CN113712527A

CN113712527A - Three-dimensional blood flow imaging method and system based on amplitude decorrelation

Info

Publication number: CN113712527A
Application number: CN202110815552.8A
Authority: CN
Inventors: 李花坤; 李鹏
Original assignee: Zhejiang University ZJU
Current assignee: Zhejiang University ZJU
Priority date: 2021-07-19
Filing date: 2021-07-19
Publication date: 2021-11-30

Abstract

The invention discloses a three-dimensional blood flow imaging method and system based on amplitude decorrelation. Collecting OCT scattering signals of scattering signal samples in a three-dimensional space through a collector; constructing a two-dimensional characteristic space by combining an established classifier with the inverse signal-to-noise ratio and the amplitude decorrelation coefficient of the OCT scattering signal, and realizing the classification of dynamic blood flow and static tissues, wherein the method specifically comprises the following steps: calculating and analyzing the amplitude part of the OCT scattering signal by adopting first-order and zero-order autocovariance to obtain two characteristics of a reciprocal signal-to-noise ratio and an amplitude decorrelation coefficient; constructing a two-dimensional feature space; and constructing a signal-to-noise ratio self-adaptive classifier based on the characteristic space based on a multivariate time series model and statistical analysis, and removing static background tissues. The invention can obviously inhibit the influence of system noise on blood flow imaging, improve the contrast of blood flow images and improve the accuracy of blood flow quantification; in addition, the method provided by the invention is suitable for an OCT system with unstable phase due to the adoption of an amplitude decorrelation coefficient as a motion contrast.

Description

Three-dimensional blood flow imaging method and system based on amplitude decorrelation

Technical Field

The present invention relates generally to the field of biomedical imaging, and more particularly to flow imaging associated with Optical Coherence Tomography (OCT) and Optical Coherence Angiography (OCTA) and signal-to-noise adaptive classifiers based on OCT scattering signal inverse signal-to-noise ratio, amplitude decorrelation coefficient feature space.

Background

Blood perfusion is an important index for measuring physiological functions and pathological states, and the current common blood vessel imaging technology in clinic needs intravenous injection of exogenous markers, so that the current common blood vessel imaging technology is not suitable for long-term and frequent tracking detection of human blood flow due to possible side effects. In recent years, an angiography technology OCTA developed on the basis of an optical coherence tomography technology replaces a traditional exogenous fluorescent marker with endogenous blood flow motion, has the characteristics of non-invasiveness and no marker, and has the capability of clearly and reliably three-dimensionally imaging a microvascular network in a biological tissue, so that the technology is developed rapidly since the invention is invented and is applied to researches such as fundus imaging and cerebral cortex vessel imaging.

In order to acquire an OCTA blood flow image, it is usually necessary to perform repeated sampling (repeated a-line scanning or B-frame scanning) at certain time intervals at each spatial position of biological tissue, the motion intensity at each signal is quantified by analyzing the temporal dynamics of OCT scattered signals, and the blood flow signal and static tissue signal are classified according to the quantified motion intensity. The classification of OCTA blood flow, which has been reported so far, is mainly based on the difference, variance or decorrelation calculation between adjacent A-line scans (or between adjacent B-line scans). The method analyzes the change of signals in a time-space window comprising three-dimensional space and time based on the operation of decorrelation, fully utilizes sample data, and has high motion contrast and anti-noise interference capability. Meanwhile, since the decorrelation measures the similarity between adjacent B scanning frames, the influence of the change of the overall light source intensity is small.

The complex decorrelation operation needs a stable complex interference signal, but the phase part of the interference signal in the swept-frequency OCT system is unstable, and thus is difficult to be used for the calculation of the complex decorrelation coefficient. Therefore, in the swept-frequency OCT system, an amplitude decorrelation coefficient is widely used as a motion contrast, but the stability of the amplitude component of the scattering signal is affected by the random noise of the system. Decorrelation artifacts introduced by amplitude fluctuations due to random noise strongly influence the visibility of blood vessels, especially in low signal-to-noise ratio regions (e.g. deep tissue). Although studies have analyzed the asymptotic relationship between the complex decorrelation coefficients and the inverse of the signal-to-noise ratio, classification boundaries derived based on complex decorrelation cannot be used directly for amplitude decorrelation. Theoretical analysis demonstrates that the relationship between the inverse signal-to-noise ratio and the amplitude decorrelation coefficient (IDa) is significant for suppressing amplitude decorrelation artifacts.

Most of the existing signal-to-noise ratio adaptive OCTA classification methods are only suitable for OCT systems with stable phase signals. The existing method based on the amplitude component of the scattering signal is mainly based on numerical analysis, lacks theoretical quantitative derivation, and cannot effectively and accurately distinguish dynamic blood flow signals from static background tissues interfered by noise.

Disclosure of Invention

In order to solve the problems existing in the background technology and overcome the defects of the prior art, the invention provides an optical coherence angiography method and system based on an OCT scattering signal inverse-amplitude decorrelation coefficient characteristic space. The invention can obviously inhibit the artifact problem caused by noise, improve the contrast ratio of blood flow and background and improve the continuity of blood vessels.

The purpose of the invention is realized by the following technical scheme:

a three-dimensional blood flow imaging method based on amplitude decorrelation comprises the following steps:

a collector is used for collecting OCT scattering signals of scattering signal samples in a three-dimensional space;

a classifier is characterized in that a two-dimensional feature space is constructed by combining the inverse signal-to-noise ratio of an OCT scattering signal and an amplitude decorrelation coefficient, the signal-to-noise ratio adaptive classifier is obtained through analysis, and classification of dynamic blood flow and static background tissue signals is achieved through the signal-to-noise ratio adaptive classifier;

and an imaging contrast step, namely generating a blood flow contrast map based on the classification result of the classifier.

The scattering signal sample is a biological tissue sample, which may be, for example, skin, brain tissue, or eye of a human or other animal.

The OCT scattering signal for collecting the scattering signal sample in the three-dimensional space comprises the following steps: and performing three-dimensional OCT scanning imaging on the scattering signal sample, repeatedly sampling the same spatial position and the nearby position at T different time points, or further combining other dimensions, such as: and carrying out parallel sampling on multiple dimensions such as wavelength, angle, polarization and the like.

The three-dimensional OCT scanning imaging is carried out on the scattering signal sample, the same spatial position and the nearby position are repeatedly sampled at T different time points, and one of the following methods is adopted: a time domain OCT imaging method for changing the optical path of the reference arm by scanning; a spectral domain OCT imaging method for recording spectral interference signals by using a spectrometer; a frequency sweep OCT imaging method for recording spectrum interference signals by utilizing a frequency sweep light source.

The classifier specifically comprises:

s1, calculating the characteristic of the OCT scattering signal: calculating and analyzing the amplitude part of the OCT scattering signals by adopting first-order and zero-order autocovariance, and performing moving average or Gaussian average on multiple dimensions such as time, space and channels (including wavelength, angle, polarization and the like) to obtain two characteristics of reciprocal signal-to-noise ratio and amplitude decorrelation coefficient of each OCT scattering signal;

s2, establishing a signal-to-noise ratio reciprocal-amplitude decorrelation coefficient (IDa) feature space, and mapping the voxels of the OCT scattering signal to the signal-to-noise ratio reciprocal-amplitude decorrelation coefficient feature space; the method comprises the steps of analyzing and processing in a signal-to-noise ratio reciprocal-amplitude decorrelation coefficient characteristic space to establish a signal-to-noise ratio adaptive classifier, classifying OCT scattering signals by using the signal-to-noise ratio adaptive classifier into dynamic blood flow signals and static tissue signals, and generating a blood flow radiography image according to a classification result of the classifier.

In S1, the characteristics of the amplitude decorrelation coefficient include: and calculating a decorrelation coefficient for the amplitude part of the OCT scattering signals obtained by scanning at T different time points at the same spatial position or the nearby position, wherein the decorrelation coefficient is used as the OCTA blood flow contrast of each OCT scattering signal, namely as the OCTA blood flow information.

The prior art solution uses complex decorrelation coefficients, whereas the present invention uses amplitude decorrelation coefficients. Compared with the prior art, the method only utilizes the amplitude information of the OCT scattering signal, so that the method is suitable for the OCT imaging system with unstable phase, such as the sweep OCT system with higher scanning speed at present, and has wider application scenes.

In step S2, a signal-to-noise ratio adaptive classifier is established in the signal-to-noise ratio reciprocal-amplitude decorrelation coefficient feature space through analysis processing, and the analysis processing specifically includes:

analyzing an asymptotic dependence relationship between an amplitude decorrelation coefficient and the inverse signal-to-noise ratio of the static signal by adopting a multivariate time sequence model, wherein the asymptotic dependence relationship is a curve, and carrying out statistical analysis to obtain a standard deviation of the static signal in the inverse signal-to-noise ratio-amplitude decorrelation coefficient characteristic space overall distribution based on a probability density function of the static signal in time and space dimension distribution so as to obtain a standard deviation of the dynamic and static signal in the IDa characteristic space overall distribution; and determining the distribution boundary of the static signal with the signal-to-noise ratio adaptation as the classification boundary of the signal-to-noise ratio adaptation classifier by combining the asymptotic dependence of the static signal amplitude decorrelation coefficient and the inverse signal-to-noise ratio and the standard deviation of the overall distribution in the feature space.

Determining a signal-to-noise adaptive static signal distribution boundary as a classification boundary of a signal-to-noise adaptive classifier by combining an asymptotic dependency relationship between a static signal amplitude decorrelation coefficient and a reciprocal of a signal-to-noise ratio and a standard deviation of total distribution in a feature space, and specifically determining by adopting the following formula:

wherein D is_c(iSNR) is the signal-to-noise ratio adaptive dynamic and static signal classification boundary, sigma, in the classifier_st0Standard deviation, σ, representing the global distribution of the stationary signal in IDa feature space with a time-space mean kernel size of 2 in amplitude decorrelation calculations_stThe standard deviation of the general static signal in the total distribution of IDa characteristic space is shown;

representing amplitude of static signalThe asymptotic dependence between the decorrelation coefficient and the inverse of the signal-to-noise ratio.

The multivariate time series model is specifically defined as follows: the collected OCT scattering signal A (m, t) is formed by superposing a back scattering light signal S (m, t) of a biological tissue sample and complex white Gaussian noise n (m, t), wherein m and t respectively represent the three-dimensional space and time of a voxel. The backscattered light signal S and the noise n satisfy the following conditions:

1. n (m, t) is white noise independent of the temporal and spatial dimensions and independent of the backscattered light S (m, t);

2. in the spatial dimension S (m, t) is a stationary, traversal time series that is mutually independent and co-distributed, and which is expected to exist, with limited variance;

3. in the time dimension, S (m, t) satisfies independent equal distribution for dynamic signals; for static signals, S (m, t) is a constant vector.

The statistical analysis of the OCT scattered signal is specifically defined as: and analyzing the distribution standard deviation of the amplitude decorrelation coefficient under the condition of any inverse signal-to-noise ratio according to the probability density function of the distribution of the static signals in the time and space dimensions.

In specific implementation, an asymptotic relation between the amplitude decorrelation coefficient of the static signal and the reciprocal of the signal-to-noise ratio is calculated according to the multivariate time series model. In order to obtain the standard deviation of the overall distribution of the static signal in the IDa characteristic space, firstly, under the condition that the time-space average kernel size is 2, the distribution standard deviation of the amplitude decorrelation coefficient under the condition of any reciprocal of the signal-to-noise ratio is obtained based on the statistical analysis of the probability density function of the distribution of the static signal in the time and space dimensions. Further, assuming that the samples of the time dimension and the space dimension are independent of each other, the distribution variance is inversely proportional to the time-space kernel size, and then the standard deviation of the static signal in the IDa characteristic space distribution under any time-space kernel size is obtained.

And (3) combining the asymptotic dependence relationship between the static signal amplitude decorrelation coefficient and the inverse signal-to-noise ratio and the standard deviation of the overall distribution in the feature space, constructing a signal-to-noise ratio self-adaptive classifier, distinguishing the dynamic blood flow signal projected into the feature space from the static background signal, and displaying the separated blood flow signal by using the amplitude decorrelation coefficient or directly setting the amplitude decorrelation coefficient to be 1.

And generating a blood flow image based on the classification result of the classifier, wherein the static tissue signals with noise signals are removed from the classification result, the rest signals are used as dynamic blood flow signals, and the amplitude decorrelation coefficient values of the dynamic blood flow signals are used for forming the blood flow image.

The generating of the blood flow contrast map based on the classification result of the classifier further comprises: after a blood flow radiography image based on a classification result of the classifier is obtained, operations such as Gaussian filtering, median filtering, Hessian filtering and the like are carried out to remove noise, and the blood flow radiography image effect is improved.

II, a three-dimensional blood flow imaging system based on amplitude decorrelation:

the OCT optical coherence tomography detection device is used for collecting OCT scattering signals of scattering signal samples in a three-dimensional space;

and one or more signal processors for acquiring and analyzing the inverse signal-to-noise ratio information and the amplitude decorrelation coefficients of the OCT scattering signals and classifying the dynamic blood flow signals and the static tissue background by combining the inverse signal-to-noise ratio information and the amplitude decorrelation coefficients of the scattering signals.

Specifically, a two-dimensional feature space is constructed by combining the inverse signal-to-noise ratio of an OCT scattering signal and an amplitude decorrelation coefficient, a signal-to-noise ratio adaptive classifier is obtained through analysis, and classification of dynamic blood flow and static background tissue signals is achieved by the aid of the signal-to-noise ratio adaptive classifier.

The OCT optical coherence tomography detection device adopts one of the following methods:

the system comprises a low-coherence light source, an interferometer and a detector, wherein the low-coherence light source and the detector are respectively connected to the interferometer;

or comprises a low-coherence light source, an interferometer and a spectrometer, wherein the low-coherence light source and the spectrometer are respectively connected to the interferometer;

or comprises a swept-bandwidth spectral light source, an interferometer and a detector, which are respectively connected to the interferometer.

The OCT optical coherence tomography detection device is optionally provided with a visible light indicating device which is used for indicating the position of an OCT detection beam and guiding the placement position of a detection target.

The visible light indicating device mainly comprises a visible light indicating light source and a collimating lens.

According to the method, asymptotic distribution of static signals in IDa characteristic space is obtained based on a multivariate time sequence model, statistical analysis is carried out based on probability density functions of the static signals in time dimension and space dimension to obtain standard deviation of overall distribution of the signals in the characteristic space, and then a signal-to-noise ratio self-adaptive IDa-OCTA classifier based on IDa characteristic space is constructed, so that accurate distinguishing of dynamic blood flow and static background signals is achieved.

The invention has the following beneficial effects and innovation points:

compared with the prior art, the method starts from the basic assumption of the statistical characteristics of the static signals in time and space dimensions, combines a statistical model, analyzes the relation between the amplitude decorrelation coefficient and the reciprocal of the signal-to-noise ratio of the static signals from the theoretical angle, and establishes the signal-to-noise ratio self-adaptive classifier. The classifier has universality, can be suitable for an OCT system with unstable phase, can effectively inhibit decorrelation artifacts introduced by random noise, and improves the quality of blood flow radiography.

Compared with the prior art, the invention has the following remarkable advantages:

1. and performing decorrelation calculation by using only amplitude components of the OCT scattering signals to obtain amplitude decorrelation coefficients. The signal-to-noise ratio self-adaptive classifier provided by the invention has no requirement on the stability of the system phase, so that the signal-to-noise ratio self-adaptive classifier has wider application scenes and can be suitable for OCT systems with unstable phases, such as frequency-sweeping OCT systems.

2. In OCTA based decorrelation calculations, noise-induced decorrelation artifacts are difficult to distinguish from dynamic blood flow in static regions of low signal-to-noise ratio due to random noise causing fluctuations in the OCT scattering signal. A common solution is to set an empirical threshold for intensity masking, which is equivalent to removing the signal in the low signal-to-noise ratio region with an intensity (signal-to-noise ratio) threshold in IDa feature space. Although simple, the method also removes possible blood flow signals in a low signal-to-noise ratio area, and influences the visibility and continuity of blood vessels in a deep area. According to theoretical derivation, the invention constructs the signal-to-noise ratio self-adaptive classifier according to the asymptotic distribution relation of the static signals in the IDa characteristic space and the standard deviation of the total distribution, can improve the accuracy of classification of the dynamic and static signals, and obtains higher blood flow contrast.

3. The classifier provided by the invention starts from the basic hypothesis of the statistical characteristics of the static signals in time and space dimensions, is analyzed by combining with a statistical model, and has solid theoretical foundation and universality.

4. Different from a numerical analysis method, the method provided by the invention strictly deduces the distribution condition of the signal in IDa characteristic space, and is beneficial to further optimization of the algorithm.

5. The decision parameters of the classifier are only determined by the noise level of the system and the size of the decorrelation calculation kernel, and other complicated calibrations for other parameters of the system or complicated corrections for a correlation algorithm are not needed.

6. Compared with the existing method, the classifier established by the invention is more reliable; meanwhile, most static and noise areas are removed, the visibility and the overall contrast of the blood flow radiography image under all signal-to-noise ratios can be improved, and the performance of the provided signal-to-noise ratio self-adaptive classifier is obviously superior to that of the traditional method through a large number of sample verification.

Drawings

FIG. 1 is a schematic diagram of the process of the present invention;

FIG. 2 is a schematic view of the apparatus of the present invention;

FIG. 3 is a schematic view of an apparatus according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of an IDa classifier;

fig. 4(a) is a scatter plot of the spatial distribution of static (crosses) and dynamic (circles) signals in the IDa feature space. Amplitude decorrelation uses a time-space kernel calculation with size N300. The black solid line represents the theoretical asymptote of the static signal distribution, and the black dotted line represents the theoretical classification boundary line D_c。

Fig. 4(b) shows the relationship between the standard deviation of the static signal distribution and the inverse signal-to-noise ratio in the case of an average kernel size N of 2, N of 15, and N of 75. Scatter is the result of numerical simulation, and the solid line corresponds to formula (8);

FIG. 5 is a graph showing the performance of the IDa-OCTA method by tissue phantom experiments;

FIG. 5(a) is a block diagram of a tissue phantom, the graph on the left being a plot of signal-to-noise ratio versus depth decay;

FIG. 5(b) is an amplitude decorrelation plot for a tissue phantom. The static area marked by the triangle represents a high decorrelation value due to random noise in a low signal-to-noise ratio area;

in fig. 5(c), voxels of the static region (the dashed rectangle in fig. 5 a) and the dynamic region (the solid rectangle in fig. 5 a) are mapped to the IDa feature space, and the dynamic and static signals are classified according to the classification boundary defined according to equation (9);

FIG. 5(d) is an IDa blood flow mask obtained after removing the static and noise signals in IDa feature space;

fig. 5(e) is an IDa-OCTA blood flow imaging result obtained by overlaying the IDa blood flow mask of fig. 5(d) on the amplitude decorrelation map of fig. 5(b), and a time-space decorrelation calculation kernel with a size of 5 × 5 × 3 × 4(z × x × y × t) is used to reduce the degree of dispersion of the signal in the feature space distribution;

fig. 5(f) is a graph of the working plots of the subjects using the proposed IDa mask method and the global threshold method at different multiples of standard deviation. In the global thresholding method, a fixed intensity threshold (mean plus three standard deviations of the noise) and a series of different global decorrelation thresholds are used to derive the blood flow mask. Further, the time-space decorrelation calculation kernel size employed here is 5 × 3 × 1 × 5(z × x × y × t);

FIG. 6 shows the results of IDa-OCTA imaging of oral mucosa in healthy subjects;

FIG. 6(a) is a cross-sectional view (x-z) of a representative structure;

FIG. 6(b) is a representative amplitude decorrelation profile (x-z);

FIG. 6(c) is the distribution of signals in IDa feature space and proposed classification boundaries in the cross-sectional view;

FIG. 6(d) a cross-sectional view obtained by IDa-OCTA blood flow imaging;

FIG. 6(e) is a maximum projection of papillary lamina flow contrast;

fig. 6(f) is a maximum value lens result chart of blood flow imaging of the reticular lamina propria.

Detailed Description

Reference will now be made in detail to the present embodiments of the invention, examples of which are illustrated in the accompanying drawings, which form a part hereof. It should be noted that the description and illustrations are exemplary only and should not be construed as limiting the scope of the invention, which is defined by the appended claims, as any variation based on the claims is intended to be within the scope of the invention.

The examples of the invention are as follows:

to facilitate an understanding of embodiments of the invention, operations are described as multiple discrete operations, but the order of description does not represent the order in which the operations are performed.

The x-y-z three-dimensional coordinate representation based on spatial direction is adopted for the sample measurement space in the description. This description is merely intended to facilitate discussion and is not intended to limit application of embodiments of the present invention. Wherein: the depth z direction is a direction along the incident optical axis; the x-y plane is a plane perpendicular to the optical axis, where x is orthogonal to y, and x denotes the OCT lateral fast scan direction and y denotes the slow scan direction.

The above m, t, S, n, etc. represent variables, merely to facilitate discussion, and are not intended to limit the application of embodiments of the present invention, and may be any number of 1,2,3, etc. For simplicity, the average discussion of the OCT system wavelength, angle, and polarization dimensions is omitted here, and the space-time dimensions are only used as examples. The actual implementation steps are the same as the operations described below in the spatio-temporal dimension.

As shown in figure 1, firstly, for a signal acquisition part, an OCT three-dimensional scanning imaging is carried out on a tissue sample, and repeated sampling 1 is carried out on the same or adjacent spatial positions at T different time points. And secondly, a signal classification part is used for constructing a two-dimensional characteristic space by combining the inverse signal-to-noise ratio of the OCT scattering signal and the amplitude decorrelation coefficient, so that classification 2 of the dynamic blood flow signal and the static tissue background is realized.

The method comprises the following specific steps:

1) based on the amplitude part of the OCT scattered signals, the relative motion of the blood flow and the surrounding tissue is analyzed to obtain the inverse signal-to-noise ratio and amplitude decorrelation coefficient characteristics 21 of each OCT scattered signal.

For a certain local area in blood flow and surrounding tissue, for the amplitude part of the OCT scattered signal of each voxel, the first and zeroth order autocovariance and amplitude decorrelation coefficients of each voxel are obtained by averaging (i.e. convolving with a high-dimensional averaging kernel) with its neighboring B-scan frames (x-z plane) of T OCT scans:

wherein, C represents the first-order autocovariance, I represents the zero-order autocovariance, namely the intensity, D represents the amplitude decorrelation coefficient, and is used as the OCTA blood flow information; a is₀(m, t) is the normalized OCT signal amplitude at time t for a certain spatial position (z, x), defined as a₀(m，t)＝a(m，t)/σ_nWhere a (m, t) is the corresponding OCT signal amplitude,

is the noise level of the OCT system; m represents the total number of high-dimensional average kernels in an x-y-z space, wherein the high-dimensional average kernels are taken when decorrelation coefficients are calculated; m represents the ordinal number of the high-dimensional average kernel in the x-y-z space when the decorrelation coefficient calculation is carried out; t represents the total number of high-dimensional average kernels in the time dimension T during the computation of the decorrelation coefficient, namely the total number of the high-dimensional average kernels in the same space in the OCT scanThe number of frames of the inter-position B scan frame; t represents the ordinal number of the high-dimensional average kernel in the time dimension when the decorrelation coefficient calculation is performed. C represents the first-order autocovariance, I represents the zero-order autocovariance, namely the intensity, D represents the amplitude decorrelation coefficient and is used as OCTA blood flow information; a is₀(m, t) is the normalized OCT signal amplitude at time t for a certain spatial position (z, x), defined as a₀(m，t＝a(m，t)/σ_nWhere a (m, t) is the corresponding OCT signal amplitude,

is the noise level of the OCT system.

In the calculation process, the first-order and zero-order autocovariance of each voxel is calculated by adopting the formula, and the average is carried out on all dimensions such as time, space and the like, so that the amplitude decorrelation values of all voxels in the scanning volume of the whole scattering signal sample are obtained, and the operation speed can be improved.

2) In OCT systems, the noise source is mainly shot noise, considered approximately constant throughout the scan volume, and can be obtained by calculating the average of the OCT signal in the top air region and the bottom noise region in the tomogram.

The inverse signal-to-noise ratio, insr, for each voxel is then calculated using the following formula, defined as follows:

wherein I represents the zeroth order autocovariance.

3) An IDa two-dimensional feature space is established in combination with the above obtained inverse signal-to-noise ratio and amplitude decorrelation coefficients, and the OCT scattering signals are projected into the feature space 22.

The asymptotic distribution 231 of the static signal in the IDa feature space is analyzed based on a multivariate time series model. In IDa feature space, the value of the asymptotic dependence of the amplitude decorrelation coefficient of the static signal on the reciprocal of the signal-to-noise ratio on the curve

And the inverse signal-to-noise ratio satisfies the following relationship:

wherein e represents a natural constant, s₀Is an integral variable parameter, similar to the integral variable x, L in a normal fixed integral_0.5Lagrange polynomials, → indicate convergence.

Convergence can only be achieved if the average kernel is sufficiently large, and the standard deviation 232 of the overall distribution of the signal is further analyzed in the case of a limited average kernel. Starting from the basic assumption of the statistical characteristics of the static signal in the time and space dimensions, and combining with a statistical model, theoretical derivation can obtain the standard deviation sigma of the total distribution of the static signal in the IDa characteristic space when the average kernel is 2_st0Satisfies the following conditions:

wherein a is₀(t) represents the normalized OCT signal amplitude for the current frame, defined as a (t)/σ_nWhere a (t) is the OCT signal amplitude, a0(t +1) represents the normalized OCT signal amplitude of the next adjacent frame, SNR represents the specific signal-to-noise ratio, insr^*Is the corresponding inverse signal-to-noise ratio, d (a)₀(t)) represents the relationship between decorrelation coefficients and normalized OCT signal amplitude for a certain signal-to-noise ratio, E [. cndot.]Representing a mathematical expectation.

Probability density function p of amplitude decorrelation coefficients for a specific signal-to-noise ratio in the equation_D|SNR*Comprises the following steps:

wherein the content of the first and second substances,

probability density function representing amplitude of backscattered light signal

Probability density function representing amplitude of OCT scattering signal collected at a certain time of amplitude of backscattered light signal, I₀(. is a first class zero order modified Bessel function; a is₀Indicating normalized OCT Signal amplitude, SNR^*Indicating a particular signal-to-noise ratio.

Furthermore, assuming that the signals acquired in the temporal and spatial dimensions are independent of each other, the variance of the spatial distribution of the signals in the IDa characteristic is inversely proportional to the size N of the time-space kernel, M · T, and the standard deviation σ of the distribution of the stationary signals_stComprises the following steps:

wherein σ_st0Standard deviation, σ, representing the global distribution of the stationary signal in IDa feature space with a time-space mean kernel size of 2 in amplitude decorrelation calculations_stThe standard deviation of the general static signal in the total distribution of IDa characteristic space is shown;

4) a classifier 24 is then constructed to classify 25 the OCT scattered signal in a two-dimensional feature space of the signal-to-noise ratio inverse-amplitude decorrelation coefficients.

The asymptotic relation of the static scattering signals in IDa characteristic space and the standard deviation of the overall distribution are obtained through the previous theoretical derivation, and the following signal-to-noise ratio adaptive classifier is constructed by taking a three-time standard deviation (3 sigma) boundary of the static signal distribution as a classification boundary based on the Lauda criterion (3 sigma criterion):

wherein D is_c(iSNR) is the signal-to-noise ratio adaptive amplitude decorrelation classification boundary in the classifier.

Specifically, as shown in fig. 4(a), the black dotted line is the classification boundary, the left side is the static background signal, and the right side is the dynamic blood flow signal.

5) Based on the results of the classifier classification, a blood flow map 26 is generated.

The amplitude decorrelation coefficient may be used to display the separated blood flow signals, or set to 1 directly, according to the classification result. And performing morphological filtering or Gaussian filtering and other processing on the preliminary classification result from the aspect of the graph to obtain a final OCTA blood flow image.

Fig. 2 is a schematic structural diagram of an acquisition device of the OCT angiography technique based on the inverse signal-to-noise ratio and amplitude decorrelation coefficient feature space in the present invention. The main structure of the low coherence interferometry part of the device is an interferometer which is composed of 11-23 parts. The light source 11 is connected to an input end of one side of the beam splitter 12, and light emitted from the light source 11 is split into two light beams by the beam splitter 12: one beam of light enters a reference arm of the interferometer through a polarization controller 13 and irradiates a reference plane mirror 15 through a reference arm collimating mirror 14; the other beam of light enters the sample arm through another polarization controller 13, and is focused on the sample 21 to be measured through the collimating lens 16 and the scanning device optical path. In the optical path of the scanning device, light beams are reflected by the two-dimensional

scanning mirror groups

17 and 18, the "4 f"

lens groups

54 and 55 and the dichroic mirror 19, and then are focused on a sample 21 to be measured through the objective lens 20. The

lens groups

54 and 55 are formed by arranging two

lenses

54 and 55 on the same optical axis, and the design of the

lens groups

54 and 55 is to ensure that the beam center of the mirror surface of the two-dimensional scanning galvanometer and the beam center of the reflecting surface of the dichromatic mirror are fixed and unchanged during scanning, so that the beam in the OCT sample arm does not influence the imaging characteristic of the objective lens during scanning.

Then, the light reflected by the reference arm and the sample arm returns to the beam splitter 12 for output through the original path, and is received by the interference signal detection device 22 after interference occurs, and the interference signal detection device 22 is connected to the signal processor module and the calculation unit 23. For the optical fiber type optical path, two polarization controllers 13 are adopted to adjust the polarization state of the light beam, and the signal interference effect is maximized.

The specific implementation is also provided with a visible light indicating device, the visible light indicating device comprises a low-power visible light source 25, a collimating lens 24 and a light filter 52, and the visible light used for indicating sequentially passes through the collimating lens 24, the light filter 52, the dichroic mirror 19 and the focusing objective lens 20 and then reaches the sample 21 to be measured.

According to different ways of detecting signals with low coherence interference, the three-dimensional blood flow imaging system apparatus based on inverse signal-to-noise ratio and amplitude decorrelation coefficient feature space shown in fig. 2 specifically includes:

1) a time domain measurement device. The light source 11 uses broadband low coherent light, the plane mirror 15 can move along the optical axis direction, and the interference signal detection device 22 is a point detector. The optical path of the reference arm is changed by moving the plane mirror 15, the interference signals of the two arms are detected by the point detector 22, and the low coherence interference detection is carried out on the scattered signals in the z direction of a certain space depth, so that a sampling body of the depth space dimension is obtained.

2) Spectral domain measuring device. The light source 11 adopts broadband low-coherence light, the plane mirror 15 is fixed, and the interference signal detection device 22 adopts a spectrometer. The interference signal passes through a line camera in the spectrometer 22 while the interference spectrum is recorded. And analyzing the interference spectrum signals by adopting a Fourier analysis method, and parallelly acquiring scattering information in the depth z direction so as to obtain a sampling body of the depth space dimension.

3) Provided is a sweep frequency measuring device. The light source 11 adopts a sweep frequency light source, the plane mirror 15 is fixed, and the interference signal detection device 22 adopts a point detector. The point detector 22 time-divisionally records the low coherence interference spectrum of the swept source. And (3) carrying out Fourier analysis on the interference spectrum signal, and obtaining the scattering information in the depth z direction in parallel, thereby obtaining a sampling body of the depth space dimension.

Fig. 3 illustrates an exemplary embodiment utilizing the present invention. The three-dimensional OCT angiography system based on the characteristic space of the inverse signal-to-noise ratio and the amplitude decorrelation coefficient comprises a broadband low-coherence light source 26, an optical circulator 27, a fiber coupler 28 with a splitting ratio of 50:50, a first polarization controller 29, a first fiber collimating device 30, a focusing lens 36, a plane mirror 37, a second polarization controller 38, a second fiber collimating device 39, two-dimensional

scanning galvanometer combinations

40 and 41, a dichroic mirror 42, a focusing objective 43, a third fiber collimating device 45, a grating 46, a focusing lens 47, a high-speed linear array camera 48, a signal processor module and calculation unit 49, a visible light indication light source 50, a collimating lens 51, and "4 f"

lens groups

56 and 57.

The broadband low-coherence light source 26 shown in this example is a superluminescent diode light source with a central wavelength of 1325nm and a bandwidth of 100nm, the focusing objective 43 is an achromatic doublet cemented lens with a focal length of 30mm, and the high-speed linear array camera 48 is a linear array scanning camera composed of 2048 voxel units. In which light emitted from a low coherence broadband light source 26 used in the present exemplary apparatus enters an input end of one side of an optical fiber coupler 28 with a splitting ratio of 50:50 after passing through an optical circulator 27, and the light emitted from the optical fiber coupler 28 is split into two sub-beams: one of the beams is connected to a first fiber collimating device 30 in the reference arm through a first polarization controller 29 by an optical fiber, passes through a collimating and focusing lens 36 and then irradiates a plane mirror 37; the other beam of light is connected to a second optical fiber collimating device 39 of the sample arm part through an optical fiber via a second polarization controller 38, and after being collimated, reflected by two scanning

galvanometers

40, 41, a

4f lens group

56, 57 and a dichroic mirror 42, is focused on a sample 44 to be measured by a focusing objective 43 and is reflected and scattered back to the back, wherein the

lens group

56, 57 is designed to ensure that the beam center of the mirror surface of the two-dimensional scanning galvanometer and the beam center of the reflecting surface of the dichroic mirror are fixed and unchanged during scanning. The light reflected by the plane mirror 37 in the reference arm interferes with the light backscattered from the sample to be measured in the sample arm at the optical fiber coupler 28, the interference light is detected and recorded by the spectrometers 45-48, and then the interference light is collected by the signal processor module and the computing unit 49 and is subjected to signal analysis and processing. The spectrometer comprises devices 45-48 which are connected in sequence, wherein the device 45 is an optical fiber coupler, the device 46 is a grating, the device 47 is a converging lens, and light split by grating dispersion is focused on a linear array detector shown by 48.

The specific implementation is also provided with a visible light indicating device, the visible light indicating device comprises a visible light indicating light source 50 and a collimating lens 51, and the visible light emitted by the visible light indicating light source 50 and used for indicating passes through the collimating lens 51, the dichroic mirror 42 and the focusing objective 43 and then reaches the sample 44 to be measured.

Fig. 4 is a verification of the asymptotic distribution of the dynamic and static signals in the IDa feature space, the standard deviation of the distribution, and the validity of the classification boundary. Static (crosses) and dynamic (circles) are computed using a spatio-temporal kernel of size 15 × 20(M × T) and projected into IDa feature space as in fig. 4 (a). The static signal is distributed around its theoretical asymptote (solid black line, equation 5), and can be effectively removed by the proposed classification boundary (dashed black line, equation 9), proving the correctness of the theoretical derivation. Fig. 4(b) analyzes the relationship between the standard deviation of the distribution of static signals and the reciprocal of the signal-to-noise ratio obtained from theoretical derivation (solid line) and simulation analysis (scatter point) for different mean kernels. The theoretical derivation is consistent with the simulation analysis, and when N is 2, the coefficient R is determined²When N is 15, the coefficient R is determined at 1.000²When N is 75 and 0.914, the coefficient R is determined²＝0.896。

FIG. 5 uses a tissue phantom experiment to verify the effectiveness of the proposed IDa-OCTA method. The tissue phantom is composed of two parts: one part adopts solid gel and low-concentration milk to simulate static tissues, and the other part adopts milk solution to simulate dynamic blood flow. The milk solution is separated from the solid gel by using a capillary tube with an inner diameter of 0.3 +/-0.1 mm. As shown in the block diagram of FIG. 5(a), four capillaries were embedded at different depth positions of the solid gel. Fig. 5(b) is a cross-sectional view of amplitude decorrelation coefficients, and in general, dynamic signals exhibit higher decorrelation values, while static regions exhibit lower decorrelation values. However, the static regions deep indicated by the triangles also exhibit large decorrelation values (i.e., noise-induced decorrelation artifacts) due to the interference of additional random noise. According to the classification boundary of the dynamic and static signals (dotted line in fig. 5c, formula 9) derived by theory, the IDa feature space is divided into two parts, dynamic and static. The distribution of voxels in the static region (dashed rectangle in fig. 5 a) and dynamic region (solid rectangle in fig. 5 a) in the feature space matches the theoretical analysis. Voxels in the feature space to the left of the classification line are zeroed out and the IDa mask shown in fig. 5(d) is obtained by digital etching. The IDa mask was overlaid on the decorrelation map to finally obtain the IDa-OCTA angiography result (FIG. 5e), and in order to quantitatively evaluate the performance of the IDa-OCTA method, the true positive rate (true positive rate) was calculated to be 82.9% and the false positive rate (false positive rate) was calculated to be 0.7%. As shown in fig. 5(f), using the IDa mask method with different multiples of standard deviation, the area under the working curve of the former subject is significantly larger, meaning more accurate classification accuracy, compared to the method using a global intensity threshold (noise mean plus three times standard deviation) and a series of decorrelation thresholds to obtain the blood vessel mask.

FIG. 6 shows IDa-OCTA imaging of oral mucosa in healthy subjects. Fig. 6(a) is a representative structural cross-section, and fig. 6(b) is a corresponding decorrelation cross-section, which shows a large decorrelation value due to interference of random noise in a deep static tissue region, and is difficult to distinguish from dynamic blood flow. With the help of the classification boundaries defined in the IDa feature space (fig. 6c), the IDa mask is obtained by digital etching, and the final angiographic result is obtained by overlaying the IDa mask in the decorrelation map (fig. 6 d). Based on the three-dimensional IDa-OCTA blood flow radiography result, a specific depth range is selected for maximum value projection, so that a depth-resolved blood vessel network can be obtained, and fig. 6e and 6f are blood vessel networks of a papillary lamina propria and a reticular lamina propria respectively.

The above experimental comparison results fully illustrate that: the optical coherent blood flow radiography method based on the signal-to-noise ratio reciprocal-amplitude decorrelation coefficient characteristic space can improve the accuracy of blood flow signal classification, realize effective enhancement of blood flow contrast and improvement of blood flow image quality, and has the outstanding and obvious technical effect.

The invention can obviously inhibit the influence of system noise on blood flow imaging, improve the contrast of blood flow images and improve the accuracy of blood flow quantification; because the amplitude decorrelation coefficient is used as the motion contrast, the method provided by the invention is suitable for the OCT system with unstable phase.

Claims

1. A method for three-dimensional blood flow imaging based on amplitude decorrelation, comprising:

a collector (1) is used for collecting OCT scattering signals of scattering signal samples in three-dimensional space;

a classifier (2) is used for constructing a two-dimensional characteristic space by combining the inverse signal-to-noise ratio of an OCT scattering signal and an amplitude decorrelation coefficient, analyzing to obtain a signal-to-noise ratio self-adaptive classifier, and classifying dynamic blood flow and static background tissue signals by using the signal-to-noise ratio self-adaptive classifier;

and an imaging contrast step of generating a blood flow contrast map (26) based on the classification result of the classifier (2).

2. The method of claim 1, wherein the amplitude decorrelation-based three-dimensional blood flow imaging method comprises: the OCT scattering signal for collecting the scattering signal sample in the three-dimensional space comprises the following steps: and performing three-dimensional OCT scanning imaging on the scattering signal sample, and repeatedly sampling the same spatial position and the nearby position at T different time points.

3. The method of claim 2, wherein the amplitude decorrelation-based three-dimensional blood flow imaging method comprises: the three-dimensional OCT scanning imaging is carried out on the scattering signal sample, the same spatial position and the nearby position are repeatedly sampled at T different time points, and one of the following methods is adopted: a time domain OCT imaging method for changing the optical path of the reference arm by scanning; a spectral domain OCT imaging method for recording spectral interference signals by using a spectrometer; a frequency sweep OCT imaging method for recording spectrum interference signals by utilizing a frequency sweep light source.

4. The method of claim 1, wherein the amplitude decorrelation-based three-dimensional blood flow imaging method comprises: the classifier (2) specifically comprises:

s1, calculating the OCT scattering signal characteristics (21): calculating and analyzing the amplitude part of the OCT scattering signals by adopting first-order and zero-order autocovariance, and performing sliding average or Gaussian average on multiple dimensions such as time, space, channels and the like to obtain two characteristics of the reciprocal of the signal-to-noise ratio and the amplitude decorrelation coefficient of each OCT scattering signal;

s2, establishing a signal-to-noise ratio reciprocal-amplitude decorrelation coefficient (IDa) feature space, and mapping the voxels of the OCT scattering signal to the signal-to-noise ratio reciprocal-amplitude decorrelation coefficient feature space (22); and analyzing and processing in the signal-to-noise ratio reciprocal-amplitude decorrelation coefficient feature space to establish a signal-to-noise ratio adaptive classifier (24), and classifying (25) the OCT scattering signals into dynamic blood flow signals and static tissue signals by using the signal-to-noise ratio adaptive classifier.

5. The method of claim 4, wherein the amplitude decorrelation-based three-dimensional blood flow imaging method comprises: in S1, the characteristics of the amplitude decorrelation coefficient include: and calculating a decorrelation coefficient for the amplitude part of the OCT scattering signals obtained by scanning the same spatial position or the positions nearby the same spatial position at T different time points, and taking the decorrelation coefficient as the OCTA blood flow contrast of each OCT scattering signal.

6. The method of claim 4, wherein the amplitude decorrelation-based three-dimensional blood flow imaging method comprises: in step S2, the signal-to-noise ratio adaptive classifier is established by analyzing and processing in the signal-to-noise ratio reciprocal-amplitude decorrelation coefficient feature space, specifically:

the method comprises the steps of analyzing an asymptotic dependency relationship (231) between an amplitude decorrelation coefficient and an inverse signal-to-noise ratio of a static signal by adopting a multivariate time series model, carrying out statistical analysis on the basis of a probability density function of the static signal in time and space dimension distribution to obtain a standard deviation (232) of the static signal in signal-to-noise ratio inverse-amplitude decorrelation coefficient feature space overall distribution, and determining a signal-to-noise ratio adaptive static signal distribution boundary as a classification boundary of a signal-to-noise ratio adaptive classifier by combining the asymptotic dependency relationship between the amplitude decorrelation coefficient and the inverse signal-to-noise ratio of the static signal and the standard deviation in the feature space overall distribution.

7. The method of claim 6, wherein the amplitude decorrelation-based three-dimensional blood flow imaging method comprises: determining a signal-to-noise adaptive static signal distribution boundary as a classification boundary of a signal-to-noise adaptive classifier by combining an asymptotic dependency relationship between a static signal amplitude decorrelation coefficient and a reciprocal of a signal-to-noise ratio and a standard deviation of total distribution in a feature space, and specifically determining by adopting the following formula:

a value on a curve representing the asymptotic dependence of the amplitude decorrelation coefficient of the static signal on the inverse of the signal-to-noise ratio.

8. An amplitude decorrelation-based three-dimensional blood flow imaging system for implementing the method according to any one of claims 1 to 7, comprising:

9. The amplitude decorrelation-based three-dimensional blood flow imaging system according to claim 8, wherein: the OCT optical coherence tomography detection device adopts one of the following methods:

10. The amplitude decorrelation-based three-dimensional blood flow imaging system according to claim 8 or 9, wherein: the OCT optical coherence tomography detection device is optionally provided with a visible light indicating device which is used for indicating the position of an OCT detection beam and guiding the placement position of a detection target.