Tumor identification and evaluation method, system and equipment based on multidimensional analysis

Abstract

The present invention discloses a tumor identification and evaluation method, system and equipment based on multidimensional analysis, and relates to the technical field of tumor image recognition. The tumor identification and evaluation method based on multidimensional analysis includes the following steps: comprehensively analyzing and judging the skin tumor identification hidden danger level according to the skin canceration hidden danger results of the functional area of ​​human skin, the human skin skin color level correction results and the ambient light intensity correction results; determining the corresponding skin tumor warning scheme according to the skin tumor identification hidden danger level. The present invention achieves the effect of improving the accuracy of the method of multidimensional analysis of skin tumor images through the analysis of the hidden danger of canceration in the functional area of ​​human skin, the analysis of the influence of human skin skin color, the analysis of the influence of ambient light intensity, the judgment of the skin tumor identification hidden danger level and the determination of the skin tumor warning scheme, and solves the problem of insufficient accuracy of the method of multidimensional analysis of skin tumor images in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of tumor image recognition, and in particular to a tumor recognition and evaluation method, system and device based on multidimensional analysis. Background Art

[0002] With the advancement of medical technology and the development of multidimensional analysis, the demand for early identification, accurate evaluation and personalized treatment of tumors is increasing. As an analytical method that combines multiple data types such as clinical data, biomarkers, and imaging information, multidimensional analysis shows great potential in tumor diagnosis, prognosis assessment, and treatment planning.

[0003] Existing tumor recognition and evaluation methods based on multidimensional analysis are implemented through the following technologies, including: data integration technology, which integrates and standardizes patients' clinical data, biomolecular data, imaging data, etc.; machine learning algorithms, which apply algorithms for feature extraction and pattern recognition.

[0004] For example, the skin cancer image recognition method, system, device and readable storage medium disclosed in the patent application with publication number: CN114881137A include: inputting the skin image to be detected into a deep convolutional neural network skin cancer image recognition model, and the skin image to be detected is processed by the input layer, asymmetric decomposition convolution layer, deep feature extraction layer and category output layer of the deep convolutional neural network skin cancer image recognition model in sequence to obtain an output result, which is the recognition result of the skin image to be detected, and the output result includes whether the skin image to be detected is a skin cancer image and the category of the skin cancer image.

[0005] For example, the invention patent with announcement number CN107403194B discloses a t-SNE-based skin cancer image recognition visualization system, which includes: a matrix acquisition unit, a linearization unit, a t-SNE calculation unit, and a drawing unit. The matrix acquisition unit is used to acquire the matrix obtained after the last convolution layer in the skin cancer image recognition system is convolved and recorded in the matrix data table; the linearization unit is used to linearize the acquired matrix so that each row of the matrix follows the previous row until the matrices become vectors; the t-SNE calculation unit is used to obtain a series of low-dimensional data through t-SNE calculation; and the drawing unit is used to draw points or two-dimensional images based on the low-dimensional data to observe the position of the input image at the drawing point or in the two-dimensional image.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0007] In the prior art, multidimensional analysis methods need to correctly extract the features of tumor images in order to perform effective analysis and identification. However, skin tumor images may be affected by factors such as lighting and skin color, resulting in feature loss during feature extraction and inaccurate subsequent identification judgment. There is a problem of insufficient accuracy in the multidimensional analysis method of skin tumor images. Summary of the invention

[0008] The embodiments of the present application solve the problem of insufficient accuracy of the multidimensional analysis method of skin tumor images in the prior art by providing a tumor recognition and evaluation method, system and equipment based on multidimensional analysis, thereby achieving the effect of improving the accuracy of the multidimensional analysis method of skin tumor images.

[0009] The embodiment of the present application provides a tumor identification and evaluation method based on multidimensional analysis, comprising the following steps: dividing human skin functional areas to obtain human skin functional area parameters, and obtaining skin cancer risk results of human skin functional areas through neural network model analysis; dividing human skin skin color levels to obtain human skin skin color parameters, and obtaining human skin skin color grade correction results by analyzing human skin skin color parameters; dividing ambient light intensity levels to obtain ambient light intensity parameters, and obtaining ambient light intensity correction results by analyzing ambient light intensity parameters; comprehensively analyzing and judging skin cancer risk results of human skin functional areas, human skin skin color grade correction results, and ambient light intensity correction results to obtain skin tumor identification risk levels; and determining corresponding skin tumor warning plans according to the skin tumor identification risk levels.

[0010] Furthermore, the specific process of determining the corresponding skin tumor warning plan according to the skin tumor identification hidden danger level is: if the skin tumor identification hidden danger level is a first-level skin tumor identification hidden danger level, no skin tumor identification abnormal hidden danger warning is issued; if the skin tumor identification hidden danger level is a second-level skin tumor identification hidden danger level, a first-level skin tumor identification abnormal hidden danger warning is issued, and the relevant medical personnel are notified to immediately re-examine the skin condition of the corresponding patient's human skin functional area.

[0011] Furthermore, the skin tumor warning scheme also includes: if the skin tumor identification risk level is a first-level skin tumor identification risk level, the corresponding personnel are reminded to display the comprehensive assessment value of the skin tumor identification risk level through the intelligent display device and popularize skin health knowledge; if the skin tumor identification risk level is a second-level skin tumor identification risk level, a first-level warning of skin tumor identification abnormal risk is issued, and a detailed pathological examination is recommended to the corresponding personnel, and the comprehensive assessment value of the skin tumor identification risk level, human skin functional area parameters, human skin skin color parameters, and ambient light intensity parameters are displayed to relevant medical personnel.

[0012] Furthermore, the specific process of comprehensively analyzing and judging the skin tumor identification risk level according to the skin cancer risk results of the human skin functional areas, the human skin skin color level correction results and the ambient light intensity correction results is as follows: the skin cancer risk results of the human skin functional areas are the skin cancer risk assessment values ​​of the human skin functional areas, and the skin cancer risk assessment values ​​of the human skin functional areas are used to describe the level of skin tumor risk identified in the human skin functional areas; the human skin skin color level correction results are the human skin skin color level correction values, and the human skin skin color level correction values ​​are used to refer to the correction level of human skin skin color difference for the impact on the identification of skin tumor risks; the ambient light intensity correction results are the ambient light comprehensive correction values, and the ambient light comprehensive correction values ​​are used to indicate the correction level of ambient light intensity difference for the impact on the identification of skin tumor risks; according to the skin cancer risk assessment values ​​of the human skin functional areas, the human skin skin color level correction results are The comprehensive evaluation value of the degree of hidden danger of skin tumor identification is calculated by comprehensive analysis of the body skin skin color level correction value and the comprehensive correction value of ambient light; the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is compared with the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification; if the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is not higher than the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification, the skin tumor identification hidden danger level is set to the first-level skin tumor identification hidden danger level; if the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is higher than the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification, the skin tumor identification hidden danger level is set to the second-level skin tumor identification hidden danger level; the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is used to describe the level of the comprehensive hidden danger of skin tumor identified; the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification is used to describe the measurement level of the judgment of the level of comprehensive hidden danger of skin tumor identification corresponding to the level of hidden danger of skin tumor identification.

[0013] Furthermore, the specific calculation formula for the comprehensive evaluation value of the skin tumor identification risk level is:

[0014]

[0015] It represents the comprehensive evaluation value of the degree of hidden danger of skin tumor identification, sets the monitoring points of the functional area of ​​human skin, and numbers the monitoring points of the functional area of ​​human skin in sequence, A0=1,2,...,A, A0 represents the number of the monitoring point of the functional area of ​​human skin, and A represents the total number of the monitoring points of the functional area of ​​human skin; It represents the evaluation value of skin cancer risk in the functional area of ​​human skin at the A0th human skin functional area monitoring point. Indicates the correction value of human skin color level. Indicates the comprehensive correction value of ambient light.

[0016] Furthermore, the specific process of obtaining the hidden danger results of skin cancer in functional areas of human skin through neural network model analysis is: extracting human skin image data through a skin cancer database of functional areas of human skin, the human skin image data including normal skin image data of each functional area of ​​human skin and image data of skin cancer patients; annotating the human skin image data to obtain human skin image annotated data, the specific content of the data annotation including human skin functional area, skin cancer color, skin cancer texture and skin cancer shape; constructing and training a neural network model through the human skin image annotated data, adjusting the neural network model parameters through a back propagation algorithm, and obtaining the hidden danger results of skin cancer in functional areas of human skin through neural network model analysis.

[0017] Furthermore, the specific process of obtaining a human skin skin color grade correction result by analyzing human skin skin color parameters is as follows: extracting a number of human skin skin color data through the shared skin cancer image data of the hospital platform, performing data normalization processing on the number of human skin skin color data to obtain a number of human skin skin color normalized data, and taking the average value of the number of human skin skin color normalized data as the human skin standard skin color value; analyzing the difference between the human skin skin color parameters and the human skin standard skin color value and the actual degree of influence on the identification of hidden dangers of human skin cancer through the principal component analysis algorithm to obtain the human skin skin color grade correction result.

[0018] Furthermore, the specific process of obtaining the ambient light intensity correction result by analyzing the ambient light intensity parameters is as follows: dividing the ambient light intensity data into different categories, each category referring to a light intensity level, and setting the ambient standard light intensity; predicting the specific light intensity level according to the ambient time factor data, the ambient climate factor data and the ambient light factor data through regression analysis; simulating the different specific light intensity level distributions in the environment through ray tracing technology, and obtaining the data of the area affected by the actual specific light intensity level distribution on the identification of skin cancer images; and obtaining the ambient light intensity correction result by combining the different actual specific light intensity level distributions with the data of the area affected by the identification of skin cancer images with the ambient standard light intensity through a clustering algorithm.

[0019] The embodiment of the present application provides a tumor identification and evaluation system based on multidimensional analysis, including: a human skin functional area cancer risk analysis module, a human skin skin color influence analysis module, an ambient light intensity influence analysis module, a skin tumor identification risk level judgment module and a skin tumor warning scheme module; wherein the human skin functional area cancer risk analysis module is used to divide the human skin functional area, obtain human skin functional area parameters, and obtain the human skin functional area skin cancer risk results through neural network model analysis; the human skin skin color influence analysis module is used to divide the human skin skin color level, obtain human skin skin color parameters, and obtain the human skin skin color level correction result by analyzing the human skin skin color parameters; the ambient light intensity influence analysis module is used to divide the ambient light intensity level, obtain the ambient light intensity parameters, and obtain the ambient light intensity correction result by analyzing the ambient light intensity parameters; the skin tumor identification risk level judgment module is used to obtain the skin tumor identification risk level through comprehensive analysis and judgment based on the skin cancer risk results of the human skin functional area, the human skin skin color level correction result and the ambient light intensity correction result; the skin tumor warning scheme module is used to determine the corresponding skin tumor warning scheme according to the skin tumor identification risk level.

[0020] An embodiment of the present application provides an electronic device, comprising: a memory and a processor, wherein the memory is used to store executable instructions of the processor; when the processor is configured to execute the instructions, the electronic device implements a tumor identification and evaluation method based on multidimensional analysis.

[0021] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0022] 1. Through the analysis of the hidden dangers of canceration in functional areas of human skin, the analysis of the influence of human skin color, the analysis of the influence of ambient light intensity, the judgment of the hidden danger level of skin tumor identification and the determination of skin tumor warning plan, the effect of improving the accuracy of the method of multidimensional analysis of skin tumor images is achieved, and the problem of insufficient accuracy of the method of multidimensional analysis of skin tumor images in the existing technology is solved.

[0023] 2. By dividing the functional areas of human skin, we obtain the parameters of the functional areas of human skin, and through the neural network model analysis, we obtain the skin cancer risk results of the functional areas of human skin, so as to evaluate different cancer risk results according to the same skin areas, thereby improving the scientific nature of the evaluation of cancer risk results.

[0024] 3. The skin tumor identification risk level is determined by comprehensively analyzing the skin cancer risk results in the functional areas of human skin, the skin skin color level correction results and the ambient light intensity correction results. Through the comprehensive analysis of multiple factors, the potential risk of skin tumors can be more accurately identified, thereby reducing the error rate of tumor identification assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A schematic diagram of a tumor identification and evaluation method based on multidimensional analysis provided in an embodiment of the present application;

[0026] Figure 2 This is a schematic diagram of an example of a comprehensive evaluation value function image of the degree of hidden danger of skin tumor identification provided in an embodiment of the present application;

[0027] Figure 3 A schematic diagram of the structure of a tumor recognition and evaluation system based on multidimensional analysis provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The embodiments of the present application solve the problem of insufficient accuracy of the methods for multidimensional analysis of skin tumor images in the prior art by providing a tumor recognition and evaluation method, system and equipment based on multidimensional analysis. The purpose of the method for multidimensional analysis of skin tumor images is achieved by analyzing the potential risks of canceration in functional areas of human skin, the influence of human skin color, the influence of ambient light intensity, judging the level of potential risks in skin tumor recognition and determining a skin tumor warning plan.

[0029] The technical solution in the embodiment of the present application is to solve the problem of insufficient accuracy of the method for multi-dimensional analysis of skin tumor images. The overall idea is as follows:

[0030] Through the analysis of the potential risks of canceration in functional areas of human skin, the influence of human skin color, and the influence of ambient light intensity, we obtained the potential risks of skin canceration in functional areas of human skin, the correction results of human skin color level, and the correction results of ambient light intensity respectively. Based on this, the potential risk level of skin tumor identification is analyzed, and the skin tumor warning plan is judged and determined according to the potential risk level of skin tumor identification, thereby achieving the effect of improving the accuracy of the method of multidimensional analysis of skin tumor images.

[0031] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0032] like Figure 1As shown, it is a schematic flow chart of a tumor identification and evaluation method based on multidimensional analysis provided in an embodiment of the present application. The method is applied to a tumor identification and evaluation system based on multidimensional analysis, and the method includes the following steps: dividing the functional areas of human skin to obtain parameters of the functional areas of human skin, and obtaining skin cancer risk results of the functional areas of human skin through neural network model analysis; dividing the skin color levels of human skin to obtain skin color parameters of human skin, and obtaining skin skin color grade correction results of human skin by analyzing the skin color parameters of human skin; dividing the ambient light intensity levels to obtain ambient light intensity parameters, and obtaining ambient light intensity correction results by analyzing the ambient light intensity parameters; comprehensively analyzing and judging the skin cancer risk results of the functional areas of human skin, the skin color grade correction results of human skin, and the ambient light intensity correction results to obtain the skin tumor identification risk level; determining the corresponding skin tumor warning plan according to the skin tumor identification risk level.

[0033] Furthermore, the specific process of determining the corresponding skin tumor warning plan according to the skin tumor identification risk level is: if the skin tumor identification risk level is a first-level skin tumor identification risk level, no skin tumor identification abnormal risk warning is issued; if the skin tumor identification risk level is a second-level skin tumor identification risk level, a first-level skin tumor identification abnormal risk warning is issued, and the relevant medical personnel are notified to immediately re-check the skin condition of the corresponding patient's human skin functional area.

[0034] In this embodiment, if the analysis result shows that the skin tumor identification risk level is level 1, it means that there is a low risk of skin cancer. In this case, no immediate emergency measures are required, but the patient is advised to continue to maintain regular skin self-examination and follow the doctor's general advice, such as avoiding prolonged exposure to the sun and using sunscreen.

[0035] If the analysis results show that the skin tumor identification risk level is level 2, this indicates a higher risk of skin cancer or some suspicious skin changes have already occurred. In this case, notify the relevant medical staff to re-examine the patient, especially for those skins that have shown skin changes or high-risk areas, including physical examination, dermoscopy or further diagnostic tests, such as skin biopsy.

[0036] Furthermore, the skin tumor warning plan also includes: if the skin tumor identification risk level is the first-level skin tumor identification risk level, the corresponding personnel are reminded through the intelligent display device to display the comprehensive assessment value of the skin tumor identification risk level and popularize skin health knowledge; if the skin tumor identification risk level is the second-level skin tumor identification risk level, a first-level warning of skin tumor identification abnormal risks is issued, and a detailed pathological examination is recommended to the corresponding personnel, and the comprehensive assessment value of the skin tumor identification risk level, human skin functional area parameters, human skin skin color parameters, and ambient light intensity parameters are displayed to relevant medical personnel.

[0037] In this embodiment, the corresponding personnel are recommended to undergo more detailed examinations, such as dermoscopy, imaging examinations, or laboratory tests. Dermoscopy is the observation of lesions on the skin through a dermatoscope, which helps to identify the type and malignancy of the tumor, and histopathology is the examination of tissue samples obtained through skin biopsy by a pathologist under a microscope.

[0038] Furthermore, the specific calculation formula for the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is:

[0039]

[0040] It represents the comprehensive evaluation value of the degree of hidden danger of skin tumor identification, sets monitoring points of human skin functional areas, and numbers the monitoring points of human skin functional areas in sequence, A0=1,2,...,A, A0 represents the number of the monitoring points of human skin functional areas, and A represents the total number of the monitoring points of human skin functional areas.

[0041] It represents the evaluation value of skin cancer risk in the functional area of ​​human skin at the A0th human skin functional area monitoring point. Indicates the correction value of human skin color level. Indicates the comprehensive correction value of ambient light.

[0042] In this embodiment, in addition to being obtained through a calculation formula, the comprehensive evaluation value of the degree of hidden danger of skin tumor identification can also be obtained by collecting a large number of skin tumor cases and evaluating the risk level of the tumor using a decision tree algorithm.

[0043] The results of skin cancer risk in functional areas of human skin are obtained through analysis of the neural network model. The results of skin cancer risk in functional areas of human skin are the evaluation values ​​of skin cancer risk in functional areas of human skin. The evaluation values ​​of skin cancer risk in functional areas of human skin are not less than 0 and not greater than 1.

[0044] If the average value of the skin cancer risk assessment values ​​of the functional areas of human skin under multiple monitoring points of the functional areas of human skin is taken as a fixed value of 0.5, if the skin color level correction value of the human skin is taken as a fixed value of 1.5, and the comprehensive correction value of the ambient light is taken as an independent variable, an example diagram of the function image of the comprehensive assessment value of the degree of risk of skin tumor identification is made, as shown in Figure 2 As shown, it is a schematic diagram of an example of a function image of a comprehensive evaluation value of the degree of hidden danger of skin tumor identification provided by an embodiment of the present application; the actual value of the comprehensive correction value of ambient light is not less than 0 and not greater than 1, so in this image, the actual value of the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is not less than 0.063 and not greater than 0.597 (take three significant decimal places), and Table 1 is an example table of the comprehensive evaluation value of the degree of hidden danger of skin tumor identification, and Table 1 is as follows:

[0045] Table 1 Example of comprehensive evaluation values ​​for skin tumor identification risk level

[0046]

[0047]

[0048] According to the function image and Table 1, the comprehensive evaluation value of the degree of hidden danger in skin tumor identification decreases with the increase of the comprehensive correction value of ambient light. When the contrast of ambient light is strong, the image contrast of the comprehensive evaluation of the degree of hidden danger in skin tumor identification increases, which is conducive to the recognition of skin cancer images and conforms to objective laws. If the comprehensive correction value of ambient light is taken as a fixed value and the correction value of human skin color level is taken as the independent variable, the comprehensive evaluation value of the degree of hidden danger in skin tumor identification decreases with the increase of the correction value of human skin color level. The darker the human skin color is, the more unfavorable it is for the comprehensive identification of the degree of hidden danger in skin tumor identification, which conforms to objective laws. The comprehensive analysis of the factor data of the comprehensive evaluation factors of the degree of hidden danger in skin tumor identification under different conditions makes the comprehensive evaluation value of the degree of hidden danger in skin tumor identification more accurate. The function image of the comprehensive evaluation value of the degree of hidden danger in skin tumor identification first decreases and then increases with the change of the correction value of human skin color level and the comprehensive correction value of ambient light, so that the change of the function image conforms to the law of change of external physical conditions, which is convenient for directly observing the change of the comprehensive evaluation of the degree of hidden danger in skin tumor identification.

[0049] Furthermore, the specific process of obtaining the hidden danger results of skin cancer in functional areas of human skin through neural network model analysis is as follows: extracting human skin image data through the skin cancer database of functional areas of human skin, the human skin image data including normal skin image data of each functional area of ​​human skin and image data of skin cancer patients; annotating the human skin image data to obtain human skin image annotated data, the specific contents of the data annotation include human skin functional area, skin cancer color, skin cancer texture and skin cancer shape; constructing and training a neural network model through the human skin image annotated data, adjusting the neural network model parameters through the back propagation algorithm, and obtaining the hidden danger results of skin cancer in functional areas of human skin through the neural network model analysis.

[0050] In this embodiment, human skin image data can also be preprocessed before data annotation: image standardization, unifying the image size, adjusting the brightness, contrast and color balance, so that the image data reaches a consistent preprocessing standard before model training; data enhancement, increasing the diversity of data by means of rotation, flipping, scaling, cropping, etc. to improve the generalization ability of the model; normalization, scaling the pixel values ​​to between 0 and 1.

[0051] Furthermore, the specific process of obtaining the human skin skin color grade correction result by analyzing the human skin skin color parameters is as follows: extracting a number of human skin skin color data through the shared skin cancer image data of the hospital platform, performing data normalization processing on the several human skin skin color data to obtain a number of human skin skin color normalized data, and taking the average value of the several human skin skin color normalized data as the human skin standard skin color value; analyzing the difference between the human skin skin color parameters and the human skin standard skin color value and the actual degree of influence on the identification of hidden dangers of human skin cancer through the principal component analysis algorithm to obtain the human skin skin color grade correction result.

[0052] In this embodiment, the shared skin cancer image data of the hospital platform can be obtained through experimental measurement or from an existing medical database. The data normalization processing of a number of human skin color data can be performed by multi-faceted or unilateral processing based on the division of parameters such as skin color hue, saturation and brightness. For example, when grayscale processing is used, the grayscale range is not less than 0 and not more than 255. If the grayscale value of the actual human skin color value ranges from 70 to 140, the average value of the normalized human skin color data is 105.

[0053] The difference between the human skin color parameter and the human skin standard skin color value and the actual human skin cancer risk identification are analyzed. Generally, the darker the skin color, the more difficult it is to identify the skin cancer image. For example, the actual area of ​​skin cancer is used as the evaluation direction, the gray value corresponding to the human skin color parameter is 140, and the difference with the human skin standard skin color value is 35. The error rate between the skin cancer identification area and the actual skin cancer area is 10%. Then, a mapping relationship is established between the difference 35 and the error rate 10%, and the human skin color level correction value is 1+10%, that is, 1.1.

[0054] In addition to analyzing the difference between skin color parameters and standard skin color values ​​through the principal component analysis algorithm, identifying the main change trends and variation points in the skin color data, and analyzing the ratio of the variation point with the largest change trend to the average value, the human skin color level correction value can also be obtained through the calculation formula. The specific calculation formula for the human skin color level correction value is:

[0055]

[0056] Indicates the human skin color level correction value, S X Indicates the human skin color tone value, S Y Represents the historical average hue of human skin color, C D Indicates the color purity value of human skin, C Y Represents the average historical color purity of human skin color, M S Indicates the color brightness value of human skin, M Y Represents the average historical color brightness of human skin color.

[0057] The HSV color space is widely used in image processing. It expresses the hue, vividness and brightness of colors very intuitively, and is convenient for color comparison. The HSV (hue, saturation, brightness) color space is often used to intuitively represent and adjust the color of images in image processing. In many image processing software and programming languages, images can be converted from RGB (red, green and blue) color space to HSV color space through built-in libraries or functions. Images are acquired through camera equipment, and the human skin color hue value, human skin color historical hue average value, human skin color color purity value, human skin color historical color purity average value, human skin color color brightness value and human skin color historical color brightness average value are obtained through library or function processing in image processing software.

[0058] The darker the human skin color, the greater the negative impact on the identification of skin cancer risk results in the functional areas of the human skin. The hue value, color purity value and color brightness value are compared with the corresponding historical average data values ​​respectively, reflecting the comparison level between the current data value and the historical data value, and evaluating the comparison level between the human skin color and the average skin color.

[0059] In actual situations, the darker the human skin color, the greater the ratio of the human skin color purity value to the average historical color purity value of the human skin color, and the value range is not less than 0 and not greater than 1, and the greater the negative impact on the identification of skin cancer risk results in functional areas of human skin.

[0060] The higher the ratio of the human skin color brightness value to the average historical color brightness of the human skin color, the higher the contrast of the human skin color. The value range is not less than 0 and not greater than 1. The higher the contrast, the greater the positive impact on the identification of skin cancer risk results in functional areas of human skin.

[0061] Comprehensive analysis of the above skin color grade correction factors makes the assessment of human skin color for skin cancer identification more accurate. The function graph of the skin color grade correction value slowly and monotonically increases within the effective definition domain, making it easy to observe the changes in the degree of skin color grade correction factors.

[0062] Furthermore, the specific process of obtaining the ambient light intensity correction result by analyzing the ambient light intensity parameters is as follows: dividing the ambient light intensity data into different categories, each category refers to a light intensity level, and setting the ambient standard light intensity; predicting the specific light intensity level according to the ambient time factor data, the ambient climate factor data and the ambient light factor data through regression analysis; simulating the different specific light intensity level distributions in the environment through ray tracing technology, and obtaining the data of the area affected by the actual specific light intensity level distribution on the identification of skin cancer images; and obtaining the ambient light intensity correction result by combining the different actual specific light intensity level distributions with the ambient standard light intensity analysis through a clustering algorithm.

[0063] In this embodiment, the ambient light intensity data is divided into different categories, each category refers to a light intensity level, a medium light environment, the light intensity is between 500-2000l ux, a high light first environment, the light intensity is between 2000-3000l ux, a high light second environment, the light intensity is between 3000-4000l ux, a high light third environment, the light intensity is between 4000-5000l ux, and an extremely high light environment, the light intensity is greater than 5000l ux.

[0064] The standard environmental light intensity is obtained by experimental simulation to obtain the light intensity with the least influence on the identification of skin cancer, or by the average light intensity in an actual medical environment.

[0065] The specific light intensity level of light intensity can be predicted based on environmental time factor data, environmental climate factor data and environmental light factor data through regression analysis. The linear relationship between environmental time factor data, environmental climate factor data and environmental light factor data and the specific light intensity level can be analyzed from historical data to predict the actual light intensity level.

[0066] In actual situations, the light intensity in an image may change significantly. By simulating the distribution of different specific light intensity levels in the environment through ray tracing technology, we can obtain the data on the area affected by the actual specific light intensity level distribution for identifying skin cancer images. For example, when the light intensity is between 3000-4000l ux, it occupies 60% of the area in the image, and when the light intensity is between 4000-5000l ux, it occupies 40% of the area in the image.

[0067] Through the clustering algorithm, the data of the affected area of ​​the skin cancer image by different actual specific light intensity level distributions are combined with the environmental standard light intensity to obtain the ambient light intensity correction result. For example, the environmental standard light intensity is set to 3000, the specific light intensity level distribution is between 3000-4000l ux, occupying 30% of the area in the image, and the light intensity is between 4000-5000l ux, occupying 70% of the area in the image. Through the clustering algorithm, the difference between the specific light intensity level distribution and the environmental standard light intensity is analyzed with the actual affected image area to obtain the ambient light comprehensive correction value.

[0068] In addition to the above analysis obtained through the clustering algorithm, the comprehensive correction value of ambient light can also be obtained through the calculation formula. The specific calculation formula of the comprehensive correction value of ambient light is:

[0069]

[0070] represents the comprehensive correction value of ambient light, e represents the natural constant, G Z Indicates the ambient light intensity value, G Y Represents the historical average of ambient light intensity, Z D Indicates the uniformity of ambient light illumination, Z Y Indicates the average historical illumination uniformity of the ambient light.

[0071] Ambient light intensity values ​​are measured using a photometer or handheld light meter. These devices can accurately measure light intensity and usually record the data for later analysis.

[0072] The historical average value of ambient light intensity is obtained by analyzing and calculating the monitoring data after long-term monitoring of the light intensity in the environment.

[0073] The stronger the ambient light intensity value, the higher the image contrast, the easier it is to recognize the image; the higher the ambient light illumination uniformity, the easier it is to recognize the image; therefore, the higher the ambient light comprehensive correction value, the easier it is to recognize the image.

[0074] Illuminance uniformity refers to the ratio of the minimum illuminance to the average illuminance on a specified surface. The more uniform the light distribution, the better the illuminance. The value is not less than 0 and not greater than 1. The closer the illuminance uniformity is to 1, the better; conversely, the closer it is to 0, the worse. Illuminance uniformity can be measured by an illuminance meter and a photometer. An illuminance meter usually has a sensor that receives light from a light source and converts it into a readable illuminance value. A photometer uses a built-in optical module to measure illuminance.

[0075] A comprehensive analysis of the data on the comprehensive influencing factors of ambient light on skin cancer identification under different lighting conditions makes the evaluation of the comprehensive correction value of ambient light more accurate. The function graph of the comprehensive correction value of ambient light is monotonically increasing, which makes the increasing rate of the function graph slower and slower, making it easier to directly observe the changes in the warning degree of the comprehensive correction of ambient light.

[0076] like Figure 3 As shown, it is a structural schematic diagram of a tumor recognition and evaluation system based on multidimensional analysis provided in an embodiment of the present application. The tumor recognition and evaluation system based on multidimensional analysis provided in an embodiment of the present application includes: the tumor recognition and evaluation system based on multidimensional analysis includes: a human skin functional area cancer risk analysis module, a human skin skin color impact analysis module, an ambient light intensity impact analysis module, a skin tumor recognition risk level judgment module and a skin tumor warning program module; wherein the human skin functional area cancer risk analysis module is used to divide the human skin functional area, obtain the human skin functional area parameters, and obtain the skin cancer risk results of the human skin functional area through neural network model analysis; the human skin skin color impact analysis module The module is used to classify human skin skin color levels, obtain human skin skin color parameters, and obtain human skin skin color level correction results by analyzing human skin skin color parameters; the ambient light intensity impact analysis module is used to classify ambient light intensity levels, obtain ambient light intensity parameters, and obtain ambient light intensity correction results by analyzing ambient light intensity parameters; the skin tumor identification hidden danger level judgment module is used to obtain the skin tumor identification hidden danger level by comprehensive analysis and judgment based on the skin cancer hidden danger results in the functional areas of human skin, the human skin skin color level correction results and the ambient light intensity correction results; the skin tumor warning plan module is used to determine the corresponding skin tumor warning plan based on the skin tumor identification hidden danger level.

[0077] Among them, an embodiment of the present application provides an electronic device, including: a memory and a processor, the memory is used to store executable instructions of the processor; when the processor is configured to execute the instructions, the electronic device implements a tumor identification and evaluation method based on multidimensional analysis.

[0078] The technical scheme in the above-mentioned embodiment of the present application has at least the following technical effects or advantages: relative to the skin cancer image recognition method, system, device and readable storage medium disclosed in the patent application with publication number: CN114881137A, the embodiment of the present application divides the functional areas of human skin to obtain the parameters of the functional areas of human skin, and obtains the skin cancer risk results of the functional areas of human skin through neural network model analysis, thereby evaluating different cancer risk results according to the same skin area, thereby achieving the improvement of the scientific nature of the evaluation of cancer risk results; relative to the skin cancer image recognition visualization system based on t-SNE announced in the invention patent with publication number: CN107403194B, the embodiment of the present application obtains the skin tumor recognition risk level through comprehensive analysis and judgment based on the skin cancer risk results of the functional areas of human skin, the human skin skin color level correction results and the ambient light intensity correction results, thereby more accurately identifying the potential risks of skin tumors through comprehensive analysis of multiple factors, thereby achieving the reduction of the error rate of tumor recognition evaluation.

[0079] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0080] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A system that specifies the functions of a box or multiple boxes.

[0081] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction system, which is implemented in the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0082] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0083] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0084] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A tumor recognition and evaluation method based on multidimensional analysis, characterized in that: The following steps are involved: Divide the functional areas of human skin to obtain the parameters of the functional areas of human skin, and obtain the results of skin cancer risk in the functional areas of human skin through neural network model analysis; Classifying human skin color levels to obtain human skin color parameters, and obtaining human skin color level correction results by analyzing the human skin color parameters; The ambient light intensity levels are divided to obtain ambient light intensity parameters, and the ambient light intensity correction result is obtained by analyzing the ambient light intensity parameters; The skin tumor identification risk level is determined by comprehensive analysis based on the skin cancer risk results of the functional area of ​​human skin, the skin color level correction results of human skin, and the ambient light intensity correction results. Determine the corresponding skin tumor warning plan according to the level of skin tumor identification risk; The specific process of obtaining the human skin color level correction result by analyzing the human skin color parameters is as follows: Extracting a number of human skin color data from the shared skin cancer image data on the hospital platform, performing data normalization processing on the number of human skin color data to obtain a number of human skin color normalized data, and taking the average value of the number of human skin color normalized data as the human skin standard skin color value; The principal component analysis algorithm is used to analyze the difference between the human skin color parameters and the human skin standard skin color value and the actual human skin cancer risk identification impact, and the human skin color level correction result is obtained; The specific process of obtaining the ambient light intensity correction result by analyzing the ambient light intensity parameters is as follows: Divide the ambient light intensity data into different categories, each category refers to a light intensity level, and set the ambient standard light intensity; The specific light intensity level is predicted based on the environmental time factor data, environmental climate factor data and environmental light factor data through regression analysis; The ray tracing technology is used to simulate the distribution of different specific light intensity levels in the environment, and the data of the area affected by the actual specific light intensity level distribution on the recognition of skin cancer images is obtained; The clustering algorithm is used to analyze the area data of skin cancer image impacted by different actual specific light intensity levels and the ambient standard light intensity to obtain the ambient light intensity correction result; the specific process of comprehensively analyzing and judging the skin cancer potential risk results of the functional area of ​​human skin, the human skin color level correction results and the ambient light intensity correction results to obtain the skin tumor recognition potential risk level is as follows: The skin cancer risk result of the functional area of ​​human skin is the skin cancer risk assessment value of the functional area of ​​human skin, and the skin cancer risk assessment value of the functional area of ​​human skin is used to describe the level of skin tumor risk identified in the functional area of ​​human skin; The human skin color level correction result is the human skin color level correction value, and the human skin color level correction value is used to refer to the correction level of human skin color difference affecting the identification of skin tumor hidden dangers; The result of the ambient light intensity correction is the ambient light comprehensive correction value, which is used to indicate the correction level of the ambient light intensity difference for the impact on the identification of skin tumor hidden dangers; The comprehensive evaluation value of the degree of hidden danger of skin tumor identification is calculated by comprehensive analysis based on the evaluation value of skin cancer risk in the functional area of ​​human skin, the correction value of human skin color level and the comprehensive correction value of environmental light; Compare and analyze the comprehensive evaluation value of the degree of hidden danger of skin tumor identification with the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification; if the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is not higher than the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification, then set the skin tumor identification hidden danger level as the first-level skin tumor identification hidden danger level; if the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is higher than the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification, then set the skin tumor identification hidden danger level as the second-level skin tumor identification hidden danger level; The comprehensive evaluation value of the degree of hidden danger of skin tumor identification is used to describe the level of the comprehensive hidden danger of skin tumor identified; The preset skin tumor identification hidden danger level comprehensive assessment threshold is used to describe the measurement level of the identified skin tumor comprehensive hidden danger level level corresponding to the skin tumor identification hidden danger level judgment.

2. The tumor identification and evaluation method based on multidimensional analysis as claimed in claim 1, characterized in that: The specific process of determining the corresponding skin tumor warning plan according to the skin tumor identification risk level is as follows: If the skin tumor identification risk level is the first level, no skin tumor identification abnormal risk warning will be issued; If the skin tumor identification hidden danger level is the second-level skin tumor identification hidden danger level, a first-level warning of skin tumor identification abnormal hidden danger will be issued, and relevant medical personnel will be notified to immediately re-examine the skin condition of the corresponding patient's human skin functional area.

3. The tumor identification and evaluation method based on multidimensional analysis according to claim 1, characterized in that: The skin tumor alert program also includes: If the skin tumor identification risk level is the first level, the corresponding personnel will be reminded through the intelligent display device to display the comprehensive assessment value of the skin tumor identification risk level and popularize skin health knowledge; If the skin tumor identification risk level is the second-level skin tumor identification risk level, a first-level warning of abnormal skin tumor identification risk will be issued, and a detailed pathological examination will be recommended to the corresponding personnel. The comprehensive assessment value of the skin tumor identification risk level, human skin functional area parameters, human skin skin color parameters, and ambient light intensity parameters will be displayed to relevant medical personnel.

4. The tumor identification and evaluation method based on multidimensional analysis according to claim 1, characterized in that: The specific calculation formula for the comprehensive evaluation value of the skin tumor identification hidden danger level is: θ represents the comprehensive evaluation value of the degree of hidden danger of skin tumor identification, and the monitoring points of the functional area of ​​human skin are set, and the monitoring points of the functional area of ​​human skin are numbered in sequence, A0=1, 2, ..., A, A0 represents the number of the monitoring point of the functional area of ​​human skin, and A represents the total number of the monitoring points of the functional area of ​​human skin; It represents the evaluation value of skin cancer risk in the functional area of ​​human skin at the A0th human skin functional area monitoring point. Indicates the correction value of human skin color level. Indicates the comprehensive correction value of ambient light.

5. The tumor recognition and evaluation method based on multidimensional analysis as claimed in claim 1, characterized in that: The specific process of obtaining the hidden danger results of skin cancer in functional areas of human skin through neural network model analysis is as follows: Extracting human skin image data from a human skin functional area skin cancer database, wherein the human skin image data includes normal skin image data of each human skin functional area and image data of skin cancer; Annotating the human skin image data to obtain human skin image annotation data, wherein the data annotation specifically includes human skin functional area, skin cancer color, skin cancer texture and skin cancer shape; The neural network model is constructed and trained through human skin image annotation data, the neural network model parameters are adjusted through the back propagation algorithm, and the results of skin cancer risk in functional areas of human skin are obtained through neural network model analysis.

6. A tumor recognition and evaluation system based on multidimensional analysis, characterized in that: The tumor recognition and evaluation system based on multi-dimensional analysis includes: a human skin functional area cancer risk analysis module, a human skin color influence analysis module, an ambient light intensity influence analysis module, a skin tumor recognition risk level judgment module and a skin tumor warning program module; The human skin functional area cancer risk analysis module is used to divide the human skin functional area, obtain the human skin functional area parameters, and obtain the skin cancer risk results of the human skin functional area through neural network model analysis; The human skin color influence analysis module is used to classify the human skin color level to obtain the human skin color parameters, and obtain the human skin color level correction result by analyzing the human skin color parameters; The ambient light intensity impact analysis module is used to classify the ambient light intensity levels to obtain ambient light intensity parameters, and obtain the ambient light intensity correction result by analyzing the ambient light intensity parameters; The skin tumor identification hidden danger level judgment module is used to comprehensively analyze and judge the skin tumor identification hidden danger level according to the skin cancer hidden danger result of the human skin functional area, the human skin skin color level correction result and the ambient light intensity correction result; The skin tumor warning scheme module is used to determine the corresponding skin tumor warning scheme according to the skin tumor identification risk level; The specific process of obtaining the human skin color level correction result by analyzing the human skin color parameters is as follows: Extracting a number of human skin color data from the shared skin cancer image data on the hospital platform, performing data normalization processing on the number of human skin color data to obtain a number of human skin color normalized data, and taking the average value of the number of human skin color normalized data as the human skin standard skin color value; The principal component analysis algorithm is used to analyze the difference between the human skin color parameters and the human skin standard skin color value and the actual human skin cancer risk identification impact, and the human skin color level correction result is obtained; The specific process of obtaining the ambient light intensity correction result by analyzing the ambient light intensity parameters is as follows: Divide the ambient light intensity data into different categories, each category refers to a light intensity level, and set the ambient standard light intensity; The specific light intensity level is predicted based on the environmental time factor data, environmental climate factor data and environmental light factor data through regression analysis; The ray tracing technology is used to simulate the distribution of different specific light intensity levels in the environment, and the data of the area affected by the actual specific light intensity level distribution on the recognition of skin cancer images is obtained; The clustering algorithm is used to analyze the area data of skin cancer image impacted by different actual specific light intensity levels and the ambient standard light intensity to obtain the ambient light intensity correction result; the specific process of comprehensively analyzing and judging the skin cancer potential risk results of the functional area of ​​human skin, the human skin color level correction results and the ambient light intensity correction results to obtain the skin tumor recognition potential risk level is as follows: The skin cancer risk result of the functional area of ​​human skin is the skin cancer risk assessment value of the functional area of ​​human skin, and the skin cancer risk assessment value of the functional area of ​​human skin is used to describe the level of skin tumor risk identified in the functional area of ​​human skin; The human skin color level correction result is the human skin color level correction value, and the human skin color level correction value is used to refer to the correction level of human skin color difference affecting the identification of skin tumor hidden dangers; The result of the ambient light intensity correction is the ambient light comprehensive correction value, which is used to indicate the correction level of the ambient light intensity difference for the impact on the identification of skin tumor hidden dangers; The comprehensive evaluation value of the degree of hidden danger of skin tumor identification is calculated by comprehensive analysis based on the evaluation value of skin cancer risk in the functional area of ​​human skin, the correction value of human skin color level and the comprehensive correction value of environmental light; Compare and analyze the comprehensive evaluation value of the degree of hidden danger of skin tumor identification with the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification; if the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is not higher than the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification, then set the skin tumor identification hidden danger level as the first-level skin tumor identification hidden danger level; if the comprehensive evaluation value of the degree of hidden danger of skin tumor identification is higher than the preset comprehensive evaluation threshold of the degree of hidden danger of skin tumor identification, then set the skin tumor identification hidden danger level as the second-level skin tumor identification hidden danger level; The comprehensive evaluation value of the degree of hidden danger of skin tumor identification is used to describe the level of the comprehensive hidden danger of skin tumor identified; The preset skin tumor identification hidden danger level comprehensive assessment threshold is used to describe the measurement level of the identified skin tumor comprehensive hidden danger level level corresponding to the skin tumor identification hidden danger level judgment.

7. An electronic device, characterized in that: include: A memory and a processor, wherein the memory is used to store executable instructions of the processor; When the processor is configured to execute the instructions, the electronic device implements the tumor recognition and evaluation method based on multidimensional analysis as described in any one of claims 1 to 5.

Images