Intelligent identification system and method of respiratory tumor cells based on deep neural network

Abstract

The present invention belongs to the technical field of medical image analysis. The present invention discloses an intelligent identification system and method for respiratory tumor cells based on a deep neural network, including: a data acquisition module: collecting cell image data of respiratory tumor cells and corresponding image description labels; an image processing module: performing adaptive energy compensation on the cell image data to obtain a pixel compensation map, performing multimodal enhancement on the pixel compensation map to obtain a detail enhancement map; a spatial enhancement module: performing structural visualization on the detail enhancement map to obtain a cell visualization map; a model construction module: constructing a tumor cell recognition model based on the cell visualization map and the corresponding image description labels, and realizing high-precision recognition of tumor cells based on the constructed tumor cell recognition model; thereby greatly improving the efficiency and accuracy of respiratory tumor cell recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image analysis, and more specifically, to a system and method for intelligently identifying respiratory tumor cells based on a deep neural network. Background Art

[0002] With the continuous development of medical imaging, molecular biology and artificial intelligence technology, tumor diagnosis, especially the early detection of respiratory tumors, has become an important research direction in the field of clinical medicine. Early detection of respiratory tumors, especially lung cancer, tracheal cancer, laryngeal cancer, etc., is crucial to improving patients' survival rate. In recent years, with the successful application of artificial intelligence technology, especially deep learning algorithms in the medical field, medical image analysis technology has made significant progress, especially in the early diagnosis of cancer, cytological analysis and disease prediction, and has achieved unprecedented results.

[0003] At present, the diagnosis of respiratory tumors mostly relies on imaging examinations, tissue biopsies, and molecular biological testing. Traditional medical imaging technologies often have problems with low image quality, high noise, and severe background interference in tumor cell detection. In particular, in low-resolution images, the details of tumor cells may be blurred or obscured. Especially when cells are in contact or adhered to each other, the blurred boundaries make it impossible for segmentation algorithms to accurately identify and distinguish different cells, resulting in difficulties in morphological analysis and further diagnosis. Traditional image processing methods, such as threshold segmentation and edge detection, usually rely on manually designed feature extraction algorithms and lack in-depth understanding and adaptability of cell morphology and complex structures, resulting in The effect is limited in practical applications. Although deep learning (especially convolutional neural networks, CNN) has made certain progress in the field of medical image processing in recent years, most existing deep learning models mainly rely on fixed network structures and lack sufficient consideration of image quality, noise interference and dynamic changes of cells. There are still problems with insufficient accuracy and robustness. In many tumor cells, the chromatin distribution of the cell nucleus is uneven, which makes the morphological and structural analysis of the cell nucleus more complicated. Variations in chromatin distribution can affect the identification of tumors, especially in the early stages of malignant tumors. Chromatin changes are not always obvious, and traditional methods have difficulty identifying these subtle differences, which in turn leads to misidentification of tumor cells.

[0004] In view of this, the present invention proposes a respiratory tumor cell intelligent identification system and method based on deep neural network to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solution: an intelligent identification system for respiratory tumor cells based on a deep neural network, comprising:

[0006] Data acquisition module: collects cell image data of respiratory tumor cells and corresponding image description labels;

[0007] Image processing module: includes a compensation unit and an enhancement unit. The compensation unit performs adaptive energy compensation on the cell image data to obtain a pixel compensation image; the enhancement unit performs multimodal enhancement on the pixel compensation image to obtain a detail enhancement image.

[0008] Spatial enhancement module: Perform structural visualization on the detail enhancement map to obtain a cell visualization map;

[0009] Model construction module: Build a tumor cell recognition model based on the cell visualization map and the corresponding image description labels, and achieve high-precision recognition of tumor cells based on the constructed tumor cell recognition model.

[0010] Furthermore, the cell image data is obtained by labeling tumor cells using a fluorescent labeling method, and performing cell imaging on the labeled cells using a fluorescent microscope to obtain the cell image data.

[0011] Furthermore, the method of performing adaptive energy compensation on the cell image data includes:

[0012] The RGB weighted average method is used to grayscale the cell image data to obtain a cell grayscale image; based on the pixel width and pixel height of the cell grayscale image, the pixel in the lower left corner of the cell grayscale image is used as the pixel origin to mark the coordinates of each pixel to obtain the pixel coordinates of each pixel; a compensation window is preset, for pixel A in the cell grayscale image, pixel A is used as the center of the compensation window, and other pixels circled in the compensation window are used as the window pixel set of pixel A; based on the window pixel set, each pixel in the cell grayscale image is balanced and defined to obtain a balance coefficient; a balance threshold is preset, and pixels with a balance coefficient greater than the balance threshold are used as detail pixels. Based on the window pixel set of the detail pixels, the grayscale probability of each pixel in the window pixel set is calculated, and the number of each pixel grayscale in the window pixel set is divided by the size of the pixel grayscale set; pixel energy compensation is performed on the detail pixels based on the grayscale probability to obtain a compensated grayscale; grayscale compensation is performed on the detail pixels based on the compensated grayscale, and the pixel grayscale of the detail pixels is adjusted to the corresponding compensated grayscale to obtain a pixel compensation image.

[0013] Furthermore, the calculation formula for balancing each pixel in the cell grayscale image is: ;in, The pixel coordinates are The equalization coefficient of the pixel at Represents the size of the window pixel set, Represents the first pixel in the window The pixel grayscale of pixels, Represents the first pixel in the window The pixel grayscale of pixels, The horizontal axis value representing the pixel coordinate, The vertical axis value representing the pixel coordinate;

[0014] The formula for performing pixel energy compensation on detail pixels is:

[0015] ;in, The pixel coordinates are The compensated grayscale of the pixel at The pixel coordinates are The pixel grayscale of the pixel at represents the energy parameter, Represents the first pixel in the window The grayscale probability of a pixel, Represents the first pixel in the window The pixel grayscale of pixels.

[0016] Furthermore, the method of performing multimodal enhancement on the pixel compensation map includes:

[0017] A standard rate-of-change operator is used to perform local analysis on each pixel in the pixel compensation map to obtain the pixel change rate. A description vector for each pixel is constructed based on the pixel change rate and pixel grayscale. For pixel A in the pixel compensation map, the pixels adjacent to pixel A in the four directions (upper, lower, left, and right) are used as the adjacent pixels of pixel A, and all the adjacent pixels of pixel A constitute the adjacent pixel set of pixel A. Each pixel in the pixel compensation map is used as a point to be evaluated, and an edge assessment is performed on the point to be evaluated based on the adjacent pixel set to obtain a boundary assessment value.

[0018] The evaluation threshold is preset, and the points to be evaluated whose boundary evaluation values ​​are greater than or equal to the evaluation threshold are taken as boundary points. Local change adjustment is performed on the boundary points. The formula for local change adjustment of the boundary points is:

[0019] ;in, Represents the enhanced change rate of the boundary point, Represents the pixel change rate of the boundary point, represents the enhancement parameter, The size of the set of adjacent pixels representing the boundary point, The first pixel in the adjacent pixel set representing the boundary point The pixel change rate of adjacent pixels, represents the similarity rating function, The first pixel in the adjacent pixel set representing the boundary point description vectors of adjacent pixels, A description vector representing the boundary point; performing an inverse operation of local analysis on the boundary point based on the enhancement change rate to obtain the boundary grayscale; adjusting the pixel grayscale of the boundary point to the corresponding boundary grayscale based on the boundary grayscale to obtain a detail enhancement map.

[0020] Furthermore, the formula for performing edge evaluation on the point to be evaluated is:

[0021] ;in, represents the boundary assessment value, represents the size of the adjacent pixel set, Represents the feature category describing the vector, represents the edge parameter, Represents the neighboring pixel set of the point to be evaluated. The description vector of the adjacent pixels The value of the feature, Represents the first The value of a feature.

[0022] Furthermore, the method of performing structural visualization on the detail enhancement map includes:

[0023] The pixel coordinate system is constructed with the pixel at the center of the detail enhancement image as the pixel origin, the vertical pixel distance and horizontal pixel distance from other pixels to the pixel origin are used as the mapping position, and the pixel grayscale is used as the attribute to map the detail enhancement image to the pixel coordinate system; the pixel coordinate system is transformed into an image spectrum to obtain the image spectrum; for the image spectrum, the spectrum amplitude and spectrum phase of the image spectrum at each spectrum horizontal component scale are recorded, and the corresponding pixels in the detail enhancement image are spectrally uniformly enhanced based on the spectrum amplitude and spectrum phase. The formula for spectrally uniformly enhancing the corresponding pixels in the detail enhancement image is as follows:

[0024] ;in, The frequency domain coordinates are The enhanced spectrum of the corresponding pixel at represents the spectrum amplitude, represents the spectral phase, represents pi, Represents the traversal from the starting frequency domain coordinate of the image spectrum to the frequency domain coordinate The image spectrum values ​​at the locations are summed; a spectrum inverse function is constructed based on the formula for image spectrum conversion, and the enhanced spectrum is calculated by the spectrum inverse function to obtain the spectrum enhancement grayscale value. Based on the spectrum enhancement grayscale value, the image grayscale is adjusted to match the corresponding spectrum enhancement grayscale value to obtain a cell visualization map.

[0025] Furthermore, the formula for performing image spectrum conversion on the mapping coordinate system is:

[0026] ;in, The frequency domain coordinates are The image spectrum at Represents the coordinates in the pixel coordinate system The pixel grayscale of the pixel at represents the imaginary unit, represents pi, Represents half the pixel height in the detail enhancement image, Represents half the pixel width in the detail enhancement image, Represents the horizontal axis value of the pixel in the pixel coordinate system, Represents the vertical axis value of the pixel in the pixel coordinate system, Represents the integration of the horizontal axis value. Represents the integration of the vertical axis value. Represents the horizontal spectrum component corresponding to the horizontal axis value, Represents the vertical component of the spectrum corresponding to the vertical axis value.

[0027] Furthermore, the tumor cell recognition model is constructed by:

[0028] Based on the cell visualization image and the corresponding image description label, the CNN model is used as the initial model of the tumor cell recognition model, the cell visualization image and the corresponding image description label are used as training data, and the training data is used as the training sample set. The CNN model is trained using the training sample set, and the cell visualization image and the corresponding image description label are used as the input data of the tumor cell recognition model, and the predicted image prediction label is used as the output data of the tumor cell recognition model; minimizing the error between the actual image description label and the image prediction label predicted by the tumor cell recognition model is used as the training goal, and the recall rate function is used as the loss function of the tumor cell recognition model. When the loss function reaches convergence, the training is stopped to obtain the tumor cell recognition model.

[0029] Intelligent identification method for respiratory tumor cells based on deep neural network, including:

[0030] S1, collecting cell image data of respiratory tumor cells and corresponding image description labels;

[0031] S2. performing adaptive energy compensation on the cell image data to obtain a pixel compensation image; performing multimodal enhancement on the pixel compensation image to obtain a detail enhancement image;

[0032] S3, perform structural visualization on the detail enhancement map to obtain a cell visualization map;

[0033] S4. Construct a tumor cell recognition model based on the cell visualization map and the corresponding image description labels, and achieve high-precision recognition of tumor cells based on the constructed tumor cell recognition model.

[0034] The technical effects and advantages of the present invention's intelligent respiratory tumor cell identification system and method based on deep neural network are as follows:

[0035] The present invention improves the overall quality of cell images by performing adaptive energy compensation on cell image data, better displays the outlines and details of cells, and especially provides clearer views of important structures such as cell nuclei and chromatin; by performing multimodal enhancement on pixel compensation images, the edge information in the image is effectively enhanced, making the difference between tumor cells and surrounding normal cells or background more obvious, which can significantly improve the clarity of cell boundaries, eliminate fuzzy areas, and effectively improve the accuracy of accurate identification of tumor cells; by performing structural visualization on detail enhancement images, the subtle morphological differences between cancer cells and normal cells can be effectively enhanced, making key information such as cell nuclei and chromatin more prominent in the image, providing higher accuracy for automated cell recognition systems, and also providing clearer reference images for manual inspection, thereby improving the reliability of tumor cell recognition; through the construction of a tumor cell recognition model, more reliable data support is provided for pathological diagnosis, treatment plan formulation, etc., greatly improving the efficiency and accuracy of tumor cell recognition, and reducing the occurrence of human errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Schematic diagram of the intelligent recognition system for respiratory tumor cells based on deep neural network of the present invention;

[0037] Figure 2 Schematic diagram of the intelligent identification method of respiratory tumor cells based on deep neural network of the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] Example 1;

[0040] See also Figure 1 As shown, the intelligent respiratory tumor cell recognition system based on deep neural network described in this embodiment includes:

[0041] Data acquisition module: collects cell image data of respiratory tumor cells and corresponding image description labels;

[0042] Image processing module: includes a compensation unit and an enhancement unit. The compensation unit performs adaptive energy compensation on the cell image data to obtain a pixel compensation image; the enhancement unit performs multimodal enhancement on the pixel compensation image to obtain a detail enhancement image.

[0043] Spatial enhancement module: structurally visualize the detail enhancement map to obtain a cell visualization map;

[0044] Model construction module: Builds a tumor cell recognition model based on the cell visualization map and the corresponding image description labels, and achieves high-precision recognition of tumor cells based on the constructed tumor cell recognition model;

[0045] Each module is connected via wired and / or wireless means to achieve data transmission between modules;

[0046] The cell image data is obtained by labeling tumor cells with a fluorescent labeling method and imaging the labeled cells using a fluorescence microscope. The image description label is the annotation information of the tumor cells corresponding to the cell image data.

[0047] Methods for adaptive energy compensation of cell image data include:

[0048] The RGB weighted average method was used to grayscale the cell image data to obtain a cell grayscale image. Based on the pixel width and pixel height of the cell grayscale image, the pixel in the lower left corner of the cell grayscale image was used as the pixel origin to mark the coordinates of each pixel to obtain the pixel coordinates of each pixel. A compensation window was preset. For pixel A in the cell grayscale image, pixel A was used as the center of the compensation window, and the other pixels circled in the compensation window were used as the window pixel set of pixel A. Based on the window pixel set, each pixel in the cell grayscale image was balanced and defined. The calculation formula for the balanced definition of each pixel in the cell grayscale image was as follows: ;in, The pixel coordinates are The equalization coefficient of the pixel at Represents the size of the window pixel set, Represents the first pixel in the window The pixel grayscale of pixels, Represents the first pixel in the window The pixel grayscale of pixels, The horizontal axis value representing the pixel coordinate, The vertical axis value represents the pixel coordinates. The smaller the equalization coefficient, the simpler the pixel area is and the more likely it is the background area. The smaller the equalization coefficient, the more complex the pixel area is and the more texture and details it contains.

[0049] A preset equalization threshold is set by a person skilled in the art based on actual conditions. Pixels with an equalization coefficient greater than the equalization threshold are regarded as detail pixels. Based on the window pixel set of the detail pixels, the grayscale probability of each pixel in the window pixel set is calculated, and the number of each pixel grayscale in the window pixel set is divided by the size of the pixel grayscale set. Pixel energy compensation is performed on the detail pixels based on the grayscale probability. The formula for performing pixel energy compensation on the detail pixels is:

[0050] ;in, The pixel coordinates are The compensated grayscale of the pixel at The pixel coordinates are The pixel grayscale of the pixel at Represents the energy parameter, which is used to adjust the sensitivity of energy compensation. Represents the first pixel in the window The grayscale probability of a pixel, Represents the first pixel in the window grayscale compensation is performed on detail pixels based on the compensation grayscale, and the pixel grayscale of the detail pixels is adjusted to the corresponding compensation grayscale to obtain a pixel compensation map;

[0051] By performing adaptive energy compensation on cell image data, the brightness of each area can be reasonably enhanced or suppressed, avoiding overcompensation or uneven illumination, and enhancing the details in the image, thereby improving the overall image quality. In particular, in cell morphology analysis, cell contours can be better displayed.

[0052] The multimodal enhancement methods for the pixel compensation map include:

[0053] A standard rate of change operator is used to perform local analysis on each pixel in the pixel compensation image (gradient operation is performed on the pixel using the rate of change operator) to obtain the pixel change rate. Common rate of change operators include the Sobel operator and the Prewitt operator. A description vector for each pixel is constructed based on the pixel change rate and the pixel grayscale. For pixel A in the pixel compensation image, the pixels adjacent to pixel A in the four directions of up, down, left, and right are used as the adjacent pixels of pixel A. All the adjacent pixels of pixel A constitute the adjacent pixel set of pixel A. Each pixel in the pixel compensation image is used as a point to be evaluated. The edge evaluation of the point to be evaluated is performed based on the adjacent pixel set. The formula for edge evaluation of the point to be evaluated is:

[0054] ;in, represents the boundary assessment value, represents the size of the adjacent pixel set, Represents the feature category describing the vector, Represents the edge parameter, which is used to adjust the scale of the boundary assessment value. Represents the neighboring pixel set of the point to be evaluated. The description vector of the adjacent pixels The value of the feature, Represents the first The value of a feature;

[0055] The evaluation threshold is preset. The evaluation threshold is set by those skilled in the art based on actual conditions. The points to be evaluated whose boundary evaluation values ​​are greater than or equal to the evaluation threshold are taken as boundary points. Local changes are adjusted for the boundary points. The formula for local changes to the boundary points is:

[0056] ;in, Represents the enhanced change rate of the boundary point, Represents the pixel change rate of the boundary point, represents the enhancement parameter, which is used to adjust the balance between the pixel change rate of the boundary point and the pixel change rate of the adjacent pixels in the adjacent pixel set. The size of the set of adjacent pixels representing the boundary point, The first pixel in the adjacent pixel set representing the boundary point The pixel change rate of adjacent pixels, Represents the similarity evaluation function. Common similarity evaluation function formulas include cosine similarity function and Euclidean distance function. The first pixel in the adjacent pixel set representing the boundary point description vectors of adjacent pixels, A description vector representing the boundary point; performing an inverse operation of local analysis on the boundary point based on the enhanced change rate (using the change rate operator to perform the inverse operation of local analysis on the boundary point) to obtain the boundary grayscale; adjusting the pixel grayscale of the boundary point to the corresponding boundary grayscale based on the boundary grayscale to obtain a detail enhancement map;

[0057] By adjusting the local changes of the boundary points, the boundary details of the pixel compensation image are significantly enhanced, which effectively improves the clarity of the cell boundaries in the pixel compensation image, eliminates the unclear boundary effects caused by blur, and can effectively improve the effects of cell segmentation and feature extraction.

[0058] Ways to visualize the structure of detail enhancement maps include:

[0059] The pixel at the center of the detail enhancement map is used as the pixel origin to construct a pixel coordinate system. The vertical and horizontal pixel distances from other pixels to the pixel origin are used as mapping positions. The pixel grayscale is used as the attribute to map the detail enhancement map to the pixel coordinate system. The image spectrum conversion is performed on the pixel coordinate system. The formula for image spectrum conversion on the mapping coordinate system is:

[0060] ;in, The frequency domain coordinates are The image spectrum at Represents the coordinates in the pixel coordinate system The pixel grayscale of the pixel at represents the imaginary unit, represents pi, Represents half the pixel height in the detail enhancement image, Represents half the pixel width in the detail enhancement image, Represents the horizontal axis value of the pixel in the pixel coordinate system, Represents the vertical axis value of the pixel in the pixel coordinate system, Represents the integration of the horizontal axis value. Represents the integration of the vertical axis value. Represents the horizontal spectrum component corresponding to the horizontal axis value, Represents the vertical component of the spectrum corresponding to the vertical axis value; for the image spectrum, record the spectrum amplitude and spectrum phase of the image spectrum at each spectrum horizontal component scale, and perform spectrum-consistent enhancement on the corresponding pixels in the detail enhancement image based on the spectrum amplitude and spectrum phase. The formula for spectrum-consistent enhancement is:

[0061] ;in, The frequency domain coordinates are The enhanced spectrum of the corresponding pixel at represents the spectrum amplitude, represents the spectral phase, represents pi, Represents the traversal from the starting frequency domain coordinate of the image spectrum to the frequency domain coordinate The image spectrum values ​​at the locations are summed; an inverse spectrum function (i.e., a function that inversely infers pixel grayscale from the image spectrum) is constructed based on the formula for image spectrum conversion. The enhanced spectrum is calculated using the inverse spectrum function to obtain a spectrum enhancement grayscale value. Based on the spectrum enhancement grayscale value, the image grayscale is adjusted to match the corresponding spectrum enhancement grayscale value to obtain a cell visualization map.

[0062] In tumor cell identification, subtle morphological differences between cancer cells and normal cells (such as abnormal nuclear morphology and uneven distribution of chromatin) are key diagnostic bases. Spectral consistency enhancement can effectively enhance subtle structural differences, thereby improving the effective information in the image (such as the texture of the cell nucleus and chromatin). This helps the automated cell recognition system and provides a more accurate reference for manual inspection, thereby improving the accuracy and reliability of diagnosis.

[0063] The methods for constructing tumor cell recognition models include:

[0064] Based on the cell visualization image and the corresponding image description label, the CNN model is used as the initial model of the tumor cell recognition model, the cell visualization image and the corresponding image description label are used as training data, and the training data is used as the training sample set. The CNN model is trained using the training sample set, and the cell visualization image and the corresponding image description label are used as the input data of the tumor cell recognition model, and the predicted image prediction label is used as the output data of the tumor cell recognition model; minimizing the error between the actual image description label and the image prediction label predicted by the tumor cell recognition model is used as the training goal, and the recall rate function is used as the loss function of the tumor cell recognition model. When the loss function reaches convergence, the training is stopped to obtain the tumor cell recognition model; based on the constructed tumor cell recognition model, high-precision recognition of tumor cells is achieved.

[0065] This embodiment improves the overall quality of cell images by performing adaptive energy compensation on cell image data, better displaying cell outlines and details, and particularly providing clearer views of important structures such as the cell nucleus and chromatin. Multimodal enhancement of the pixel compensation map effectively enhances edge information in the image, making the distinction between tumor cells and surrounding normal cells or background more distinct. This significantly improves the clarity of cell boundaries, eliminates fuzzy areas, and effectively improves the accuracy of tumor cell identification. Structural visualization of the detail enhancement map effectively enhances the subtle morphological differences between cancer cells and normal cells, making key information such as the cell nucleus and chromatin more prominent in the image, providing higher accuracy for automated cell recognition systems and providing clearer reference images for manual inspection, thereby improving the reliability of tumor cell identification. The construction of a tumor cell recognition model provides more reliable data support for pathological diagnosis, treatment plan formulation, etc., significantly improving the efficiency and accuracy of tumor cell identification and reducing the occurrence of human error.

[0066] Example 2;

[0067] See also Figure 2 As shown, for the parts not described in detail in this embodiment, please refer to the description of Example 1. The method for intelligent identification of respiratory tumor cells based on a deep neural network includes:

[0068] S1, collecting cell image data of respiratory tumor cells and corresponding image description labels;

[0069] S2. performing adaptive energy compensation on the cell image data to obtain a pixel compensation image; performing multimodal enhancement on the pixel compensation image to obtain a detail enhancement image;

[0070] S3, perform structural visualization on the detail enhancement map to obtain a cell visualization map;

[0071] S4. Construct a tumor cell recognition model based on the cell visualization map and the corresponding image description labels, and achieve high-precision recognition of tumor cells based on the constructed tumor cell recognition model.

[0072] Example 3;

[0073] This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the operation mode of the above-mentioned method for intelligent identification of respiratory tumor cells based on deep neural networks is implemented.

[0074] Since the electronic device described in this embodiment is an electronic device used to implement the method for intelligent identification of respiratory tumor cells based on a deep neural network in the embodiment of this application, based on the method for intelligent identification of respiratory tumor cells based on a deep neural network described in the embodiment of this application, those skilled in the art can understand the specific implementation of the electronic device of this embodiment and its various variations, so how the electronic device implements the method in the embodiment of this application will not be described in detail here. As long as those skilled in the art implement the electronic device used by the method for intelligent identification of respiratory tumor cells based on a deep neural network in the embodiment of this application, it falls within the scope of protection of this application.

[0075] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0076] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the principles of the present invention are within the scope of protection of the present invention. It should be noted that for users of ordinary skill in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. Intelligent identification system for respiratory tumor cells based on deep neural network, characterized by: include: Data acquisition module: collects cell image data of respiratory tumor cells and corresponding image description labels; Image processing module: including compensation unit and enhancement unit. The compensation unit performs adaptive energy compensation on cell image data to obtain pixel compensation map. The enhancement unit performs multimodal enhancement on the pixel compensation image to obtain a detail enhancement image; Spatial enhancement module: Perform structural visualization on the detail enhancement map to obtain a cell visualization map; Model construction module: Builds a tumor cell recognition model based on the cell visualization map and the corresponding image description labels, and achieves high-precision recognition of tumor cells based on the constructed tumor cell recognition model; The method of performing adaptive energy compensation on cell image data includes: The cell image data is grayed out using the RGB weighted average method to obtain a cell grayscale image; based on the pixel width and pixel height of the cell grayscale image, the pixel in the lower left corner of the cell grayscale image is used as the pixel origin to mark the coordinates of each pixel to obtain the pixel coordinates of each pixel; a compensation window is preset, and for pixel A in the cell grayscale image, pixel A is used as the center of the compensation window, and other pixels circled in the compensation window are used as the window pixel set of pixel A; each pixel in the cell grayscale image is balanced based on the window pixel set to obtain a balance coefficient; a balance threshold is preset, and pixels with a balance coefficient greater than the balance threshold are used as detail pixels, and based on the window pixel set of the detail pixels, the grayscale probability of each pixel in the window pixel set is calculated, and the number of each pixel grayscale in the window pixel set is divided by the size of the pixel grayscale set; pixel energy compensation is performed on the detail pixels based on the grayscale probability to obtain a compensated grayscale; grayscale compensation is performed on the detail pixels based on the compensated grayscale, and the pixel grayscale of the detail pixels is adjusted to the corresponding compensated grayscale to obtain a pixel compensation image; The method of performing structural visualization on the detail enhancement map includes: The pixel coordinate system is constructed with the pixel at the center of the detail enhancement image as the pixel origin, the vertical pixel distance and horizontal pixel distance from other pixels to the pixel origin are used as the mapping position, and the pixel grayscale is used as the attribute to map the detail enhancement image to the pixel coordinate system; the pixel coordinate system is transformed into an image spectrum to obtain the image spectrum; for the image spectrum, the spectrum amplitude and spectrum phase of the image spectrum at each spectrum horizontal component scale are recorded, and the corresponding pixels in the detail enhancement image are spectrally uniformly enhanced based on the spectrum amplitude and spectrum phase. The formula for spectrally uniformly enhancing the corresponding pixels in the detail enhancement image is as follows: ;in, The frequency domain coordinates are The enhanced spectrum of the corresponding pixel at represents the spectrum amplitude, represents the spectral phase, represents pi, Represents the traversal from the starting frequency domain coordinate of the image spectrum to the frequency domain coordinate The image spectrum values ​​at the locations are summed; a spectrum inverse function is constructed based on the formula for image spectrum conversion, and the enhanced spectrum is calculated by the spectrum inverse function to obtain the spectrum enhancement grayscale value. Based on the spectrum enhancement grayscale value, the image grayscale is adjusted to match the corresponding spectrum enhancement grayscale value to obtain a cell visualization map.

2. The deep neural network-based intelligent identification system for respiratory tumor cells according to claim 1, characterized in that: The cell image data is obtained by labeling tumor cells with a fluorescent labeling method, and performing cell imaging on the labeled cells using a fluorescent microscope to obtain the cell image data.

3. The deep neural network-based intelligent identification system for respiratory tumor cells according to claim 2, characterized in that: The calculation formula for balancing each pixel in the cell grayscale image is: ;in, The pixel coordinates are The equalization coefficient of the pixel at Represents the size of the window pixel set, Represents the first pixel in the window The pixel grayscale of pixels, Represents the first pixel in the window The pixel grayscale of pixels, The horizontal axis value representing the pixel coordinate, The vertical axis value representing the pixel coordinate; The formula for performing pixel energy compensation on detail pixels is: ;in, The pixel coordinates are The compensated grayscale of the pixel at The pixel coordinates are The pixel grayscale of the pixel at represents the energy parameter, Represents the first pixel in the window The grayscale probability of a pixel, Represents the first pixel in the window The pixel grayscale of pixels.

4. The deep neural network-based intelligent identification system for respiratory tumor cells according to claim 3, characterized in that: The method of performing multimodal enhancement on the pixel compensation map includes: A standard rate-of-change operator is used to perform local analysis on each pixel in the pixel compensation map to obtain the pixel change rate. A description vector for each pixel is constructed based on the pixel change rate and pixel grayscale. For pixel A in the pixel compensation map, the pixels adjacent to pixel A in the four directions (upper, lower, left, and right) are used as the adjacent pixels of pixel A, and all the adjacent pixels of pixel A constitute the adjacent pixel set of pixel A. Each pixel in the pixel compensation map is used as a point to be evaluated, and an edge assessment is performed on the point to be evaluated based on the adjacent pixel set to obtain a boundary assessment value. The evaluation threshold is preset, and the points to be evaluated whose boundary evaluation values ​​are greater than or equal to the evaluation threshold are taken as boundary points. Local change adjustment is performed on the boundary points. The formula for local change adjustment of the boundary points is: ;in, Represents the enhanced change rate of the boundary point, Represents the pixel change rate of the boundary point, represents the enhancement parameter, The size of the set of adjacent pixels representing the boundary point, The first pixel in the adjacent pixel set representing the boundary point The pixel change rate of adjacent pixels, represents the similarity rating function, The first pixel in the adjacent pixel set representing the boundary point description vectors of adjacent pixels, A description vector representing the boundary point; performing an inverse operation of local analysis on the boundary point based on the enhancement change rate to obtain the boundary grayscale; adjusting the pixel grayscale of the boundary point to the corresponding boundary grayscale based on the boundary grayscale to obtain a detail enhancement map.

5. The deep neural network-based intelligent identification system for respiratory tumor cells according to claim 4, characterized in that: The formula for edge evaluation of the point to be evaluated is: ;in, represents the boundary assessment value, represents the size of the adjacent pixel set, Represents the feature category describing the vector, represents the edge parameter, Represents the neighboring pixel set of the point to be evaluated. The description vector of the adjacent pixels The value of the feature, Represents the first The value of a feature.

6. The deep neural network-based intelligent identification system for respiratory tumor cells according to claim 5, characterized in that: The formula for performing image spectrum conversion on the pixel coordinate system is: ;in, The frequency domain coordinates are The image spectrum at Represents the coordinates in the pixel coordinate system The pixel grayscale of the pixel at represents the imaginary unit, represents pi, Represents half the pixel height in the detail enhancement image, Represents half the pixel width in the detail enhancement image, Represents the horizontal axis value of the pixel in the pixel coordinate system, Represents the vertical axis value of the pixel in the pixel coordinate system, Represents the integration of the horizontal axis value. Represents the integration of the vertical axis value. Represents the horizontal spectrum component corresponding to the horizontal axis value, Represents the vertical component of the spectrum corresponding to the vertical axis value.

7. The deep neural network-based intelligent identification system for respiratory tumor cells according to claim 6, characterized in that: The method for constructing the tumor cell recognition model includes: Based on the cell visualization image and the corresponding image description label, the CNN model is used as the initial model of the tumor cell recognition model, the cell visualization image and the corresponding image description label are used as training data, and the training data is used as the training sample set. The CNN model is trained using the training sample set, and the cell visualization image and the corresponding image description label are used as the input data of the tumor cell recognition model, and the predicted image prediction label is used as the output data of the tumor cell recognition model; minimizing the error between the actual image description label and the image prediction label predicted by the tumor cell recognition model is used as the training goal, and the recall rate function is used as the loss function of the tumor cell recognition model. When the loss function reaches convergence, the training is stopped to obtain the tumor cell recognition model.

8. A method for intelligently identifying respiratory tract tumor cells based on a deep neural network, which is implemented based on the intelligent system for intelligently identifying respiratory tract tumor cells based on a deep neural network according to any one of claims 1 to 7, comprising: S1, collecting cell image data of respiratory tumor cells and corresponding image description labels; S2. performing adaptive energy compensation on the cell image data to obtain a pixel compensation image; performing multimodal enhancement on the pixel compensation image to obtain a detail enhancement image; S3, perform structural visualization on the detail enhancement map to obtain a cell visualization map; S4. Construct a tumor cell recognition model based on the cell visualization map and the corresponding image description labels, and achieve high-precision recognition of tumor cells based on the constructed tumor cell recognition model.

Images