Rapid detection method and system for dermatophyte

Abstract

The invention relates to the technical field of physical analysis, in particular to a rapid detection method and system for dermatophyte. Obtaining longitudinal offset estimation values and transverse offset estimation values of all target molecule types of the upper envelope line and the lower envelope line of the infrared absorption spectrum, and judging a relative content maximum estimation value of the target molecule types after infrared absorption peak section calibration; the method comprises the following steps: acquiring an adaptive decomposition scale of each similar fluctuation region, performing signal decomposition on the similar fluctuation regions, acquiring a contribution ratio mean value of component signals with similar signal expressions, acquiring the actual relative content of a molecular type in combination with a maximum estimated value of the relative content, and judging whether fungal infection exists in a sample or not. According to the invention, signal overlapping and noise pollution caused by the fact that a skin sample is often suffered from interferon of bacteria and skin components are avoided, and the detection accuracy and the sensitivity to low-component-content fungi or early infection are improved.

Description

Technical Field

[0001] The present invention relates to the field of physical analysis technology, and in particular to a method and system for rapid detection of skin fungi. Background Art

[0002] Fungal skin infections are common skin diseases caused by a variety of fungi (such as dermatophytes, Candida species, and molds). Common types include tinea pedis, tinea corporis, and candidal infections. With the increasing number of cases, rapid and accurate diagnosis is crucial for early treatment, preventing complications, and controlling spread. Prompt identification and treatment of fungal infections is particularly important in high-risk groups, such as immunosuppressed patients and those with diabetes.

[0003] Existing rapid skin fungus detection technology is based on infrared absorption spectroscopy, which identifies fungal infections by extracting characteristic infrared absorption spectra from skin samples. However, skin samples are often interfered with by bacteria and skin components (such as the stratum corneum, sebum, sweat, etc.), resulting in signal overlap and noise pollution, which affects the accuracy of detection and reduces the sensitivity to low-component fungi or early-stage infections. Therefore, improving the sensitivity and anti-interference ability of rapid detection results is the main challenge facing this technology. Summary of the Invention

[0004] The present invention provides a method and system for rapid detection of skin fungi to solve the existing problems.

[0005] The present invention provides a rapid detection method and system for skin fungi using the following technical solutions:

[0006] One embodiment of the present invention provides a method for rapid detection of dermatophytes, the method comprising the following steps:

[0007] Obtaining infrared absorption spectra of skin samples;

[0008] According to the absorbance in the infrared absorption spectrum, the maximum estimated value of the relative content of the infrared absorption peak segment of the target molecule type is determined;

[0009] Obtaining the upper and lower envelopes of the infrared absorption spectrum, and determining a longitudinal shift estimate and a lateral shift estimate of the infrared absorption peak segment of the target molecule type based on a difference between the upper and lower envelopes; obtaining a maximum relative content estimate of the infrared absorption peak segment of the target molecule type after calibration based on the longitudinal shift estimate and the lateral shift estimate;

[0010] Obtaining the local disorder degree of each data point based on the number of extreme values ​​within the infrared absorption peak segment of the target molecule type and the slope of the data point; dividing the infrared absorption peak segment of the target molecule type into a plurality of similar fluctuation regions based on the local disorder degree; decomposing each similar fluctuation region into a plurality of components to obtain the contribution ratio of the component signals; determining the actual relative content of the target molecule type based on the contribution ratio of the component signals and the calibrated maximum relative content estimate;

[0011] By inputting the actual relative content of the target molecule type, a neural network algorithm is used to determine whether there is fungal infection in the skin sample.

[0012] Furthermore, the determination of the maximum estimated value of the relative content of the infrared absorption peak segment of the target molecule type includes the following specific steps:

[0013] Obtaining a wavenumber interval of any molecular type in the infrared absorption spectrum database of skin fungi, and obtaining an infrared absorption peak segment within the wavenumber interval in the infrared absorption spectrum as the infrared absorption peak segment of the i-th target molecular type;

[0014] Within the infrared absorption peak section of the i-th target molecule type, the range of the absorbance of all data points is taken as the maximum estimated value of the relative content of the infrared absorption peak section of the i-th target molecule type.

[0015] Furthermore, the step of determining the longitudinal offset estimate and the lateral offset estimate of the infrared absorption peak segment of the target molecule type includes the following specific steps:

[0016] Within the wavenumber interval of the infrared absorption peak of the i-th target molecule type, calculate the ratio of the slope of the m-th data point on the upper envelope to the slope of the m-th data point on the lower envelope, then calculate the cumulative multiplication of the ratios of the slopes of all data points with the same sequence number on the upper envelope and the lower envelope, and record the absolute value of the difference between the cumulative multiplication result and 1 as the longitudinal offset estimate V of the infrared absorption peak of the i-th target molecule type. i ;

[0017] According to the phase information of the upper and lower envelope frequency domain data in the wavenumber interval where the infrared absorption peak segment of the i-th target molecule type is located in the infrared absorption spectrum, the lateral offset estimation value of the infrared absorption peak segment of the i-th target molecule type is determined.

[0018] Furthermore, the lateral shift estimation value of the infrared absorption peak segment of the target molecule type includes the following specific steps:

[0019] Obtain the phase information of the upper and lower envelope frequency domain data within the wave number interval of the infrared absorption peak segment of the i-th target molecule type in the infrared absorption spectrum;

[0020] The absolute value of the difference between the upper and lower envelope frequency domain data phase information is recorded as the lateral offset estimate E of the infrared absorption peak segment of the i-th target molecule type i .

[0021] Furthermore, the step of obtaining the maximum estimated value of the relative content of the infrared absorption peak segment of the target molecule type after calibration includes the following specific steps:

[0022] The maximum estimated value G of the relative content of the infrared absorption peak segment of the i-th target molecule type and the estimated value V of the longitudinal offset of the infrared absorption peak segment of the i-th target molecule type are combined. i and the estimated lateral shift E of the infrared absorption peak segment of the i-th target molecule type i The ratio of the sum of the values ​​is recorded as the maximum estimated value of the relative content after calibration of the infrared absorption peak of the i-th target molecule type R i .

[0023] Furthermore, the specific steps of obtaining the local disorder degree of each data point include the following:

[0024] Set the threshold c, take the jth data point in the infrared absorption peak segment of the i-th target molecule type as the starting point, and construct a window with a step length of j;

[0025] The slope variance S(j) of all data points in the jth data point window in the infrared absorption peak segment of the i-th target molecule type is calculated by multiplying the number of extreme points N in the jth data point window in the infrared absorption peak segment of the i-th target molecule type by the slope variance S(j) of all data points in the jth data point window in the infrared absorption peak segment of the i-th target molecule type. i,j The product of is recorded as the local disorder degree j of the jth data point in the infrared absorption peak segment of the jth target molecule type. ,j .

[0026] Furthermore, the infrared absorption peak segment of the target molecule type is divided into a number of similar fluctuation regions, including the following specific steps:

[0027] Calculate the normalized value of the absolute value of the difference between the local disorder degree of the jth and j+1th data points in the infrared absorption peak segment of the jth target molecule type as the dissimilarity Q of the jth and j+1th data points j,j+1 ;

[0028] Set the threshold T, in the infrared absorption peak section of the i-th target molecule type, when Q j,j+1 When ≤T, continue to judge Q j+1,j+2 Is it less than or equal to T, until Q j+n-1,h+n>T, divide the area between the hth to h+n-1th data points into the first similar fluctuation area, and then start from the j+nth data point, according to the method of obtaining the first similar fluctuation area, obtain the second similar fluctuation area, and so on, obtain several similar fluctuation areas, where Q j+1,j+2 , Q j+n-1,j+n They respectively represent the dissimilarity between the j+1th and j+2th data points, and the j+n-1th and j+nth data points in the infrared absorption peak segment of the i-th target molecule type.

[0029] Furthermore, the steps of decomposing each similar fluctuation region into several components and obtaining the contribution ratio of the component signals include the following:

[0030] Obtain the traditional signal decomposition scale of the infrared absorption peak segment of the i-th target molecule type;

[0031] The product of the mean of the local disorder degree of all data points in the r-th similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type and the traditional signal decomposition scale of the infrared absorption peak segment of the i-th target molecule type is recorded as the adaptive decomposition scale of the r-th similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type;

[0032] According to the adaptive decomposition scale of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type, several component signals of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type are obtained;

[0033] Obtain the contribution ratio of each component signal to the original signal of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type.

[0034] Furthermore, the determination of the actual relative content of the target molecule type includes the following specific steps:

[0035] In the infrared absorption peak segment of the i-th target molecule type, obtaining the maximum value of the contribution ratios of all component signals of the r-th similar fluctuation region in the original signal as the maximum contribution ratio of the r-th similar fluctuation region;

[0036] The product of the normalized value of the mean of the maximum contribution ratios of all similar fluctuation regions and the maximum estimated value of the relative content after calibration of the infrared absorption peak segment of the i-th target molecule type is recorded as the actual relative content R' of the i-th target molecule type. i .

[0037] The present invention also proposes a rapid detection system for skin fungi, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned rapid detection method for skin fungi.

[0038] The technical solution of the present invention has the following beneficial effects: obtaining a longitudinal offset estimate of the infrared absorption peak segment of the target molecule type based on the upper and lower envelopes of the infrared absorption spectrum; obtaining a lateral offset estimate of the infrared absorption peak segment of all target molecule types based on the phase information of the upper and lower envelope frequency domain data; obtaining a calibrated maximum relative content estimate of the infrared absorption peak segment of the target molecule type based on the longitudinal offset estimate and the lateral offset estimate; avoiding the influence of skin tissue when acquiring infrared absorption spectrum data by FTIR, correcting the maximum relative content of each absorption peak, and improving the accuracy of rapid fungal detection; constructing a window for the infrared absorption peak segment, obtaining the local disorder degree of the data points in the window, and thereby obtaining a similar fluctuation region; obtaining a component disorder scale of the similar fluctuation region, thereby obtaining an adaptive decomposition scale of each similar fluctuation region, and performing signal decomposition on the similar fluctuation region to obtain the contribution ratio of the component signal, thereby obtaining a maximum relative content estimate and the actual relative content of the molecule type; and determining whether a sample is infected with a fungus based on a neural network, avoiding signal overlap and noise contamination caused by bacteria and interferon of skin components that are often present in skin samples, thereby improving detection accuracy and sensitivity to low-component fungi or early-stage infections. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 This is a flowchart of the steps of a rapid detection method for skin fungi of the present invention;

[0041] Figure 2 It is a module for the rapid detection system of skin fungi;

[0042] Figure 3 This is the RLMD decomposition effect diagram. DETAILED DESCRIPTION

[0043] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail a method and system for rapid detection of dermatophytes according to the present invention, including its specific implementation, structure, features, and effectiveness. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0044] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0045] The specific scheme of the rapid detection method and system for skin fungi provided by the present invention is described in detail below with reference to the accompanying drawings.

[0046] See also Figure 1 , which shows a flowchart of a method for rapid detection of dermatophytes provided by one embodiment of the present invention, the method comprising the following steps:

[0047] Step S001: obtaining an infrared absorption spectrum of a skin sample;

[0048] The rapid detection system for skin fungi described in this embodiment is as follows Figure 2 The following modules are shown: sample collection module, detection module, data processing and analysis module, result analysis and diagnosis module; the data processing and analysis module includes different molecular content restriction units and high-content molecular analysis units, and the result analysis and diagnosis module is used to display the analyzed data in the form of a report.

[0049] In the sample collection module, fungal test samples are collected from the patient's skin using sampling tools such as sterile swabs and scrapers; the collected fungal test samples are pre-processed, such as removing surface impurities, extracting fungal cells or DNA, and filtering impurities;

[0050] In the detection module, the processed sample is placed on the infrared absorption spectrum collection area of ​​the FTIR system to ensure that the infrared light can be evenly irradiated on the skin sample;

[0051] Using an interferometer (preferably a Michelson interferometer) to collect infrared absorption interference signals of skin sample molecules; performing Fourier transform on the infrared absorption interference signals to obtain an infrared absorption spectrum of the skin sample;

[0052] The measured infrared absorption spectrum data is transmitted to the subsequent data processing module in real time through the data interface and communication protocol;

[0053] It should be noted that the horizontal axis of the spectrum is wave number (describing infrared radiation in different frequency bands), and the vertical axis is absorbance (the degree to which infrared radiation of that wavelength is absorbed by the sample); the FTIR (Fourier transform infrared absorption spectroscopy) system is an analytical instrument used in the fields of chemistry, biology, materials science, environmental science and technology, and resource science and technology. It analyzes the chemical composition and structure of a substance by detecting its absorption of infrared light.

[0054] Step S002: determining the maximum estimated value of the relative content of the infrared absorption peak segment of the target molecule type according to the absorbance in the infrared absorption spectrum;

[0055] According to the analysis, different types of biological molecules in skin tissue have different absorption characteristics of infrared light, resulting in characteristic absorption peaks in different bands of the infrared absorption spectrum curve of the skin sample, corresponding to different types and contents of structural molecules;

[0056] Obtaining the wavenumber interval of any molecular type from the infrared absorption spectrum database of skin fungi, and obtaining the infrared absorption peak segment within the wavenumber interval in the infrared absorption spectrum as the infrared absorption peak segment of the i-th target molecular type;

[0057] For example, for skin collagen, the infrared absorption peak segment within the wavenumber interval corresponding to the type of molecule in the infrared absorption spectrum is obtained as the infrared absorption peak segment of the molecule type;

[0058] In the infrared absorption peak section of the i-th target molecule type, the range of the absorbance of all data points is taken as the maximum estimated value of the relative content of the infrared absorption peak section of the i-th target molecule type, which is recorded as G.

[0059] It should be noted that the infrared absorption spectrum database of skin fungi contains the wavenumber intervals corresponding to the molecular types of various skin fungi in the infrared absorption spectrum, which is a well-known technology and will not be described in detail.

[0060] Step S003: Obtaining the upper and lower envelopes of the infrared absorption spectrum, and determining an estimated longitudinal offset and an estimated transverse offset of the infrared absorption peak segment of the target molecule type based on the difference between the upper and lower envelopes; obtaining an estimated maximum relative content of the infrared absorption peak segment of the target molecule type after calibration based on the estimated longitudinal offset and the estimated transverse offset;

[0061] According to the analysis, the skin has multi-level structural heterogeneity, including the epidermis, dermis and subcutaneous tissue. The differences in chemical composition and tissue structure of each layer lead to different infrared absorption spectral responses, resulting in abnormal changes in the infrared absorption spectrum curve such as displacement and deformation of local absorption peaks. The infrared absorption peaks corresponding to molecules with different structures may undergo longitudinal (intensity changes) or transverse (wavenumber shift, peak shape expansion) shifts. The longitudinal and transverse shifts of the absorption peaks will introduce quantitative analysis errors, usually resulting in an inflated maximum estimate of the relative content of the molecules, thereby affecting the quantitative analysis of the relative content of each structural molecule.

[0062] The extreme value method is used to obtain all the maximum and minimum points in the infrared absorption spectrum, and the maximum points (upper envelope) and minimum points (lower envelope) are connected by cubic spline interpolation to obtain the upper and lower envelopes of the infrared absorption spectrum; the slope of the mth data point on the upper envelope within the wavenumber interval of the infrared absorption peak of the i-th target molecule type is obtained. And the slope of the mth data point on the lower envelope

[0063] It should be noted that: H1 and H2 represent the upper envelope and the lower envelope respectively; the extreme value method and the cubic spline difference method are well-known technologies, among which the extreme value method is used to extract the maximum and minimum values ​​in the curve, and the cubic spline difference method is used to extract the upper and lower envelopes in the curve, which will not be repeated here.

[0064] Thus, the longitudinal shift estimate V of the infrared absorption peak segment of the i-th target molecule type is obtained i :

[0065]

[0066] Among them, M i Indicates the number of data points on the infrared absorption peak segment of the i-th target molecule type.

[0067] What needs to be explained is: when the denominator is 0, let its value be 1 to prevent it from being meaningless; The greater the difference between the value and 1, the more data points there are in the infrared absorption peak segment of the target molecule type. At the same time, the slopes of the corresponding points of the upper and lower envelopes are significantly different, indicating that there is a significant longitudinal shift in the infrared absorption peak segment of the target molecule type. The estimated longitudinal shift value V of the infrared absorption peak segment of the i-th target molecule type is i It describes that in skin samples, the differences in the chemical composition of each layer of tissue lead to different relative molecular contents; when the infrared radiation emitted by the FTIR system passes through each layer of skin tissue, the propagation path of the infrared light changes due to the differences in the refractive index and molecular composition of each layer of tissue, resulting in a longitudinal shift of the infrared absorption peak corresponding to the target molecule.

[0068] The fast Fourier transform algorithm is used to convert the upper and lower envelope data within the wavenumber interval of the infrared absorption peak of the i-th target molecule type in the infrared absorption spectrum into frequency domain data; the phase information of the upper and lower envelope frequency domain data is obtained respectively; the absolute value of the difference between the phase information of the upper and lower envelope frequency domain data is calculated, thereby obtaining the lateral offset estimate E of the infrared absorption peak of the i-th target molecule type. i .

[0069] It should be noted that the fast Fourier transform algorithm is a well-known technology used to convert time domain data into frequency domain data, which will not be described in detail. When infrared light passes through the skin, surface impurities change the scattering characteristics of the light, causing the infrared light to be absorbed and scattered to varying degrees when returning. These effects cause lateral shifts in the infrared absorption peak, reflecting the interference of surface impurities on the propagation of infrared light.

[0070] Thus, the maximum estimated value R of the relative content after calibration of the infrared absorption peak segment of the i-th target molecule type is obtained. i :

[0071]

[0072] Where G represents the maximum estimated value of the relative content of the infrared absorption peak segment of the current i-th target molecule type;

[0073] What needs to be explained is: R i It represents the sum of the relative contents of multiple structural molecules within a single overlapping absorption peak segment. Subsequently, it is only necessary to analyze the proportion of each molecular component within the mixed absorption peak band to obtain the accurate structural molecule type and corresponding relative content in the skin sample tissue;

[0074] Step S004: obtaining the local disorder degree of each data point based on the number of extreme values ​​within the infrared absorption peak segment of the target molecule type and the slope of the data point; dividing the infrared absorption peak segment of the target molecule type into a plurality of similar fluctuation regions based on the local disorder degree; decomposing each similar fluctuation region into a plurality of components, obtaining the contribution ratio of the component signals; determining the actual relative content of the target molecule type based on the contribution ratio of the component signals and the maximum estimated relative content value;

[0075] Analysis shows that the interweaving and overlapping of the infrared absorption bands of multiple structural molecules in skin samples not only shifts the absorption peaks to varying degrees, but also makes it impossible to accurately distinguish the specific relative content of each structural molecule within a single absorption peak, thus affecting subsequent fungal detection in skin samples. In the infrared absorption spectrum data of skin samples, different structural molecules have different specific absorption bands, and their influence on the target absorption peak varies.

[0076] Set a threshold c, and construct a window with a step size of c, starting from the jth data point in the infrared absorption peak segment of the i-th target molecule type;

[0077] The slope variance S(j) of all data points in the jth data point window in the infrared absorption peak segment of the i-th target molecule type is calculated by multiplying the number of extreme points N in the jth data point window in the infrared absorption peak segment of the i-th target molecule type by the slope variance S(j) of all data points in the jth data point window in the infrared absorption peak segment of the i-th target molecule type. i,j The product of is recorded as the local disorder degree D of the jth data point in the infrared absorption peak segment of the i-th target molecule type. i,j ;

[0078] What needs to be explained is: set the threshold c = 9; if there are insufficient data points in the window, the remaining data points are considered to be in the same area; D i,j It describes that in the infrared absorption peak section of structural molecules, the absorption peaks of multiple different molecules will overlap to form a composite absorption peak, resulting in a significant peak in the infrared absorption peak section. This peak superposition causes the infrared absorption spectrum data containing rich structural information to show significant and drastic changes in its fluctuation characteristics. i,j The larger the value, the more dramatic the fluctuation of the infrared absorption spectrum curve within the jth data point window in the infrared absorption peak segment of the i-th target molecule type, which usually means that it contains richer structural molecular information.

[0079] Calculate the absolute value of the difference between the local disorder degree of the jth and j+1th data points in the infrared absorption peak segment of the i-th target molecule type, and normalize it using the norm function, which is recorded as the dissimilarity Q between the jth and j+1th data points j,j+1 ;

[0080] Set the threshold T, in the infrared absorption peak section of the i-th target molecule type, when Q j,j+1 When ≤T, continue to judge Q j+1,j+2 Is it less than or equal to T, until Q j+n-1,j+n >T, divide the area between the jth to j+n-1th data points into the first similar fluctuation area, and then, starting from the j+nth data point, obtain the second similar fluctuation area according to the method of obtaining the first similar fluctuation area, and so on, obtain several similar fluctuation areas;

[0081] What needs to be explained is: Q j+1,j+2 , Q j+n-1,j+n They represent the dissimilarity between the j+1th and j+2th data points, and the j+n-1th and j+nth data points in the infrared absorption peak segment of the i-th target molecule type;

[0082] For the rth similar fluctuation region among the above similar fluctuation regions, calculate the local disorder mean μ(Dr ), denoted as the component disorder scale of the current r-th similar fluctuation region;

[0083] Based on the RLMD algorithm, the traditional signal decomposition scale Y of the infrared absorption peak segment of the i-th target molecule type is obtained. o ;

[0084] It should be noted that the RLMD algorithm is a robust local mode decomposition method (Robust Local Mode Decomposition), which is mainly used for signal processing and will not be described in detail.

[0085] Get the adaptive decomposition scale Y of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type i,r :

[0086] Y i,r =Y o ×μ(D r )

[0087] It should be noted that: the threshold T is set to 0.3; the norm function is a linear normalization function that normalizes the data values ​​to the interval [0, 1]; the RLMD (Signal Matrix Decomposition) algorithm is a low-rank matrix decomposition algorithm that aims to recover the true component signals from overlapping mixed signals;

[0088] Based on the RLMD algorithm, the adaptive decomposition scale Y of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type is combined i,r Decompose the signal to obtain several component signals of the rth similar fluctuation area in the infrared absorption peak segment of the i-th target molecule type, such as Figure 3 As shown;

[0089] What needs to be explained is: Figure 3 The figure in the figure shows the signal data before and after signal decomposition of the similar fluctuation region of the infrared absorption peak segment of the target molecule type using the RLMD algorithm, obtaining the original signal and the f1, f2, f3, f4, f5, and f6 component signals. The horizontal axis represents the sampling sequence of the signal, and the vertical axis represents the amplitude of the signal or the amplitude of the modal component after decomposition.

[0090] According to the RLMD algorithm, the contribution ratio of any component signal to the original signal of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type is obtained;

[0091] Obtaining the maximum value of the contribution ratios of all component signals of the r-th similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type in the original signal as the maximum contribution ratio of the r-th similar fluctuation region;

[0092] Get the mean of the maximum contribution ratios of all similar fluctuation regions in the infrared absorption peak segment of the i-th target molecule type, recorded as

[0093] Thus, the actual relative content R' of the i-th target molecule type is obtained i :

[0094]

[0095] Among them, R i Indicates the maximum estimated value of the relative content of the infrared absorption peak segment of the i-th target molecule type after calibration;

[0096] It should be noted that: F( ) represents a sum normalization function, i.e., obtaining the sum normalized value of the mean of the maximum contribution ratio of all component signals relative to the mean of other contribution ratios;

[0097] Step S005: By inputting the actual relative content of the target molecule type, a neural network algorithm is used to determine whether there is fungal infection in the skin sample.

[0098] According to R' i The actual relative content of each molecular type in the infrared absorption spectrum database of skin fungi is obtained by the acquisition method;

[0099] By collecting a large number of skin samples with known fungal infections and non-infections, a multidimensional training dataset was constructed, which included the relative content of various molecular types (such as lipids, carbohydrates, proteins, etc.);

[0100] A neural network (DNN) is used to model the training data based on supervised learning methods.

[0101] The actual relative content of each molecular type in the infrared absorption spectrum data of the sample to be tested is input into the trained model. The model outputs the classification result of whether the sample is infected with fungi based on the actual content of each molecular type obtained from the infrared absorption spectrum.

[0102] Finally, a diagnostic report is generated based on the test results, which includes the type of fungal infection, the degree of infection, and related recommendations.

[0103] It should be noted that deep neural network (DNN) is an advanced machine learning algorithm and will not be explained in detail. During the training process, cross-validation method is used to optimize the performance of the model.

[0104] So far, the present invention is completed.

[0105] In summary, in an embodiment of the present invention, a longitudinal offset estimate of the infrared absorption peak segment of the target molecule type is obtained based on the upper and lower envelopes of the infrared absorption spectrum; phase information of the upper and lower envelope frequency domain data is obtained to obtain the lateral offset estimate of the infrared absorption peak segment of all target molecule types; based on the longitudinal offset estimate and the lateral offset estimate, a calibrated maximum relative content estimate of the infrared absorption peak segment of the target molecule type is obtained; the influence of skin tissue when FTIR acquires infrared absorption spectrum data is avoided, and the maximum relative content of each absorption peak is corrected, thereby improving the accuracy of rapid fungal detection; a window is constructed for the infrared absorption peak segment to obtain the local disorder degree of the data points in the window, and thereby obtain a similar fluctuation area; the component disorder scale of the similar fluctuation area is obtained, thereby obtaining an adaptive decomposition scale for each similar fluctuation area, and the similar fluctuation area is subjected to signal decomposition to obtain the contribution ratio of the component signal, thereby obtaining the maximum relative content estimate and the actual relative content of the molecule type; based on a neural network, whether a sample is infected with fungi is determined, thereby avoiding signal overlap and noise pollution caused by interferon of bacteria and skin components that are often present in skin samples, thereby improving the accuracy of detection and the sensitivity to low-component fungi or early-stage infections.

[0106] The present invention also provides a rapid detection system for skin fungi, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned rapid detection method for skin fungi.

[0107] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A rapid detection method for skin fungi, characterized in that: The method comprises the following steps: Obtaining infrared absorption spectra of skin samples; According to the absorbance in the infrared absorption spectrum, the maximum estimated value of the relative content of the infrared absorption peak segment of the target molecule type is determined; Obtaining the upper and lower envelopes of the infrared absorption spectrum, and determining a longitudinal shift estimate and a lateral shift estimate of the infrared absorption peak segment of the target molecule type based on a difference between the upper and lower envelopes; obtaining a maximum relative content estimate of the infrared absorption peak segment of the target molecule type after calibration based on the longitudinal shift estimate and the lateral shift estimate; Obtaining the local disorder degree of each data point based on the number of extreme values ​​within the infrared absorption peak segment of the target molecule type and the slope of the data point; dividing the infrared absorption peak segment of the target molecule type into a plurality of similar fluctuation regions based on the local disorder degree; decomposing each similar fluctuation region into a plurality of components to obtain the contribution ratio of the component signals; determining the actual relative content of the target molecule type based on the contribution ratio of the component signals and the calibrated maximum relative content estimate; By inputting the actual relative content of the target molecule type, a neural network algorithm is used to determine whether there is fungal infection in the skin sample.

2. A rapid detection method for skin fungi according to claim 1, characterized in that: The specific steps of determining the maximum estimated value of the relative content of the infrared absorption peak segment of the target molecule type are as follows: Obtaining a wavenumber interval of any molecular type in the infrared absorption spectrum database of skin fungi, and obtaining an infrared absorption peak segment within the wavenumber interval in the infrared absorption spectrum as the infrared absorption peak segment of the i-th target molecular type; Within the infrared absorption peak section of the i-th target molecule type, the range of the absorbance of all data points is taken as the maximum estimated value of the relative content of the infrared absorption peak section of the i-th target molecule type.

3. A rapid detection method for skin fungi according to claim 1, characterized in that: The specific steps of determining the longitudinal offset estimation value and the lateral offset estimation value of the infrared absorption peak segment of the target molecule type include the following: Within the wavenumber interval of the infrared absorption peak of the i-th target molecule type, calculate the ratio of the slope of the m-th data point on the upper envelope to the slope of the m-th data point on the lower envelope, then calculate the cumulative multiplication of the ratios of the slopes of all data points with the same sequence number on the upper envelope and the lower envelope, and record the absolute value of the difference between the cumulative multiplication result and 1 as the longitudinal offset estimate V of the infrared absorption peak of the i-th target molecule type. i ; According to the phase information of the upper and lower envelope frequency domain data in the wavenumber interval where the infrared absorption peak segment of the i-th target molecule type is located in the infrared absorption spectrum, the lateral offset estimation value of the infrared absorption peak segment of the i-th target molecule type is determined.

4. A rapid detection method for skin fungi according to claim 3, characterized in that: The lateral shift estimation value of the infrared absorption peak segment of the target molecule type includes the following specific steps: Obtain the phase information of the upper and lower envelope frequency domain data within the wave number interval of the infrared absorption peak segment of the i-th target molecule type in the infrared absorption spectrum; The absolute value of the difference between the upper and lower envelope frequency domain data phase information is recorded as the lateral offset estimate E of the infrared absorption peak segment of the i-th target molecule type i .

5. A rapid detection method for skin fungi according to claim 1, characterized in that: The specific steps of obtaining the maximum estimated value of the relative content of the target molecule type after calibration of the infrared absorption peak segment are as follows: The maximum estimated value G of the relative content of the infrared absorption peak segment of the i-th target molecule type and the estimated value V of the longitudinal offset of the infrared absorption peak segment of the i-th target molecule type are combined. i and the estimated lateral shift E of the infrared absorption peak segment of the i-th target molecule type i The ratio of the sum of the values ​​is recorded as the maximum estimated value of the relative content after calibration of the infrared absorption peak of the i-th target molecule type R i .

6. A rapid detection method for skin fungi according to claim 1, characterized in that: The specific steps of obtaining the local disorder degree of each data point are as follows: Set the threshold c, take the jth data point in the infrared absorption peak segment of the i-th target molecule type as the starting point, and construct a window with a step size of c; The slope variance S(j) of all data points in the jth data point window in the infrared absorption peak segment of the i-th target molecule type is calculated by multiplying the number of extreme points N in the jth data point window in the infrared absorption peak segment of the i-th target molecule type by the slope variance S(j) of all data points in the jth data point window in the infrared absorption peak segment of the i-th target molecule type. i,j The product of Denote the local disorder degree D of the jth data point in the infrared absorption peak segment of the i-th target molecule type i,j .

7. A rapid detection method for skin fungi according to claim 1, characterized in that: The specific steps of dividing the infrared absorption peak segment of the target molecule type into several similar fluctuation regions are as follows: Calculate the normalized value of the absolute value of the difference between the local disorder degree of the jth and j+1th data points in the infrared absorption peak segment of the i-th target molecule type as the dissimilarity Q of the jth and j+1th data points j,j+1 ; Set the threshold T, in the infrared absorption peak section of the i-th target molecule type, when Q j,j+1 When ≤T, continue to judge Q j+1,j+2 Is it less than or equal to T, until Q j+n-1,j+n >T, divide the area between the jth to j+n-1th data points into the first similar fluctuation area, and then start from the j+nth data point, according to the method of obtaining the first similar fluctuation area, obtain the second similar fluctuation area, and so on, obtain several similar fluctuation areas, where Q j+1,j+2 , Q j+n-1,j+n They respectively represent the dissimilarity between the j+1th and j+2th data points, and the j+n-1th and j+nth data points in the infrared absorption peak segment of the i-th target molecule type.

8. A rapid detection method for skin fungi according to claim 1, characterized in that: The specific steps of decomposing each similar fluctuation region into several components and obtaining the contribution ratio of the component signals are as follows: Obtain the traditional signal decomposition scale of the infrared absorption peak segment of the i-th target molecule type; The product of the mean of the local disorder degree of all data points in the r-th similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type and the traditional signal decomposition scale of the infrared absorption peak segment of the i-th target molecule type is recorded as the adaptive decomposition scale of the r-th similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type; According to the adaptive decomposition scale of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type, several component signals of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type are obtained; Obtain the contribution ratio of each component signal to the original signal of the rth similar fluctuation region in the infrared absorption peak segment of the i-th target molecule type.

9. A rapid detection method for skin fungi according to claim 1, characterized in that: The specific steps of determining the actual relative content of the target molecule type include the following: In the infrared absorption peak segment of the i-th target molecule type, obtaining the maximum value of the contribution ratios of all component signals of the r-th similar fluctuation region in the original signal as the maximum contribution ratio of the r-th similar fluctuation region; The product of the normalized value of the mean of the maximum contribution ratios of all similar fluctuation regions and the maximum estimated value of the relative content after calibration of the infrared absorption peak segment of the i-th target molecule type is recorded as the actual relative content R' of the i-th target molecule type. i .

10. A rapid detection system for skin fungi, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the computer program realizes the functions as claimed in claims 1-9. The steps of any one of the rapid detection systems for skin fungi.

Images