Respiratory tract reconstruction method and device in respiratory dynamic state

Abstract

The invention relates to the technical field of biomedical imaging and image processing, in particular to a respiratory tract reconstruction method and device.The method comprises the steps that respiratory tract cross section images of different axial positions of a respiratory tract and exceeding one respiratory cycle are collected, lumen segmentation is conducted on the basis of the respiratory tract cross section images, and the respiratory tract cross section images are obtained; and extracting image data of the cross section of the respiratory tract at each axial position at the end of expiration and the end of inspiration by utilizing the segmented contour area or inner diameter change information based on the lumen structure so as to obtain data with consistent respiratory dynamic states, and performing respiratory tract reconstruction according to the data with consistent respiratory dynamic states. And obtaining a respiratory tract model with consistent breathing states. Therefore, the problems that in the related technology, the respiratory dynamic influence is not considered, the shape and the inner diameter of the respiratory tract lumen are changed due to the change of the respiratory state of the patient within the pullback collection time, so that the reconstruction model is inaccurate, and diagnosis of doctors is influenced are solved.

Description

Technical Field

[0001] The present application relates to the field of biomedical imaging and image processing technology, and in particular to a method and device for airway reconstruction in dynamic respiratory conditions. Background Art

[0002] Among the related technologies, endoscopic OCT (Optical Coherence Tomography) airway reconstruction is based on the principle of low-coherence interference. It does not have the ionizing radiation of CT (Computed Tomography) and the metal foreign body risk of MRI (Magnetic Resonance Imaging). It does not require a special equipment room and can provide high-resolution tissue depth information and real-time airway imaging.

[0003] However, in the related technology, the endoscopic OCT respiratory reconstruction method does not take into account the dynamic impact of breathing. Due to the changes in the patient's respiratory state during the pullback acquisition time, the shape and inner diameter of the respiratory tract lumen change, making the reconstructed model inaccurate, which in turn affects the doctor's diagnosis and needs to be improved urgently. Summary of the invention

[0004] The present application provides a method and device for airway reconstruction in respiratory dynamics to solve the problem that the related technology does not take into account the impact of respiratory dynamics, and the changes in the patient's respiratory state during the pullback acquisition time lead to changes in the shape and inner diameter of the airway lumen, making the reconstructed model inaccurate, which in turn affects the doctor's diagnosis.

[0005] A first aspect of the present application provides a method for reconstructing the respiratory tract in respiratory dynamics, comprising the following steps: acquiring respiratory tract cross-sectional images at different axial positions of the respiratory tract and exceeding a preset respiratory cycle; performing lumen segmentation based on the respiratory tract cross-sectional images to obtain the lumen structure of the respiratory tract; extracting the respiratory state based on the lumen structure to obtain a cross-sectional image consistent with the respiratory dynamics; and reconstructing the respiratory tract based on the cross-sectional image consistent with the respiratory dynamics to obtain a respiratory tract model consistent with the respiratory state.

[0006] Through the above technical scheme, the embodiment of the present application can comprehensively capture the dynamic change information of the respiratory tract during the breathing process by collecting data covering different axial positions and more than one respiratory cycle, thereby providing a rich data basis for accurate reconstruction; the lumen structure is obtained through lumen segmentation, which is helpful to accurately analyze the internal morphology of the respiratory tract; the end-expiratory and end-inspiratory data are extracted through contour area or inner diameter changes to obtain respiratory dynamic consistent data, effectively eliminating the interference of inconsistent respiratory state on reconstruction and improving model accuracy; reconstruction is performed through respiratory dynamic consistent data, and finally a respiratory tract model with consistent respiratory state is obtained, which can provide a more reliable basis for the accurate diagnosis of respiratory diseases, and effectively assist clinicians in accurately judging and effectively treating patients' conditions.

[0007] Optionally, in one embodiment of the present application, the acquisition of respiratory cross-sectional images at different axial positions of the respiratory tract and exceeding a preset breathing cycle includes: during the rotation of the light beam, at the same or similar axial position, controlling the light beam to rotate multiple times, so that when the light beam rotates to the same angle, it irradiates the same or similar position on the surface of the lumen, so as to acquire the respiratory cross-sectional images at the same or similar axial position under different breathing states.

[0008] Through the above technical solution, the embodiment of the present application can control the light beam to rotate multiple times at the same or similar axial position during the light beam rotation process, so as to achieve the effect of irradiating the same position or a similar position on the lumen surface at different time points, thereby collecting cross-sectional images to generate respiratory cross-sectional images, and being able to obtain data at multiple time points at the same position or a similar position, which is extremely critical for capturing dynamic changes in breathing. Since the state of the respiratory tract changes with time during the breathing process, multiple acquisitions at the same position or a similar position can better record the changing pattern of the shape of the respiratory tract lumen with the respiratory cycle, obtain richer information, and provide a more comprehensive and accurate data basis for subsequent operations such as lumen segmentation and respiratory state extraction, which helps to improve the quality of the respiratory tract model with a consistent respiratory state that is finally reconstructed, so that the model can more accurately reflect the true state of the respiratory tract in respiratory dynamics, and thus provide a more reliable basis for the diagnosis of respiratory diseases.

[0009] Optionally, in one embodiment of the present application, the lumen segmentation is performed based on the respiratory tract cross-sectional image to obtain the lumen structure of the respiratory tract, including: based on a preset deep learning U-Net network model, extracting high-level features of the respiratory tract cross-sectional image by stepwise downsampling, and restoring the spatial resolution of the respiratory tract cross-sectional image by stepwise upsampling to obtain the lumen structure of the respiratory tract.

[0010] Through the above technical scheme, the embodiment of the present application can be based on the deep learning U-Net network model, extract high-level features of the image by step-by-step downsampling, effectively focus on the key information of the lumen in the image, accurately capture the deep-level features related to the respiratory lumen structure, and restore the image spatial resolution by step-by-step upsampling, so that while acquiring key features, the spatial detail information of the image is not lost, ensuring that the final respiratory lumen structure has both accurate morphological contours and retains rich details, thereby improving the accuracy and reliability of lumen segmentation.

[0011] Optionally, in one embodiment of the present application, the respiratory state is extracted based on the tubular structure to obtain a cross-sectional image consistent with the respiratory dynamics, including: using the information of the change of the contour area or inner diameter over time after segmentation to extract image data of the respiratory cross-section at each position at the end of exhalation and the end of inhalation, so as to obtain the cross-sectional image consistent with the respiratory dynamics.

[0012] Through the above technical solution, the embodiment of the present application can extract the image data of the end of exhalation and the end of inspiration through the information of the change of the contour area or inner diameter over time after segmentation, accurately capture the switching point of respiratory dynamics, make full use of the periodic law of respiratory movement, and do not rely on complex external equipment. It has strong adaptability and accuracy. The extracted image data of the end of exhalation and the end of inspiration of different respiratory cross-sections provide data consistent with respiratory dynamics, providing a more comprehensive and accurate data basis for subsequent operations such as rotation correction and respiratory reconstruction, ensuring the consistency of respiratory dynamics of data at different axial positions of the respiratory reconstruction model, and laying a solid foundation for finally obtaining a high-quality, consistent respiratory state and structurally accurate respiratory model, which helps to improve the reliability and accuracy of respiratory disease diagnosis, thereby improving the quality and accuracy of the entire respiratory reconstruction model.

[0013] Optionally, in one embodiment of the present application, the method according to claim 1 is characterized in that the respiratory reconstruction based on the cross-sectional image consistent with respiratory dynamics includes: converting the cross-sectional image consistent with respiratory dynamics into an unfolded image with a coordinate as an angle and a coordinate as a distance from a pixel to a center; based on the unfolded image, each image uses the unfolded image consistent with respiratory dynamics at a previous position as a template to perform sliding window matching on the image translation expansion image; based on the sliding window matching, the similarity between the unfolded images at different rotation angles or the distance error of the contours are obtained to determine the corrected error angle; the cross-sectional image consistent with respiratory dynamics is rotationally corrected according to the corrected error angle, and then a three-dimensional respiratory model is reconstructed.

[0014] Through the above technical scheme, the embodiment of the present application can convert the respiratory dynamics consistent cross-sectional image into an expanded image with a coordinate as an angle and a coordinate as a distance from a pixel to a center. The translation of the expanded image can be regarded as a rotation of the cross-sectional image, and the complex rotation calculation is converted into a translation calculation; based on the expanded image, each image uses the expanded image consistent with the respiratory dynamics at the previous position as a template to perform sliding window matching on the image translation expansion map, and integrates the rotation information into the translation of the two-dimensional image space; based on sliding matching, the correction error angle is determined based on the similarity between images at different rotation angles or the distance error of the contour, which makes full use of the characteristic information of the image itself, accurately quantifies the rotation deviation, does not rely on complex external equipment or additional marking information, and has strong adaptability and accuracy; the respiratory dynamics consistent data is rotationally corrected according to the determined correction error angle, effectively solving the image rotation angle error problem caused by the hardware limitations of endoscopic OCT imaging, ensuring the consistency of the rotation angle between each axial position data, and improving the accuracy of each axial position data.

[0015] A second aspect of the present application provides a device for reconstructing the respiratory tract in respiratory dynamics, including: an acquisition module for acquiring respiratory tract cross-sectional images at different axial positions of the respiratory tract and exceeding a preset respiratory cycle; a segmentation module for performing lumen segmentation based on the respiratory tract cross-sectional images to obtain the lumen structure of the respiratory tract; an extraction module for extracting the respiratory state based on the lumen structure to obtain a cross-sectional image consistent with the respiratory dynamics; and a reconstruction module for performing respiratory tract reconstruction based on the cross-sectional image consistent with the respiratory dynamics to obtain a respiratory tract model consistent with the respiratory state.

[0016] Through the above technical scheme, the embodiment of the present application can comprehensively capture the dynamic change information of the respiratory tract during the breathing process by collecting data covering different axial positions and more than one respiratory cycle, thereby providing a rich data basis for accurate reconstruction; the lumen structure is obtained by lumen segmentation, which is helpful to accurately analyze the internal morphology of the respiratory tract; the end-expiratory and end-inspiratory data are extracted by contour area or inner diameter changes to obtain dynamic consistent respiratory data, effectively eliminating the interference of inconsistent respiratory state on reconstruction and improving model accuracy; by reconstructing the respiratory tract with dynamic consistent respiratory data, a respiratory tract model with consistent respiratory state is finally obtained, which can provide a more reliable basis for the accurate diagnosis of respiratory diseases, and effectively assist clinicians in accurately judging and effectively treating patients' conditions.

[0017] Optionally, in one embodiment of the present application, the acquisition module includes: during the rotation of the light beam, at the same or similar axial position, controlling the light beam to rotate multiple times, so that when the light beam rotates to the same angle, it irradiates the same position or a similar position on the surface of the lumen, so as to acquire the respiratory cross-sectional image at the same or similar axial position under different respiratory states.

[0018] Through the above technical solution, the embodiment of the present application can control the light beam to rotate multiple times at the same or similar axial position during the light beam rotation process, so as to achieve the effect of irradiating the same position or a similar position on the lumen surface at different time points, thereby collecting cross-sectional images to generate respiratory cross-sectional images, and being able to obtain data at multiple time points at the same position or a similar position, which is extremely critical for capturing dynamic changes in breathing. Since the state of the respiratory tract changes with time during the breathing process, multiple acquisitions at the same position or a similar position can better record the changing pattern of the shape of the respiratory tract lumen with the respiratory cycle, obtain richer information, and provide a more comprehensive and accurate data basis for subsequent operations such as lumen segmentation and respiratory state extraction, which helps to improve the quality of the respiratory tract model with a consistent respiratory state that is finally reconstructed, so that the model can more accurately reflect the true state of the respiratory tract in respiratory dynamics, and thus provide a more reliable basis for the diagnosis of respiratory diseases.

[0019] Optionally, in one embodiment of the present application, the segmentation module includes: an image processing unit, which is used to extract high-level features of the respiratory tract cross-sectional image by step-by-step downsampling based on a preset deep learning U-Net network model, and to restore the spatial resolution of the respiratory tract cross-sectional image by step-by-step upsampling to obtain the tubular structure of the respiratory tract.

[0020] Through the above technical scheme, the embodiment of the present application can be based on the deep learning U-Net network model, extract high-level features of the image by step-by-step downsampling, effectively focus on the key information of the lumen in the image, accurately capture the deep-level features related to the respiratory lumen structure, and restore the image spatial resolution by step-by-step upsampling, so that while acquiring key features, the spatial detail information of the image is not lost, ensuring that the final respiratory lumen structure has both accurate morphological contours and retains rich details, thereby improving the accuracy and reliability of lumen segmentation.

[0021] Optionally, in one embodiment of the present application, the extraction module includes: using the information of the change of the contour area or inner diameter over time after segmentation to extract the image data of the respiratory cross-section at each position at the end of exhalation and the end of inspiration, so as to obtain a cross-sectional image consistent with the respiratory dynamics.

[0022] Through the above technical solution, the embodiment of the present application can extract the image data of the end of exhalation and the end of inspiration through the information of the change of the contour area or inner diameter over time after segmentation, accurately capture the switching point of respiratory dynamics, make full use of the periodic law of respiratory movement, and do not rely on complex external equipment. It has strong adaptability and accuracy. The extracted image data of the end of exhalation and the end of inspiration of different respiratory cross-sections provide data consistent with respiratory dynamics, providing a more comprehensive and accurate data basis for subsequent operations such as rotation correction and respiratory reconstruction, ensuring the consistency of respiratory dynamics of data at different axial positions of the respiratory reconstruction model, and laying a solid foundation for finally obtaining a high-quality, consistent respiratory state and structurally accurate respiratory model, which helps to improve the reliability and accuracy of respiratory disease diagnosis, thereby improving the quality and accuracy of the entire respiratory reconstruction model.

[0023] Optionally, in one embodiment of the present application, the reconstruction module includes: a conversion unit, used to convert the cross-sectional image consistent with respiratory dynamics into an expanded image with a coordinate as an angle and a coordinate as a distance from a pixel to a center; a matching unit, used to perform sliding window matching on the image translation expansion image based on the expanded image, with each image using a previous expanded image consistent with respiratory dynamics as a template; a determination unit, used to obtain the similarity between the expanded images at different rotation angles or the distance error of the contours based on the sliding window matching, so as to determine the corrected error angle; and a correction unit, used to perform rotation correction on the cross-sectional image consistent with respiratory dynamics according to the corrected error angle.

[0024] Through the above technical scheme, the embodiment of the present application can convert the respiratory dynamics consistent cross-sectional image into an expanded image with a coordinate as an angle and a coordinate as a distance from a pixel to a center. The translation of the expanded image can be regarded as a rotation of the cross-sectional image, and the complex rotation calculation is converted into a translation calculation; based on the expanded image, each image uses the expanded image consistent with the respiratory dynamics at the previous position as a template to perform sliding window matching on the image translation expansion map, and integrates the rotation information into the translation of the two-dimensional image space; based on sliding matching, the correction error angle is determined based on the similarity between images at different rotation angles or the distance error of the contour, which makes full use of the characteristic information of the image itself, accurately quantifies the rotation deviation, does not rely on complex external equipment or additional marking information, and has strong adaptability and accuracy; the respiratory dynamics consistent data is rotationally corrected according to the determined correction error angle, effectively solving the image rotation angle error problem caused by the hardware limitations of endoscopic OCT imaging, ensuring the consistency of the rotation angle between each axial position data, and improving the accuracy of each axial position data.

[0025] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for airway reconstruction in respiratory dynamics as described in the above embodiment.

[0026] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above method for airway reconstruction in respiratory dynamics.

[0027] The fifth aspect of the present application provides a computer program product, including a computer program, which, when executed, is used to implement the above method for airway reconstruction in respiratory dynamics.

[0028] The embodiments of the present application can collect data covering multiple axial positions and more than one respiratory cycle, and use lumen segmentation, extract consistent respiratory dynamics data based on contour features, and perform rotation correction and reconstruction to comprehensively capture the dynamic changes of respiratory breathing in the respiratory tract, optimize the model quality, and provide a reliable basis for the diagnosis of respiratory diseases; during data collection, the light beam is rotated multiple times at the same or similar axial position, and the same position or similar positions are collected multiple times to obtain rich information, thereby improving the model's ability to reflect the true state of the respiratory tract; based on the preset deep learning network model, downsampling is used to extract features and upsampling is used to restore spatial resolution to improve the accuracy of lumen segmentation; respiratory states are extracted based on the segmented lumen structure to obtain data consistent with respiratory dynamics, thereby improving the state consistency of the respiratory reconstruction model; the correction error angle is determined by image features, the respiratory dynamics data is converted into an expanded image, and the error angle is calculated using sliding window matching to solve the rotation problem caused by hardware limitations, ensure the consistency and accuracy of the model rotation angle, and improve the overall quality and diagnostic reliability of the respiratory reconstruction model.

[0029] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0031] Figure 1 A flowchart of a method for reconstructing the respiratory tract in a dynamic respiratory state provided according to an embodiment of the present application;

[0032] Figure 2 The structure and optical path diagram of a swept source OCT system according to an embodiment of the present application;

[0033] Figure 3A schematic diagram of the optical path of a spectral domain OCT system according to an embodiment of the present application;

[0034] Figure 4 A schematic diagram of the structure and working principle of a proximal rotating scanning endoscope probe according to an embodiment of the present application;

[0035] Figure 5 A schematic diagram of the structure and working principle of a remote rotating scanning endoscope probe according to an embodiment of the present application;

[0036] Figure 6 This is a schematic diagram of the principle of repeated B-scanning at the same axial position according to an embodiment of the present application;

[0037] Figure 7 A schematic diagram of a multiple acquisition mode at the same position according to an embodiment of the present application;

[0038] Figure 8 A schematic diagram of the axial dense B-scanning principle according to an embodiment of the present application;

[0039] Fig. 9 A schematic diagram of a U-Net network structure according to an embodiment of the present application;

[0040] Fig.10 A time-area curve diagram of luminal respiration state extraction according to one embodiment of the present application;

[0041] Fig.11 A schematic diagram of a process of collecting images according to an embodiment of the present application;

[0042] Fig.12 This is a schematic diagram of the principle of rotation correction based on lumen contour according to one embodiment of the present application;

[0043] Fig.13 The figure is a comparison diagram of the airway reconstruction model without and with respiratory dynamic correction according to one embodiment of the present application.

[0044] Fig.14 It is a structural schematic diagram of a respiratory dynamic airway reconstruction device provided according to an embodiment of the present application;

[0045] Fig.15 The figure is a structural example diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0046] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0047] The following is a description of the method and device for reconstructing the respiratory tract in the respiratory dynamics of the embodiment of the present application with reference to the accompanying drawings. In view of the related technologies mentioned in the above background technology center, the dynamic influence of respiratory movement is not considered. Due to the change of the patient's respiratory state during the pullback acquisition time, the shape and inner diameter of the respiratory tract lumen change, which makes the reconstruction model inaccurate, thereby affecting the doctor's diagnosis and other problems. The present application provides a method for reconstructing the respiratory tract in the respiratory dynamics, in which the dynamic change information of the respiratory tract during the respiratory process can be fully captured by collecting data covering different axial positions and more than one respiratory cycle, providing a rich data basis for accurate reconstruction; the lumen structure is obtained by lumen segmentation, which helps to accurately analyze the internal morphology of the respiratory tract; the end-expiratory and end-inspiratory data are extracted by contour area or inner diameter changes to obtain respiratory dynamic consistent data, effectively eliminating the interference of inconsistent respiratory state on reconstruction, and improving the accuracy of the model; by reconstructing the respiratory tract on the respiratory dynamic consistent data, a respiratory tract model with consistent respiratory state is finally obtained, which can provide a more reliable basis for the accurate diagnosis of respiratory diseases, and effectively assist clinicians in making accurate judgments and effective treatments on the patient's condition. This solves the problem in related technologies that the dynamic effects of breathing are not taken into account, and that changes in the patient's breathing state during the pullback acquisition time lead to changes in the shape and inner diameter of the respiratory lumen, making the reconstructed model inaccurate, which in turn affects the doctor's diagnosis.

[0048] Specifically, Figure 1 A flowchart of a method for reconstructing the respiratory tract during dynamic breathing provided in an embodiment of the present application.

[0049] like Figure 1 As shown, the method for airway reconstruction in respiratory dynamics includes the following steps:

[0050] In step S101 , respiratory tract cross-sectional images at different axial positions of the respiratory tract and exceeding a preset respiratory cycle are acquired.

[0051] Those skilled in the art should understand that OCT is a three-dimensional tomography method based on the principle of low-coherence interference, which has the advantages of no labeling, high resolution and large field of view, and can be used for two-dimensional cross-sectional and three-dimensional volume imaging of tissues. Endoscopic OCT realizes real-time imaging of biological cavities through the combination of endoscopes and OCT systems, thereby imaging the respiratory tract. The guide rail controlled by the motor pulls the endoscope probe to slowly translate in the lumen, and the light beam emitted from the probe rotates at a certain rate, thereby realizing the scanning of the lumen. When the light beam scans a position on the surface of the lumen, the OCT system immediately obtains an interference signal that varies with the wavelength. After transforming the interference signal, a sequence of complex numbers that varies with depth is obtained, reflecting the scattering information of different tissue depths. By calculating the intensity of this set of complex sequences, images of tissue structures at different depths below this surface position can be obtained. In endoscopic OCT sampling, A-line scanning can obtain one-dimensional information along the depth direction (that is, the beam direction and radial direction), and B-scan consists of multiple A-lines obtained after rotating the beam 360 degrees, which can obtain a two-dimensional cross-sectional image. The C-scan is constructed from multiple B-scan cross-sectional images at different axial positions to obtain a three-dimensional image.

[0052] Optionally, currently commonly used OCT systems are divided into swept source optical coherence tomography and spectral domain optical coherence tomography.

[0053] Specifically, the principle of swept source optical coherence tomography is as follows Figure 2 As shown, the OCT system includes a swept-frequency light source, a coupler, a circulator, a reference optical path, a sample optical path, and a photodetector. Coupler 1 realizes the splitting effect, and can divide the light beam into two beams of a certain proportion as required, one beam enters the reference optical path, and the other beam enters the sample optical path. After the light beam of the reference optical path enters port 1 of circulator 1, it is emitted from port 2 of circulator 1, and after being reflected by the reflector, it returns from port 2 of circulator 1, and is emitted from port 3 of circulator 1 and enters coupler 2. After the light beam of the sample optical path enters port 1 of circulator 2, it is emitted from port 2 of circulator 2, enters the endoscope probe, and irradiates the tissue sample. The light scattered or reflected by the tissue sample is collected by the endoscope probe, returned to port 2 of circulator 2, and is emitted from port 3 of circulator 2 and enters coupler 2. The light beams from the reference optical path and the sample optical path interfere in coupler 2, and the interference light is detected by the photodetector. The swept-frequency light source outputs a single-wavelength (or narrow-band) laser at each moment, and the photodetector can detect the intensity of interference light of different wavelengths. After obtaining the interference signal that varies with wavelength, the interference signal is transformed to obtain a sequence of complex numbers that varies with depth, reflecting the scattering information at different tissue depths.

[0054] The spectral domain OCT imaging system is based on a continuous wide spectrum light source. The laser output by the light source passes through the isolator and is split by the coupler. Part of the light enters the sample light path, and part of the light enters the reference light path. Figure 3 As shown. The light from the reference optical path is reflected by the reflector and returns to the coupler. The light entering the sample optical path enters the endoscope probe and irradiates the tissue sample. The light scattered or reflected by the tissue sample is collected by the endoscope probe and returns to the coupler. The light beams returned from the sample optical path and the reference optical path interfere in the coupler. After the interference signal passes through the lens, it is split by the grating to separate the interference light of different wavelengths in space. After passing through the lens, it is photoelectrically converted on the camera. After obtaining the interference signal that varies with wavelength, the interference signal is transformed to obtain a sequence of complex numbers that varies with depth, reflecting the scattering information at different tissue depths.

[0055] Furthermore, the light beam emitted by the endoscope probe irradiates the lumen wall radially. In the lumen, the light beam is rotated and dragged to achieve circular scanning and spiral scanning of the light beam, thereby obtaining a two-dimensional cross-sectional image and a three-dimensional stereoscopic image of the tissue. In order to achieve the translation of the light beam, a one-dimensional translation rail can be used to drag the optical fiber and the endoscope probe. In order to achieve the rotation of the light beam, two methods can be adopted: proximal rotation scanning with the overall rotation of the probe and distal rotation scanning with the rotation of the reflector inside the probe.

[0056] In some embodiments, the structure of the proximal rotation scanning method in which the probe rotates as a whole is as follows: Figure 4 As shown, after the light beam from the light source enters the sample optical path, it enters from the stator end of the fiber slip ring and is output from the rotor end of the fiber slip ring. After passing through the converging lens in the probe, the light beam is reflected by the reflector in the probe and irradiates the wall of the lumen. The optical fiber at the stator end does not rotate, while the optical fiber at the rotor end is driven by a motor and a belt to rotate. The optical fiber at the rotor end is connected to the endoscope probe module, driving the endoscope probe to rotate together, thereby realizing rotational scanning. The motor and the fiber slip ring are fixed on the electric translation slide rail, and the slide rail drags the motor, the fiber slip ring, and the endoscope probe module to perform linear motion.

[0057] In some other embodiments, the structure of the distal end rotation scanning method of the rotation of the reflector in the probe is as follows: Figure 5 As shown in the figure, after the light beam from the light source enters the sample optical path, it passes through the converging lens in the endoscope probe, is reflected by the reflector in the probe, and irradiates the lumen wall. The rotating micromotor in the probe drives the reflector in the probe to rotate, thereby changing the exit angle of the light beam and realizing the rotation scanning of the light beam. During this process, the probe as a whole does not rotate, only the reflector in the probe rotates. The electric translation slide drags the translation of the endoscope probe through the optical fiber.

[0058] In the actual implementation process, during the rotation of the light beam, the light beam is controlled to rotate multiple times at the same or similar axial position, so that when the light beam rotates to the same angle, it irradiates the same or similar position on the surface of the lumen, so as to collect respiratory cross-sectional images at the same or similar axial position under different respiratory states.

[0059] Specifically, Figure 6 As shown, in order to make the light beam illuminate the same position on the lumen surface at different time points during the rotation process, the light beam can be rotated N times (at different times) at the same axial position during the rotation process. When the light beam rotates to the same angle, it illuminates the same position on the lumen surface, achieving N sampling times with different times and the same position (the same position means the same axial position and the same scanning angle). This method can be called repeated B scanning at the same axial position.

[0060] Furthermore, the image acquisition mode can use a mode of multiple acquisitions at the same position, such as Figure 7 As shown. The probe is controlled by a rotating motor to keep rotating at a certain speed to obtain a B-scan, that is, a cross-sectional two-dimensional image. When the guide rail controlled by the motor pulls the endoscope probe to translate in the lumen, the probe stops at each axial position at equal intervals and obtains a cross-sectional image of the entire respiratory cycle at each position. This dwell time is determined by the respiratory cycle time of the specific organism to be tested. For example, a rabbit usually has a respiratory cycle time of 2 seconds.

[0061] Furthermore, the image acquisition mode can also use a dense sampling mode. In order to make the light beam irradiate approximately the same position on the lumen surface at different time points during the rotation process, the axial position can be changed during the rotation of the light beam to form a spiral scan. During the rotation of the light beam, the axial position is changed, and the light beam rotates N circles (at different times). When the light beam rotates to the same angle, the movement distance of the light beam center of two adjacent circles is less than or approximately equal to the beam diameter, so that N sampling times are different and positions are approximately the same (similar positions mean similar axial positions and the same scanning angles). This method can be called axial dense B scanning. Figure 8 As shown in the figure, when dense sampling occurs, the distance between adjacent positions is small, and its effect is similar to the mode of multiple sampling at the same position. However, in the case of long-distance sampling, dense sampling will have the problem of too much data.

[0062] The embodiments of the present application can realize two-dimensional cross-sectional and three-dimensional volumetric imaging of tissues by adopting OCT, a three-dimensional tomographic imaging method based on the principle of low-coherence interference; two-dimensional and three-dimensional images can be obtained by different scanning methods such as circular scanning and spiral scanning, and there are multiple effective acquisition modes when collecting cross-sectional images of the respiratory tract, such as the multiple acquisition mode at the same position, which can acquire cross-sectional images by staying at each axial position according to the respiratory cycle of the biological being tested, and the dense sampling mode such as axial dense B scanning, which can make the light beam irradiate approximately the same position of the lumen at different time points; the rotational scanning of the endoscopic probe has two modes: proximal rotational scanning of the overall rotation of the probe and distal rotational scanning of the rotation of the reflector inside the probe, which enriches the means of implementation, and there is a corresponding control method when the light beam rotates to collect cross-sectional images, to ensure effective sampling at the same position of the lumen at the same angle.

[0063] In step S102, lumen segmentation is performed based on the respiratory tract cross-sectional image to obtain the lumen structure of the respiratory tract.

[0064] It is understandable that lumen segmentation algorithms are currently mainly divided into traditional methods and deep learning methods. Traditional methods are divided into two-dimensional and three-dimensional segmentation methods, among which three-dimensional segmentation methods can avoid the influence of some artifacts in two-dimensional segmentation, such as pixel oversaturation, speckle noise and mirroring, and make up for the lack of two-dimensional information through the association between three-dimensional information. Deep learning methods are better able to deal with special situations than traditional methods, such as sputum in the lumen, catheter wall adhesion and artifacts, etc.

[0065] Optionally, in one embodiment of the present application, lumen segmentation is performed based on the respiratory tract cross-section to obtain the lumen structure of the respiratory tract, including: based on a preset deep learning U-Net network model, extracting high-level features of the respiratory tract cross-sectional image by step-by-step downsampling, and restoring the spatial resolution of the respiratory tract cross-sectional image by step-by-step upsampling to obtain the lumen structure of the respiratory tract.

[0066] Specifically, Fig. 9 As shown in the figure, the U-Net network encoder includes multiple layers of convolution and pooling operations, which can extract high-level features of the image by gradually downsampling; the decoder includes multiple layers of deconvolution and upsampling operations, which can restore the spatial resolution of the image by gradually upsampling to obtain the segmentation result; the jump link can combine deep information and underlying information to help the decoder recover more detailed information.

[0067] The embodiment of the present application can perform lumen segmentation through a deep learning U-Net network model, extract high-level features through gradual downsampling, accurately grasp key lumen information, and then restore resolution through gradual upsampling to obtain a clear and accurate lumen structure, effectively avoiding artifact problems of traditional methods, giving full play to the advantages of deep learning, and laying a solid foundation for subsequent respiratory reconstruction operations, thereby improving the reliability and effectiveness of the respiratory reconstruction model.

[0068] In step S103, the respiratory state is extracted based on the lumen structure to obtain a cross-sectional image consistent with the respiratory dynamics.

[0069] It is understandable that breathing usually consists of two states, one is inhalation and the other is exhalation. When inhaling, the cavity will expand, the inner diameter will increase, and the lumen area will also increase; when exhaling, the cavity will contract, the inner diameter will decrease, and the lumen area will also decrease. Respiratory state extraction can extract the turning point of the respiratory state based on the change of lumen area or the maximum and minimum inner diameter changes. The state turning point corresponds to the end of inspiration and the end of exhalation.

[0070] In the actual implementation process, the image data of each respiratory cross section at the end of exhalation and the end of inspiration are extracted using the information of the change of the contour area or inner diameter after segmentation over time to obtain a cross-sectional image consistent with the respiratory dynamics.

[0071] Specifically, the state extraction algorithm based on lumen segmentation can be based on the lumen inner diameter or the lumen area, or a combination of both. The embodiment of the present application uses the change in lumen area to extract the respiratory state. When inhaling, the lumen will expand and the lumen area will increase; when exhaling, the lumen will contract and the lumen area will decrease. Fig.10 As shown in the figure, the lumen time-area curve is collected at the same position, and the maximum and minimum values ​​of the lumen area can be obtained. The maximum value corresponds to the state of transition from inspiration to exhalation, that is, the end of inspiration; the minimum value corresponds to the state of transition from exhalation to inspiration, that is, the end of exhalation.

[0072] The embodiment of the present application can extract the respiratory state based on the change of the lumen area. By analyzing the lumen time-area curve collected at the same position, according to the rule that the lumen area increases during inhalation and decreases during exhalation, it is accurately determined that the maximum value of the lumen area corresponds to the end of inspiration and the minimum value corresponds to the end of exhalation, thereby accurately extracting key data of the respiratory state.

[0073] In step S104, the respiratory tract is reconstructed based on the cross-sectional image consistent with the respiratory dynamics to obtain a respiratory tract model consistent with the respiratory state.

[0074] It should be understood by those skilled in the art that, due to various hardware limitations of endoscopic OCT, it is difficult to ensure a constant rotation speed of the endoscope during imaging, whether it is a proximal scanning system or a distal scanning system, so there will be rotation angle errors between images during the acquisition process, such as Fig.11 As shown in the figure, since the number of cross-sectional images between each axial position data after respiratory state extraction may be relatively large, the accumulated rotation angle error may be large, resulting in a very obvious error, so rotation correction is required. For rotation correction, the error angle to be corrected can be calculated using contour-based distance error or image similarity.

[0075] Optionally, in one embodiment of the present application, rotation correction is performed between each axial position based on the respiratory dynamics consistent data, including: converting the cross-sectional image consistent with the respiratory dynamics into an unfolded image with a coordinate as an angle and a distance from a pixel to a center; based on the unfolded image, each image uses the unfolded image consistent with the respiratory dynamics at the previous position as a template to perform sliding window matching on the image translation expansion image; based on the sliding window matching, the similarity or distance error between the unfolded images at different rotation angles is obtained to determine the corrected error angle; rotation correction is performed on the cross-sectional image consistent with the respiratory dynamics according to the corrected error angle, thereby reconstructing a three-dimensional respiratory tract model.

[0076] Specifically, the rotation correction of the rotation angle error can be based on the distance error of the lumen contour or on the similarity of different rotation angles of the image. The embodiment of the present application uses the distance error based on the lumen contour for rotation correction, and obtains the lumen contour through lumen segmentation, which lays the foundation for using the lumen contour to calculate the distance error. Because the respiratory endoscopy OCT image itself has weak texture, the contour-based rotation correction does not require a strong texture pattern. The rotation of the image can be converted into an expanded image with the horizontal coordinate as the angle and the vertical coordinate as the distance from the contour to the center, such as Fig.12 As shown. Each image (nth position) uses the unfolded contour image with consistent breathing dynamics of the previous position (n-1th position) as a template. The template will be matched with a sliding window on the image translation and expansion contour image. The matching principle is to calculate the minimum contour distance error. The horizontal coordinate of the sliding offset of the window corresponding to the minimum contour error is the rotation angle. The corresponding angle of the image rotation at the nth position can be the same as the rotation angle at the n-1th position.

[0077] Furthermore, the 3D reconstruction method is implemented in code or 3D reconstruction software. The embodiment of the present application uses a 3D reconstruction program written in C++ code based on VTK, which performs 3D reconstruction of the respiratory tract of the rabbit after the above process. Fig.13As shown, the reconstructed models without and with respiratory dynamics correction are shown. It can be found that the respiratory tract model reconstructed with respiratory correction is very smooth, without the bumps and bumps caused by respiratory dynamics.

[0078] The embodiment of the present application can improve the accuracy and smoothness of the respiratory model through a rotation correction method based on the lumen contour distance error, utilize specific unfolded images and sliding window matching to calculate the correction angle, and combine it with VTK's C++ 3D reconstruction program, thereby avoiding the impact of respiratory dynamics and providing a reliable basis for the diagnosis of respiratory diseases.

[0079] According to the method for reconstructing the respiratory tract in the dynamic state of breathing proposed in the embodiment of the present application, the dynamic change information of the respiratory tract in the breathing process can be comprehensively captured by collecting data covering each axial position and more than one respiratory cycle, thereby providing a rich data basis for accurate reconstruction; the lumen structure is obtained by lumen segmentation, which is helpful to accurately analyze the internal morphology of the respiratory tract; the end-expiratory and end-inspiratory data are extracted by contour area or inner diameter changes to obtain consistent data of respiratory dynamics, effectively eliminating the interference of inconsistent respiratory state on reconstruction and improving model accuracy; through rotation correction and reconstruction using the corrected data, the model quality is further optimized, and finally a respiratory tract model with consistent respiratory state is obtained, which can provide a more reliable basis for the accurate diagnosis of respiratory diseases, and effectively assist clinicians in accurately judging and effectively treating patients' conditions.

[0080] Next, the device for airway reconstruction in respiratory dynamics proposed in accordance with an embodiment of the present application will be described with reference to the accompanying drawings.

[0081] Fig.14 It is a block diagram of an apparatus for airway reconstruction in respiratory dynamics according to an embodiment of the present application.

[0082] like Fig.14 As shown, the apparatus 10 for reconstructing the respiratory tract in dynamic breathing includes: an acquisition module 100 , a segmentation module 200 , an extraction module 300 and a reconstruction module 400 .

[0083] Specifically, the acquisition module 100 is used to acquire respiratory cross-sectional images at different axial positions of the respiratory tract and exceeding a preset respiratory cycle.

[0084] The segmentation module 200 is used to perform lumen segmentation based on the respiratory tract cross-sectional image to obtain the lumen structure of the respiratory tract.

[0085] The extraction module 300 is used to extract the respiratory state based on the lumen structure to obtain a cross-sectional image consistent with the respiratory dynamics.

[0086] The reconstruction module 400 is used to reconstruct the respiratory tract according to the cross-sectional image with consistent respiratory dynamics, so as to obtain a respiratory tract model with consistent respiratory state.

[0087] Optionally, in one embodiment of the present application, the acquisition module 100 includes: during the rotation of the light beam, at the same or similar axial position, controlling the light beam to rotate multiple times, so that when the light beam rotates to the same angle, it irradiates the same position or a similar position on the surface of the lumen, so as to acquire respiratory cross-sectional images at the same or similar axial position under different respiratory states.

[0088] Optionally, in one embodiment of the present application, the segmentation module 200 includes: an image processing unit, which is used to extract high-level features of the respiratory tract cross-sectional image by step-by-step downsampling based on a preset deep learning U-Net network model, and to restore the spatial resolution of the respiratory tract cross-sectional image by step-by-step upsampling to obtain the tubular structure of the respiratory tract.

[0089] Optionally, in one embodiment of the present application, the extraction module 300 includes: using the information of the change of the segmented contour area or inner diameter over time to extract the image data of the respiratory cross-section at each position at the end of exhalation and the end of inspiration, so as to obtain a cross-sectional image consistent with the respiratory dynamics.

[0090] Optionally, in one embodiment of the present application, the reconstruction module 400 includes: a conversion unit, a matching unit, a determination unit and a correction unit.

[0091] The conversion unit is used to convert the cross-sectional image consistent with respiratory dynamics into an unfolded image with a coordinate being an angle and a coordinate being a distance from a pixel to a center.

[0092] The matching unit is used for performing sliding window matching on the image translation expansion map based on the expansion map, with each image taking a previous expansion map consistent with the respiratory dynamics as a template.

[0093] A determination unit is used to obtain the similarity between the unfolded images or the distance error of the contours at different rotation angles based on the sliding window matching, so as to determine the error angle to be corrected.

[0094] A correction unit is used to perform rotation correction on the cross-sectional image consistent with respiratory dynamics according to the correction error angle.

[0095] It should be noted that the above explanations of the embodiment of the method for airway reconstruction in dynamic breathing are also applicable to the apparatus for airway reconstruction in dynamic breathing of this embodiment, and will not be repeated here.

[0096] According to the device for reconstructing the respiratory tract in the dynamic state of breathing proposed in the embodiment of the present application, the dynamic change information of the respiratory tract in the breathing process can be comprehensively captured by collecting data covering each axial position and more than one breathing cycle, thereby providing a rich data basis for accurate reconstruction; the lumen structure is obtained by lumen segmentation, which is helpful to accurately analyze the internal morphology of the respiratory tract; the end-expiratory and end-inspiratory data are extracted by contour area or inner diameter changes to obtain consistent data of respiratory dynamics, effectively eliminating the interference of inconsistent respiratory state on reconstruction and improving the accuracy of the model; the model quality is further optimized by rotation correction and reconstruction using the corrected data, and finally a respiratory tract model with consistent respiratory state is obtained, which can provide a more reliable basis for the accurate diagnosis of respiratory diseases and effectively assist clinicians in accurately judging and effectively treating patients' conditions.

[0097] Fig.15 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0098] A memory 1501 , a processor 1502 , and a computer program stored in the memory 1501 and executable on the processor 1502 .

[0099] When the processor 1502 executes the program, the method for airway reconstruction in respiratory dynamics provided in the above embodiment is implemented.

[0100] Furthermore, the electronic device further comprises:

[0101] The communication interface 1503 is used for communication between the memory 1501 and the processor 1502 .

[0102] The memory 1501 is used to store computer programs that can be executed on the processor 1502 .

[0103] The memory 1501 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0104] If the memory 1501, the processor 1502 and the communication interface 1503 are implemented independently, the communication interface 1503, the memory 1501 and the processor 1502 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.15 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0105] Optionally, in a specific implementation, if the memory 1501, the processor 1502 and the communication interface 1503 are integrated on a chip, the memory 1501, the processor 1502 and the communication interface 1503 can communicate with each other through an internal interface.

[0106] The processor 1502 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0107] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above method for airway reconstruction in respiratory dynamics.

[0108] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, is used to implement the above method for airway reconstruction in respiratory dynamics.

[0109] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0110] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0111] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0112] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0113] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0114] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0115] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0116] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for airway reconstruction in respiratory dynamics, characterized in that: The following steps are involved: Acquire cross-sectional images of the respiratory tract at different axial positions of the respiratory tract and over a preset respiratory cycle; Performing lumen segmentation based on the respiratory tract cross-sectional image to obtain the lumen structure of the respiratory tract; Extracting the respiratory state based on the lumen structure to obtain a cross-sectional image consistent with the respiratory dynamics; Respiratory tract reconstruction is performed based on the cross-sectional image consistent with respiratory dynamics to obtain a respiratory tract model consistent with respiratory state.

2. The method according to claim 1, characterized in that The collecting of respiratory tract cross-sectional images at different axial positions of the respiratory tract and exceeding a preset respiratory cycle includes: During the rotation of the light beam, the light beam is controlled to rotate multiple times at the same or similar axial position so that when the light beam rotates to the same angle, it irradiates the same or similar position on the surface of the lumen to collect the cross-sectional images of the respiratory tract at the same or similar axial position under different respiratory states.

3. The method according to claim 1, characterized in that The lumen segmentation based on the respiratory tract cross-sectional image to obtain the lumen structure of the respiratory tract includes: Based on a preset deep learning U-Net network model, high-level features of the respiratory tract cross-sectional image are extracted by step-by-step downsampling, and the spatial resolution of the respiratory tract cross-sectional image is restored by step-by-step upsampling to obtain the tubular structure of the respiratory tract.

4. The method according to claim 1, characterized in that: The extracting of the respiratory state based on the lumen structure to obtain a cross-sectional image consistent with the respiratory dynamics includes: The image data of the respiratory cross section at the end of exhalation and the end of inspiration of each position are extracted by using the information of the change of the contour area or inner diameter over time after segmentation, so as to obtain the cross-sectional image consistent with the respiratory dynamics.

5. The method according to claim 1, characterized in that The respiratory tract reconstruction according to the cross-sectional image consistent with respiratory dynamics includes: Convert the cross-sectional image consistent with respiratory dynamics into an expanded image with a coordinate being an angle and a coordinate being a distance from a pixel to a center; Based on the expanded image, each image uses the expanded image with consistent respiratory dynamics at the previous position as a template to perform sliding window matching on the image translation expansion image; Based on the sliding window matching, the similarity between the unfolded images at different rotation angles or the distance error of the contours are obtained to determine the corrected error angle; The cross-sectional image consistent with respiratory dynamics is rotationally corrected according to the corrected error angle, thereby reconstructing a three-dimensional respiratory tract model.

6. A device for airway reconstruction during respiratory dynamics, characterized in that: include: An acquisition module, used for acquiring cross-sectional images of the respiratory tract at different axial positions of the respiratory tract and exceeding a preset respiratory cycle; A segmentation module, used for performing lumen segmentation based on the respiratory tract cross-sectional image to obtain the lumen structure of the respiratory tract; An extraction module, used for extracting the respiratory state based on the lumen structure to obtain a cross-sectional image consistent with the respiratory dynamics; The reconstruction module is used to reconstruct the respiratory tract according to the cross-sectional image with consistent respiratory dynamics, so as to obtain a respiratory tract model with consistent respiratory state.

7. The device according to claim 6, characterized in that The acquisition module comprises: During the rotation of the light beam, the light beam is controlled to rotate multiple times at the same or similar axial position so that when the light beam rotates to the same angle, it irradiates the same or similar position on the surface of the lumen to collect the cross-sectional images of the respiratory tract at the same or similar axial position under different respiratory states.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for airway reconstruction in dynamic breathing as described in any one of claims 1-5.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for airway reconstruction in respiratory dynamics as described in any one of claims 1-5.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed to implement the method for airway reconstruction in dynamic breathing according to any one of claims 1-5.

Images