 2023-09-26 | Method, apparatus and storage medium for automatically extracting multiple fluorescence reference spectra |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117218650A true CN117218650A (en) | 2023-12-12 |
Family
ID=89042279
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311243184.XA Pending CN117218650A (en) | 2023-09-26 | 2023-09-26 | Method, apparatus and storage medium for automatically extracting multiple fluorescence reference spectra |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117218650A (en) |
-
2023
- 2023-09-26 CN CN202311243184.XA patent/CN117218650A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10943346B2 (en) | Multi-sample whole slide image processing in digital pathology via multi-resolution registration and machine learning | |
| JP6862579B2 (en) | Acquisition of image features | |
| Otálora et al. | Staining invariant features for improving generalization of deep convolutional neural networks in computational pathology | |
| CN111524137B (en) | Cell identification counting method and device based on image identification and computer equipment | |
| CN111915596A (en) | Method and device for predicting benign and malignant pulmonary nodules | |
| US20160358335A1 (en) | Group sparsity model for image unmixing | |
| EP4258283A1 (en) | Method for training artificial neural network providing determination result of pathological specimen, and computing system for performing same | |
| CN115410050B (en) | Tumor cell detection equipment based on machine vision and method thereof | |
| US20120219206A1 (en) | High-Throughput Biomarker Segmentation Utilizing Hierarchical Normalized Cuts | |
| Bai et al. | Nhl pathological image classification based on hierarchical local information and googlenet-based representations | |
| Han et al. | Learning generative models of tissue organization with supervised GANs | |
| US20240054639A1 (en) | Quantification of conditions on biomedical images across staining modalities using a multi-task deep learning framework | |
| CN112703531A (en) | Generating annotation data for tissue images | |
| CN114358279A (en) | Image recognition network model pruning method, device, equipment and storage medium | |
| US11636924B2 (en) | Computer device for detecting an optimal candidate compound and methods thereof | |
| Pastore et al. | Efficient unsupervised learning of biological images with compressed deep features | |
| Zhu et al. | Scene text detection via extremal region based double threshold convolutional network classification | |
| CN111383207A (en) | Musculoskeletal ultrasonic image processing method, system and device and readable storage medium | |
| CN116342628A (en) | Pathological image segmentation method, pathological image segmentation device and computer equipment | |
| CN117218650A (en) | Method, apparatus and storage medium for automatically extracting multiple fluorescence reference spectra | |
| Díaz et al. | Histopathological image classification using stain component features on a plsa model | |
| Aljuhani et al. | Whole slide imaging: deep learning and artificial intelligence | |
| Mansour et al. | Kidney segmentations using cnn models | |
| Ye et al. | A multi-attribute controllable generative model for histopathology image synthesis | |
| CN113222887A (en) | Deep learning-based nano-iron labeled neural stem cell tracing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |