An inflammatory lesion detection method and system based on parity spectrum decorrelation OCT

Abstract

This invention proposes a method and system for detecting inflammatory lesions based on odd-even spectral decorrelation OCT. The method utilizes the positive complementary characteristics of odd and even sampling points in the frequency domain of the interference spectrum for signal processing. By solving the decorrelation features between operator spectra, it can effectively extract dynamic scattering signals within the tissue, thereby enhancing functional information contrast while maintaining the system's axial resolution at the theoretical limit. Furthermore, by combining intra-frame and inter-frame decorrelation characteristics in static, vascular, and lymphatic regions, it overcomes the interference of highly scattering edema fluid in tissues without the need for exogenous contrast agents, accurately capturing the dilated morphology of weakly scattering lymphatic vessels, and achieving decoupling and independent imaging of blood vessels and lymphatic vessels.

Description

Technical Field

[0001] This invention relates to the fields of optical detection and biomedical functional imaging, and in particular to a method and system for detecting inflammatory lesions based on odd-even spectrum decorrelation OCT. Background Technology

[0002] In the field of optical imaging, optical coherence tomography (OCT) has become an ideal "optical biopsy" tool due to its advantages such as micron-level high resolution, non-invasiveness, and non-destructiveness. Currently, in clinical skin examinations, confocal microscopy has limited imaging depth (usually less than 100 μm), while methods such as skin biopsies are invasive and prone to scarring. OCT technology, by detecting backscattered light within biological tissues, can non-invasively reconstruct the tissue's microstructure.

[0003] However, current research is largely limited to the characterization of tissue morphology and microvascular networks. The microcirculation system, composed of both blood and lymphatic circulation, plays a crucial role in maintaining skin homeostasis, antigen presentation, and inflammatory drainage. In inflammatory lesions, increased vascular permeability leads to interstitial fluid accumulation, causing compensatory dilation of lymphatic vessels. Because lymph fluid is primarily composed of water and a small amount of protein, lacking strong scattering particles such as red blood cells, its optical scattering coefficient is much lower than that of blood. This results in lymphatic vessels often appearing as "transparent" signal holes in conventional OCT images, making them difficult to identify effectively. Existing lymphangiography methods, such as radionuclide imaging and MR lymphangiography, typically require the injection of exogenous contrast agents, posing invasiveness and safety risks. Summary of the Invention

[0004] To address the above issues, this invention proposes an inflammatory lesion detection method based on odd-even spectrum decorrelation OCT. By utilizing the positive interactive complementarity of odd and even sampling points in the frequency domain of the interference spectrum for signal processing, and by solving the decorrelation features between operator spectra, the dynamic scattering signal within the tissue can be effectively extracted. This enhances the contrast of functional information while maintaining the axial resolution of the system at the theoretical limit.

[0005] This invention is achieved through the following technical solution: Firstly, this invention provides temporally continuous multi-frame interferometric spectral data based on suspected inflammatory lesion sites, including:

[0006] Intra-frame decorrelation imaging, inter-frame decorrelation imaging, differential imaging, and reconstructed imaging; Intra-frame decorrelation imaging performs the following steps: S101: Odd-even spectrum extraction is performed on the interference spectral data of each frame to obtain the odd-numbered interference spectrum and even-numbered interference spectrum corresponding to each frame; S102: Based on the intensity values ​​of odd-numbered and even-numbered interference spectra in the same spatial neighborhood within the same frame, calculate the first intensity decorrelation coefficient of the sub-spectrum of the same frame to obtain the first decorrelation matrix; S103: After performing binarization masking based on the first decorrelation matrix, the data is processed with the corresponding interference spectral data to obtain a lymphocyte imaging map; Inter-frame decorrelation imaging performs the following steps: S201: Based on the intensity values ​​of the interference spectra in adjacent time frames at the same lateral position, calculate the second intensity decorrelation coefficient between adjacent time frames to obtain the second decorrelation matrix; S202: After performing binarization masking based on the second decorrelation matrix, time-dispersive processing is performed on the corresponding interference spectral data to obtain a time-decorrelation image; Differential operational imaging performs the following steps: S301: Adaptive weight masking is applied to the temporal decorrelation image based on the lymphocyte imaging image to obtain a vascular imaging image; The reconstructed imaging process performs the following steps: S401: The lymphatic imaging image and the vascular imaging image are reconstructed to obtain a vascular-lymphatic image of the suspected inflammatory lesion site.

[0007] Furthermore, after step S103, the method further includes: S104: The lymphocyte imaging map is processed by the Frangi filtering algorithm to obtain a continuous lymphocyte imaging map.

[0008] Further, step S102 calculates each first intensity decorrelation coefficient in the first decorrelation matrix using the following formula:

[0009] in, The average window size in space. For odd-numbered interference spectra within the window The intensity value of each pixel. For even-numbered interference spectra within the window, the first... The intensity value of each pixel.

[0010] Further, step S201 calculates each second intensity decorrelation coefficient in the second decorrelation matrix using the following formula:

[0011] in, for Time of the first The intensity value of each pixel. for +1 moment Intensity value of each pixel This represents the system noise term.

[0012] Furthermore, the adaptive weight masking process is completed using the following formula:

[0013] in, for t The temporal decorrelation image after adaptive weight masking at time step 1 z The intensity value of each pixel. To suppress the weighting factor.

[0014] Furthermore, after obtaining the vascular imaging image in step S301, the process also includes: S302: Perform spatial noise filtering on the vascular imaging image to obtain a continuous vascular imaging image.

[0015] Furthermore, the spatial noise filtering algorithm is either median filtering or connected component labeling algorithm.

[0016] On the other hand, the present invention also provides an inflammatory lesion detection system based on odd-even spectrum decorrelation OCT, for performing the inflammatory lesion detection method described in any of the above claims, the inflammatory lesion detection system comprising: Intra-frame decorrelation imaging unit, inter-frame decorrelation imaging unit, differential operation imaging unit, and reconstructed imaging unit; The intra-frame decorrelation imaging unit includes: an odd-even spectrum subunit: used to extract odd-even spectra from the interference spectral data of each frame to obtain the odd-numbered interference spectrum and even-numbered interference spectrum corresponding to each frame; The first decorrelation matrix calculation subunit is used to calculate the first intensity decorrelation coefficient of the sub-spectrum of the same frame based on the intensity values ​​of the odd-numbered and even-numbered interference spectra in the same spatial neighborhood, and to obtain the first decorrelation matrix. Lymphatic imaging subunit: used to perform binarization masking based on the first decorrelation matrix and then process it with the corresponding interference spectral data to obtain a lymphatic imaging map; The inter-frame decorrelation imaging unit includes: a second decorrelation matrix calculation subunit: used to calculate the second intensity decorrelation coefficient between adjacent time frames based on the intensity values ​​of the interference spectrum at the same lateral position in adjacent time frames, and obtain the second decorrelation matrix; Temporal decorrelation image imaging subunit: used to perform time-spectral separation processing on the corresponding interference spectral data after performing binarization mask processing based on the second decorrelation matrix, to obtain a temporal decorrelation image; Differential operational imaging unit: used to perform adaptive weight masking processing on the temporal decorrelation image based on the lymphocyte imaging map to obtain a vascular imaging map; Reconstruction imaging unit: used to reconstruct the lymphatic imaging image and the vascular imaging image to obtain a vascular-lymphatic image of the suspected inflammatory lesion site.

[0017] Furthermore, the intra-frame decorrelation imaging unit further includes: Frangi filtering subunit: used to process the lymphatic imaging map using the Frangi filtering algorithm to obtain a continuous lymphatic imaging map.

[0018] Furthermore, the differential operational imaging unit further includes: Spatial filtering subunit: used to perform spatial noise filtering on the vascular imaging image to obtain a continuous vascular imaging image.

[0019] The present invention relates to an inflammatory lesion detection method and system based on odd-even spectral decorrelation OCT. This method utilizes the positive complementary characteristics of odd and even sampling points in the frequency domain of the interference spectrum for signal processing. By solving the decorrelation features between operator spectra, it can effectively extract dynamic scattering signals within tissues. This enhances the contrast of functional information while maintaining the system's axial resolution at the theoretical limit. Furthermore, it combines intra-frame and inter-frame decorrelation characteristics in static, vascular, and lymphatic regions. Without the need for exogenous contrast agents, this invention overcomes the interference of highly scattering edema fluid in tissues, accurately capturing the dilated morphology of weakly scattering lymphatic vessels, and achieving decoupling and independent imaging of blood vessels and lymphatic vessels. Furthermore, the intensity-based decorrelation algorithm of this invention is not limited by the Nyquist sampling theorem regarding phase detection speed, maintaining imaging accuracy and integrity even when the flow rate exceeds 1.5 times the system's Doppler detection limit. Combined with large field-of-view stitching technology, this invention can cover the complete dermal vascular architecture, providing an objective, accurate, and non-destructive testing method for vascular remodeling, drug efficacy evaluation, and early diagnosis in inflammatory lesions.

[0020] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0021] Figure 1 This is a structural diagram of the SD-OCT imaging system used in this invention to acquire spectral data; Figure 2 This is a structural block diagram of an exemplary inflammatory lesion detection system based on parity-even spectrum decorrelation OCT according to the present invention; Figure 3 To execute Figure 2 The flowchart of the inflammatory lesion detection method of the inflammatory lesion detection system shown is as follows; Figure 4 This is a structural block diagram of an intra-frame decorrelation imaging unit in a preferred embodiment of the present invention; Figure 5To execute Figure 4 The flowchart of the execution method of the intra-frame decorrelation imaging unit is shown; Figure 6 The images show the simulated vascular imaging results of samples with different diameters based on the milk-agar capillary sampling tube simulation model in a specific embodiment of the present invention; (a) original structure diagram; (b) original spectrum decorrelation diagram; (c) OMAG diagram; (d) split spectrum decorrelation diagram. Scale bar = 500 μm. Detailed Implementation

[0022] As mentioned above, due to the low scattering properties of lymph fluid, it appears as a transparent void signal in the spectral signal. Therefore, in OCT, it is impossible to perform non-invasive OCT imaging of the inflamed area without adding contrast fluid. To address this issue, the inventors analyzed the spectral signal and discovered that the original interference spectrum acquired by the linear CCD... Along the spectrum Data was extracted from the direction and split into odd-numbered interferometric spectra consisting of odd-numbered data. and even-numbered interference spectra composed of even-numbered data. .because and Originating from the same spectral scan, the true intensity signals of the samples are completely synchronized and consistent in both time and space. However, because the two sets of sub-spectrums occupy completely different physical spectral channels, the detector thermal noise and photon noise exhibit randomly distributed white noise characteristics in the frequency domain, which makes the reconstructed random noise... and It exhibits extremely strong statistical independence. At this point, analyzing the correlation between the two in the static tissue region, vascular region and lymphatic region, it was found that due to the low scattering of the lymphatic region, the strong correlation of the true intensity in the odd and even spectral spatial neighborhood and the strong independence of random noise, only the lymphatic region has extremely strong decorrelation. Therefore, it provides a theoretical basis for non-destructive imaging of lymphatic vessels without contrast agents.

[0023] Based on the above analysis, this invention proposes a method and system for detecting inflammatory lesions based on odd-even spectrum decorrelation OCT, wherein the inflammatory lesion detection system based on odd-even spectrum decorrelation OCT is similar to... Figure 1 The spectrometer connections for the SD-OCT imaging system are shown. Please refer to [link / reference]. Figure 2The system includes an intra-frame decorrelation imaging unit 10, an inter-frame decorrelation imaging unit 20, a differential operation imaging unit 301, and a reconstruction imaging unit 401. The intra-frame decorrelation imaging unit 10 performs odd-even spectral separation processing based on the acquired multi-frame temporal continuous interference spectra, and performs signal processing based on the positive complementary characteristics of odd and even sampling points in the frequency domain. By solving the decorrelation features between the spectra, it can effectively extract dynamic scattering signals within the tissue, achieving lymphatic vessel imaging. The inter-frame decorrelation imaging unit 20 extracts the correlation of consecutive frames of the original spectral data in the time dimension, obtaining a temporal decorrelation image containing signals from the blood flow region and lymphatic vessels. Subsequently, the differential operation imaging unit 301 performs adaptive weight masking processing on the temporal decorrelation image and the lymphatic imaging image to obtain a vascular imaging image. Finally, the reconstruction imaging unit 401 reconstructs the vascular imaging image and the lymphatic imaging image to obtain an image of a non-invasive, non-contrast-free inflammatory area. Please refer to [link to relevant documentation]. Figure 3 The working process of each component in the inflammatory lesion detection system based on parity-even spectrum decorrelation OCT includes: The intra-frame decorrelation imaging unit 10 performs intra-frame decorrelation imaging, including an odd-even spectrum subunit 101 for performing step S101: performing odd-even spectrum extraction on the interference spectral data of each frame to obtain the odd interference spectrum and even interference spectrum corresponding to each frame.

[0024] Interference spectral data by Figure 1 The SD-OCT imaging system shown acquires the following data. Light emitted from a broadband light source (center wavelength 1310 nm, wavelength bandwidth 55 nm) is transmitted through an optical fiber into a fiber coupler (splitting ratio 50:50), where it is split into two beams. One beam enters the reference arm, is reflected by a mirror, and returns to the fiber coupler along the same path. The other beam enters the sample arm, where it is scanned by a two-dimensional galvanometer. The backscattered light from the sample returns to the optical fiber along the same path; hence, this branch is called the sample arm. The returning light is combined in the fiber coupler and finally received by a spectrometer integrating a beam splitter and a linear CCD array (with 1024 detector elements and a maximum acquisition frequency of 91911 Hz), forming an interference spectrum. Six consecutive frames of image data were acquired through six B-scans.

[0025] Subsequently, the raw interference spectrum acquired by the linear CCD was analyzed. Along the spectrum Data was extracted from the direction and split into odd-numbered interferometric spectra consisting of odd-numbered data. and even-numbered interference spectra composed of even-numbered data. This processing is not simply frequency domain segmentation, but rather, while maintaining the physical bandwidth, it reduces the system phase's sensitivity to axial motion by remapping the sampling frequency. When dealing with scattering media with large speckle characteristics, such as biological tissues, this mechanism can convert minute axial vibrations into tolerable phase drifts, thereby smoothing out unsteady noise in the interference spectrum.

[0026] The first decorrelation matrix calculation subunit 102 is used to perform step S102: based on the intensity values ​​of the odd-numbered and even-numbered interference spectra in the same spatial neighborhood within the same frame, calculate the first intensity decorrelation coefficient of the sub-spectrum of the same frame to obtain the first decorrelation matrix.

[0027] Fourier transforms were performed on both odd-numbered and even-numbered interference spectra to reconstruct the reflectance profile within the sample. This eliminated phase fluctuation interference and focused on the scattering intensity contrast. and Become and Its expression is as follows:

[0028]

[0029] in, Represents the true intensity signal of the sample. and This refers to the random noise carried by each sub-spectral channel.

[0030] At this point, the intensity signals of odd-numbered interference spectra and even-numbered interference spectra are structurally equivalent, that is:

[0031] Where z is the pixel coordinate and t is the frame number.

[0032] The noise statistical characteristics (autocorrelation characteristics) are as follows: ，

[0033] Where z is the pixel coordinate and t is the frame number.

[0034] Channel independence is:

[0035] Where z is the pixel coordinate and t is the frame number.

[0036] The correlation of lymphatic areas is as follows:

[0037] Where z is the pixel coordinate and t is the frame number.

[0038] Based on the above analysis, calculations were performed. and The decorrelation values ​​between them enable the identification and classification of targets with different scattering energy levels within the tissue. Therefore, the following formula is constructed to calculate each first intensity decorrelation coefficient in the first decorrelation matrix:

[0039] in, The average window size in space. For odd-numbered interference spectra within the window The intensity value of each pixel. For even-numbered interference spectra within the window, the first... The intensity value of each pixel.

[0040] In the above formula, the numerator term (cross-correlation term) is affected by odd-numbered subspectral noise. With even subspectral noise They are statistically independent, and the noise is distinct from the true signal. They are also uncorrelated, and the expected values ​​of all terms containing noise in their expansions tend to zero, i.e.

[0041] The denominator (the sum of autocorrelation terms) includes the energy contribution of noise in the autocorrelation operation, and its result is derived from the sum of signal power and noise variance. Together they constitute:

[0042] At this point, the decorrelation coefficient model approximates:

[0043] Substitute the signal strength and noise variance at each point to calculate the decorrelation coefficient for each point in the frame image, thus obtaining the first decorrelation matrix.

[0044] The lymphatic imaging subunit 103 is used to perform step S103: after performing binarization masking based on the first decorrelation matrix, it is processed with the corresponding interference spectral data to obtain a lymphatic imaging map.

[0045] Based on the approximate decorrelation model and the characteristics of different tissue regions, it was found that in static tissue regions, due to the strong backscattered signal, the signal power... Under these conditions, the denominator is dominated by the signal term, and the decorrelation value is removed. This indicates that the odd-even spectrum exhibits extremely high structural consistency in the static high-scattering region.

[0046] Although red blood cells are in a flowing state in the vascular region, due to their high backscattering coefficient, the power of the extracted odd and even sub-spectral signals within a single frame is still much greater than the noise variance. Therefore, the decorrelation values ​​calculated between sub-spectrums also satisfy the decorrelation value... 。 This feature ensures that the spectral segmentation algorithm can effectively suppress pseudo-decorrelation interference generated by blood flow within a single frame.

[0047] In the lymphatic region, there is very little scattering medium within the lumen, and the actual signal intensity approaches zero. ,at this time 1 Therefore, after binarizing and masking the first decorrelation matrix, it is calculated with the corresponding interference spectrum data. The signal in the lymphatic vessel region is preserved, while the signal in the other regions is filtered out, thus achieving the stripping of the lymphatic vessel signal. Finally, the lymphatic imaging map is obtained through Fourier transform.

[0048] In another preferred embodiment, considering the exponential attenuation of the actual OCT signal with increasing detection depth, the signal-to-noise ratio (SNR) of deep tissue signals decreases significantly. Beyond a certain depth, the decorrelation results generated by background random noise are highly similar to the decorrelation characteristics of lymphatic vessel signals, making accurate separation difficult to achieve solely based on intensity information or correlation criteria. Conventional processing schemes typically employ depth-limited projection (e.g., projecting only data within a certain threshold range below the skin surface) or hard threshold segmentation, resulting in deep noise confounding or lymphatic vessel structure breakage, leading to discontinuous lymphatic imaging. Therefore, this invention also introduces a Frangi filter subunit 104 in the intra-frame decorrelation imaging unit 10. Please refer to... Figure 4 and Figure 5 The Frangi filter subunit 104 is used to perform step S104: process the lymphatic imaging map with the Frangi filter algorithm to obtain a continuous lymphatic imaging map.

[0049] The Frangi filtering algorithm analyzes the eigenvalues ​​of the local space of an image by calculating the second derivative matrix (Hessian matrix), enhancing structures with linear continuity while removing unstructured random noise points from the background. Through this process, the algorithm transforms areas that would otherwise be considered "noise" or "voids" in a structural image into functionally meaningful imaging signals. This effectively solves the problem that traditional OCT cannot resolve "dark" blood vessels in strong tissue backgrounds, achieving label-free imaging of lymphatic vessels and decoupling the physical mechanisms of blood vessels and lymphatic vessels into imaging.

[0050] Inter-frame decorrelation imaging unit 20 performs inter-frame decorrelation imaging, including a second decorrelation matrix calculation subunit 201 for performing step S201: calculating the second intensity decorrelation coefficient between adjacent time frames based on the intensity values ​​of the interference spectra at the same lateral position within adjacent time frames, and obtaining the second decorrelation matrix.

[0051] The core of open-circuit angiography (OCTA) lies in extracting time-series signal changes caused by red blood cell flow. Vascular imaging requires calculating the changes in adjacent time frames (the first and second frames) at the same lateral position. Frame and the This is achieved by intensity decorrelation between frames of signals, and each second intensity decorrelation coefficient in the second decorrelation matrix is ​​calculated by constructing the following formula:

[0052] in, for Time of the first The intensity value of each pixel. for +1 moment Intensity value of each pixel This represents the system noise term. In actual live imaging, due to the unavoidable overall displacement of the sample, this volumetric motion causes global intensity fluctuations.

[0053] The temporal decorrelation image imaging subunit 202 is used to perform step S202: after performing binarization masking processing based on the second decorrelation matrix, it performs temporal spectrum processing with the corresponding interference spectral data to obtain a temporal decorrelation image.

[0054] In the static region, without considering particle displacement, the intensity signal between two consecutive frames remains essentially constant. At this point, the decorrelation coefficient is mainly limited by the system noise level. Therefore, it can be seen that in static regions with strong intensity signals, It approaches 0; however, when the signal is weak, It will increase as the signal-to-noise ratio decreases. This noise-induced decorrelation signal produces artifacts similar to blood flow, severely interfering with the identification of blood vessels.

[0055] In the blood flow region, the rapid movement of erythrocytes causes dynamic evolution of the speckle field, with significant differences in intensity signals between consecutive frames. .

[0056] at this time ,lead to Because the signal intensity in the blood flow region is much higher than the noise background, its decorrelation value can reflect the dynamic characteristics of the fluid, resulting in high detection sensitivity. Although It can effectively identify blood flow, but in lymphatic areas, due to its extremely low signal intensity, Similarly, the time decorrelation value will approach 1 due to noise dominance. This means that traditional time decorrelation algorithms cannot directly distinguish between blood vessels and lymph nodes. Therefore, the obtained time decorrelation images contain mixed information of blood flow, lymphatic vessels, and system noise, which are difficult to separate through time decorrelation.

[0057] The differential imaging unit 301 is used to perform step S301: perform adaptive weight masking processing on the temporal decorrelation image based on the lymphocyte imaging image to obtain a vascular imaging image.

[0058] analyze and In (1) static organization region: Because this region has a stable backscattering intensity and no particle displacement, its signal power is much greater than the noise variance. The physical response maintains a high degree of statistical correlation across both adjacent time frames and the subspectral dimension within the same frame, and satisfies the following:

[0059] This indicates that the static background exhibits extremely low numerical characteristics in both decorrelational dimensions.

[0060] (2) Vascular region: The rapid flow of red blood cells causes drastic evolution of the speckle field between adjacent time frames. However, due to the high scattering properties of red blood cells, the power of the odd and even subset signals extracted within a single frame is still sufficient to suppress random noise. Therefore, the vascular region exhibits significant temporal decorrelation and spatial subspectral correlation characteristics, and its physical response satisfies: .

[0061] This unique characteristic of being "temporally uncorrelated but spatially correlated" constitutes the sole physical criterion for separating vascular signals from a complex background.

[0062] (3) Lymphatic region: Due to the extremely weak scattering within the lumen ( The intensity signal in this region is dominated by independent random noise terms in both the temporal and spatial dimensions. Based on the noise independence characteristic described in Section 3.2.1, the lymphatic region exhibits a completely uncorrelated physical response across all dimensions, i.e.: ,In summary, and The significant difference exists only in the blood flow region. In static tissue and lymphatic regions, the signal responses are consistent. Therefore, a scaling factor can be set. To determine blood flow signals: When When; determined as blood flow: When When the flow is considered non-blood flow, it is determined to be non-blood flow. In the algorithm implementation presented in this paper, based on experimental experience, the following settings are used: .

[0063] Therefore, the following adaptive weight mask processing is constructed:

[0064] This two-dimensional decorrelation method can effectively preserve high-contrast blood flow signals while eliminating artifacts caused by lymphatic vessels and depth attenuation noise, thereby achieving accurate angiography of the vascular system.

[0065] Furthermore, the vascular images obtained after adaptive masking still retain isolated high-value speckle noise, which affects the continuity of vascular anatomy. Therefore, the differential imaging unit 301 of this system also includes: The spatial filtering subunit 302 is used to perform step S302 to perform spatial noise filtering on the vascular imaging image to obtain a continuous vascular imaging image. Spatial noise filtering can be median filtering or a connected component labeling algorithm.

[0066] The reconstructed imaging unit 401 is used to perform step S401: reconstructing the lymphatic imaging image and the vascular imaging image to obtain a vascular-lymphatic image of the suspected inflammatory lesion site.

[0067] Finally, the continuous imaging of blood vessels and lymph nodes is reconstructed to obtain a blood vessel-lymph node map of the suspected inflammatory lesion site.

[0068] Based on the aforementioned method and system for detecting inflammatory lesions using parity-even spectrum decorrelation OCT, a simulated vascular imaging experiment was conducted in a prepared milk-agar capillary sampling model. A high-concentration milk mixture was selected as the background to simulate the high scattering characteristics of biological tissues, thereby constructing an imaging background noise close to that of real biological tissues. To simulate the strong scattering characteristics of blood, a 15V / V% milk solution diluted with water was pumped into all three capillary tubes. The results are as follows. Figure 6 As can be seen from the above: 1) Comparison of structural diagrams: exist Figure 6 In (a), since diluted milk is pumped into the pipe, its scattering coefficient is similar to that of the background (milk-agar), resulting in extremely low contrast between the fluid and the background. It is difficult to clearly distinguish the pipe wall boundary with the naked eye, especially for a 0.1 mm micro-pipe, which is basically submerged in the background noise.

[0069] 2) Differences in algorithm performance: Figure 6(b) Although the original spectrum decorrelation algorithm can extract fluid signals, the image background is not pure enough due to the influence of background Brownian motion and system noise, and the ROI signal is easily confused with the background.

[0070] Figure 6 (c) The OMAG algorithm (c) has good sensitivity to strongly scattering fluids, but it exhibits some signal broadening at microtube structures and its longitudinal resolution is slightly inferior to the decorrelation algorithm.

[0071] Figure 6 (d) The spectral decorrelation algorithm (i.e., the method proposed in this study) performs best. Thanks to the reduced sensitivity to speckle noise caused by spectral processing, this result significantly improves the resolution of fine structures while maintaining a high signal-to-noise ratio. In the figure, the three pipes with diameters of 0.1 mm, 0.3 mm, and 0.5 mm have continuous morphology and clear boundaries, and the background is effectively suppressed.

[0072] Experimental Conclusions: Experimental results demonstrate that, under strong scattering backgrounds, the algorithm proposed in this study exhibits a higher signal-to-noise ratio and better microvascular resolution compared to traditional algorithms in angiography. It can effectively extract simulated vessel signals with diameters as low as 0.1 mm, and the vessel wall boundaries are sharp. This confirms that the algorithm possesses excellent lateral resolution and sensitivity, enabling non-invasive and angiography-free detection of inflammatory lesions.

[0073] In summary, this invention proposes a method and system for detecting inflammatory lesions based on odd-even spectral decorrelation OCT. It utilizes the positive complementary characteristics of odd and even sampling points in the frequency domain of the interference spectrum for signal processing. By solving the decorrelation features between operator spectra, it can effectively extract dynamic scattering signals within the tissue, thereby enhancing the contrast of functional information while maintaining the system's axial resolution at the theoretical limit. Furthermore, by combining intra-frame and inter-frame decorrelation characteristics in static, vascular, and lymphatic regions, it overcomes the interference of highly scattering edema fluid in tissues without the need for exogenous contrast agents, accurately capturing the dilated morphology of weakly scattering lymphatic vessels, and achieving decoupling and independent imaging of blood vessels and lymphatic vessels. Moreover, the intensity-based decorrelation algorithm of this invention is not limited by the Nyquist sampling theorem regarding phase detection speed, maintaining imaging accuracy and integrity even when the flow rate exceeds 1.5 times the system's Doppler detection limit. Combined with large field-of-view stitching technology, this invention can cover the complete skin vascular architecture, providing an objective and accurate non-destructive detection method for vascular remodeling, drug efficacy evaluation, and early diagnosis of inflammatory lesions.

[0074] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and the present invention also intends to include these modifications and variations.

Claims

1. A method for detecting inflammatory lesions based on odd-even spectral decorrelation OCT, characterized by: ... (the method is described in the original text, but the provided text is incomplete and cannot be accurately translated.) include: Intra-frame decorrelation imaging, inter-frame decorrelation imaging, differential imaging, and reconstructed imaging; Intra-frame decorrelation imaging performs the following steps: S101: Odd-even spectrum extraction is performed on the interference spectral data of each frame to obtain the odd-numbered interference spectrum and even-numbered interference spectrum corresponding to each frame; S102: Based on the intensity values ​​of odd-numbered and even-numbered interference spectra in the same spatial neighborhood within the same frame, calculate the first intensity decorrelation coefficient of the sub-spectrum of the same frame to obtain the first decorrelation matrix; S103: After performing binarization masking based on the first decorrelation matrix, the data is processed with the corresponding interference spectral data to obtain a lymphocyte imaging map; Inter-frame decorrelation imaging performs the following steps: S201: Based on the intensity values ​​of the interference spectra in adjacent time frames at the same lateral position, calculate the second intensity decorrelation coefficient between adjacent time frames to obtain the second decorrelation matrix; S202: After performing binarization masking based on the second decorrelation matrix, time-dispersive processing is performed on the corresponding interference spectral data to obtain a time-decorrelation image; Differential operational imaging performs the following steps: S301: Adaptive weight masking is applied to the temporal decorrelation image based on the lymphocyte imaging image to obtain a vascular imaging image; The reconstructed imaging process performs the following steps: S401: The lymphatic imaging image and the vascular imaging image are reconstructed to obtain a vascular-lymphatic image of the suspected inflammatory lesion site.

2. The method for detecting inflammatory lesions based on odd-even spectrum decorrelation OCT according to claim 1, characterized in that, The method further includes the following after step S103: S104: The lymphocyte imaging map is processed by the Frangi filtering algorithm to obtain a continuous lymphocyte imaging map.

3. The method for detecting inflammatory lesions based on odd-even spectrum decorrelation OCT according to claim 2, characterized in that, Step S102 calculates each first intensity decorrelation coefficient in the first decorrelation matrix using the following formula: in, The average window size in space. For odd-numbered interference spectra within the window The intensity value of each pixel. For even-numbered interference spectra within the window, the first... The intensity value of each pixel.

4. The method for detecting inflammatory lesions based on parity-even spectrum decorrelation OCT according to claim 3, characterized in that, Step S201 calculates each second intensity decorrelation coefficient in the second decorrelation matrix using the following formula: in, for Time of the first The intensity value of each pixel. for +1 moment Intensity value of each pixel This represents the system noise term.

5. The method for detecting inflammatory lesions based on parity-even spectrum decorrelation OCT according to claim 4, characterized in that, The adaptive weight masking process is completed using the following formula: in, for t The temporal decorrelation image after adaptive weight masking at time step 1 z The intensity value of each pixel. To suppress the weighting factor.

6. The method for detecting inflammatory lesions based on parity-even spectrum decorrelation OCT according to claim 5, characterized in that, After obtaining the vascular imaging image in step S301, the following steps are also included: S302: Perform spatial noise filtering on the vascular imaging image to obtain a continuous vascular imaging image.

7. The method for detecting inflammatory lesions based on parity-even spectrum decorrelation OCT according to claim 6, characterized in that, The spatial noise filtering algorithm is either median filtering or connected component labeling algorithm.

8. An inflammatory lesion detection system based on odd-even spectrum decorrelation OCT, characterized in that, include: Intra-frame decorrelation imaging unit, inter-frame decorrelation imaging unit, differential operation imaging unit, and reconstructed imaging unit; The intra-frame decorrelation imaging unit includes: an odd-even spectrum subunit: used to extract odd-even spectra from the interference spectral data of each frame to obtain the odd-numbered interference spectrum and even-numbered interference spectrum corresponding to each frame; The first decorrelation matrix calculation subunit is used to calculate the first intensity decorrelation coefficient of the sub-spectrum of the same frame based on the intensity values ​​of the odd-numbered and even-numbered interference spectra in the same spatial neighborhood, and to obtain the first decorrelation matrix. Lymphatic imaging subunit: used to perform binarization masking based on the first decorrelation matrix and then process it with the corresponding interference spectral data to obtain a lymphatic imaging map; The inter-frame decorrelation imaging unit includes: a second decorrelation matrix calculation subunit: used to calculate the second intensity decorrelation coefficient between adjacent time frames based on the intensity values ​​of the interference spectrum at the same lateral position in adjacent time frames, and obtain the second decorrelation matrix; Temporal decorrelation image imaging subunit: used to perform time-spectral separation processing on the corresponding interference spectral data after performing binarization mask processing based on the second decorrelation matrix, to obtain a temporal decorrelation image; Differential operational imaging unit: used to perform adaptive weight masking processing on the temporal decorrelation image based on the lymphocyte imaging map to obtain a vascular imaging map; Reconstruction imaging unit: used to reconstruct the lymphatic imaging image and the vascular imaging image to obtain a vascular-lymphatic image of the suspected inflammatory lesion site.

9. The inflammatory lesion detection system based on odd-even spectrum decorrelation OCT according to claim 8, characterized in that, The intra-frame decorrelation imaging unit further includes: Frangi filtering subunit: used to process the lymphatic imaging map using the Frangi filtering algorithm to obtain a continuous lymphatic imaging map.

10. The inflammatory lesion detection system based on odd-even spectrum decorrelation OCT according to claim 9, characterized in that, The differential imaging unit further includes: Spatial filtering subunit: used to perform spatial noise filtering on the vascular imaging image to obtain a continuous vascular imaging image.

Images