WO2022104648A1 - 一种超分辨成像方法及系统 - Google Patents

一种超分辨成像方法及系统 Download PDF

Info

Publication number
WO2022104648A1
WO2022104648A1 PCT/CN2020/130109 CN2020130109W WO2022104648A1 WO 2022104648 A1 WO2022104648 A1 WO 2022104648A1 CN 2020130109 W CN2020130109 W CN 2020130109W WO 2022104648 A1 WO2022104648 A1 WO 2022104648A1
Authority
WO
WIPO (PCT)
Prior art keywords
contrast agent
image
correlation coefficient
signals
cross
Prior art date
Application number
PCT/CN2020/130109
Other languages
English (en)
French (fr)
Inventor
马腾
雷爽
王丛知
肖杨
张琪
郑海荣
Original Assignee
深圳先进技术研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳先进技术研究院 filed Critical 深圳先进技术研究院
Priority to PCT/CN2020/130109 priority Critical patent/WO2022104648A1/zh
Publication of WO2022104648A1 publication Critical patent/WO2022104648A1/zh

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B8/00Diagnosis using ultrasonic, sonic or infrasonic waves
    • A61B8/12Diagnosis using ultrasonic, sonic or infrasonic waves in body cavities or body tracts, e.g. by using catheters

Definitions

  • the present application relates to the field of ultrasound technology, and in particular, to a super-resolution imaging method and system.
  • Endoscopic Ultrasonography System is a medical device that integrates ultrasound and endoscopy. After the endoscope enters the body cavity, tomography scan is performed on the internal organ wall or adjacent organs under the direct vision of the endoscope to obtain ultrasound images of the layers below the mucosa of the internal organ wall and surrounding adjacent organs, such as the mediastinum, pancreas, bile duct and
  • ultrasound Doppler technology or contrast-enhanced ultrasound (CEUS) technology can be used to image blood flow, which has great advantages in staging of gastrointestinal tumors and judging the nature of tumors originating from the intestinal wall.
  • both the ultrasound Doppler technology and the contrast-enhanced ultrasound technology have the problem that the imaging resolution is limited by the ultrasound diffraction limit, which makes it impossible to obtain a large imaging depth and high resolution when imaging blood flow.
  • the embodiments of the present application provide a super-resolution imaging method and system, so as to achieve the purpose of improving the imaging resolution and ensuring a large imaging depth.
  • the technical solutions are as follows:
  • a super-resolution imaging method is applied to an endoscopic annular array ultrasonic transducer, the method comprising:
  • the contrast agent signal is separately separated from each of the echo reflection signals, and based on each of the contrast agent signals, the determination is made. Contrast image;
  • the first area corresponding to the largest cross-correlation coefficient in each second area in the contrast agent image is used as the target area, and the center coordinates of each target area are respectively used as the center coordinates of each contrast agent, and the second area is used as the center coordinate of each contrast agent.
  • a region contains at least one of said first regions;
  • a super-resolution image is obtained by accumulating the positions corresponding to the center coordinates of each of the contrast agents in the multiple frames of the contrast agent images.
  • the method further includes:
  • the cross-correlation coefficient corresponding to the first region is smaller than the set cross-correlation coefficient threshold, the first region is deleted from the contrast agent image.
  • the first area corresponding to the largest cross-correlation coefficient in each of the second areas in the contrast agent image is used as the target area, including:
  • the first region corresponding to the largest cross-correlation coefficient in each of the second regions in each of the sub-contrast agent images is used as the target region.
  • the calculating, respectively, the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image includes:
  • f represents the contrast agent image
  • t represents the PSF template image
  • c represents the cross-correlation coefficient
  • x represents the abscissa
  • y represents the ordinate
  • f(x, y) represents the pixel value at the (x, y) coordinate in the contrast agent image
  • u , v represent the translation amount of the PSF template image along the x-axis and y-axis in the contrast agent image, respectively
  • f u, v represent when the PSF template image is translated in the contrast agent image, the contrast agent image
  • the contrast agent signal is separated from each of the echo reflection signals, including:
  • n x represents each quadrature demodulation signal along the horizontal axis
  • n z represents the data obtained by sampling each of the quadrature demodulation signals along the vertical axis direction
  • n t represents the number of the plurality of the quadrature demodulation signals
  • Substitute the updated singular value matrix into the singular value decomposition relationship S U ⁇ V * , calculate the updated two-dimensional space-time matrix, and convert the updated two-dimensional space-time matrix into a data matrix S(n x ,n z ,n t ), the signal corresponding to the converted data matrix S(n x ,n z ,n t ) is used as the contrast agent signal.
  • a super-resolution imaging system applied to an endoscopic annular array ultrasonic transducer comprising:
  • the signal emission and acquisition module is used to transmit an ultrasonic signal into the imaging target by using the ring array ultrafast imaging algorithm at every set time and collect the ultrasonic signal when an ultrasonic contrast agent is injected into the imaging target.
  • a signal separation module configured to separate the contrast agent signal from each of the echo reflection signals based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals
  • a first determination module for determining a contrast agent image based on each of the contrast agent signals, respectively;
  • the first calculation module is configured to obtain a pre-calibrated PSF template image, and for each frame of the contrast agent image, respectively calculate the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image, the The cross-correlation coefficient is used to characterize the similarity between the PSF template image and the first region in the contrast agent image;
  • the second determination module is configured to use the first area corresponding to the largest cross-correlation coefficient in each second area in the contrast agent image as the target area, and use the center coordinate of each target area as the center coordinate of each contrast agent. center coordinates, the second area includes at least one of the first areas;
  • An imaging module configured to obtain a super-resolution image by accumulating the positions corresponding to the center coordinates of each of the contrast agents in the multiple frames of the contrast agent images.
  • the system also includes:
  • a comparison module configured to respectively compare the cross-correlation coefficient corresponding to each first region in the contrast agent image and set the cross-correlation coefficient threshold
  • a deletion module configured to delete the first region from the contrast agent image if the cross-correlation coefficient corresponding to the first region is smaller than the set cross-correlation coefficient threshold.
  • the first computing module is specifically used for:
  • the second determining module is specifically used for:
  • the first region corresponding to the largest cross-correlation coefficient in each of the second regions in each of the sub-contrast agent images is used as the target region.
  • the first computing module is specifically used for:
  • f represents the contrast agent image
  • t represents the PSF template image
  • c represents the cross-correlation coefficient
  • x represents the abscissa
  • y represents the ordinate
  • f(x, y) represents the pixel value at the (x, y) coordinate in the contrast agent image
  • u , v represent the translation amount of the PSF template image along the x-axis and y-axis in the contrast agent image, respectively
  • f u, v represent when the PSF template image is translated in the contrast agent image, the contrast agent image
  • the signal separation module is specifically used for:
  • n x represents each quadrature demodulation signal along the horizontal axis
  • n z represents the data obtained by sampling each of the quadrature demodulation signals along the vertical axis direction
  • n t represents the number of the plurality of the quadrature demodulation signals
  • Substitute the updated singular value matrix into the singular value decomposition relationship S U ⁇ V * , calculate the updated two-dimensional space-time matrix, and convert the updated two-dimensional space-time matrix into a data matrix S(n x ,n z ,n t ), the signal corresponding to the converted data matrix S(n x ,n z ,n t ) is used as the contrast agent signal.
  • the ring array ultrafast imaging algorithm is used to transmit ultrasonic signals into the imaging target, and the echo reflection signals of the ultrasonic signals are collected.
  • the echo reflection signal in a larger range can be collected at a higher frame rate, and the lateral resolution during imaging can be improved by increasing the number of virtual point sources.
  • the contrast agent signal can characterize the blood in the imaging target.
  • the contrast agent signal is separated from the echo reflection signal, and the pre-calibrated PSF template image is obtained, and the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image is calculated respectively, and the The first area corresponding to the largest cross-correlation coefficient in each second area in the contrast agent image is taken as the target area, the area with the highest similarity with the PSF template image is determined, the center coordinate of the area is taken as the center coordinate of the contrast agent, and the reduction
  • the space of each contrast agent in the image ensures the high resolution of the super-resolution image, and can improve the positioning accuracy of each contrast agent to ensure the accuracy of the super-resolution image.
  • the frequency of the ultrasound transducer is traded for high resolution and, therefore, a large imaging depth can be guaranteed.
  • Embodiment 1 is a flowchart of a super-resolution imaging method provided in Embodiment 1 of the present application;
  • Fig. 2 is the spherical wave composite schematic diagram of setting three virtual point sources provided by this application;
  • FIG. 3 is a schematic diagram of beam projection with the center of the ring array as the coordinate origin provided by the present application;
  • FIG. 5 is a flowchart of a super-resolution imaging method provided in Embodiment 3 of the present application.
  • FIG. 6 is a schematic diagram of a kind of contrast agent image angular division provided by the present application.
  • FIG. 7 is a flowchart of a super-resolution imaging method provided in Embodiment 4 of the present application.
  • FIG. 8 is a schematic diagram of a logical structure of the super-resolution imaging system provided by the present application.
  • FIG. 1 a flowchart of a super-resolution imaging method provided in Embodiment 1 of the present application is applied to an endoscopic annular array ultrasonic transducer. As shown in FIG. 1 , the method may include but is not limited to The following steps:
  • Step S11 In the case where an ultrasonic contrast agent is injected into the imaging target, use the ring array ultrafast imaging algorithm to transmit an ultrasonic signal into the imaging target every set time, and collect the echo reflection signal of the ultrasonic signal. .
  • an ultrasonic contrast agent when using the endoscopic annular array ultrasonic transducer to image blood flow, an ultrasonic contrast agent needs to be injected into the imaging target. After the ultrasound contrast agent is injected into the imaging target, the ultrasound contrast agent flows with the blood of the imaging target, so the signal of the ultrasound contrast agent can represent the blood flow signal.
  • a virtual point source is set internally, and the time for the virtual spherical wave to reach different array elements can be calculated according to the distance from the virtual point source to different array elements.
  • the spherical wave formed by the signals emitted by all the array elements can be regarded as the spherical wave emitted from the virtual point source position, so its forward propagation time can be determined by Accurate calculation, by using the delay stacking method to beamform the echo signals of different receiving array elements, the radio frequency signal image of the entire imaging plane can be obtained.
  • radio frequency signal images obtained from different divergent waves are coherently superimposed, so that the different divergent waves emitted each time can be coherently superimposed at any position in the imaging plane to form a synthetic focus.
  • Figure 2 for setting the three virtual point sources Schematic diagram of spherical wave compounding.
  • the sound field distribution at the imaging point P along the vertical direction connecting the origin and the imaging point can be obtained as:
  • N is the number of virtual point sources, and li is the distance from the ith virtual point source to the imaging point p.
  • N virtual point sources are distributed on a circle with radius r, the spatial spectrum range of the composite sound field at the imaging point P along the vertical direction of the line connecting the origin and the imaging point is:
  • the ring array ultrafast imaging algorithm is used to transmit ultrasonic signals into the imaging target, and the ultrasonic signals are collected.
  • the echo reflects the signal, which can improve the lateral resolution during imaging.
  • Step S12 based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals, separate a contrast agent signal from each of the echo reflection signals, respectively, and based on each of the contrast agent signals signal to determine the contrast agent image.
  • the tissue in the imaging target exists stably, and the contrast agent flows with the blood, there is no stable relationship between the contrast agents, so the correlation between the tissue signals is greater than the correlation between the contrast agent signals.
  • the echo reflection signal includes the tissue signal and the contrast agent signal, so it can be based on the relationship between the two, specifically based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals , the contrast agent signal is separated from the echo reflection signal.
  • Determining a contrast agent image based on the contrast agent signal can be understood as: extracting contrast agent elements from the contrast agent signal, and combining the extracted contrast agent elements into a contrast agent image.
  • the contrast agent signal is separately separated from each of the echo reflection signals, which may include but not limited to At:
  • S122 Sampling a plurality of the orthogonal demodulation signals, and forming the sampled data into a data matrix S(n x , n z , n t ), where n x represents the measurement of each orthogonal solution along the horizontal axis The data obtained by sampling the modulated signal, nz represents the data obtained by sampling each of the quadrature demodulation signals along the vertical axis direction, and nt represents the number of the plurality of the quadrature demodulation signals.
  • the data matrix S(n x ,n z ,n t ) can be understood as C(n x ,n z ,n t )+B(n x ,n z ,n t )+N(n x ,n z , n t ), C(n x ,n z ,n t ) denotes the set of data obtained by sampling the tissue signal, and B(n x ,n z ,n t ) denotes the set of data obtained by sampling the contrast agent signal , N(n x ,n z ,n t ) represents the set of data obtained by sampling the noise signal.
  • the singular value in the singular value matrix Based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals, it can be determined that the larger the singular value in the singular value matrix, the larger the singular value indicates that the singular value corresponds to the tissue signal, therefore, the singular value can be deleted.
  • the singular value in the matrix is greater than the set singular value threshold, and the updated singular value matrix is obtained.
  • the singular values in the updated singular value matrix correspond to the contrast agent signal.
  • the process of separating the contrast agent signal from the echo reflection signal may also be: based on the correlation between tissue signals.
  • the correlation information and the high-pass filtering method that the correlation is greater than the correlation between the contrast agent signals separates the contrast agent signal from the echo reflection signal.
  • Step S13 Acquire a pre-calibrated PSF template image, and for each frame of the contrast agent image, respectively calculate the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image, and the cross-correlation coefficient uses to characterize the similarity between the PSF template image and the first region in the contrast agent image.
  • the process of PSF (point spread function, point spread function) template image calibration may be: placing a thin line perpendicular to the imaging plane of the endoscopic ring array ultrasonic transducer to act as an ideal point scatterer, The image acquired by the endoscopic ring array ultrasonic transducer is used as the PSF template image.
  • the independent contrast agent image also shows the point spread function (PSF) of the ultrasound system, so it can be judged that the PSF template image is similar to the contrast agent. degrees to determine the center coordinates of the contrast agent. Specifically, the cross-correlation coefficient of each first region in the PSF template image and the contrast agent image is calculated respectively.
  • PSF point spread function
  • the calculation of the cross-correlation coefficients between the PSF template image and the first regions in the contrast agent image may include, but is not limited to:
  • the PSF template image is translated in the contrast agent image according to the set step size, and the cross-correlation coefficient between the PSF template image and the first region in the contrast agent image is calculated once for each translation.
  • the first area can be understood as: the area in the contrast agent image that overlaps with the figure (eg, circumscribed rectangle or circumscribed circle) containing the PSF template image .
  • the contrast agent image can also be divided in advance to obtain a plurality of first regions, and the cross-correlation coefficients between the PSF template image and the first regions in the contrast agent image are calculated respectively.
  • the calculation of the cross-correlation coefficients between the PSF template image and the first regions in the contrast agent image may include, but is not limited to:
  • f represents the contrast agent image
  • t represents the PSF template image
  • c represents the cross-correlation coefficient
  • x represents the abscissa
  • y represents the ordinate
  • f(x, y) represents the pixel value at the (x, y) coordinate in the contrast agent image
  • u , v represent the translation amount of the PSF template image along the x-axis and y-axis in the contrast agent image, respectively
  • f u, v represent when the PSF template image is translated in the contrast agent image, the contrast agent image
  • Step S14 taking the first area corresponding to the largest cross-correlation coefficient in each second area in the contrast agent image as the target area, and taking the center coordinates of each target area as the center coordinates of each contrast agent, respectively.
  • the second region includes at least one of the first regions.
  • the first area corresponding to the largest cross-correlation coefficient can be selected from the second area, Take the selected first area as the target area.
  • Step S15 Obtain a super-resolution image by accumulating the positions corresponding to the center coordinates of each of the contrast agents in the multiple frames of the contrast agent images.
  • a super-resolution image is obtained, which can be understood as: adding the center of each of the contrast agents in the multiple frames of the contrast agent images
  • the positions corresponding to the coordinates are accumulated to obtain the movement trajectory of the contrast agent, and the image containing the movement trajectory of the contrast agent is used as a super-resolution image.
  • the ring array ultrafast imaging algorithm is used to transmit ultrasonic signals into the imaging target, and the echo reflection signals of the ultrasonic signals are collected,
  • the echo reflection signal in a large range can be collected, and the lateral resolution during imaging can be improved based on the echo reflection signal in a large range.
  • the contrast agent signal can represent the blood flow signal in the imaging target, Therefore, the contrast agent signal is separated from the echo reflection signal, and a pre-calibrated PSF template image is obtained, the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image is calculated respectively, and the contrast agent The first area corresponding to the largest cross-correlation coefficient in each second area in the image is used as the target area, and the area with the highest similarity to the PSF template image is determined. On this basis, the space of each contrast agent in the image is reduced to ensure the high resolution of the super-resolution image. At the same time, it is not necessary to increase the frequency of the endoscopic ring array ultrasonic transducer in exchange for High resolution, therefore, can ensure a large imaging depth.
  • FIG. 4 it is a flowchart of Embodiment 2 of a super-resolution imaging method provided by the present application.
  • This embodiment is mainly an extension of the super-resolution imaging method described in Embodiment 1 above.
  • Scheme, as shown in Figure 4 the method can include but is not limited to the following steps:
  • Step S21 In the case where an ultrasonic contrast agent is injected into the imaging target, use the ring array ultrafast imaging algorithm to transmit an ultrasonic signal into the imaging target every set time, and collect the echo reflection signal of the ultrasonic signal. .
  • Step S22 Based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals, separate a contrast agent signal from each of the echo reflection signals, and determine based on the contrast agent signal. Contrast image.
  • Step S23 Acquire a pre-calibrated PSF template image, and for each frame of the contrast agent image, respectively calculate the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image, and the cross-correlation coefficient uses to characterize the similarity between the PSF template image and the first region in the contrast agent image.
  • steps S21-S23 can be related to the related introduction of steps S11-S13 in Embodiment 1, and details are not repeated here.
  • Step S24 respectively comparing the cross-correlation coefficient corresponding to each first region in the contrast agent image with the set cross-correlation coefficient threshold.
  • the cross-correlation coefficient threshold can be determined and set based on the fact that the cross-correlation coefficient between the region composed of noise pixels and the PSF template image is smaller than the cross-correlation coefficient between the region containing the contrast agent and the PSF template image.
  • Step S25 if the cross-correlation coefficient corresponding to the first region is smaller than the set cross-correlation coefficient threshold, delete the first region from the contrast agent image.
  • the cross-correlation coefficient corresponding to the first area is smaller than the set cross-correlation coefficient threshold, it indicates that the first area is an area composed of noise pixels, so the first area is deleted from the contrast agent image to delete the contrast agent image noise pixels in .
  • Step S26 taking the first area corresponding to the largest cross-correlation coefficient in each second area in the contrast agent image as the target area, and taking the center coordinate of each target area as the center coordinate of each contrast agent, so
  • the second region includes at least one of the first regions.
  • the contrast agent image in this step can be understood as: the contrast agent image after the noise pixels are deleted.
  • Step S27 Obtain a super-resolution image by accumulating the positions corresponding to the center coordinates of each of the contrast agents in the multiple frames of the contrast agent images.
  • step S27 For the detailed process of step S27, reference may be made to the relevant introduction of step S15 in Embodiment 1, and details are not repeated here.
  • the first region is deleted from the contrast agent image, the noise pixels in the contrast agent image are deleted, and the corresponding maximum cross-correlation coefficient in each second region in the contrast agent image from which the noise pixel points are deleted will be deleted.
  • the first area is used as a target area, and the center coordinates of each target area are used to further improve the accuracy of the center coordinates of the contrast agent and further improve the resolution of the super-resolution image.
  • FIG. 5 it is a flowchart of Embodiment 3 of a super-resolution imaging method provided by the present application. This embodiment mainly details the super-resolution imaging method described in Embodiment 1 above. As shown in Figure 5, the method can include but is not limited to the following steps:
  • Step S31 In the case where an ultrasound contrast agent is injected into the imaging target, use the ring-array ultrafast imaging algorithm to transmit an ultrasound signal into the imaging target every set time, and collect echo reflection signals of the ultrasound signal. ;
  • Step S32 based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals, separate the contrast agent signal from each of the echo reflection signals respectively, and separately based on each of the contrast agent signals signal to determine the contrast agent image.
  • steps S31-S32 For the detailed process of steps S31-S32, reference may be made to the related introduction of steps S11-S12 in Embodiment 1, and details are not repeated here.
  • Step S33 acquiring pre-calibrated PSF template images of different angles.
  • the endoscopic ring array ultrasonic transducer transmits ultrasonic signals and collects echo reflection signals in the form of a ring array
  • the contrast agent signals in the different echo reflection signals collected by the endoscopic ring array ultrasonic transducer The angles are different, therefore, the PSF template images can be pre-calibrated for different angles.
  • Step S34 Divide each of the contrast agent images into sub-contrast agent images of different angles respectively.
  • dividing the contrast agent image into sub-contrast agent images with different angles may include, but is not limited to: dividing the contrast agent image into sub-contrast agent images with four angles. As shown in FIG. 6 , the contrast agent image is divided into sub-contrast agent images of four angles with the endoscopic annular array ultrasonic transducer as the center.
  • Step S35 for each sub-contrast agent image of the angle, select the PSF template image corresponding to the angle of the sub-contrast agent image from the PSF template images of different angles, as the target PSF template image, and calculate the target respectively.
  • Cross-correlation coefficients between the PSF template image and each first region in the sub-contrast agent image are cross-correlation coefficients between the PSF template image and each first region in the sub-contrast agent image.
  • Steps S33-S35 are a specific implementation of step S13 in Embodiment 1.
  • Step S36 taking the first area corresponding to the maximum cross-correlation coefficient in each second area in each sub-contrast agent image as the target area, and taking the center coordinates of each target area as the center of each contrast agent respectively Coordinates, the second area includes at least one of the first areas.
  • Step S36 is a specific implementation of step S14 in Embodiment 1.
  • Step S37 obtaining a super-resolution image by accumulating the positions corresponding to the center coordinates of each of the contrast agents in the multiple frames of the contrast agent images.
  • step S37 For the detailed process of step S37, reference may be made to the relevant introduction of step S15 in Embodiment 1, and details are not repeated here.
  • the PSF template image corresponding to the angle of the sub-contrast agent image is selected from the PSF template image, as the target PSF template image, and the mutual relationship between the target PSF template image and each first region in the sub-contrast agent image is calculated respectively. It can reduce the positioning deviation and further improve the accuracy of the center coordinates of the contrast agent, thereby further improving the resolution of the super-resolution image.
  • FIG. 7 is a flowchart of Embodiment 4 of a super-resolution imaging method provided by the present application
  • this embodiment is mainly an extension of the super-resolution imaging method described in Embodiment 1 above Scheme, as shown in Figure 7, the method can include but is not limited to the following steps:
  • Step S41 in the case where an ultrasonic contrast agent is injected into the imaging target, use the ring array ultrafast imaging algorithm to transmit an ultrasonic signal into the imaging target once every set time, and collect the echo reflection signal of the ultrasonic signal .
  • Step S42 based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals, separate the contrast agent signal from each of the echo reflection signals, respectively, and based on each of the contrast agent signals signal to determine the contrast agent image.
  • Step S43 Acquire a pre-calibrated PSF template image, and for each frame of the contrast agent image, respectively calculate the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image, and the cross-correlation coefficient uses to characterize the similarity between the PSF template image and the first region in the contrast agent image.
  • Step S44 taking the first area corresponding to the maximum cross-correlation coefficient in each second area in the contrast agent image as the target area, and taking the center coordinate of each target area as the center coordinate of each contrast agent, so
  • the second region includes at least one of the first regions.
  • Step S45 obtaining a super-resolution image by accumulating the positions corresponding to the center coordinates of each of the contrast agents in the multiple frames of the contrast agent images.
  • steps S41-S45 For the detailed process of steps S41-S45, reference may be made to the related introduction of steps S11-S15 in Embodiment 1, and details are not repeated here.
  • Step S46 use the nearest neighbor matching algorithm to match the positions corresponding to the center coordinates of the contrast agent in the adjacent two frames of super-resolution images, and calculate the distance between the positions corresponding to the center coordinates of the two matched contrast agents. the distance.
  • the nearest neighbor matching algorithm may be the nearest neighbor matching algorithm (Kuhn-Munkras matching algorithm) in Particle Tracking Velocimetry (PTV).
  • Step S47 Calculate the movement speed of the contrast agent based on the distance between the positions corresponding to the center coordinates of the two contrast agents.
  • the moving speed of the contrast agent can be used as the blood flow speed in the imaging target.
  • the blood flow velocity in the imaging target is further calculated, the function of the endoscopic annular array ultrasonic transducer is improved, more information is provided for medical examination, and medical progress is promoted.
  • the super-resolution imaging system is applied to an endoscopic ring array ultrasonic transducer.
  • the super-resolution imaging system includes: a signal emission and acquisition module 100, a signal separation module 200, a first determination module 300, a first calculation module 400, The second determination module 500 and the imaging module 600 .
  • the signal emission and acquisition module is used to transmit an ultrasonic signal into the imaging target by using the ring array ultrafast imaging algorithm at every set time and collect the ultrasonic signal when an ultrasonic contrast agent is injected into the imaging target.
  • a signal separation module configured to separate the contrast agent signal from each of the echo reflection signals, respectively, based on the relationship information that the correlation between the tissue signals is greater than the correlation between the contrast agent signals
  • a first determination module for determining a contrast agent image based on each of the contrast agent signals, respectively;
  • the first calculation module is configured to obtain a pre-calibrated PSF template image, and for each frame of the contrast agent image, respectively calculate the cross-correlation coefficient between the PSF template image and each first region in the contrast agent image, the The cross-correlation coefficient is used to characterize the similarity between the PSF template image and the first region in the contrast agent image;
  • the second determination module is configured to use the first area corresponding to the largest cross-correlation coefficient in each second area in the contrast agent image as the target area, and use the center coordinate of each target area as the center coordinate of each contrast agent. center coordinates, the second area includes at least one of the first areas;
  • An imaging module configured to obtain a super-resolution image by accumulating the positions corresponding to the center coordinates of each of the contrast agents in the multiple frames of the contrast agent images.
  • the super-resolution imaging system may further include:
  • a comparison module configured to respectively compare the cross-correlation coefficient corresponding to each first region in the contrast agent image and set the cross-correlation coefficient threshold
  • a deletion module configured to delete the first region from the contrast agent image if the cross-correlation coefficient corresponding to the first region is smaller than the set cross-correlation coefficient threshold.
  • the first computing module can be specifically used for:
  • the second determining module can be specifically used for:
  • the first region corresponding to the largest cross-correlation coefficient in each of the second regions in each of the sub-contrast agent images is used as the target region.
  • the first computing module can be specifically used for:
  • f represents the contrast agent image
  • t represents the PSF template image
  • c represents the cross-correlation coefficient
  • x represents the abscissa
  • y represents the ordinate
  • f(x, y) represents the pixel value at the (x, y) coordinate in the contrast agent image
  • u , v represent the translation amount of the PSF template image along the x-axis and y-axis in the contrast agent image, respectively
  • f u, v represent when the PSF template image is translated in the contrast agent image, the contrast agent image
  • the signal separation module can be specifically used for:
  • n x represents each quadrature demodulation signal along the horizontal axis
  • n z represents the data obtained by sampling each of the quadrature demodulation signals along the vertical axis direction
  • n t represents the number of the plurality of the quadrature demodulation signals
  • Substitute the updated singular value matrix into the singular value decomposition relationship S U ⁇ V * , calculate the updated two-dimensional space-time matrix, and convert the updated two-dimensional space-time matrix into a data matrix S(n x ,n z ,n t ), the signal corresponding to the converted data matrix S(n x ,n z ,n t ) is used as the contrast agent signal.
  • each embodiment focuses on the differences from other embodiments, and the same and similar parts between the various embodiments may be referred to each other.
  • the apparatus type embodiment since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant part, please refer to the partial description of the method embodiment.

Landscapes

  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Pathology (AREA)
  • Radiology & Medical Imaging (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Ultra Sonic Daignosis Equipment (AREA)

Abstract

一种超分辨成像方法及系统,应用于内窥环阵超声换能器。该方法通过获取预先标定的PSF模板图像,分别计算PSF模板图像与造影剂图像中各个第一区域的互相关系数,将造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,确定出与PSF模板图像相似度最高的区域,将该区域的中心坐标作为造影剂的中心坐标,能够提高每个造影剂定位的准确性,在此基础上,缩小每个造影剂在图像中的空间,保证超分辨图像的高分辨率,同时,由于并不需要提高内窥环阵超声换能器的频率来换取高分辨率,因此,能够保证大的成像深度。

Description

一种超分辨成像方法及系统 技术领域
本申请涉及超声技术领域,特别涉及一种超分辨成像方法及系统。
背景技术
超声波内窥镜(Endoscopic Ultrasonography System,EUS)是一种集超声波与内镜检查为一身的医疗设备。当内镜进入体腔后,在内镜直视下对内脏器官壁或邻近脏器进行断层扫描,获得内脏器官壁黏膜以下各层次和周围邻近脏器的超声图像,如纵膈、胰腺、胆管及淋巴结等,并且可采用超声多普勒技术或对比增强超声(CEUS)技术对血流进行成像,在胃肠道肿瘤的分期及判断肠壁起源肿瘤的性质方面具有极大的优势。
但是,超声多普勒技术或对比增强超声技术,均存在成像分辨率受到超声衍射极限限制的问题,导致对血流进行成像时无法获得大的成像深度和高分辨率。
发明内容
为解决上述技术问题,本申请实施例提供一种超分辨成像方法及系统,以达到提高成像分辨率的同时,保证大的成像深度的目的,技术方案如下:
一种超分辨成像方法,应用于内窥环阵超声换能器,该方法包括:
在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号;
基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,并分别基于每个所述造影剂信号,确定造影剂图像;
获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的相似度;
将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域;
通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
所述分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数之后,还包括:
分别比较所述造影剂图像中各个第一区域对应的互相关系数与设定互相关系数阈值;
若所述第一区域对应的互相关系数小于所述设定互相关系数阈值,则从所述造影剂图像中删除所述第一区域。
所述获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,包括:
获取预先标定的不同角度的PSF模板图像;
将每帧所述造影剂图像划分为不同角度的子造影剂图像;
对每个所述角度的子造影剂图像,从不同角度的PSF模板图像中选择与所述子造影剂图像的角度对应的PSF模板图像,作为目标PSF模板图像,分别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数;
所述将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,包括:
将每个所述子造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域。
所述分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,包括:
利用关系式
Figure PCTCN2020130109-appb-000001
分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数;
其中,f表示所述造影剂图像,t表示所述PSF模板图像,
Figure PCTCN2020130109-appb-000002
表示所述造影 剂图像的均值,
Figure PCTCN2020130109-appb-000003
表示所述PSF模板图像的均值,c表示互相关系数,x表示横坐标,y表示纵坐标,f(x,y)表示所述造影剂图像中(x,y)坐标处的像素值,u,v分别表示所述PSF模板图像在所述造影剂图像中沿x轴和y轴的平移量,f u,v表示当所述PSF模板图像在所述造影剂图像中平移时,所述造影剂图像在所述PSF模板图像覆盖区域中的像素值。
所述基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,包括:
对每个所述回波反射信号进行波束合成,得到波束合成信号,并对每个所述波束合成信号进行正交解调,得到正交解调信号;
对多个所述正交解调信号进行采样,将采样得到的数据组成数据矩阵S(n x,n z,n t),n x表示沿横轴方向对每个所述正交解调信号采样得到的数据,n z表示沿纵轴方向对每个所述正交解调信号采样得到的数据,n t表示多个所述正交解调信号的个数;
将所述数据矩阵S(n x,n z,n t)转换为以Casorati矩阵形式重新排列的二维时空矩阵;
利用奇异值分解关系式S=UΔV *,对所述二维时空矩阵进行奇异值分解,得到奇异值矩阵,其中,S表示所述二维时空矩阵,U和V分别表示不同的正交矩阵,U等于(n x×n z,n x×n z),V等于(n t,n t),*表示共轭转置,Δ表示奇异值矩阵;
基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,删除所述奇异值矩阵中大于设定奇异值阈值的奇异值,得到更新后的奇异值矩阵;
将所述更新后的奇异值矩阵代入所述奇异值分解关系式S=UΔV *,计算得到更新后的二维时空矩阵,将更新后的二维时空矩阵转换为数据矩阵S(n x,n z,n t),将转换得到的数据矩阵S(n x,n z,n t)对应的信号作为造影剂信号。
一种超分辨成像系统,应用于内窥环阵超声换能器,该系统包括:
信号发射及采集模块,用于在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号;
信号分离模块,用于基于组织信号之间的相关性大于造影剂信号之间的相 关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号;
第一确定模块,用于分别基于每个所述造影剂信号,确定造影剂图像;
第一计算模块,用于获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的相似度;
第二确定模块,用于将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域;
成像模块,用于通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
所述系统还包括:
比较模块,用于分别比较所述造影剂图像中各个第一区域对应的互相关系数与设定互相关系数阈值;
删除模块,用于若所述第一区域对应的互相关系数小于所述设定互相关系数阈值,则从所述造影剂图像中删除所述第一区域。
所述第一计算模块,具体用于:
获取预先标定的不同角度的PSF模板图像;
将每帧所述造影剂图像划分为不同角度的子造影剂图像;
对每个所述角度的子造影剂图像,从不同角度的PSF模板图像中选择与所述子造影剂图像的角度对应的PSF模板图像,作为目标PSF模板图像,分别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数;
所述第二确定模块,具体用于:
将每个所述子造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域。
所述第一计算模块,具体用于:
利用关系式
Figure PCTCN2020130109-appb-000004
分别计算所 述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数;
其中,f表示所述造影剂图像,t表示所述PSF模板图像,
Figure PCTCN2020130109-appb-000005
表示所述造影剂图像的均值,
Figure PCTCN2020130109-appb-000006
表示所述PSF模板图像的均值,c表示互相关系数,x表示横坐标,y表示纵坐标,f(x,y)表示所述造影剂图像中(x,y)坐标处的像素值,u,v分别表示所述PSF模板图像在所述造影剂图像中沿x轴和y轴的平移量,f u,v表示当所述PSF模板图像在所述造影剂图像中平移时,所述造影剂图像在所述PSF模板图像覆盖区域中的像素值。
所述信号分离模块,具体用于:
对每个所述回波反射信号进行波束合成,得到波束合成信号,并对每个所述波束合成信号进行正交解调,得到正交解调信号;
对多个所述正交解调信号进行采样,将采样得到的数据组成数据矩阵S(n x,n z,n t),n x表示沿横轴方向对每个所述正交解调信号采样得到的数据,n z表示沿纵轴方向对每个所述正交解调信号采样得到的数据,n t表示多个所述正交解调信号的个数;
将所述数据矩阵S(n x,n z,n t)转换为以Casorati矩阵形式重新排列的二维时空矩阵;
利用奇异值分解关系式S=UΔV *,对所述二维时空矩阵进行奇异值分解,得到奇异值矩阵,其中,S表示所述二维时空矩阵,U和V分别表示不同的正交矩阵,U等于(n x×n z,n x×n z),V等于(n t,n t),*表示共轭转置,Δ表示奇异值矩阵;
基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,删除所述奇异值矩阵中大于设定奇异值阈值的奇异值,得到更新后的奇异值矩阵;
将所述更新后的奇异值矩阵代入所述奇异值分解关系式S=UΔV *,计算得到更新后的二维时空矩阵,将更新后的二维时空矩阵转换为数据矩阵S(n x,n z,n t),将转换得到的数据矩阵S(n x,n z,n t)对应的信号作为造影剂信号。
与现有技术相比,本申请的有益效果为:
在本申请中,由于内窥环阵超声换能器的声场能量较高,因此利用环阵超快成像算法向所述成像目标内发射超声信号,并采集所述超声信号的回波 反射信号,能够以较高帧率采集到较大范围内的回波反射信号,并通过增加虚拟点源个数能够提高成像时的横向分辨率,在此基础上,由于造影剂信号可以表征成像目标内血流的信号,因此从回波反射信号中分离出造影剂信号,并获取预先标定的PSF模板图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,确定出与PSF模板图像相似度最高的区域,将该区域的中心坐标作为造影剂的中心坐标,缩小每个造影剂在图像中的空间,保证超分辨图像的高分辨率,并能够提高每个造影剂定位的准确性,保证超分辨图像的准确性,同时,由于并不需要提高内窥环阵超声换能器的频率来换取高分辨率,因此,能够保证大的成像深度。
附图说明
为了更清楚地说明本申请实施例中的技术方案,下面将对实施例描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动性的前提下,还可以根据这些附图获得其他的附图。
图1是本申请实施例1提供的一种超分辨成像方法的流程图;
图2是本申请提供的设置三个虚拟点源的球面波复合示意图;
图3是本申请提供的以环阵中心为坐标原点的波束投影示意图;
图4是本申请实施例2提供的一种超分辨成像方法的流程图;
图5是本申请实施例3提供的一种超分辨成像方法的流程图;
图6是本申请提供的一种造影剂图像分角度划分的示意图;
图7是本申请实施例4提供的一种超分辨成像方法的流程图;
图8是本申请提供的超分辨成像系统的一种逻辑结构示意图。
具体实施方式
下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本申请一部分实施例, 而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
为使本发明的上述目的、特征和优点能够更加明显易懂,下面结合附图和具体实施方式对本发明作进一步详细的说明。
参照图1,为本申请实施例1提供的一种超分辨成像方法的流程图,该方法应用于内窥环阵超声换能器,如图1所示,该方法可以包括但并不局限于以下步骤:
步骤S11、在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号。
本实施例中,在利用内窥环阵超声换能器对血流进行成像时,需要向成像目标内注射超声造影剂。超声造影剂注射到成像目标内后,超声造影剂随成像目标的血液流动,因此超声造影剂的信号可以表征血流信号。
本实施例中,环阵超快成像算法的具体原理为:
内部设置虚拟点源,根据虚拟点源到不同阵元的距离,可以计算出虚拟的球面波到达不同阵元的时间。以该时间间隔依次激发环阵上的所有阵元,那么最终由所有阵元发射信号所形成的球面波可以看作是从虚拟点源位置发射出的球面波,因此其正向传播时间可以被精确计算,通过对不同接收阵元的回波信号采用延时叠加方法进行波束形成,可以得到整个成像平面的射频信号图像。通过在环阵内部设置多个虚拟点源,形成不同的发散球面波并得到相应的射频信号图像。最后将不同发散波得到的射频信号图像进行相干叠加,使得每次发射的不同发散波在成像平面内任意位置都能相干叠加形成合成聚焦,如图2所示,为设置三个虚拟点源的球面波复合示意图。
如图3所示,在以环阵中心为坐标原点的极坐标系中,由虚拟点源S(r,β)发出的发散波波阵面到达成像点
Figure PCTCN2020130109-appb-000007
处时,可以将波束k 0向原点与成像点连线的垂直方向(切向)做投影,记为k τ。由几何关系可知:
Figure PCTCN2020130109-appb-000008
当存在多个虚拟点源的时候,将多个虚拟点源产生的声场在成像点P处相干叠加之后,可以得到成像点P处沿原点与成像点连线的垂直方向的声场分布为:
Figure PCTCN2020130109-appb-000009
N为虚拟点源的个数,l i为第i个虚拟点源到成像点p的距离。当N个虚拟点源都分布在半径为r的圆环上时,成像点P处沿原点与成像点连线的垂直方向的复合声场的空间频谱范围为:
Figure PCTCN2020130109-appb-000010
因此由上述理论分析可知,虚拟点源所在的圆环半径r越大,复合声场的空间频率范围越大,空间频谱宽度也越大,理论上图像的横向分辨率也就越高。而虚拟点源个数越多,对空间频谱的采样率就越高,而且声场能量越高,因此利用环阵超快成像算法向所述成像目标内发射超声信号,并采集所述超声信号的回波反射信号,可以提高成像时的横向分辨率。
步骤S12、基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,并分别基于每个所述造影剂信号,确定造影剂图像。
由于成像目标内的组织是稳定存在的,而造影剂是随血液流动的,造影剂之间不存在稳定的关系,因此组织信号之间的相关性大于造影剂信号之间的相 关性。
可以理解的是,回波反射信号中包含组织信号和造影剂信号,因此可以基于两者之间的关系,具体可以基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,从所述回波反射信号中分离出造影剂信号。
基于所述造影剂信号,确定造影剂图像,可以理解为:从所述造影剂信号中提取造影剂元素,将提取出的造影剂元素,组成造影剂图像。
本实施例中,所述基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,可以包括但不局限于:
S121、对每个所述回波反射信号进行波束合成,得到波束合成信号,并对每个所述波束合成信号进行正交解调,得到正交解调信号;
S122、对多个所述正交解调信号进行采样,将采样得到的数据组成数据矩阵S(n x,n z,n t),n x表示沿横轴方向对每个所述正交解调信号采样得到的数据,n z表示沿纵轴方向对每个所述正交解调信号采样得到的数据,n t表示多个所述正交解调信号的个数。
其中,数据矩阵S(n x,n z,n t)可以理解为C(n x,n z,n t)+B(n x,n z,n t)+N(n x,n z,n t),C(n x,n z,n t)表示对组织信号进行采样得到的数据的集合,B(n x,n z,n t)表示对造影剂信号进行采样得到的数据的集合,N(n x,n z,n t)表示对噪声信号进行采样得到的数据的集合。
S123、将所述数据矩阵S(n x,n z,n t)转换为以Casorati矩阵形式重新排列的二维时空矩阵;
S124、利用奇异值分解关系式S=UΔV *,对所述二维时空矩阵进行奇异值分解,得到奇异值矩阵,其中,S表示所述二维时空矩阵,U和V分别表示不同的正交矩阵,U等于(n x×n z,n x×n z),V等于(n t,n t),*表示共轭转置,Δ表示奇异值矩阵;
S125、基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,删除所述奇异值矩阵中大于设定奇异值阈值的奇异值,得到更新后的奇异值矩阵。
基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,可以确定奇异值矩阵中奇异值越大,表征该奇异值对应的是组织信号,因此,可以删除所述奇异值矩阵中大于设定奇异值阈值的奇异值,得到更新后的奇异值矩阵。更新后的奇异值矩阵中的奇异值对应的是造影剂信号。
S126、将所述更新后的奇异值矩阵代入所述奇异值分解关系式S=UΔV *,计算得到更新后的二维时空矩阵,将更新后的二维时空矩阵转换为数据矩阵S(n x,n z,n t),将转换得到的数据矩阵S(n x,n z,n t)对应的信号作为造影剂信号。
当然,所述基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,从所述回波反射信号中分离出造影剂信号的过程也可以为:基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息及高通滤波方法,从所述回波反射信号中分离出造影剂信号。
步骤S13、获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的相似度。
本实施例中,PSF(点扩展函数,point spread function)模板图像标定的过程,可以为:将一根细线垂直于内窥环阵超声换能器成像平面放置,以充当理想点散射体,内窥环阵超声换能器采集到的图像作为PSF模板图像。
由于造影剂的大小远小于内窥环阵超声换能器的超声波长,因此独立的造影剂图像也表现为超声系统的点扩散函数(PSF),因此可以通过判断PSF模板图像与造影剂的相似度,来确定造影剂的中心坐标。具体地,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数。
本实施例中,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,可以包括但不局限于:
将PSF模板图像按照设定步长在造影剂图像中平移,每平移一次,计算一次PSF模板图像与造影剂图像中第一区域的互相关系数。
在将PSF模板图像按照设定步长在造影剂图像中平移的情况下,第一区域可以理解为:造影剂图像中与包含PSF模板图像的图形(如,外接矩形或外接圆)重叠的区域。
当然,也可以预先对造影剂图像进行划分,得到多个第一区域,并分别计算所述PSF模板图像与造影剂图像中各个第一区域的互相关系数。
需要说明的是,互相关系数越大,PSF模板图像与第一区域的相似度越高。
本实施例中,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,可以包括但不局限于:
利用关系式
Figure PCTCN2020130109-appb-000011
分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数;
其中,f表示所述造影剂图像,t表示所述PSF模板图像,
Figure PCTCN2020130109-appb-000012
表示所述造影剂图像的均值,
Figure PCTCN2020130109-appb-000013
表示所述PSF模板图像的均值,c表示互相关系数,x表示横坐标,y表示纵坐标,f(x,y)表示所述造影剂图像中(x,y)坐标处的像素值,u,v分别表示所述PSF模板图像在所述造影剂图像中沿x轴和y轴的平移量,f u,v表示当所述PSF模板图像在所述造影剂图像中平移时,所述造影剂图像在所述PSF模板图像覆盖区域中的像素值。
步骤S14、将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域。
在第二区域包含至少一个第一区域的情况下,第二区域中每个第一区域均对应有一个互相关系数,则可以从第二区域中选取出最大互相关系数对应的第 一区域,将选取出的第一区域作为目标区域。
步骤S15、通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像,可以理解为:将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到造影剂运动轨迹,将包含造影剂运动轨迹的图像作为超分辨图像。
在本申请中,由于内窥环阵超声换能器的声场能量较高,因此利用环阵超快成像算法向所述成像目标内发射超声信号,并采集所述超声信号的回波反射信号,能够采集到较大范围内的回波反射信号,基于较大范围内回波反射信号能够提高成像时的横向分辨率,在此基础上,由于造影剂信号可以表征成像目标内血流的信号,因此从回波反射信号中分离出造影剂信号,并获取预先标定的PSF模板图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,确定出与PSF模板图像相似度最高的区域,将该区域的中心坐标作为造影剂的中心坐标,能够提高每个造影剂定位的准确性,在此基础上,缩小每个造影剂在图像中的空间,保证超分辨图像的高分辨率,同时,由于并不需要提高内窥环阵超声换能器的频率来换取高分辨率,因此,能够保证大的成像深度。
作为本申请另一可选实施例,参照图4,为本申请提供的一种超分辨成像方法实施例2的流程图,本实施例主要是对上述实施例1描述的超分辨成像方法的扩展方案,如图4所示,该方法可以包括但并不局限于以下步骤:
步骤S21、在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号。
步骤S22、基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,并基于所述造影剂信号,确定造影剂图像。
步骤S23、获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的相似度。
步骤S21-S23的详细过程可以实施例1中步骤S11-S13的相关介绍,在此不再赘述。
步骤S24、分别比较所述造影剂图像中各个第一区域对应的互相关系数与设定互相关系数阈值。
可以理解的是,由于造影剂信号中可能会存在噪声信号,因此造影剂图像中也会相应的存在噪声像素点。这种情况下,可以基于噪声像素点组成的区域与PSF模板图像之间的互相关系数小于包含造影剂的区域与PSF模板图像之间的互相关系数,确定设定互相关系数阈值。
步骤S25、若所述第一区域对应的互相关系数小于所述设定互相关系数阈值,则从所述造影剂图像中删除所述第一区域。
若所述第一区域对应的互相关系数小于所述设定互相关系数阈值,则表征第一区域为噪声像素点组成的区域,因此从造影剂图像中删除第一区域,以删除造影剂图像中的噪声像素点。
步骤S26、将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域。
本步骤中的造影剂图像可以理解为:删除噪声像素点之后的造影剂图像。
步骤S27、通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
步骤S27的详细过程可以参见实施例1中步骤S15的相关介绍,在此不再赘述。
本实施例中,通过分别比较所述造影剂图像中各个第一区域对应的互相关系数与设定互相关系数阈值,若所述第一区域对应的互相关系数小于所述设定互相关系数阈值,则从所述造影剂图像中删除所述第一区域,实现删除造影剂图像中的噪声像素点,并将删除噪声像素点的造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐 标,进一步提高造影剂的中心坐标的准确性,进一步提高超分辨图像的分辨率。
作为本申请另一可选实施例,参照图5,为本申请提供的一种超分辨成像方法实施例3的流程图,本实施例主要是对上述实施例1描述的超分辨成像方法的细化方案,如图5所示,该方法可以包括但并不局限于以下步骤:
步骤S31、在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号;
步骤S32、基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,并分别基于每个所述造影剂信号,确定造影剂图像。
步骤S31-S32的详细过程可以参见实施例1中步骤S11-S12的相关介绍,在此不再赘述。
步骤S33、获取预先标定的不同角度的PSF模板图像。
由于内窥环阵超声换能器是以环阵的形式发射超声信号及采集回波反射信号的,因此内窥环阵超声换能器采集到的不同的回波反射信号中的造影剂信号的角度是不同的,因此,可以针对不同角度,预先标定PSF模板图像。
步骤S34、分别将每个所述造影剂图像划分为不同角度的子造影剂图像。
本实施例中,将所述造影剂图像划分为不同角度的子造影剂图像,可以包括但不局限于:将造影剂图像划分为四个角度的子造影剂图像。如图6所示,以内窥环阵超声换能器为中心,将造影剂图像划分为四个角度的子造影剂图像。
步骤S35、对每个所述角度的子造影剂图像,从不同角度的PSF模板图像中选择与所述子造影剂图像的角度对应的PSF模板图像,作为目标PSF模板图像,分别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数。
分别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数的具体过程可以参见实施例1中分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数的相关介绍,在此不再赘述。
步骤S33-S35为实施例1中步骤S13的一种具体实施方式。
步骤S36、将每个所述子造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域。
步骤S36为实施例1中步骤S14的一种具体实施方式。
步骤S37、通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
步骤S37的详细过程可以参见实施例1中步骤S15的相关介绍,在此不再赘述。
本实施例中,通过获取预先标定的不同角度的PSF模板图像,及将所述造影剂图像划分为不同角度的子造影剂图像,及对每个所述角度的子造影剂图像,从不同角度的PSF模板图像中选择与所述子造影剂图像的角度对应的PSF模板图像,作为目标PSF模板图像,分别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数,能够降低定位偏差,进一步提高造影剂的中心坐标的准确性,从而进一步提高超分辨图像的分辨率。
作为本申请另一可选实施例,参照图7,为本申请提供的一种超分辨成像方法实施例4的流程图,本实施例主要是对上述实施例1描述的超分辨成像方法的扩展方案,如图7所示,该方法可以包括但并不局限于以下步骤:
步骤S41、在成像目标内注射有超声造影剂的情况下,每个设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号。
步骤S42、基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,并分别基于每个所述造影剂信号,确定造影剂图像。
步骤S43、获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的 相似度。
步骤S44、将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域。
步骤S45、通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
步骤S41-S45的详细过程可以参见实施例1中步骤S11-S15的相关介绍,在此不再赘述。
步骤S46、利用最邻近匹配算法,对相邻的两帧超分辨图像中所述造影剂的中心坐标对应的位置进行匹配,并计算匹配的两个所述造影剂的中心坐标对应的位置之间的距离。
本实施例中,最邻近匹配算法可以为粒子追踪测速(PTV)中的最邻近匹配算法(Kuhn-Munkras匹配算法)。
步骤S47、基于匹配的两个所述造影剂的中心坐标对应的位置之间的距离,计算所述造影剂的运动速度。
本实施例中,造影剂的运动速度可以作为成像目标内的血流速度。
本实施例中,在得到超分辨图像之后,进一步计算出成像目标内的血流速度,完善内窥环阵超声换能器的功能,为医疗检查提供更多的信息,推进医疗进步。
接下来对本申请提供的超分辨成像系统进行介绍,下文介绍的超分辨成像系统与上文介绍的超分辨成像方法可相互对应参照。
请参见图8,超分辨成像系统应用于内窥环阵超声换能器,超分辨成像系统包括:信号发射及采集模块100、信号分离模块200、第一确定模块300、第一计算模块400、第二确定模块500和成像模块600。
信号发射及采集模块,用于在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号;
信号分离模块,用于基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号
第一确定模块,用于分别基于每个所述造影剂信号,确定造影剂图像;
第一计算模块,用于获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的相似度;
第二确定模块,用于将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域;
成像模块,用于通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
本实施例中,超分辨成像系统还可以包括:
比较模块,用于分别比较所述造影剂图像中各个第一区域对应的互相关系数与设定互相关系数阈值;
删除模块,用于若所述第一区域对应的互相关系数小于所述设定互相关系数阈值,则从所述造影剂图像中删除所述第一区域。
本实施例中,所述第一计算模块,具体可以用于:
获取预先标定的不同角度的PSF模板图像;
将每帧所述造影剂图像划分为不同角度的子造影剂图像;
对每个所述角度的子造影剂图像,从不同角度的PSF模板图像中选择与所述子造影剂图像的角度对应的PSF模板图像,作为目标PSF模板图像,分别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数;
所述第二确定模块,具体可以用于:
将每个所述子造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域。
所述第一计算模块,具体可以用于:
利用关系式
Figure PCTCN2020130109-appb-000014
分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数;
其中,f表示所述造影剂图像,t表示所述PSF模板图像,
Figure PCTCN2020130109-appb-000015
表示所述造影剂图像的均值,
Figure PCTCN2020130109-appb-000016
表示所述PSF模板图像的均值,c表示互相关系数,x表示横坐标,y表示纵坐标,f(x,y)表示所述造影剂图像中(x,y)坐标处的像素值,u,v分别表示所述PSF模板图像在所述造影剂图像中沿x轴和y轴的平移量,f u,v表示当所述PSF模板图像在所述造影剂图像中平移时,所述造影剂图像在所述PSF模板图像覆盖区域中的像素值。
所述信号分离模块,具体可以用于:
对每个所述回波反射信号进行波束合成,得到波束合成信号,并对每个所述波束合成信号进行正交解调,得到正交解调信号;
对多个所述正交解调信号进行采样,将采样得到的数据组成数据矩阵S(n x,n z,n t),n x表示沿横轴方向对每个所述正交解调信号采样得到的数据,n z表示沿纵轴方向对每个所述正交解调信号采样得到的数据,n t表示多个所述正交解调信号的个数;
将所述数据矩阵S(n x,n z,n t)转换为以Casorati矩阵形式重新排列的二维时空矩阵;
利用奇异值分解关系式S=UΔV *,对所述二维时空矩阵进行奇异值分解,得到奇异值矩阵,其中,S表示所述二维时空矩阵,U和V分别表示不同的正交矩阵,U等于(n x×n z,n x×n z),V等于(n t,n t),*表示共轭转置,Δ表示奇异值矩阵;
基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,删除所述奇异值矩阵中大于设定奇异值阈值的奇异值,得到更新后的奇异值矩阵;
将所述更新后的奇异值矩阵代入所述奇异值分解关系式S=UΔV *,计算得到更新后的二维时空矩阵,将更新后的二维时空矩阵转换为数据矩阵S(n x,n z,n t),将转换得到的数据矩阵S(n x,n z,n t)对应的信号作为造影剂信号。
需要说明的是,每个实施例重点说明的都是与其他实施例的不同之处,各个实施例之间相同相似的部分互相参见即可。对于装置类实施例而言,由于其与方法实施例基本相似,所以描述的比较简单,相关之处参见方法实施例的部分说明即可。
最后,还需要说明的是,在本文中,诸如第一和第二等之类的关系术 语仅仅用来将一个实体或者操作与另一个实体或操作区分开来,而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括所述要素的过程、方法、物品或者设备中还存在另外的相同要素。
为了描述的方便,描述以上装置时以功能分为各种单元分别描述。当然,在实施本申请时可以把各单元的功能在同一个或多个软件和/或硬件中实现。
通过以上的实施方式的描述可知,本领域的技术人员可以清楚地了解到本申请可借助软件加必需的通用硬件平台的方式来实现。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品可以存储在存储介质中,如ROM/RAM、磁碟、光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本申请各个实施例或者实施例的某些部分所述的方法。
以上对本申请所提供的一种超分辨成像方法及系统进行了详细介绍,本文中应用了具体个例对本申请的原理及实施方式进行了阐述,以上实施例的说明只是用于帮助理解本申请的方法及其核心思想;同时,对于本领域的一般技术人员,依据本申请的思想,在具体实施方式及应用范围上均会有改变之处,综上所述,本说明书内容不应理解为对本申请的限制。

Claims (10)

  1. 一种超分辨成像方法,其特征在于,应用于内窥环阵超声换能器,该方法包括:
    在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号;
    基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,并分别基于每个所述造影剂信号,确定造影剂图像;
    获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的相似度;
    将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域;
    通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
  2. 根据权利要求1所述的方法,其特征在于,所述分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数之后,还包括:
    分别比较所述造影剂图像中各个第一区域对应的互相关系数与设定互相关系数阈值;
    若所述第一区域对应的互相关系数小于所述设定互相关系数阈值,则从所述造影剂图像中删除所述第一区域。
  3. 根据权利要求1所述的方法,其特征在于,所述获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,包括:
    获取预先标定的不同角度的PSF模板图像;
    将每帧所述造影剂图像划分为不同角度的子造影剂图像;
    对每个所述角度的子造影剂图像,从不同角度的PSF模板图像中选择与所述子造影剂图像的角度对应的PSF模板图像,作为目标PSF模板图像,分 别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数;
    所述将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,包括:
    将每个所述子造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域。
  4. 根据权利要求1或2所述的方法,其特征在于,所述分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,包括:
    利用关系式
    Figure PCTCN2020130109-appb-100001
    分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数;
    其中,f表示所述造影剂图像,t表示所述PSF模板图像,
    Figure PCTCN2020130109-appb-100002
    表示所述造影剂图像的均值,
    Figure PCTCN2020130109-appb-100003
    表示所述PSF模板图像的均值,c表示互相关系数,x表示横坐标,y表示纵坐标,f(x,y)表示所述造影剂图像中(x,y)坐标处的像素值,u,v分别表示所述PSF模板图像在所述造影剂图像中沿x轴和y轴的平移量,f u,v表示当所述PSF模板图像在所述造影剂图像中平移时,所述造影剂图像在所述PSF模板图像覆盖区域中的像素值。
  5. 根据权利要求1所述的方法,其特征在于,所述基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号,包括:
    对每个所述回波反射信号进行波束合成,得到波束合成信号,并对每个所述波束合成信号进行正交解调,得到正交解调信号;
    对多个所述正交解调信号进行采样,将采样得到的数据组成数据矩阵S(n x,n z,n t),n x表示沿横轴方向对每个所述正交解调信号采样得到的数据,n z表示沿纵轴方向对每个所述正交解调信号采样得到的数据,n t表示多个所述正交解调信号的个数;
    将所述数据矩阵S(n x,n z,n t)转换为以Casorati矩阵形式重新排列的二维时空矩阵;
    利用奇异值分解关系式S=UΔV *,对所述二维时空矩阵进行奇异值分解, 得到奇异值矩阵,其中,S表示所述二维时空矩阵,U和V分别表示不同的正交矩阵,U等于(n x×n z,n x×n z),V等于(n t,n t),*表示共轭转置,Δ表示奇异值矩阵;
    基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,删除所述奇异值矩阵中大于设定奇异值阈值的奇异值,得到更新后的奇异值矩阵;
    将所述更新后的奇异值矩阵代入所述奇异值分解关系式S=UΔV *,计算得到更新后的二维时空矩阵,将更新后的二维时空矩阵转换为数据矩阵S(n x,n z,n t),将转换得到的数据矩阵S(n x,n z,n t)对应的信号作为造影剂信号。
  6. 一种超分辨成像系统,其特征在于,应用于内窥环阵超声换能器,该系统包括:
    信号发射及采集模块,用于在成像目标内注射有超声造影剂的情况下,每隔设定时间利用环阵超快成像算法向所述成像目标内发射一次超声信号,并采集所述超声信号的回波反射信号;
    信号分离模块,用于基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,分别从每个所述回波反射信号中分离出造影剂信号;
    第一确定模块,用于分别基于每个所述造影剂信号,确定造影剂图像;
    第一计算模块,用于获取预先标定的PSF模板图像,并对每帧所述造影剂图像,分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数,所述互相关系数用于表征所述PSF模板图像与所述造影剂图像中第一区域的相似度;
    第二确定模块,用于将所述造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域,将每个所述目标区域的中心坐标,分别作为每个造影剂的中心坐标,所述第二区域包含至少一个所述第一区域;
    成像模块,用于通过将多帧所述造影剂图像中每个所述造影剂的中心坐标对应的位置累加,得到超分辨图像。
  7. 根据权利要求6所述的系统,其特征在于,所述系统还包括:
    比较模块,用于分别比较所述造影剂图像中各个第一区域对应的互相关系数与设定互相关系数阈值;
    删除模块,用于若所述第一区域对应的互相关系数小于所述设定互相关系 数阈值,则从所述造影剂图像中删除所述第一区域。
  8. 根据权利要求6所述的系统,其特征在于,所述第一计算模块,具体用于:
    获取预先标定的不同角度的PSF模板图像;
    将每帧所述造影剂图像划分为不同角度的子造影剂图像;
    对每个所述角度的子造影剂图像,从不同角度的PSF模板图像中选择与所述子造影剂图像的角度对应的PSF模板图像,作为目标PSF模板图像,分别计算所述目标PSF模板图像与所述子造影剂图像中各个第一区域的互相关系数;
    所述第二确定模块,具体用于:
    将每个所述子造影剂图像中各个第二区域中最大互相关系数对应的第一区域作为目标区域。
  9. 根据权利要求6或7所述的系统,其特征在于,所述第一计算模块,具体用于:
    利用关系式
    Figure PCTCN2020130109-appb-100004
    分别计算所述PSF模板图像与所述造影剂图像中各个第一区域的互相关系数;
    其中,f表示所述造影剂图像,t表示所述PSF模板图像,
    Figure PCTCN2020130109-appb-100005
    表示所述造影剂图像的均值,
    Figure PCTCN2020130109-appb-100006
    表示所述PSF模板图像的均值,c表示互相关系数,x表示横坐标,y表示纵坐标,f(x,y)表示所述造影剂图像中(x,y)坐标处的像素值,u,v分别表示所述PSF模板图像在所述造影剂图像中沿x轴和y轴的平移量,f u,v表示当所述PSF模板图像在所述造影剂图像中平移时,所述造影剂图像在所述PSF模板图像覆盖区域中的像素值。
  10. 根据权利要求6所述的系统,其特征在于,所述信号分离模块,具体用于:
    对每个所述回波反射信号进行波束合成,得到波束合成信号,并对每个所述波束合成信号进行正交解调,得到正交解调信号;
    对多个所述正交解调信号进行采样,将采样得到的数据组成数据矩阵S(n x,n z,n t),n x表示沿横轴方向对每个所述正交解调信号采样得到的数据,n z表 示沿纵轴方向对每个所述正交解调信号采样得到的数据,n t表示多个所述正交解调信号的个数;
    将所述数据矩阵S(n x,n z,n t)转换为以Casorati矩阵形式重新排列的二维时空矩阵;
    利用奇异值分解关系式S=UΔV *,对所述二维时空矩阵进行奇异值分解,得到奇异值矩阵,其中,S表示所述二维时空矩阵,U和V分别表示不同的正交矩阵,U等于(n x×n z,n x×n z),V等于(n t,n t),*表示共轭转置,Δ表示奇异值矩阵;
    基于组织信号之间的相关性大于造影剂信号之间的相关性的关系信息,删除所述奇异值矩阵中大于设定奇异值阈值的奇异值,得到更新后的奇异值矩阵;
    将所述更新后的奇异值矩阵代入所述奇异值分解关系式S=UΔV *,计算得到更新后的二维时空矩阵,将更新后的二维时空矩阵转换为数据矩阵S(n x,n z,n t),将转换得到的数据矩阵S(n x,n z,n t)对应的信号作为造影剂信号。
PCT/CN2020/130109 2020-11-19 2020-11-19 一种超分辨成像方法及系统 WO2022104648A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/130109 WO2022104648A1 (zh) 2020-11-19 2020-11-19 一种超分辨成像方法及系统

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/130109 WO2022104648A1 (zh) 2020-11-19 2020-11-19 一种超分辨成像方法及系统

Publications (1)

Publication Number Publication Date
WO2022104648A1 true WO2022104648A1 (zh) 2022-05-27

Family

ID=81708171

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/130109 WO2022104648A1 (zh) 2020-11-19 2020-11-19 一种超分辨成像方法及系统

Country Status (1)

Country Link
WO (1) WO2022104648A1 (zh)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6004270A (en) * 1998-06-24 1999-12-21 Ecton, Inc. Ultrasound system for contrast agent imaging and quantification in echocardiography using template image for image alignment
JP3347385B2 (ja) * 1992-03-27 2002-11-20 オリンパス光学工業株式会社 内視鏡画像処理装置
CN102551793A (zh) * 2010-12-30 2012-07-11 深圳迈瑞生物医疗电子股份有限公司 超声高帧率组织多普勒成像方法及装置
CN103876776A (zh) * 2012-12-24 2014-06-25 深圳迈瑞生物医疗电子股份有限公司 一种超声造影成像方法及装置
CN107361791A (zh) * 2017-07-21 2017-11-21 北京大学 一种快速超分辨血流成像方法
US20190365355A1 (en) * 2017-01-18 2019-12-05 Technion Research & Development Foundation Ltd. Sparsity-based ultrasound super-resolution imaging
WO2020002952A1 (en) * 2018-06-29 2020-01-02 King's College London Ultrasound method and apparatus
CN110772285A (zh) * 2019-10-31 2020-02-11 南京景瑞康分子医药科技有限公司 一种超声超分辨成像方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3347385B2 (ja) * 1992-03-27 2002-11-20 オリンパス光学工業株式会社 内視鏡画像処理装置
US6004270A (en) * 1998-06-24 1999-12-21 Ecton, Inc. Ultrasound system for contrast agent imaging and quantification in echocardiography using template image for image alignment
CN102551793A (zh) * 2010-12-30 2012-07-11 深圳迈瑞生物医疗电子股份有限公司 超声高帧率组织多普勒成像方法及装置
CN103876776A (zh) * 2012-12-24 2014-06-25 深圳迈瑞生物医疗电子股份有限公司 一种超声造影成像方法及装置
US20190365355A1 (en) * 2017-01-18 2019-12-05 Technion Research & Development Foundation Ltd. Sparsity-based ultrasound super-resolution imaging
CN107361791A (zh) * 2017-07-21 2017-11-21 北京大学 一种快速超分辨血流成像方法
WO2020002952A1 (en) * 2018-06-29 2020-01-02 King's College London Ultrasound method and apparatus
CN110772285A (zh) * 2019-10-31 2020-02-11 南京景瑞康分子医药科技有限公司 一种超声超分辨成像方法

Similar Documents

Publication Publication Date Title
US11096662B2 (en) Method and apparatus to produce ultrasonic images using multiple apertures
CN112450983B (zh) 一种超分辨成像方法及系统
JP5627890B2 (ja) 最適なスペックル追跡のための二重経路処理
US9345446B2 (en) Ultrasonic diagnosis apparatus, image processing apparatus, and image processing method
US6464638B1 (en) Ultrasound imaging system and method for spatial compounding
JP2801396B2 (ja) エコーグラフィック画像生成装置及び方法
US7758508B1 (en) Ultrasound-imaging systems and methods for a user-guided three-dimensional volume-scan sequence
US20050277835A1 (en) Ultrasound imaging by nonlinear low frequency manipulation of high frequency scattering and propagation properties
WO2006009469A2 (en) Ultrasound imaging using non-linear manipulation of forward propagation
JP2009505771A (ja) 実時間空間合成によるフローイメージングのための超音波画像形成システム及び方法
JP2005152648A (ja) 運動適応型空間合成の方法及びシステム
CN104272134A (zh) 超声成像系统中的杂波抑制
US11766297B2 (en) Apparatus and method for detecting an interventional tool
CN110101409B (zh) 波束合成方法、超声成像方法、装置及设备
JP2016153005A (ja) 超音波診断装置、画像処理装置及び画像処理プログラム
CN107278271A (zh) 超声成像系统中的杂波抑制
US9149256B2 (en) Ultrasound strain imaging based on lateral displacement compensation
WO2022104648A1 (zh) 一种超分辨成像方法及系统
CN112450973B (zh) 一种基于行列寻址环形超声换能器的成像方法及装置
WO2022104677A1 (zh) 一种基于行列寻址环形超声换能器的成像方法及装置
Chen et al. A 3D motion compensation method for high frame rate volumetric ultrasound imaging based on velocity vector estimation: a simulation study
JP2020526268A (ja) 超音波信号データを処理する方法及び装置
JP7345678B2 (ja) 3dベクトルフロー場を取得するための方法及びシステム
US20220287685A1 (en) Method and system for estimating motion from overlapping multiline acquisitions of successive ultrasound transmit events
WO2017137807A1 (en) Method and system for generating a compound image

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20961939

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20961939

Country of ref document: EP

Kind code of ref document: A1