JP2006068351A - Medical image processing method, medical image processing program, and medical image processing device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device, program, and device for medical image processing capable of automatically extracting point row data of the line core of all volume data on a time axis. <P>SOLUTION: A volume data group plurally present in a time axis direction is stored, volume data are taken out at a first time and a second time from the same, the core line point row data at the first time of the tubular structure of a subject are stored, thereby tomogram data of the tubular structure in the same cross section as a perpendicular cross section are acquired from the tomogram data of the tubular structure in the perpendicular cross section in a tangential direction of the core line and the volume data at the second time on the respective core points of the core line point row data at the first time, and a vector in which the correlation of the tomogram data at the first time and the second time of the tubular structure becomes maximum is calculated, the respective points of the core line point row data at the first time are referred to as the points corresponding to the core line point row data as the second time by the calculated vector, and the core line point row data at the second time are obtained by determining moving vectors on all the points. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus, and in particular, a lumen such as a luminal organ using three-dimensional volume data of a subject obtained from an imaging apparatus such as an X-ray CT apparatus or a magnetic resonance diagnostic apparatus (MRI apparatus). The present invention relates to a medical image processing method, a medical image processing method, and the like having a function of performing shape display and quantitative analysis.

  In recent medical diagnosis, it is required to quantitatively analyze the shape and movement of a luminal organ such as a blood vessel of a subject. In these cases, the three-dimensional volume data of the part is acquired by a medical modality using an X-ray CT apparatus or an MRI apparatus, and image processing is performed.

  In such image processing, it is necessary to know the position of a three-dimensional path line (for example, the center line of a blood vessel) of a hollow organ. As a method of extracting the core line of the luminal organ using the three-dimensional volume data, a method of searching the inside of the organ using information on luminance values (see Non-Patent Document 1), or a thinned area of the luminal organ is thinned. And the like (see Non-Patent Document 2).

  However, in the method described in Non-Patent Document 1, if there is a low-contrast portion or a small-diameter portion of the luminal organ during the search, there is a high possibility that the search will be unsuccessful because the direction is shifted there. There is.

  Further, in the method described in Non-Patent Document 2, if the accuracy of extracting the region of the luminal organ is low, the accuracy of the core line obtained by thinning is reduced, or much time is required for thinning in proportion to the number of pixels in the extraction region. There is a problem that requires.

  The present applicant has previously filed an application for an image processing apparatus that sets the core wire of a tubular structure from stereoscopic image data (Japanese Patent Application No. 2003-25260).

However, when using the method of obtaining the core point sequence data of a desired blood vessel region from one reconstructed volume data, the operator designates the start point and end point of the blood vessel for all volume data in the time axis direction. There must be. This is because, when a coronary artery or the like is a target, the blood vessel position moves with the movement of the heart, so the start point and end point specified by one volume data cannot be applied to other volume data on the time axis. This operation by the operator has a very time-consuming problem.
O. Wink, WJ Niessen, MA Viergever, “Fast Delineation and Visualization of Vessels in 3-D Angiographic Images” IEEE Trans. Med. Imaging, Vol. 19, No. 4, pp. 337-346, Apr., 2000. GD Rubin, DS Paik, PC Johnston, S. Napel, "Measurement of the Aorta and Its Branches with Helical CT" Radiology, Vol. 206, No. 3, pp. 823-829, Mar., 1998.

  The present invention has been made in view of the problem of the conventional method for three-dimensionally extracting the core line of the above-described luminal organ. Among the reconstructed volume data groups continuous in the time axis direction, volume data is provided. Provided are a medical image processing apparatus, a medical image processing method, and the like that can automatically extract core line point sequence data of all volume data on a time axis by only obtaining the core line point sequence data of one desired blood vessel above. For the purpose.

  According to claim 1 of the present invention, a volume data storage step for storing a plurality of volume data groups in the time axis direction, and volume data acquisition for extracting volume data at the first time and the second time from the volume data group. A core point sequence data storage step for storing the core point sequence data at the first time of the tubular structure of the subject, and the core point at the first time based on the volume data at the first time For each core point of the row data, the cross section of the tubular structure in the same cross section as the orthogonal cross section based on the cross sectional image data of the tubular structure in the tangential cross section of the core wire and the volume data at the second time An orthogonal cross-sectional image acquisition step for obtaining image data; and the disconnection of the tubular structure at the first time and the second time. A moving vector calculating step for calculating a vector that maximizes the correlation of the image data, and the points calculated in the first point of the core point sequence data by using the vector calculated in the moving vector calculating step. A second time core point sequence data calculation step for obtaining the core point sequence data at the second time by obtaining the movement vector for all points as the corresponding points of the point sequence data. A medical image processing method is provided.

  According to claim 2 of the present invention, a volume data storage step for storing a plurality of volume data groups in the time axis direction, a tn volume data acquisition step for extracting volume data at a predetermined time tn from the volume data group, A t (n + 1) volume data acquisition step for extracting volume data at a time t (n + 1) next to a predetermined time tn from the volume data group, and a core wire for storing the core point sequence data at the time tn of the tubular structure of the subject Based on the point sequence data storage step and the volume data at the time tn, for each core point Ptni of the core line point sequence data at the time tn, the cross-sectional image data of the tubular structure in the cross section perpendicular to the tangential direction of the core line is obtained. tn orthogonal cross-sectional image acquisition step and the time t (n + 1) T (n + 1) orthogonal cross-sectional image acquisition step for obtaining cross-sectional image data of the tubular structure in the same cross section as the orthogonal cross-section based on the volume data of the tn, the tn orthogonal cross-sectional image acquisition step and the t (n + 1) orthogonal cross-sectional image A movement vector calculation step for calculating a correlation function of cross-sectional image data of the tubular structure obtained at the acquisition step at time tn and time t (n + 1) and calculating a maximum vector thereof, and calculation by this movement vector calculation step It is assumed that each point Ptni of the core line point sequence data at time tn corresponds to the core line point sequence data Pt (n + 1) i at time t (n + 1) by the calculated vector, and the movement vector is obtained for all points to obtain time t T (n + 1) skeleton point sequence data calculation step for obtaining skeleton point sequence data at (n + 1) , To provide a medical image processing method characterized by comprising a.

  According to claim 4 of the present invention, a volume data storage step for storing a plurality of volume data groups in the time axis direction in a computer, and volume data at a first time and a second time are extracted from the volume data group. Based on the volume data acquisition step, the core point sequence data storage step for storing the core point sequence data at the first time of the tubular structure of the subject, and the volume data at the first time, the first time For each core point of the core point sequence data, the tubular structure in the same cross section as the orthogonal cross section based on the cross-sectional image data of the tubular structure in the cross section orthogonal to the tangential direction of the core wire and the volume data at the second time An orthogonal cross-sectional image acquisition step for obtaining cross-sectional image data of an object, the first time of the tubular structure, and the second A movement vector calculation step for calculating a vector that maximizes the correlation of the cross-sectional image data at the time, and each point of the core point sequence data at the first time by the vector calculated by the movement vector calculation step. A second time core line point sequence data calculating step for obtaining the core point sequence data at the second time by obtaining the movement vector for all points as corresponding points of the core point sequence data at the time, and executing A medical image processing program is provided.

  According to claim 5 of the present invention, volume data storage means for storing a plurality of volume data groups in the time axis direction, and volume data acquisition for extracting volume data at the first time and the second time from the volume data group Means, a core point sequence data storage unit for storing the core point sequence data at the first time of the tubular structure of the subject, and the core point at the first time based on the volume data at the first time For each core point of the row data, the cross section of the tubular structure in the same cross section as the orthogonal cross section based on the cross sectional image data of the tubular structure in the tangential cross section of the core wire and the volume data at the second time Cross-sectional image acquisition means for obtaining image data, and correlation between cross-sectional image data at the first time and the second time of the tubular structure Correspondence between the movement vector calculation means for calculating the maximum vector, and the points of the core point sequence data at the first time corresponding to the points calculated at the second time by the vector calculated by the movement vector calculation means And a second time core point sequence data calculation means for obtaining the core point sequence data at the second time by obtaining the movement vector for all points. A processing device is provided.

  According to claim 6 of the present invention, volume data storage means for storing a plurality of volume data groups in the time axis direction, tn volume data acquisition means for extracting volume data at a predetermined time tn from the volume data, and the volume T (n + 1) volume data acquisition means for extracting volume data at a time t (n + 1) next to a predetermined time tn from the data group, and a core point for storing the core point sequence data at the time tn of the tubular structure of the subject Based on the column data storage means and the volume data at the time tn, for each core point Ptni of the core line point sequence data at the time tn, the cross-sectional image data of the tubular structure in the cross section orthogonal to the tangential direction of the core line is obtained. The orthogonal cross-section image acquisition means and the volume data at time t (n + 1) Next, the t (n + 1) orthogonal cross-sectional image acquisition means for obtaining the cross-sectional image data of the tubular structure in the same cross section as the orthogonal cross-section, the tn orthogonal cross-sectional image acquisition means, and the t (n + 1) orthogonal cross-sectional image acquisition means. A moving vector calculating means for calculating a correlation vector of cross-sectional image data at time tn and time t (n + 1) of the tubular structure thus obtained and calculating a maximum vector thereof, and a vector calculated by the moving vector calculating means. Assume that each point Ptni of the skeleton point sequence data at time tn corresponds to the skeleton point sequence data Pt (n + 1) i at time t (n + 1), and the movement vectors are obtained for all points at time t (n + 1). And t (n + 1) core point sequence data calculation means for obtaining core point sequence data, To provide.

  According to the present invention, it can be obtained simply by extracting one core line on all volume data in the time axis direction necessary for creating a continuous image along the core line of a desired blood vessel. A medical image processing apparatus, a medical image processing method, and the like that do not cost can be obtained.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 shows a configuration example of a medical image processing apparatus according to an embodiment of the present invention. In this example, the case where the luminal organ is a blood vessel will be described.

  As will be described later, the image processing apparatus 10 includes a four-dimensional core line point sequence data storage unit 11 in which point sequence data obtained in advance is stored, and a time tn from the four-dimensional core line point sequence data storage unit 11. Tn core point sequence data storage unit 12 that extracts and stores the point sequence data of the core line, a four-dimensional volume data storage unit 13 that stores four-dimensional volume data, and a time tn from the four-dimensional volume data storage unit 13 Tn volume data storage unit 14 for retrieving and storing the volume data of t, tn + 1 volume data storage unit 15 for retrieving and storing the volume data at time tn + 1 from the four-dimensional volume data storage unit 13, and a core line orthogonal section of the core point Pi at time tn An orthogonal cross-section image creation unit 16 that creates image data, and the orthogonal cross-section image creation unit 16 The two-section correlation function calculation unit 17 that calculates the correlation function r between the slice images at the times tn and tn + 1 created in this way and calculates the vector v (l, m) that maximizes the correlation function r, and the point of one core line at this time A single core point specifying unit 18 to be specified and a tn + 1 core point sequence data storage unit 19 for storing the core point Pi at the time tn + 1 specified thereby as the core point at the time tn + 1 are included.

  The four-dimensional volume data (a group of volume data having a plurality in the time axis direction) handled in the present invention will be described here with reference to FIG. It scans 4D raw data including time elements and reconstructs it to create 4D volume data. Specifically, in FIG. 2, the raw data of Pi1 at times t1 to tmax is acquired and reconstructed to obtain the three-dimensional volume data F (X, Y, Z, t) at each time in the Δt time interval. create.

  The three-dimensional volume data is, for example, a predetermined number (for example, 512) of three-dimensional data in the X, Y, and Z directions, and these obtained at each time of t1 to tn to tmax are added with an element of time t. 4D volume data. Therefore, the three-dimensional volume data at time t1 can be expressed as, for example, F (X, Y, Z, t1). Similarly, the three-dimensional volume data at time tn is expressed as F (X, Y, Z, tn). Furthermore, the three-dimensional volume data at time tmax can be expressed as F (X, Y, Z, tmax). Such four-dimensional volume data is stored in the four-dimensional volume data storage unit 13 shown in FIG.

  On the other hand, the four-dimensional core line point sequence data storage unit 11 stores in advance the point sequence data of the core lines P1 to Pmax at time tn.

  In the present invention, since it is assumed that the original core wire is obtained in advance, an example of how to obtain the point sequence data of the original core wire will be described with reference to FIG.

  3 includes a stereoscopic image data storage unit 31, a start point / end point designation unit 32, a region extraction unit 33, a blood vessel core group extraction unit 34, a contour extraction unit 39, and a composite display unit 40d, and includes a blood vessel core extraction unit. 34 includes a core line search unit 35 having an initial search direction determination unit 36, a core line selection / connection unit 37, and a core line smoothing unit 38, and the contour extraction unit 39 includes a contour smoothing unit 40. In the blood vessel core group extracting unit 34, one core data is first extracted.

  The stereoscopic image data storage unit 31 stores three-dimensional image data obtained from an imaging apparatus such as an X-ray CT apparatus or an MRI apparatus. The start point / end point designation unit 32 is functionally configured with an arithmetic device, a display device, and an input device as main elements. In the start point / end point designation unit 32, the operator inputs the coordinates of one start point and one or more end points designated as the blood vessel extraction target range on the stereoscopic image data, and stores them.

  For example, one start point S and two end points E1 and E2 in the Y-shaped blood vessel are designated. Even when three or more end points are designated, it can be applied by slightly changing the processing of the core line selection / connection unit 37 described later. As a modification, it is possible to provide not only the start point / end point but also a start point / end point / pass point specifying unit for specifying a plurality of pass points within the extraction target range of the luminal organ. It can be applied by making a slight change to the processing of the part.

  In the region extraction unit 33, a blood vessel region is three-dimensionally extracted using the three-dimensional image data as input data. That is, three-dimensional image data binarized inside and outside the blood vessel region is created and stored.

  Specifically, this automatic extraction of the blood vessel region is specifically described in, for example, PCNN (Pulse Coupled Neural Networks) described in the paper “T. Lindblad and J. Kinser,“ Image Processing using Pulse Coupled Neural Networks ”Springer, 1998”. Or simple threshold method.

  As an example of the simple threshold method, there is a processing method in which the CT value of the input volume data is 1 (blood vessel region) when the threshold value range is represented by the following formula, and 0 otherwise.

k _min · min (V _s , V _E1 , V _E2 ) ≦ V ≦ k _max · max (V _s , V _E1 , V _E2 )
Here, V is the CT value of a certain pixel, V _s , V _E1 , and V _E2 are the CT values of the start and end points (S, E1, E2) that the operator first specified, and k _min , k _max are These are constants that determine the lower limit and the upper limit, respectively.

  The blood vessel core line group extraction unit 34 includes a core line search unit 35, a core line selection / connection unit 37, and a core line smoothing unit 38, and assigned processes are executed in this order. The core line search unit 35 includes an initial search direction determination unit 36. The core line search unit 35 uses the designated points designated by the start point / end point designation unit 32 as a starting point, and the three-dimensional volume data of the blood vessel region as the luminal organ region extracted by the region extraction unit 33. A search process to be described later is performed, and the search result is selected and connected by the core selection / connection unit 37. As a result, the core line of the I-shaped lumen organ or the Y-shaped lumen organ is extracted.

  The core line search unit 35 searches for a point group of the core line of the luminal organ from the starting point to the target point. Specifically, the initial search direction determination unit 36 determines a direction to search first (initial search direction vector), and searches the inside of the blood vessel extraction region by the iterative search process.

  That is, the initial search direction determination unit 36 sets the unit vector in the three-dimensional all-angle (4π) direction from the start point / end point and the vector in the opposite direction as the initial search vector.

  Next, the core line search unit 35 repeatedly performs the search process. In this embodiment, two methods that can be selected are prepared for this search process.

  The first method is a method of obtaining the position of the center of gravity of the blood vessel extraction region in the cross section orthogonal to the search vector as the position of the blood vessel core line. In the second method, the position where the Euclidean distance in the blood vessel extraction region in the cross section orthogonal to the search vector is the maximum is obtained as the blood vessel core line position. Specifically, in the cross-sectional image of the region extraction data that is orthogonal to the search vector and includes the vector tip position, a 1-pixel region connected to the pixel at the vector tip position is extracted and a hole ( Euclidean distance conversion is performed after the (0-pixel area) is subjected to the hole filling process. This Euclidean distance transformation is, for example, “Toyofumi Saito, Junichiro Toriwaki” Euclidean distance transformation for 3D digital images ”, IEICE Transactions, Vol.J76-D-II, No.3, pp.445-453, Mar, 1993 ”.

  This Euclidean distance conversion is a conversion in which the distance (Euclidean distance) between a pixel in the extraction region and a pixel outside the extraction region that is closest to the pixel is the value of the pixel. The value is the maximum value in the extraction area. The position of the pixel having the maximum value is obtained as the position of the core line. When the pixel at the vector tip position is 0-pixel, an area including 1-pixel at the position closest to the vector tip position is used.

  When the search for the core line ends in this way, the core line search unit 35 obtains the next vector (next orthogonal cross section).

  The calculation of the position of the blood vessel core line and the calculation of the next vector are repeatedly executed. In this iterative process, when the distance between the last core line point and the target point is within a predetermined threshold, the iterative process is stopped because the search is successful.

  The next core line selection / connection unit 37 selects and connects the core line that is the result of the search processing by the core line search unit 35, with each designated point designated by the start point / end point designation unit 32 as a starting point. The core line is also extracted from the tract lumen organ or the Y-shaped tract lumen organ.

  The core line smoothing unit 38 has a function of smoothing the three-dimensional core line extracted by the core line selecting / connecting unit 37 into a smooth curve. That is, the three-dimensional core line extracted by the above-described case division process is appropriately smoothed and shaped into a smooth curve. In the case of extraction between three points, the smoothing process is performed for each branch point.

  The contour extraction unit 39 also includes a contour smoothing unit 40. The contour extraction unit 39 is information on luminance values on a straight line extending in the radial direction from the position of the core line on the cross section orthogonal to the core line extracted by the blood vessel core line extraction unit 34 or the boundary of the region extracted by the region extraction unit 33 Also, the position of the blood vessel contour is detected.

  The contour extracted by the contour extraction unit is smoothed by the contour smoothing unit 40 and shaped into a three-dimensional blood vessel contour having a smooth curve. That is, similar to the smoothing of the core wire, a smoothing process is performed on a series of obtained contour points to form a smooth curve.

  In the composite display unit 40d, the data of the core line extracted by the blood vessel core line extraction unit 34 and the data of the contour extracted by the contour extraction unit 39 are stored in the three-dimensional image data storage unit 31 or its 3D image data. The image data calculated using the dimensional image data is combined and displayed. As described above, one core line of a blood vessel that exists three-dimensionally in a subject at a specific time is obtained as point sequence data.

  Returning to FIG. 1, the subsequent processing after obtaining the original core point sequence data in the image processing apparatus according to the embodiment of the present invention will be described. FIG. 4 shows a flow of processing for obtaining the core point sequence data sequentially from the original core point sequence data, the core point sequence data after Δt time, the core point sequence data after the next Δt time,... It was.

  The three-dimensional volume data at time tn is extracted from the four-dimensional volume data storage unit 13 storing the four-dimensional volume data shown in FIG. 2, and stored in the tn volume data storage unit 14 (step S101). In the next step S101, the three-dimensional volume data at time t (n + 1) is read from the four-dimensional volume data storage unit 13 and stored in the t (n + 1) volume data storage unit 15.

  Next, the core line point sequence data at time tn stored in the four-dimensional core line point sequence data storage unit is extracted and stored in the tn core line point sequence data storage unit 12. The core line point sequence data stored in the tn core line point sequence data storage unit 12 is P (Xtn1, Ytn1, Ztn1) to P (Xtni, Ytni, Ztni) to P (Xtnpmax, Ytnpmax, Ztnpmax). Next, the four-dimensional tn volume data F (X, Y, Z, tn) is extracted from the tn volume data storage unit 14.

  FIG. 5A shows the relationship between the four-dimensional tn volume data F (X, Y, Z, tn) and the core point sequence data at the time tn. In steps S103 to S110, the point P of the core line at time tn is sequentially changed (P1 to Pmax), and the following processing is performed. Here, processing performed as Pi will be described.

  In step S103, one point P (Xtni, Ytni, Ztni) (hereinafter sometimes simply referred to as Ptni) in the core point sequence data at time tn is specified. Next, a tangent vector v (X, Y, Z) that passes through the point Ptni and touches the core lines Ptn1 to Ptnpmax, that is, the core lines P (Xtn1, Ytn1, Ztn1) to P (Xtnpmax, Ytnpmax, Ztnpmax) is obtained (step S104).

  Next, from the tn volume data obtained from the tn volume data storage unit 14, the cross-sectional image data ftni (which passes through the point Ptni and is orthogonal to the tangent vector v (X, Y, Z) in the orthogonal cross-section image creation unit 16. x, y) is acquired (step S105). The x-axis is taken in the direction perpendicular to the tangent at the point Ptni of the core lines Ptn1 to Ptnpmax, and the y-axis is taken in the direction perpendicular to the tangent and the x-axis. An orthogonal cross section with respect to the tangent vector v (X, Y, Z) at the point Ptni is as shown in FIG.

  Next, t (n + 1) volume data F (X, Y, Z, tn + 1) as shown in FIG. 5C is obtained from the t (n + 1) volume data storage unit 15, and a point P (Xtni, Ytni, Ztni) is obtained. The cross-sectional image data orthogonal to the tangent vector v (X, Y, Z) is obtained by the orthogonal cross-sectional image creation unit 16 (step S106).

  As shown in FIG. 5D, the cross-sectional image data ft (n + 1) i (x, y) is the cross-sectional image data obtained at time t (n + 1) on the same orthogonal cross section obtained in the previous step S105. Will be expressed. The origins of these cross-sectional images are all P (Xtni, Ytni, Ztni) = ftni (0, 0).

  The two cross-sectional image data (FIGS. 5B and 5D) obtained in this way are input to the two-section correlation function calculation unit 17, and their correlation function r is obtained (step S107), and the maximum A set (l1, m1) of values (l, m) is obtained (step S108).

The correlation function r (l, m) of the two-dimensional cross-sectional image data shown in FIGS. 5B and 5D is expressed by the following equation.

  In the above equation (1), N is a matrix size in the x and y directions when a two-dimensional cross-sectional image is cut out from three-dimensional volume data, and is, for example, about three times the diameter of the blood vessel cross section.

  Cross-sectional image data at the same position and in the same direction of the two volume data at time tn and time t (n + 1) are compared, and the blood vessel cross-section on the two-dimensional cross-section from time tn to time t (n + 1) at time Δt. In order to identify the direction of movement, the correlation function r is determined in a range where l is -N / 2 to N / 2 and m is in a range -N / 2 to N / 2.

  Finding the set (l1, m1) of (l, m) that maximizes the correlation function r means that the cross-sectional image data at time tn is most similar to the image data moved in this direction. On the two-dimensional cross section, it can be estimated that the blood vessel has moved in the direction of the set (l1, m1). Eventually, it can be estimated that the i-th core point Ptni existing on the two-dimensional cross section has also moved in the direction of (l1, m1).

  After (l1, m1) is obtained as described above, (11, m1) on the ftni (x, y) plane is converted into the (X, Y, Z) coordinate system in the one-core point specifying unit 18. (Step S109). The point P (Xt (n + 1) i, Yt (n + 1) i, Zt (n + 1) i) converted into the (X, Y, Z) coordinate system by the one-core point specifying unit 18 is changed to i in the t (n + 1) volume data. The first core point Pt (n + 1) i is stored in the t (n + 1) core point sequence data storage unit 19. Then, again, the core point Pt (n + 1) (i + 1) after Δt time is obtained from the four-dimensional core line sequence data storage unit 11 for the (i + 1) th core line point Ptn (i + 1) in the tn volume data. In this way, it is checked whether all the core line point sequence data Pt (n + 1) 1 to Pt (n + 1) pmax of the t (n + 1) volume data are obtained in step S110, and if not all have been obtained yet, in step S111. Similarly, i is incremented by 1, and the next point sequence data of the core line in the t (n + 1) volume data is obtained.

  When the core point sequence data Pt (n + 1) 1 to Pt (n + 1) pmax in the t (n + 1) volume data is obtained by repeating steps S103 to S109 in this way, the obtained core point sequence data is then obtained. By repeating steps S102 to S112, the core point sequence data after the next Δt time, that is, at time t (n + 2) is obtained.

  As described above, the core point point sequence data at the next time t (n + 1) is obtained based on the core point point sequence data at time tn, and the core point point sequence data at time t (n + 2) is obtained based on the core point point sequence data. . When all the volume data from time t1 to time tmax have not been obtained yet, n is incremented by 1 in step S113 to obtain the next core point sequence data. By sequentially performing this processing, all volume data from time t1 to time tmax can be obtained, and stored in the four-dimensional skeleton line data storage unit 11.

  By the way, in the description of the above embodiment of the present invention, processing is performed on the premise that four-dimensional volume data at times t1 to tmax continuous in the time axis direction is obtained in advance and stored in the four-dimensional volume data storage unit 13 in advance. I was going.

  However, in the present invention, it is also possible to obtain the next core point sequence data every time data is collected and reconstructed. That is, in this case, a vector that maximizes the correlation function between the previous volume data and the next volume data on the time axis is obtained as described above, and new core point sequence data is immediately obtained. It is also possible to provide the latest core point sequence data to the outside in real time in conjunction with scanning by correcting.

  Further, in the above embodiment, the correlation function of the adjacent time image is calculated for a two-dimensional cross section orthogonal to the tangent vector of each point of the core point sequence data, the maximum value is obtained, and the next core line is calculated from the core sequence data. I was looking for a point. However, according to the present invention, it is also possible to calculate the correlation function for the three-dimensional cross section of the tangent vector at each point of the core point sequence data and obtain the maximum value.

The expression for the correlation function r at this time is as follows.

  Furthermore, the present invention can also obtain the correlation function by extending the dimension in the same manner.

In the description of the above embodiment, the case of obtaining the core point sequence data based on the configuration shown in FIG. 3 has been described. However, the present invention is not limited to obtaining original core point sequence data, and may be obtained by other methods.

The figure which shows the functional structural example of one Embodiment of this invention. The figure for demonstrating the four-dimensional volume data in one Embodiment of this invention. The figure which shows the functional structural example which calculates | requires the core line point sequence data based on one Embodiment of this invention. The figure for demonstrating the flow of the whole process in one Embodiment of this invention. The figure for demonstrating calculating | requiring 1 point after (DELTA) t time corresponding to 1 point of core line point sequence data in one Embodiment of this invention.

Explanation of symbols

10 Image processing apparatus,
11... 4D core point sequence data storage unit,
12 ... tn core point sequence data storage unit,
13... 4D volume data storage unit,
14 ... tn volume data storage unit,
15 ... t (n + 1) volume data storage unit 16 ... orthogonal section image creation unit,
17... Two-section correlation function calculation unit,
18 ... 1 core point identification part,
19... T (n + 1) core point sequence data storage unit.

Claims (6)

A volume data storage step for storing a plurality of volume data groups in the time axis direction;
A volume data acquisition step of extracting volume data at a first time and a second time from the volume data group;
A core point sequence data storage step for storing the core point sequence data at the first time of the tubular structure of the subject;
Based on the volume data at the first time, for each core point of the core point sequence data at the first time, the cross-sectional image data of the tubular structure in the cross section perpendicular to the tangential direction of the core line and the second time An orthogonal cross-sectional image acquisition step for obtaining cross-sectional image data of the tubular structure in the same cross section as the orthogonal cross section based on the volume data of
A movement vector calculating step for calculating a vector that maximizes the correlation between the cross-sectional image data at the first time and the second time of the tubular structure;
Based on the vector calculated in the movement vector calculation step, each point of the core point sequence data at the first time is set as a corresponding point of the core point sequence data at the second time, and the movement vector is obtained for all points. A second time core point sequence data calculation step for obtaining the core point sequence data at the second time,
A medical image processing method comprising: A volume data storage step for storing a plurality of volume data groups in the time axis direction;
A tn volume data acquisition step of extracting volume data at a predetermined time tn from the volume data group;
A t (n + 1) volume data acquisition step of extracting volume data at a time t (n + 1) next to a predetermined time tn from the volume data group;
A core point sequence data storage step for storing the core point sequence data at the time tn of the tubular structure of the subject;
A tn orthogonal cross-sectional image acquisition step for obtaining cross-sectional image data of the tubular structure in a cross-section perpendicular to the tangential direction of the core wire for each core point Ptni of the core wire point sequence data at the time tn based on the volume data at the time tn; ,
T (n + 1) orthogonal cross-sectional image acquisition step for obtaining cross-sectional image data of the tubular structure in the same cross section as the orthogonal cross-section based on the volume data at the time t (n + 1);
A correlation vector of the cross-sectional image data at the time tn and the time t (n + 1) of the tubular structure obtained by the tn orthogonal cross-sectional image acquisition step and the t (n + 1) orthogonal cross-sectional image acquisition step is calculated and the vector becomes the maximum A movement vector calculation step for calculating
It is assumed that each point Ptni of the core line point sequence data at the time tn corresponds to the core line point sequence data Pt (n + 1) i at the time t (n + 1) by the vector calculated by the movement vector calculation step, and the movement is performed for all points. T (n + 1) skeleton point sequence data calculation step for obtaining skeleton point sequence data at time t (n + 1) by obtaining a vector;
A medical image processing method comprising:   The medical image processing method according to claim 2, wherein the cross-sectional image obtained by the tn orthogonal cross-sectional image acquisition step and the cross-sectional image obtained by the t (n + 1) orthogonal cross-sectional image acquisition step are three-dimensional images or more. On the computer,
A volume data storage step for storing a plurality of volume data groups in the time axis direction, a volume data acquisition step for extracting volume data at the first time and the second time from the volume data group, and a tubular structure of the subject. Based on the core point sequence data storage step for storing the core point sequence data at the first time and the volume data at the first time, for each core point of the core point sequence data at the first time, this core line An orthogonal cross-sectional image acquisition step for obtaining cross-sectional image data of the tubular structure in the same cross section as the orthogonal cross section based on the cross-sectional image data of the tubular structure in the orthogonal cross section in the tangential direction and the volume data of the second time; The correlation between the cross-sectional image data at the first time and the second time of the tubular structure is maximized. A movement vector calculation step for calculating a vector, and each point of the skeleton point sequence data at the first time is set as a corresponding point of the skeleton point sequence data at the second time by the vector calculated by the movement vector calculation step. A medical image processing program for executing a second time core point point sequence data calculation step for obtaining core point point sequence data at a second time by obtaining the movement vector for all points. Volume data storage means for storing a plurality of volume data groups in the time axis direction;
Volume data acquisition means for extracting volume data at a first time and a second time from the volume data group;
A core point sequence data storage means for storing the core point sequence data at the first time of the tubular structure of the subject;
Based on the volume data at the first time, for each core point of the core point sequence data at the first time, the cross-sectional image data of the tubular structure in the cross section perpendicular to the tangential direction of the core line and the second time Orthogonal cross-sectional image acquisition means for obtaining cross-sectional image data of the tubular structure in the same cross section as the orthogonal cross section based on the volume data of
Movement vector calculation means for calculating a vector that maximizes the correlation between the cross-sectional image data at the first time and the second time of the tubular structure;
Based on the vector calculated by the movement vector calculation means, each point of the core point sequence data at the first time is set as a corresponding point of the core point sequence data at the second time, and the movement vector is obtained for all points. Second time core point sequence data calculating means for obtaining the core point sequence data at the second time,
A medical image processing apparatus comprising: Volume data storage means for storing a plurality of volume data groups in the time axis direction;
Tn volume data acquisition means for extracting volume data at a predetermined time tn from the volume data group;
T (n + 1) volume data acquisition means for extracting volume data at a time t (n + 1) next to a predetermined time tn from the volume data group;
Core point sequence data storage means for storing the core point sequence data at the time tn of the tubular structure of the subject;
Tn orthogonal cross-section image obtaining means for obtaining cross-sectional image data of the tubular structure in a cross-section perpendicular to the tangential direction of the core wire for each core point Ptni of the core wire point sequence data at the time tn based on the volume data at the time tn; ,
T (n + 1) orthogonal cross-section image acquisition means for obtaining cross-sectional image data of the tubular structure in the same cross section as the orthogonal cross section based on the volume data at the time t (n + 1);
The correlation function of the cross-sectional image data at the time tn and the time t (n + 1) of the tubular structure obtained by the tn orthogonal cross-sectional image acquisition means and the t (n + 1) orthogonal cross-sectional image acquisition means is calculated, and the vector becomes the maximum Movement vector calculation means for calculating
It is assumed that each point Ptni of the core line point sequence data at the time tn corresponds to the core line point sequence data Pt (n + 1) i at the time t (n + 1) by the vector calculated by the movement vector calculation means, and the movement is performed for all points. T (n + 1) skeleton point sequence data calculating means for obtaining skeleton point sequence data at time t (n + 1) by obtaining a vector;
A medical image processing apparatus comprising:

Images