wherein w is₀ Representing the final scaling, being a specified constant, e representing the natural logarithm, n representing the number of divisions;

the selected region images are separated in sequence and ordered in time sequence.

Preferably, in S1, the detection of immature DC cells on the selected area image based on the support vector machine sequentially includes:

reading and classifying a selected area image preselected as a sample, wherein the selected area image comprises a cell culture image in which immature DC cells are not induced and in which immature DC cells are induced and artificially located;

extracting features of the selected area image by adopting a local binary pattern, and converting the image into an integer value matrix;

obtaining an LBP histogram based on the integer value matrix;

training the support vector machine model by using the LBP histogram, and outputting a target class and a non-target class;

and based on the trained support vector machine model, judging the target class and the non-target class of the selected area image to be detected, and finishing the detection of the immature DC cells.

Preferably, the second acquisition frequency F₂ Is greater than the first acquisition frequency F₁ 。

Preferably, the enhancing processing in S2 specifically includes:

filtering the DC cell culture image by using a Gaussian low-pass filter to generate a noise reduction image;

and carrying out histogram equalization operation on the noise reduction image.

Preferably, the enhancing processing in S2 specifically includes:

constructing a bilateral filter based on an improved bilateral filtering algorithm to filter the DC cell culture image, and generating a noise reduction image;

and carrying out histogram equalization operation on the noise reduction image.

Preferably, the construction of the bilateral filter based on the improved bilateral filtering algorithm is as follows:

f(x)＝h(x)+d(x)，

g(x)＝h(x)+Kd(x)，

wherein f (x) represents an original image, g (x) represents a noise reduction image, h (x) represents a base layer of the original image, d (x) represents a detail layer of the original image, and K represents an adjustment factor;

the base layer is calculated as follows:

wherein x represents the pixel position, y represents the neighborhood pixel value position of x, Ω= [ -p, p]² Represents a filtering window centered on (0, 0), σ represents a scale parameter, and the range is [0,255]The term, |·| represents the euclidean norm;

and (3) performing saliency map feature calculation:

S(x)＝||f_G (x)-μ||² ，

wherein S (x) represents a saliency map, μ represents an average pixel value of the CIE-Lab spatial input image, f_G (x) Gaussian filtering of f (x) is represented, and after filtering, filtering results are converted into CIE-Lab space;

the significance map is linearly normalized to [0,1];

setting a threshold T, and if S (x) > T, identifying one pixel as saliency, thereby dividing the image into a salient region and a non-salient region, and performing range mapping:

σ(x)＝Ψλ_,T (S(x))，

wherein ψ is_λ,T (T) represents a mapping function, λ represents a slope parameter of sigmoid, and T represents a displacement parameter of sigmoid for distinguishing a high-saliency region from a low-saliency region;

the final adjustment is as follows:

t epsilon [0.2, [ 05] and sigma epsilon [50,80] are set.

Preferably, the S3 specifically includes:

acquiring image data, adding a class to an immature DC cell image and adding a class and a position data tag to the mature DC cell image to obtain a data set;

constructing a deep learning model based on DC cell detection, and training through the data set;

image set P to be identified through trained deep learning model based on DC cell detection₁ And (3) detecting mature DC cells in the image to be identified, and outputting a detection result.

Preferably, the acquiring image data specifically includes:

obtaining an image of immature DC cells in an in-vitro culture process of the DC cells by an electron microscope, wherein the image of immature DC cells is provided with immature DC cells and is not provided with mature DC cells, and the immature DC cells have the morphology of the immature DC cells and are verified to be an immature DC phenotype by a flow cytometer;

acquiring a mature DC cell image in the in-vitro culture process of the DC cells through an electron microscope, wherein the mature DC cell image is used as a second original image, and the mature DC cells have the form of mature DC cells and are verified to be in a mature DC phenotype through a flow cytometer;

acquiring a rotated image by rotating the first original image and the second original image from 0 ° to 270 ° around the center points thereof;

selecting different values according to T epsilon [0.2, [ 05] and sigma epsilon [50,80], and constructing a bilateral filter based on an improved bilateral filtering algorithm to filter a first original image, a second original image and a rotating image, so as to realize enhancement of the original images to different degrees and obtain enhanced images;

respectively inputting two similar first original images, second original images, rotating images or enhanced images into a coder Vencoder to obtain two compliant normal distributions;

randomly acquiring and processing the two normal distributions to obtain a 100-dimensional feature vector, inputting the 100-dimensional feature vector into a decoder Vdecoder, and fusing to obtain a cell superposition image;

the first original image, the second original image, the rotation image, the enhancement image and the cell superposition image are proportionally formed into a training set, a verification set and a test set.

In a second aspect, embodiments of the present invention provide a DC cell detection system comprising a memory and a processor, the memory and the processor being coupled, the memory being for storing a computer program capable of running on the processor, the processor being for performing the DC cell detection method of any one of claims 1-9 when the computer program is run.

One of the above technical solutions has the following beneficial effects:

the method provided by the embodiment of the invention provides a DC cell detection method, realizes monitoring of DC cell cultivation by means of artificial intelligence, avoids manpower waste and inaccuracy, can accurately classify the DC cell state, and improves accurate data support for quantitative analysis of the DC cell growth process.

[ description of the drawings ]

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of S1-S4 of a DC cell detection method according to an embodiment of the present invention;

FIG. 2 is a block diagram of a DC cell detection system according to an embodiment of the present invention.

[ detailed description ] of the invention

In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. It will be apparent that the described embodiments are only some, but not all, embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Please refer to fig. 1, which is a flow chart of the DC cell detection methods S1-S4 according to an embodiment of the present invention. As shown in fig. 1, the DC cell detection method according to the embodiment of the present invention includes:

By adopting the technical scheme, the first acquisition frequency F is adopted₁ The method comprises the steps of obtaining cell culture images, sequentially separating out selected area images, sequentially detecting immature DC cells in the selected area images based on a support vector machine, wherein the processing speed can be increased, the processing time can be shortened, the precision can be increased, and the immature DC cells can be rapidly detected by using the support vector machine; when immature DC cells are detected, to achieve more accurate monitoring of the DC cellsBy a second acquisition frequency F₂ Acquiring a DC cell culture image, and performing enhancement treatment to improve classification accuracy to obtain an image set P to be identified₁ The method comprises the steps of carrying out a first treatment on the surface of the Constructing a deep learning model based on DC cell detection, and identifying an image set P to be identified through the deep learning model based on DC cell detection₁ Mature DC cell detection is carried out on the image to be identified, so that the detection of the mature DC cell can be more accurately realized; acquiring an original DC cell culture image of the detected mature DC cells, and adding a time mark to obtain a mature DC cell image set P₂ Thereby completing the complete monitoring process of the mature DC cell culture; according to the invention, the monitoring of the DC cell culture is realized by means of artificial intelligence, the manpower waste and inaccuracy are avoided, the DC cell state can be accurately classified, and the accurate data support is improved for quantitative analysis of the DC cell growth process.

For some possible design ideas, the separating the selected area image in S1 specifically includes:

wherein S represents the area of the dynamic rectangular region, (x'₁ ,y′₁ ) Representing the upper left corner position coordinates of the dynamic rectangular region, (x'₂ ,y′₂ ) Represents the position coordinate of the right lower corner of the dynamic rectangular area, w represents the scaling, 0 < w < 1, S₀ Representing the area of the cell culture image, (x)₁ ,y₁ ) Representing the upper left corner position coordinates of the cell culture image, (x)₂ ,y₂ ) Representing the lower right corner of the cell culture imageSetting coordinates;

It should be noted that, as the culture proceeds, the greater the probability of monocytes and peripheral blood cells inducing immature DC cells, and at the same time, the complete detection of immature DC cells is not required. Therefore, the area of the separated selected area image is increased along with the segmentation times, so that the processing and detection speed is increased, and the later missing detection of immature DC cells is avoided.

For some possible design ideas, the step S1 of sequentially performing immature DC cell detection on the selected area image based on the support vector machine specifically includes:

obtaining an LBP histogram based on the integer value matrix;

For some possible design ideas, the support vector machine model is constructed as follows:

establishing a sample set { x }_i The method comprises the steps of carrying out a first treatment on the surface of the i=1, …, N }, N representing the total number of samples;

the relaxation factor ζ is introduced, constrained as follows:

(x_i -a)(x_i -a)^T ≤R² +ξ_l ，

ξ_l ≥0，

wherein R represents a target value, ζ represents a relaxation factor,_T representing a transpose of representations;

the minimum extremum is found by the following function:

wherein, C represents a set penalty factor, and the following Lagrangian equation is established:

wherein, xi_i Represents the ith relaxation factor, a_i And gamma_i Represents the Lagrangian multiplier and a_i Not less than 0 and gamma_i ≥0，x_i Representing the input data;

when the partial derivative is zero, the support vector machine model satisfies the constraint condition:

∑_l a_i ＝1，

0≤a_i ≤C，

in consideration of the limiting conditions, for a_l Maximization is performed:

when z is less than the target value, then sample z satisfies the requirement:

the training set is described by the above formula, and the support object is a_l Not equal to 0, in the case where C.ltoreq.1 and the support object is equal to C, the sample is considered a non-target class.

As monocytes and peripheral blood cells are induced to form immature DC cells, the cell shape changes, and the cells form dendritic projections, so that the irregularity of the cell shape increases to cause the change of texture, the change is recorded by an LBP histogram and is reflected to pixel concentration, therefore, the LBP histogram is used for training a support vector machine model, the change can be rapidly detected, the detection of the immature DC cells is completed, and the support vector machine can be used for rapidly detecting the immature DC cells.

It should be noted that the support vector machine model of the present invention may adopt the model designed in the above-mentioned manner, so that the detection speed is faster, but may also adopt other existing support vector machine models as required.

For some possible design considerations, the second acquisition frequency F₂ Is greater than the first acquisition frequency F₁ 。

For some possible design ideas, the enhancing processing in S2 specifically includes:

and carrying out histogram equalization operation on the noise reduction image.

The invention effectively removes noise by using the Gaussian low-pass filter, simultaneously retains texture information, generates a noise reduction image, and realizes image enhancement by normalizing the image by applying a histogram equation method in order to compensate for potential negative effects by considering that image contrast or brightness may be different.

performing histogram equalization operation on the noise reduction image usage;

the construction of the bilateral filter based on the improved bilateral filtering algorithm is as follows:

f(x)＝h(x)+d(x)，

g(x)＝h(x)+Kd(x)，

the base layer is calculated as follows:

and (3) performing saliency map feature calculation:

S(x)＝||f_G (x)-μ||² ，

the significance map is linearly normalized to [0,1];

σ(x)＝Ψ_λ,T (S(x))，

the final adjustment is as follows:

t epsilon [0.2, [ 05] and sigma epsilon [50,80] are set.

According to the invention, the bilateral filter is constructed based on the improved bilateral filtering algorithm to filter the DC cell culture image, the enhancement is realized rapidly by using the saliency map, the noise is removed, the details of the non-salient region are not amplified, and the texture information is reserved.

For some possible design considerations, S3 specifically includes:

image set P to be identified through trained deep learning model based on DC cell detection₁ Detecting mature DC cells of the image to be identified, and outputting a detection result;

the deep learning model based on DC cell detection is a deep convolutional neural network model based on AlexNet, and sequentially comprises an input layer, a convolutional layer 1, a maximum pooling layer 1, a convolutional layer 2, a maximum pooling layer 2, a convolutional layer 3, a convolutional layer 4, a convolutional layer 5, a maximum pooling layer 3, a full-connection layer 1, a full-connection layer 2 and an output layer; the size of an input image of the input layer is 227 x 3, the convolution layer 1 comprises 96 convolution kernels, the size of each convolution kernel is 11 x 3, the stride is 4, and the convolution is followed by batch normalization to form a 55 x 55 feature map; the convolution layer 2 comprises 256 convolution kernels, the size of each convolution kernel is 5×5×96, the stride is 1, and after convolution, batch normalization is carried out to form a 27×27 feature map; the convolution layer 3 and the convolution layer 4 comprise 384 convolution kernels, wherein each convolution kernel has a size of 3×3×256 and a stride of 1, and a feature map of 13×13 is formed; the convolution layer 5 comprises 256 convolution kernels, the size of each convolution kernel is 3 x 192, the stride is 1, and after convolution, batch normalization is carried out to form a feature map of 13 x 13; the maximum pooling layer 1, the maximum pooling layer 2 and the maximum pooling layer 3 adopt the maximum pooling with the steps of 2 and the size of 3*3; the full-connection layer 1 and the full-connection layer 2 comprise 4096 neurons, dropout operation is carried out, the output layer is a Softmax classifier, and the output class is 2.

Image set P to be identified through trained deep learning model based on DC cell detection₁ The image to be identified is used for detecting mature DC cells, the accuracy is high, and the robustness of the deep learning model is good. Mature DC cells are in a typical dendritic cell form, namely, the surface of the mature DC cells is visible with more extended burr-like protrusions, and the immature DC cells are easy to ignore because of being developed into the mature DC cells, so that the mature DC cells can be automatically detected by using a computer-aided detection method, the detection is accurate, and the negligence and the burden of personnel can be reduced.

It should be noted that the deep learning model of the present invention may adopt the model designed in the above-mentioned manner, so that the detection speed is more accurate, but may also adopt other existing deep learning models as required.

For some possible design ideas, the acquiring image data specifically includes:

wherein, the morphology of the immature DC cells is that the cells form dendritic projections, the immature DC phenotypes are that the positive rate of CD40, CD80, CD86 and HLA-DR is high, and the CD83 is low;

the mature DC cells are in the form of burr-shaped protrusions with more stretching surface, and the positive rates of the mature DC phenotypes of CD40, CD80, CD83, CD86 and HLA-DR are all obviously high-expressed;

the first original image, the second original image, the rotation image, the enhancement image and the cell superposition image are formed into a training set, a verification set and a test set in proportion;

training a model through a training set and verifying the model through a verification set in the training process;

the performance of the model is tested by the test set.

The invention realizes increasing the data set through the operation, thereby increasing the training condition of the deep learning model and improving the detection accuracy.

The embodiment of the invention further provides a system embodiment for realizing the steps and the method in the method embodiment.

Referring to fig. 2, which is a block diagram of a DC cell detection system according to an embodiment of the invention, the DC cell detection system 100 may include a processor 110, a machine-readable storage medium 120 (storage), a bus 130, and acommunication unit 140.

The processor 110 may perform various suitable actions and processes based on programs stored in the machine-readable storage medium 120, such as the program instructions associated with the DC cell detection methods described in the foregoing embodiments. The processor 110, the machine-readable storage medium 120, and thecommunication unit 140 communicate signals over the bus 130.

In particular, the processes described in the above exemplary flowcharts may be implemented as computer software programs, in accordance with embodiments of the present invention. For example, embodiments of the present invention include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flowcharts. In such an embodiment, the computer program may be downloaded and installed from a network via thecommunication unit 140, which, when executed by the processor 110, performs the above-described functions defined in the method of the embodiment of the invention.

Still another embodiment of the present invention provides a computer-readable storage medium having stored therein computer-executable instructions which, when executed by a processor, are configured to implement the DC cell detection method according to any of the above embodiments.

The computer readable medium of the present invention may be a computer readable signal medium, a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (LAM), a read-only memory (LOM), an erasable programmable read-only memory (EPLOM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-LOM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, LM (radio frequency), etc., or any suitable combination of the foregoing.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device.

The computer-readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to perform the methods shown in the above-described embodiments.

Yet another embodiment of the invention provides a computer program product comprising a computer program which, when executed by a processor, implements a DC cell detection method as described in any of the above embodiments.

The terms "first," "second," "third," "fourth" and the like in the description and in the claims and in the above drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented, for example, in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "includes" and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.

Claims

1. A method of DC cell detection, the method comprising:

2. The method according to claim 1, wherein the step of separating the image of the selected region in S1 comprises:

3. The method for detecting DC cells according to claim 2, wherein the step S1 of sequentially performing immature DC cell detection on the selected region image based on a support vector machine comprises:

obtaining an LBP histogram based on the integer value matrix;

4. The method of claim 1, wherein the second acquisition frequency F₂ Is greater than the first acquisition frequency F₁ 。

5. The method for detecting DC cells according to claim 1, wherein the step of performing enhancement treatment in S2 comprises:

and carrying out histogram equalization operation on the noise reduction image.

6. The method for detecting DC cells according to claim 1, wherein the step of performing enhancement treatment in S2 comprises:

and carrying out histogram equalization operation on the noise reduction image.

7. The method of claim 6, wherein the construction of the bilateral filter based on the modified bilateral filtering algorithm is as follows:

f(x)＝h(x)+d(x)，

g(x)＝h(x)+Kd(x)，

the base layer is calculated as follows:

and (3) performing saliency map feature calculation:

S(x)＝||f_G (x)-μ||² ，

the significance map is linearly normalized to [0,1];

σ(x)＝Ψ_λ,T (S(x))，

wherein ψ is_λ,T (T) represents a mapping function, λ represents a slope parameter of sigmoid, and T represents a displacement parameter of sigmoid forDistinguishing between high and low saliency regions;

the final adjustment is as follows:

t epsilon [0.2, [ 05] and sigma epsilon [50,80] are set.

8. The method for detecting a DC cell according to any one of claims 6 or 7, wherein S3 specifically comprises:

9. The method for detecting a DC cell according to claim 8, wherein the acquiring image data specifically comprises:

10. A DC cell detection system, characterized in that the system comprises a memory and a processor, the memory and the processor being coupled, the memory being for storing a computer program capable of running on the processor, the processor being for performing the DC cell detection method according to any one of claims 1-9 when the computer program is run.