Disclosure of Invention
In order to solve the technical problems that the existing method does not consider the cooperative change condition of oral cavity components, has defects in the aspect of oral cavity health state identification precision and further has poor effect on assisting in oral cavity health state monitoring, the invention aims to provide an oral cavity health monitoring system based on laser irradiation and a data processing method thereof, and the adopted technical scheme is as follows:
in a first aspect, the present invention provides a data processing method of an oral health monitoring system based on laser irradiation, comprising:
acquiring spectral data of each oral cavity region of an oral cavity in different wave bands, wherein the spectral data comprises first spectral data of teeth and second spectral data of saliva in the same oral cavity region;
According to the characteristic distribution of each dimension of the spectral data of each oral cavity region in each wave band and the data deviation condition under the health state, obtaining the spectral residual contrast of the spectral data of each oral cavity region in each wave band;
According to the difference distribution of spectrum residual contrast between the first spectrum data and the second spectrum data in different wave bands in each oral cavity region, combining the relevance of the characteristic distribution of the corresponding first spectrum data and second spectrum data in each dimension to obtain a cooperative change index of the first spectrum data and the second spectrum data in each oral cavity region in each wave band;
According to the fluctuation condition of the cooperative change index of the first spectrum data and the second spectrum data in each oral cavity region under each wave band, combining the spectrum residual contrast of the first spectrum data under each wave band to obtain the abnormal attention degree of the first spectrum data of each oral cavity region under each wave band;
the dental health of each oral area is monitored based on the abnormal attention.
Preferably, the obtaining the spectral residual contrast of the spectral data of each oral area under each band according to the characteristic distribution of each dimension of the spectral data of each oral area under each band and the data deviation condition under the health state specifically includes:
according to the difference of the characteristic distribution of each dimension between the spectral data of each oral cavity region in each wave band and the spectral data of the same oral cavity region in the same wave band under the health state, obtaining a spectral deviation value corresponding to the spectral data of each oral cavity region in each wave band;
Determining an equalization characteristic value under each wave band based on the equalization distribution of the spectrum deviation values corresponding to the same kind of spectrum data in all oral cavity areas under each wave band;
and determining the spectral residual contrast of the spectral data of each oral cavity region in each wave band based on the difference duty ratio between the spectral deviation value corresponding to the spectral data of each oral cavity region in each wave band and the equalization characteristic value of the corresponding spectral data in the same wave band.
Preferably, the obtaining a spectral deviation value corresponding to the spectral data of each oral cavity region in each band according to the difference of the characteristic distribution of each dimension between the spectral data of each oral cavity region in each band and the spectral data of the same oral cavity region in the same band in the health state specifically includes:
For any spectrum data of any oral cavity region under any wave band, acquiring peak value data, wavelength, peak width and peak area of the spectrum data, and constructing a characteristic distribution sequence of the spectrum data;
and taking the difference distance between the characteristic distribution sequence of each kind of spectral data of each oral cavity region in each wave band and the characteristic distribution sequence of the same kind of spectral data of the same oral cavity region in the same wave band in the health state as a spectral deviation value corresponding to the spectral data of each oral cavity region in each wave band.
Preferably, the obtaining the index of the cooperative variation of the first spectral data and the second spectral data in each oral area in each band according to the difference distribution of the spectral residual contrast between the first spectral data and the second spectral data in different bands in each oral area and the correlation of the feature distribution of the corresponding first spectral data and the corresponding second spectral data in each dimension specifically includes:
Obtaining the drifting consistency degree of each oral cavity region between different wave bands according to the difference condition of spectrum residual contrast between the first spectrum data and the second spectrum data in different wave bands in each oral cavity region;
Obtaining the spectrum correlation degree of each oral cavity region between different wave bands according to the similarity between the characteristic distribution sequence of the first spectrum data of each wave band and the characteristic distribution sequence of the second spectrum data of each wave band in the same oral cavity region;
and determining the product of the drift consistency degree and the spectrum correlation degree as a cooperative variation index of the first spectrum data and the second spectrum data in each oral cavity region under each wave band.
Preferably, the obtaining the drift consistency degree of each oral cavity region between different bands according to the difference of the spectrum residual contrast between the first spectrum data and the second spectrum data in different bands in each oral cavity region specifically includes:
For any oral cavity region, the first spectrum data under any one wave band is recorded as a first characteristic spectrum, and the second spectrum data under any one wave band is recorded as a second characteristic spectrum;
and determining the drift consistency degree between the first characteristic spectrum and the second characteristic spectrum based on the negative correlation coefficient of the difference between the spectrum residual contrast corresponding to the first characteristic spectrum and the spectrum residual contrast corresponding to the second characteristic spectrum.
Preferably, the obtaining the spectrum correlation degree of each oral cavity region between different bands according to the similarity between the characteristic distribution sequence of the first spectrum data of each band and the characteristic distribution sequence of the second spectrum data of each band in the same oral cavity region specifically includes:
And taking the pearson correlation coefficient between the characteristic distribution sequence corresponding to the first characteristic spectrum and the characteristic distribution sequence corresponding to the second characteristic spectrum as the spectrum correlation degree between the first characteristic spectrum and the second characteristic spectrum.
Preferably, the obtaining the abnormal attention degree of the first spectrum data of each oral cavity region under each band according to the fluctuation condition of the cooperative variation index of the first spectrum data and the second spectrum data under each band and combining the spectrum residual contrast of the first spectrum data under each band specifically includes:
Obtaining abnormal credibility between the first spectrum data of each wave band and the second spectrum data of each wave band in each oral cavity region according to the cooperative change index between the first spectrum data of each wave band and the second spectrum data of each wave band in each oral cavity region and the deviation degree between the overall distribution conditions of all oral cavity regions;
Obtaining the response deviation degree of the first spectrum data of each oral cavity region under each wave band according to the abnormal credibility and the index of the cooperative change between the first spectrum data of each wave band and the second spectrum data of all wave bands in each oral cavity region;
and taking the product of the response deviation degree and the spectral residual contrast of the first spectral data of the corresponding oral cavity region in the same wave band as the abnormal attention degree of the first spectral data of each oral cavity region in each wave band.
Preferably, the obtaining the abnormal credibility between the first spectrum data of each band and the second spectrum data of each band in each oral area according to the deviation degree between the index of the cooperative variation between the first spectrum data of each band and the second spectrum data of each band in each oral area and the overall distribution condition of all oral areas specifically includes:
Taking any one oral cavity area as a selected oral cavity area, and taking any two wave bands as a first wave band and a second wave band respectively;
calculating the average value of the cooperative change indexes between the first spectrum data of the first wave band and the second spectrum data of the second wave band in all the oral cavity areas to obtain a cooperative characteristic value;
And taking the difference between the cooperative variation index of the first spectral data of the first wave band and the second spectral data of the second wave band in the selected oral cavity area and the cooperative characteristic value as abnormal credibility between the first spectral data of the first wave band and the second spectral data of the second wave band in the selected oral cavity area.
Preferably, the obtaining the response deviation degree of the first spectrum data of each oral area under each band according to the abnormal reliability and the index of the cooperative variation between the first spectrum data of each band and the second spectrum data of all bands in each oral area specifically includes:
for the selected oral cavity region, normalizing the index of the cooperative change between the first spectrum data of the first wave band and the second spectrum data of each wave band to obtain the characteristic weight of the second spectrum data of each wave band;
And carrying out weighted averaging on abnormal credibility between the first spectrum data of the first wave band and the second spectrum data of each wave band by utilizing the characteristic weight to obtain the response deviation degree of the first spectrum data of the selected oral cavity region in the first wave band.
In a second aspect, the present invention provides a laser-based oral health monitoring system for implementing the steps of a data processing method of a laser-based oral health monitoring system, the laser-based oral health monitoring system comprising:
the data acquisition module is used for acquiring spectral data of each oral cavity region of the oral cavity under different wave bands, wherein the spectral data comprise first spectral data of teeth and second spectral data of saliva in the same oral cavity region;
the residual error comparison module is used for obtaining the spectral residual error contrast of the spectral data of each oral cavity region under each wave band according to the characteristic distribution of each dimension of the spectral data of each oral cavity region under each wave band and the data deviation condition under the health state;
The collaborative analysis module is used for obtaining collaborative variation indexes of the first spectrum data and the second spectrum data in each oral cavity region under each wave band according to the difference distribution of spectrum residual contrast between the first spectrum data and the second spectrum data in different wave bands in each oral cavity region and combining the relevance of the characteristic distribution of the corresponding first spectrum data and second spectrum data in each dimension;
The abnormality analysis module is used for obtaining the abnormal attention degree of the first spectrum data of each oral cavity region under each wave band by combining the spectrum residual contrast of the first spectrum data under each wave band according to the fluctuation condition of the cooperative variation index of the first spectrum data and the second spectrum data under each wave band in each oral cavity region;
and the health monitoring module is used for monitoring the tooth health condition of each oral cavity area based on the abnormal attention degree.
The embodiment of the invention has at least the following beneficial effects:
The invention firstly collects the local range data of the oral cavity region and provides a data basis for the subsequent process of analyzing the cooperative relationship of teeth and saliva of different local regions and different wave bands. Then, the first aspect analyzes the deviation condition of the actually collected spectrum data and the spectrum data in the health state, primarily quantifies the abnormal condition and the deviation condition of the actually collected spectrum data, and obtains the spectrum residual contrast. In the second aspect, the difference and correlation of the deviation conditions of the spectrum data of the teeth and saliva in each wave band are analyzed, the cooperative change relation of the spectrum signal change between the teeth and saliva in each wave band in the local oral cavity area is evaluated, the single data source can be mutually verified through the matching analysis, and the error caused by single data noise or individual difference is reduced. And finally, evaluating the abnormal condition corresponding to the teeth by combining the analysis result of the cooperative change relation and the comparison result of the deviation condition, obtaining the abnormal attention, improving the recognition accuracy of the oral health state, and providing an auxiliary diagnosis result with good diagnosis effect for medical staff.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description refers to the specific implementation, structure, characteristics and effects of an oral health monitoring system based on laser irradiation and a data processing method thereof according to the present invention.
Before describing the embodiments of the present application, some of the words used in the present application are explained in order to facilitate understanding of those skilled in the art, and are not intended to limit the present application.
The data acquisition process of oral cavity laser irradiation specifically comprises the following steps:
1. Instrument preparation:
the laser excitation source and the spectrometer output mode is pulse, and the adjustable power is 1-10mW. The spectrometer comprises a resolution less than or equal to 1nm and a detection range of 400-900nm, a probe, a two-in-one optical fiber probe, a positioning device, an adjustable XYZ triaxial displacement table or an intraoral positioning fixture, a controller and a triggering laser, wherein the probe is aligned to a dental surface or saliva film, the positioning device is used for ensuring consistent repeated measurement positions, and the triggering laser and the spectrometer are synchronously collected.
2. Tooth spectrum data acquisition:
the patient is prepared by rinsing the face of the tooth with water (or gently blowing dry) to remove saliva residue, and fixing the soft tissue with a mouth mirror or retractor to expose the scan area.
The probe alignment is to keep the fiber optic probe perpendicular to the tooth surface at a distance of about 2nm to 3mm and cover the entire tooth surface in a grid or line scan (e.g., move every 1 mm).
Laser irradiation and signal acquisition, namely laser 655nm continuous irradiation, pulse width of 10ms and interval of 100ms, setting integration time of 50ms to 100ms by a spectrometer, acquiring for 3 times, averaging, reducing random error, and storing original spectrum (wavelength and intensity) and photographing record scanning position.
And (3) marking data, namely associating each group of spectrums with tooth numbers and scanning coordinates.
Saliva sample preparation, taking small amounts of saliva (about 10-20 μl) naturally secreted in the mouth, and immediately after tapping the gums at different target areas, the sample can be formed directly on a slide, or placed in a detection well of a microfluidic chip.
The probe is positioned and irradiated by using the same or switched to a 405nm laser source, the probe lightly touches the surface of the saliva film, keeps in contact with the sample but is not compressed, and the laser power is reduced to 1mW to 5mW, so that sample flickering and thermal effects are avoided.
In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
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.
The following specifically describes a specific scheme of an oral health monitoring system based on laser irradiation and a data processing method thereof provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a data processing method of an oral health monitoring system based on laser irradiation according to an embodiment of the present invention is shown, and the method includes the following steps:
Step S100, spectrum data of each oral cavity area of the oral cavity under different wave bands are obtained, wherein the spectrum data comprise first spectrum data of teeth and second spectrum data of saliva in the same oral cavity area.
First, based on the population cavity structure, the oral cavity is divided into a plurality of functional areas, and as a specific example, each tooth can correspond to one oral cavity area, so as to provide a data basis for the subsequent characteristic analysis process aiming at the abnormal data conditions of different teeth.
For the distribution of the spectral data of the teeth and saliva in the same oral cavity region, as a specific example, the present embodiment obtains the first spectral data of the teeth and the second spectral data of the saliva in a band range from 400nm to 900nm, and uniformly divides a plurality of different bands in the band range, and for example, the wavelength range of each band is the same, and may be divided by taking a wavelength range of 100nm as an example, that is, each wavelength range corresponds to one band, and in the present embodiment, 5 different bands are included in total. It should be appreciated that the present embodiment is to collect spectral data of teeth and saliva by laser induced fluorescence and reflectance spectroscopy techniques, which are well known to those skilled in the art and will not be described in any greater detail herein.
To this end, two kinds of spectral data, one of which is the spectral data of the teeth, in this embodiment denoted as first spectral data, and one of which is the spectral data of saliva, denoted as second spectral data, can be obtained in each oral cavity region in each band.
Step S200, according to the characteristic distribution of each dimension of the spectral data of each oral cavity region in each wave band and the data deviation condition under the health state, obtaining the spectral residual contrast of the spectral data of each oral cavity region in each wave band.
When there is an oral health problem, there is an abnormality in the spectral response of each component in the oral cavity, thereby assisting in diagnosing oral health. In the process of extracting abnormal components in the oral cavity according to the spectrum data, the most obvious deviation information from the healthy baseline often corresponds to early pathological changes or potential difference information, for example, a partial comparison chart of the first spectrum data of a certain tooth and the corresponding healthy baseline spectrum data is shown in fig. 2, and the expression degree of the deviation condition of each component is obtained through comparing the actually obtained spectrum with the healthy baseline. In fig. 2, the abscissa indicates wavelength, the ordinate indicates signal intensity of spectral data, the dotted line indicates a healthy baseline spectral curve, and the solid line indicates spectral data with oral problems, so that the difference of spectral data between corresponding bands can be reflected.
As a specific example, as shown in fig. 3, the acquisition method of the spectral residual contrast may be implemented by steps S201 to S203.
Step S201, according to the difference of the characteristic distribution of each dimension between the spectral data of each oral cavity region in each wave band and the spectral data of the same oral cavity region in the same wave band in the health state, obtaining a spectral deviation value corresponding to the spectral data of each oral cavity region in each wave band.
Firstly, the oral cavity data of human health needs to be acquired as a basis for data comparison and analysis, and different normal databases can be respectively constructed for the healthy oral cavities of different age layers in practice. It should be understood that, for the stomatology or the stomatology hospital, the internal system stores the relevant information of the patient, the information system acquires and counts the spectrum data of the patient of the same age layer in the tooth health state, for each same oral area, the average value of the tooth spectrum data of all the health samples of the same age layer is calculated to form the first health data of the teeth in the corresponding oral area, and likewise, the average value of the saliva spectrum data of all the health samples of the same age layer is calculated to form the second health data of the saliva in the corresponding oral area. It should be noted that, the age group may enable the related medical personnel to divide according to the specific implementation scenario.
So far, one oral cavity area corresponds to a group of spectral data under the health state under each wave band, and the spectral data comprise the first health data corresponding to the teeth and the second health data corresponding to saliva. For convenience of description, in this embodiment, feature analysis is performed by default with data to be monitored of a patient whose age level has been determined, and corresponding first health data and second health data are acquired in a database for analysis.
The method comprises the steps of firstly, acquiring peak value data, wavelength of a peak value, peak width and peak area of spectrum data for any spectrum data of any oral cavity region under any wave band, and constructing a characteristic distribution sequence of the spectrum data.
Wherein the peak data refers to the value of the maximum signal intensity on the selected spectrum data, the peak width refers to the half-height peak width of the peak on the selected spectrum, namely the width of the peak at the half position of the peak data, and the specific acquisition method can acquire the signal intensity equal to the peak dataThe length of the interval between two wavelengths corresponding to the signal intensity is taken as the peak width. The peak area can be calculated by means of integration, which is not described in detail herein for the known art.
It should be understood that the feature extraction is performed according to the same method on each spectral data of each band in each oral cavity region, and at the same time, the feature extraction is performed according to the same method on the spectral data in a health state to obtain the feature distribution sequence of each spectral data of each oral cavity region in each band.
And secondly, taking the difference distance between the characteristic distribution sequence of each spectrum data of each oral cavity region in each wave band and the characteristic distribution sequence of the same spectrum data of the same oral cavity region in the same wave band in the health state as a spectrum deviation value corresponding to the spectrum data of each oral cavity region in each wave band.
In this embodiment, the analysis process is described by taking the oral characteristic data of the teeth and saliva corresponding to any one of the oral areas as an example, and specifically, the first spectral data of the ith band and the second spectral data of the ith band included in any one of the oral areas are described by taking as an example.
And calculating the DTW distance of the characteristic distribution sequence between the first spectral data of the ith wave band and the first health data of the same oral cavity region in the ith wave band according to the first spectral data corresponding to the teeth, and taking the DTW distance as a spectral deviation value corresponding to the first spectral data of the ith wave band.
And calculating the DTW distance of the characteristic distribution sequence between the second spectral data of the ith wave band and the second health data of the same oral cavity region in the ith wave band according to the second spectral data corresponding to saliva, and taking the DTW distance as a spectral deviation value corresponding to the second spectral data of the ith wave band.
The calculation method of the spectrum deviation value is the same for each spectrum data. The spectral deviation value characterizes the characteristic difference and the degree of performance of the characteristic deviation condition between each spectrum data and the corresponding spectrum data in the health state.
Step S202, based on the balanced distribution of the spectrum deviation values corresponding to the same kind of spectrum data in all oral cavity areas in each wave band, the balanced characteristic value in each wave band is determined.
And regarding the first spectrum data of the teeth in the oral cavity region, taking the average value of spectrum deviation values corresponding to the first spectrum data of all the oral cavity regions in the ith wave band as an equalization characteristic value corresponding to the first spectrum data in the ith wave band, and representing the overall distribution condition of the first spectrum data of the teeth in the same wave band deviating from the health state in all the oral cavity regions of the whole oral cavity.
It should be understood that the calculation and analysis are also performed according to the same method for the second spectrum data of saliva, and the average value of the spectrum deviation values corresponding to the second spectrum data of all oral cavity areas in the ith wave band is used as the equalization characteristic value corresponding to the second spectrum data in the ith wave band, so that the overall distribution condition of the second spectrum data of saliva deviating from the health state in all oral cavity areas of the whole oral cavity in the same wave band is represented.
Step S203, determining a spectral residual contrast of the spectral data of each oral area in each band based on a difference ratio between the spectral deviation value corresponding to the spectral data of each oral area in each band and the equalization characteristic value of the corresponding spectral data in the same band.
By comparing laser spectrums of a plurality of areas in the oral cavity of the same individual, each tooth spectrum of the same individual and each wave band in saliva spectrums obtained by each area are respectively compared with saliva spectrums of other teeth and other areas, local pathological signals are amplified, and residual contrast of each wave band is extracted.
For the first spectral data of teeth in the oral cavity area, calculating a difference absolute value between a spectral deviation value corresponding to the first spectral data in the ith wave band and an equalization characteristic value corresponding to the first spectral data in the ith wave band, taking the ratio of the difference absolute value to the equalization characteristic value as a difference duty ratio, and recording the difference duty ratio as a spectral residual contrast of the first spectral data in the ith wave band, wherein the difference absolute value and the equalization characteristic value can be expressed as follows by using the formula:
;
Wherein, theRepresenting the spectral residual contrast of the first spectral data of the oral region at the ith band,A spectral deviation value representing the first spectral data of the oral region at the ith band,Representing the equalization characteristic of the first spectral data of the oral region at the ith band.
It should be understood that the method for calculating the spectral residual contrast corresponding to the second spectral data of saliva in each band in the oral cavity region is the same as the method for calculating the spectral residual contrast of the first spectral data of teeth in each band in the oral cavity region, and only the data involved in calculation and the related data in the same dimension of the corresponding kind of spectral data need to be replaced.
The spectral residual contrast corresponding to each kind of spectral data in the oral cavity region under each wave band represents the residual distribution condition of the difference between each kind of spectral data and the oral cavity integral characteristic distribution, and the expression degree of the local pathological signals can be reflected preliminarily.
Step S300, according to the difference distribution of spectrum residual contrast between the first spectrum data and the second spectrum data in different wave bands in each oral cavity region, combining the relevance of the characteristic distribution of the corresponding first spectrum data and second spectrum data in each dimension to obtain the cooperative change index of the first spectrum data and the second spectrum data in each oral cavity region in each wave band.
In the oral spectrum, signals of main components such as water, protein and the like often occupy most of the dynamic range, trace metabolites or mineral changes caused by pathology are submerged in a large signal background, tooth saliva spectrum cooperative analysis is needed, and whether pathological abnormalities actually exist or not is judged, so that the method has higher credibility than that of independently observing a spectrum signal. Therefore, it is necessary to study the data components of teeth and saliva in the same oral cavity region and analyze the synergistic relationship between the spectra of each band in teeth and saliva.
Based on this, as shown in fig. 4, the method of analyzing the index of the cooperative variation of teeth and saliva in each oral area at each band can be implemented by steps S301 to S303.
Step S301, obtaining the drift consistency degree of each oral cavity region between different wave bands according to the difference condition of the spectrum residual contrast between the first spectrum data and the second spectrum data in different wave bands in each oral cavity region.
When an oral problem exists, the spectral responses of a plurality of wave bands deviate, the residual contrast of the oral spectral data wave bands is correspondingly changed, and if the wave bands are interacted, the change of the spectral residual contrast is also correlated. And obtaining the spectrum drift consistency of the spectrum data of the teeth and saliva between the wave bands through the spectrum residual contrast between the wave bands.
Specifically, for any one of the oral cavity regions, the first spectral data in any one of the bands is recorded as a first characteristic spectrum, the second spectral data in any one of the bands is recorded as a second characteristic spectrum, as a specific example, the first spectral data of the teeth in the nth oral cavity region in the nth band is recorded as the first characteristic spectrum, the second spectral data of the saliva in the mth oral cavity region in the mth band is recorded as the second characteristic spectrum, it should be understood that n and m are serial numbers of the bands and can be the same or different, and the values are used for representing the distribution of the data differences between the corresponding different spectral data in the same or different bands in the same oral cavity region.
Further, a degree of drift agreement between the first and second characteristic spectra is determined based on a negative correlation coefficient of a difference between the spectral residual contrast corresponding to the first characteristic spectrum and the spectral residual contrast corresponding to the second characteristic spectrum. As a specific example, the degree of agreement of the shift between the first spectral data of the tooth in the nth band and the second spectral data of the saliva in the mth band for the xth oral region can be expressed by the formula:
;
Wherein, theIndicating the degree of agreement of the shift between the first spectral data of the tooth in the nth band and the second spectral data of the saliva in the mth band, i.e. the degree of agreement of the shift between the first characteristic spectrum and the second characteristic spectrum in the xth oral region.The spectral residual contrast of the first spectral data of the xth oral region at the nth band is represented, i.e., the spectral residual contrast of the first characteristic spectrum of the xth oral region.The spectral residual contrast of the second spectral data of the xth oral region at the mth band is represented, i.e., the spectral residual contrast of the second characteristic spectrum of the xth oral region.Represents an exponential function based on a natural constant e, usingIs inversely related to the difference in terms of (a).
Representing the difference in spectral residual contrast between the first characteristic spectrum and the second characteristic spectrum, when the ratioThe smaller the difference from 1, the closer the spectrum residual contrast between the two is, namely the smaller the difference is, the more consistent the variation degree of the signal deviation is, and the larger the corresponding value of the drift consistency degree is. The degree of uniformity of drift reflects the degree of uniformity of signal deviation variations of spectral data of teeth and saliva in various bands within the same oral cavity region.
Step S302, according to the similarity between the characteristic distribution sequence of the first spectrum data of each wave band and the characteristic distribution sequence of the second spectrum data of each wave band in the same oral cavity region, the spectrum correlation degree of each oral cavity region between different wave bands is obtained.
When the oral cavity has health problems, the spectral characteristics of the related wave bands are changed, and the credibility of the abnormal wave bands can be improved through the linkage reaction of the related wave bands among different types of spectral data. And obtaining the correlation between the bands of the spectral data of the teeth and saliva according to the spectral manifestations of the bands in the oral spectral data set.
Specifically, a pearson correlation coefficient between a characteristic distribution sequence corresponding to the first characteristic spectrum and a characteristic distribution sequence corresponding to the second characteristic spectrum is used as a spectrum correlation degree between the first characteristic spectrum and the second characteristic spectrum. The spectral correlation reflects the similarity of the characteristic distribution of spectral data of teeth and saliva in each band in the same oral cavity region.
Step S303, determining the product of the drift consistency degree and the spectrum correlation degree as a cooperative variation index of the first spectrum data and the second spectrum data in each oral cavity region under each wave band.
And comprehensively evaluating the cooperative change relation of the spectral data of the teeth and saliva in the oral cavity region in each wave band through the consistency of the characteristic correlation relation of the teeth and saliva in the same oral cavity region in the wave band and the signal deviation change.
When the spectrum correlation degree between the first characteristic spectrum and the second characteristic spectrum in the oral cavity area is larger, the data characteristic distribution between the first characteristic spectrum and the second characteristic spectrum is similar, so that the spectrum data distribution of teeth and saliva under corresponding wave bands is similar, meanwhile, the drift consistency degree between the first characteristic spectrum and the second characteristic spectrum in the oral cavity area is larger, the consistency of signal deviation change between the first characteristic spectrum and the second characteristic spectrum is higher, the deviation change of the spectrum data of teeth and saliva under corresponding wave bands is similar, and the cooperative change relationship between the first characteristic spectrum and the second characteristic spectrum in the oral cavity area is stronger.
Step S400, according to fluctuation conditions of the cooperative change indexes of the first spectrum data and the second spectrum data in each oral cavity region in each wave band, combining the spectrum residual contrast of the first spectrum data in each wave band to obtain the abnormal attention degree of the first spectrum data in each oral cavity region in each wave band.
The tooth and saliva can reflect a certain data cooperative relationship in terms of physiological functions and component distribution in terms of spectral response. By collecting the spectrum data of teeth and saliva around the teeth and analyzing the matching condition, when the teeth and saliva are matched with each other or a specific wave band deviates from a healthy base line, the reliability of abnormality identification can be obviously improved, weak but truly existing pathological signals are amplified, the extracted features have auxiliary reference value, and the sensitivity and the specificity of oral health monitoring are obviously improved.
Based on this, as shown in fig. 5, the method of acquiring the abnormal attention of the first spectral data of each oral area at each band may be realized by steps S401 to S403.
Step S401, obtaining abnormal credibility between the first spectrum data of each wave band and the second spectrum data of each wave band in each oral cavity area according to the degree of deviation between the cooperative variation index between the first spectrum data of each wave band and the second spectrum data of each wave band in each oral cavity area and the overall distribution condition of all oral cavity areas.
The larger the difference of the collaborative variation relationship of the oral cavity regions with problems is, the higher the abnormal reliability of the spectrum signals of the corresponding wave bands can be explained, namely the higher the characteristic abnormal degree is shown. And obtaining the abnormal credibility of the spectrum signals of the wave band according to the difference of the cooperative change relations of different areas.
As a specific example, the present embodiment uses the y-th oral area as the selected oral area, the i-th band as the first band, the j-th band as the second band, and i and j are both serial numbers of the bands, which may be the same or different, based on the same reason that the n-th band and the m-th band are the same, and subsequently used to represent spectral data in the corresponding bands of teeth and saliva, respectively.
And step two, calculating the average value of the cooperative change indexes between the first spectrum data of the first wave band and the second spectrum data of the second wave band in all the oral cavity areas to obtain a cooperative characteristic value.
And thirdly, taking the difference between the cooperative variation index of the first spectral data of the first wave band and the second spectral data of the second wave band in the selected oral cavity area and the cooperative characteristic value as the abnormal credibility between the first spectral data of the first wave band and the second spectral data of the second wave band in the selected oral cavity area.
As a specific example, the calculation formula of the abnormal reliability may be expressed as: , wherein,Representing an abnormal degree of reliability between the first spectral data in the ith band and the second spectral data in the j bands in the jth oral area,Indicating a co-variation index between the first spectral data in the ith band and the second spectral data in the j bands in the jth oral region,The mean value of the index of the cooperative change between the first spectrum data in the ith wave band and the second spectrum data in the j wave bands in all the oral cavity areas is indicated, namely the cooperative characteristic value.
The difference of the cooperative change relation between the first spectrum data in the ith wave band and the second spectrum data in the j wave bands in the y-th oral cavity area and the whole oral cavity areas is reflected, and the larger the value is, the more serious the cooperative change relation in the y-th area deviates from the whole distribution, the larger the degree of abnormality is, and the larger the value of the corresponding abnormal credibility is.
Step S402, obtaining the response deviation degree of the first spectrum data of each oral cavity region under each wave band according to the abnormal reliability and the index of the cooperative change between the first spectrum data of each wave band and the second spectrum data of all wave bands in each oral cavity region.
When the oral cavity problem exists, the abnormal situation of the dental spectrum signal of the oral cavity region has a reference value, the spectral characteristics possibly have weak conditions due to various interferences, and the dental abnormal problem in each oral cavity region can be weighted and comprehensively evaluated by amplifying the spectral abnormal condition of the wave band through the cooperative change relation of the teeth and saliva under each wave band.
And normalizing the index of the cooperative variation between the first spectrum data of the first wave band and the second spectrum data of each wave band for the selected oral cavity area to obtain the characteristic weight of the second spectrum data of each wave band. The normalization method is a well-known technique, and will not be described here too much.
And secondly, weighting and averaging abnormal credibility between the first spectrum data of the first wave band and the second spectrum data of each wave band by utilizing the characteristic weight to obtain the response deviation degree of the first spectrum data of the selected oral cavity region in the first wave band.
Through the cooperative change relation between the spectrum data corresponding to the teeth and saliva, the abnormal credibility of the spectrum signals is weighted and averaged, and the spectrum data corresponding to two wave bands with tighter cooperative change relation has auxiliary reference value as the wave band deviation is more. As a specific example, the degree of response deviation may be formulated as:
;
Wherein, theIndicating the degree of deviation in response of the first spectral data at the ith band in the jth oral area,Representing an abnormal degree of reliability between the first spectral data in the ith band and the second spectral data in the j bands in the jth oral area,Indicating a co-variation index between the first spectral data in the ith band and the second spectral data in the j bands in the jth oral region,Representing the number of bands, norm is the normalization function.
When the value of the cooperative change index is larger, the stronger the cooperative relation of the spectrum signal change between the first spectrum data in the first wave band and the second spectrum data in the second wave band is, and the higher the reference value of the data abnormal condition correspondingly represented by the first spectrum data in the first wave band and the second spectrum data in the second wave band is.
Step S403, taking the product of the response deviation degree and the spectral residual contrast of the first spectral data of the corresponding oral cavity region in the same band as the abnormal attention degree of the first spectral data of each oral cavity region in each band.
The response deviation degree of the first spectrum data in the ith wave band in the (y) th oral cavity region reflects the abnormal degree of deviation of the spectrum data of the teeth in the corresponding oral cavity region in the wave band, and the spectrum residual contrast of the first spectrum data in the ith wave band in the (y) th oral cavity region reflects the abnormal degree of the spectrum data of the teeth in the corresponding oral cavity region in the wave band compared with the health state. By combining the characteristic performances of the two aspects, the degree of abnormality of the teeth in the ith wave band in the (y) th oral cavity area in the aspect of spectrum signals can be accurately represented, and the greater the degree of abnormality is, the greater the corresponding degree of abnormality attention is, namely, the more important observation is needed on the teeth in the oral cavity area, so that the purpose of assisting oral cavity diagnosis is achieved.
Step S500, monitoring the tooth health condition of each oral cavity area based on the abnormal attention.
When the value of the abnormal attention degree of the first spectrum data corresponding to the teeth in each wave band in each oral cavity area is larger, the fact that the spectrum signals of the teeth in the corresponding wave band in the oral cavity area have certain abnormal conditions is indicated, and the attention degree of the teeth in the oral cavity area needs to be increased.
In some embodiments, the threshold value judgment can be used for reminding medical staff to pay important attention to the teeth in the oral cavity area, so that the purpose of carrying out auxiliary diagnosis on the oral cavity teeth by utilizing the spectral data is achieved. As a specific example, the abnormal attention degree is normalized, and when the normalized abnormal attention degree is greater than or equal to a preset abnormal threshold value, medical staff is reminded to pay attention to the teeth in the oral area. When the normalized abnormal attention is smaller than the preset abnormal threshold, the possibility that the abnormal teeth exist in the oral cavity area is small, and excessive attention is not needed, namely reminding operation is not needed. Wherein the anomaly threshold value can be set to 0.7, and the implementer can set according to the specific implementation scenario. Meanwhile, the normalization method is a well-known technique, and will not be described here too much.
In some embodiments, the neural network may also be used to output abnormal or normal results of teeth, and to reference data to medical personnel to assist in their formal diagnostic procedure. That is, when the value of the abnormal attention degree of the first spectral data corresponding to the teeth in each band in each oral area is larger, the attention weight value in the corresponding dimension is larger, and the health state or the abnormal state of the teeth in the corresponding oral area is directly output by using the neural network.
As a specific example, a mixed architecture of a convolutional neural network and a transducer encoder is adopted, namely, firstly, a convolutional layer is used for extracting micro-characteristics from an oral spectrum sequence, then, a multi-head self-attention transducer module is used for adaptively distributing attention weights to each band position, amplifying band responses most relevant to lesions, and finally, a classification head is connected after global pooling to realize oral health or abnormality judgment.
The first spectrum data of teeth under all wavebands are input into a convolution layer after noise removal, baseline correction and normalization, and convolution kernels slide between adjacent wavelength points, so that the model can sensitively capture micro-characteristics, and the screening of local chemical composition or structure is carried out in each wavelet band, thereby enhancing the preliminary response to micro pathological signals.
The convolution output is remodeled along the wavelength axis, added with position codes and then sent to a multi-layer transducer encoder, a multi-head self-attention mechanism of the multi-layer transducer encoder can automatically allocate weights to the positions of each wave band, and in self-attention calculation, the abnormal attention degree of the first spectrum data of each wave band obtained through calculation is overlapped to be an attention scoring bias term. The attention weight is automatically biased to a wavelength position with more diagnostic value when the attention weight is transmitted forwards, namely, the attention weight is given to be high only when certain wavelengths are abnormal in the cooperative analysis of teeth and saliva spectra, so that occasional backgrounds and individual differences are eliminated.
The transducer finally outputs a characteristic diagram of each wave band after being weighted by attention, the characteristic diagram is pooled and converged into a semantic vector with fixed dimension through global average, the position difference is eliminated, a nonlinear decision boundary between health and abnormality is learned through a multi-layer perceptron (MLP), and a health or abnormality result is output, so that the monitoring of oral health risks is realized.
In summary, in the process of analyzing the oral health status through the laser spectrum data, the concentration and proportion of the components in the oral cavity change slightly, so that the spectrum characteristic change is not obvious, and the monitoring of the oral health status is further affected. Therefore, the invention extracts key characteristics reflecting subtle biochemical and structural changes through collaborative analysis of the spectrum data of teeth and saliva, thereby realizing early, accurate and noninvasive monitoring of the health state of the oral cavity, improving the detection sensitivity and accuracy, reducing the risk of misdiagnosis and providing reliable basis for personalized oral health management.
The method comprises the steps of comparing standard spectrums under a healthy state, detecting tiny component proportion change, providing an early warning signal for an early pathological state, primarily screening out a spectrum region with larger difference from normal data, finely analyzing different regions (such as teeth and saliva local regions) in an oral cavity, accurately positioning the distribution region of abnormal components, further quantifying the abnormal degree by calculating the spectrum contrast between the abnormal components and a normal background, and improving the definition and reliability of diagnostic signals, wherein under the oral health state, biochemical components in hard tissues of teeth and saliva should present certain cooperative consistency, and the single data source can be mutually verified through the matching analysis, so that errors caused by single data noise or individual difference are reduced, thereby effectively distinguishing the normal state from the abnormal state, reducing the misdiagnosis rate, and improving the accuracy and the robustness of monitoring the oral health state.
As shown in fig. 6, the present invention further provides a system for monitoring oral health based on laser irradiation, which is used for implementing the steps of a data processing method of the oral health monitoring system based on laser irradiation, and the oral health monitoring system based on laser irradiation includes:
the data acquisition module is used for acquiring spectral data of each oral cavity region of the oral cavity under different wave bands, wherein the spectral data comprise first spectral data of teeth and second spectral data of saliva in the same oral cavity region;
the residual error comparison module is used for obtaining the spectral residual error contrast of the spectral data of each oral cavity region under each wave band according to the characteristic distribution of each dimension of the spectral data of each oral cavity region under each wave band and the data deviation condition under the health state;
The collaborative analysis module is used for obtaining collaborative variation indexes of the first spectrum data and the second spectrum data in each oral cavity region under each wave band according to the difference distribution of spectrum residual contrast between the first spectrum data and the second spectrum data in different wave bands in each oral cavity region and combining the relevance of the characteristic distribution of the corresponding first spectrum data and second spectrum data in each dimension;
The abnormality analysis module is used for obtaining the abnormal attention degree of the first spectrum data of each oral cavity region under each wave band by combining the spectrum residual contrast of the first spectrum data under each wave band according to the fluctuation condition of the cooperative variation index of the first spectrum data and the second spectrum data under each wave band in each oral cavity region;
and the health monitoring module is used for monitoring the tooth health condition of each oral cavity area based on the abnormal attention degree.
Since the specific implementation process of the data processing method of the oral health monitoring system based on laser irradiation has been described in detail, the detailed description is omitted here.
Although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the foregoing embodiments may be modified or equivalents may be substituted for some of the features thereof, and that the modification or substitution does not depart from the scope of the embodiments of the present application.