JP3467948B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image display technique, and more particularly to a method and an apparatus for displaying surface information of three-dimensional image data. More specifically, X-ray CT and MR
The present invention relates to image display processing on medical images using I (nuclear magnetic resonance) CT, ultrasonic CT, or the like. 2. Description of the Related Art At present, three-dimensional image processing is becoming mainstream for images used in medical image processing. Display methods of three-dimensional image data are roughly classified into a Volume Rendering method (hereinafter, VR method) and a Surface Rendering method (hereinafter, SR method). In the SR method, only the surface information of three-dimensional data is displayed, so that the amount of data handled in comparison with VR can be reduced. Therefore, when the amount of data such as medical image data is large, the display process is generally performed at high speed. Although possible, it is necessary to extract surface information of the image. [0004] As such an image display device, there is one proposed in Japanese Patent Application Laid-Open No. 62-37782. [0005] X-ray CT, MRIC
In a medical image using T, ultrasonic CT, etc., since it is input data from a non-contact sensor, there are generally many noise components such as speckle noises, and a case where actually necessary information is missing. Therefore, 3
There has been a problem that dimensional surface information cannot be extracted. In view of the above problems, the image processing apparatus of the present invention removes a noise component based on nearby information, interpolates necessary information, and enables appropriate extraction of three-dimensional surface information. With the goal. [0007] In order to solve the above-mentioned problems, an image processing apparatus according to the present invention configures grid points in a three-dimensional space region composed of x, y, and z coordinate axes orthogonal to each other. The feature amount observed at each grid point corresponding to the spatial distribution of the physical properties of the target object existing in the region is set as input three-dimensional pixel data, and all grid points are determined based on the three-dimensional pixel data. In the image processing apparatus that performs processing on the pixel data in and outputs an image, a line segment on the constant x, y, z axis having the grid point as an end point is a line segment perpendicular to the x, y, z coordinate axes. Rotate by a certain angle on each of the three planes including the end points, select a grid point sequence representing the line segments located at each of them, and select a grid point sequence corresponding to the line segment discrimination information for discriminating multiple line segments. A first selector for outputting pixel data of the included grid points, A first integration device that integrates pixel data included in the grid point sequence input from the first selector to obtain a pixel data value of the line segment and outputs the pixel data value together with the line segment identification information; The pixel data values of the plurality of line segments output are compared, and if the comparison result is larger than a predetermined threshold value, the pixel data value of the line segment corresponding to the line segment discrimination information is determined from the larger one. A second selector to be output, and a direction component of the end point is calculated from the plurality of line segment discrimination information input from the second selector, and a plurality of the plurality of input components from the second selector are calculated. A second integration device that integrates pixel data values of line segments and outputs the pixel information at the end points as pixel information, and the direction component and the pixel information input from the second selector as grid point information of the grid points. Accumulate and all Based on the information after a child point information is obtained providing an image output means for performing an image output of an image processing apparatus according to claim. FIG. 1 is a diagram showing an image processing apparatus according to a first embodiment of the present invention. The operation of this device will be described below. In the following description, the input three-dimensional pixel data 1 includes xmax, ymax, zmax in the respective directions of the x, y, and z axes.
The following description is made on the assumption that it is composed of lattice points. First, the coordinates of the center pixel are set as x = 3, y = 3, and z = 3 as representative grid points to be processed. In the image processing apparatus of the present embodiment, when the pixel data of each grid point is processed with the three-dimensional pixel data 1 as an input, the grid point is set as the center, and the xy plane, the yz plane, and the zx A point of 5 × 5 pixels near the plane is sent to the first selector 2. Therefore, when x = 3, y = 3, and z = 3, the following data is sent. Pixel data of x = 3, y = 1-5, z = 1,5 x = 1-5, y = 3, pixel data of z = 1-5 x = 1-5, y = 1-5 , z = 3, in the first selector 2, in the neighborhood 5 × 5 around the central pixel transmitted as shown in FIG. 2 in each plane, each xy plane, yz as shown in FIG. Plane, z-
Grid points constituting eight types, that is, a total of 24 types of search line segments for three planes with respect to the x plane are selected, and the first integration device 3 integrates the selected search points to obtain an average value. The filter processing is performed, and the results for the 24 types of search line segments are sent to the second selector 4. That is, in FIG. 3, the pixel of the search point f33 on the xy plane representing each search line segment is located on the xy plane (z =
X = 3, y = 3 pixels at the center position of 3), and the search point f13 on the same plane
Indicates a relationship such as x = 1, y = 3, and by integrating these search points, the pixel data value of each search line is obtained. The second selector 4 compares the results of the received 24 types of search line segments, and searches for the type of search line segment having a value larger than a predetermined threshold value and the pixel data thereof. The value is sent to the second integrating device 5. If the 24 kinds of pixel data values are smaller than the threshold value, it is determined that no data exists as the minimum luminance value, that is, 0, and the information is sent to the second integrating device 5. The second integrating device 5 integrates the pixel data values of the sent search line to calculate an average value, sets this value as the center pixel value, and further integrates the direction value of the search line. An average value is obtained, and this value is used as a normal direction component of the central pixel, as a corresponding grid point (in this case, x = 3, y = 3, z =
The information is output to the three-dimensional image output means 6 as a result of the process 3), and the three-dimensional image output means 6 stores the information. In the above processing, the center pixel is x = 3, y = 3, z =
As described above, pixel data of all grid points is processed by updating x, y, and z one by one. The following is a simple description using equations. for (z = 3; k <zmax-2; z ++) {for (y = 3; k <ymax-2; y ++) {for (x = 3; k <xmax-2; x ++) {Filter (x, y , z);-> Neighborhood processing described so far}}} By repeating such operations, almost all three-dimensional pixel data can be processed. Then, the three-dimensional image output means 6 outputs an image after obtaining all results. Although the formulas shown here are not applied to the periphery of the calculation area, there is often little important information in the periphery, so that the input image is used as it is. Is possible. Accordingly, when the above-described processing is performed on the input three-dimensional image data, an average value is obtained from the central pixel peripheral information, and pixel data appearing irrespective of the neighborhood such as speckle noise is obtained. A threshold value is set to remove noise information, and interpolation can be performed when there is data in the vicinity and the data at the center pixel itself has disappeared. Furthermore, the extraction of the normal vector as the three-dimensional surface information of the image used for the three-dimensional image output means is performed with noise reduction and interpolation processing, so that the influence of noise can be minimized. . In the present filter, for simplicity of explanation, search points representing search line segments are averaged using pixel data itself at grid points, but as shown in FIG. Pixel data at grid points subdivided from the distances 101 and 102 from the center 100 are accurately considered.
The same can be achieved by considering (sub-pixel conversion). The straight line 104 corresponds to the processing of the filter 2 in FIG. 3. The 4 × 4 sub-pixels 107 and 108 including the intersections 105 and 106 of the straight lines 104 and 101 and 102 correspond to the pixels f24 and f15, respectively. The sub-pixel 4x4 including the center of corresponds to f33. By doing this, it is possible to make detailed settings in consideration of the degree of calculation processing in the direction as shown in FIG. (Embodiment 2) FIG. 6 shows an example of extracting a normal vector in the present invention. An input image binarized with a two-dimensional image will be described below for simplification. FIG. 6A shows two-dimensional binarized input pixel data, which is a hollow circle. FIG. 6B shows a line segment search for the line segment having pixel data equal to or larger than the threshold value, which is indicated by a vector. The length indicates the size of the pixel data of the line segment. FIG. 6C shows a result obtained by proportionally adding the direction vector of the line segment and the size of the pixel data thereof and combining them. FIG. 6D shows a case where the magnitude of the obtained combined vector is equal to or larger than the threshold value. Using these combined vectors makes it possible to obtain image surface normal data at the time of image display. Circle 8 is the outer surface 10, 11, 12, 13, 14, 15, 16, 1, 1 of the hollow circle represented by the composite vector above the threshold.
7,18,19,20, circle 9 is the inner surface 21,22,
It is a hollow circular image composed of 23 and 24. It should be noted that configuring the functions described in the above embodiment as a computer program is also one of the practical methods of realizing the present invention. That is, a program for causing a well-known computer device having a processor, a memory, a display device, and the like to execute the above functions is created. This program,
This can be easily realized by recording the program on a computer-readable recording medium such as a floppy disk, a CD-ROM, or the like, reading the program into a computer memory as needed, and executing the program. As described above, according to the present invention, filtering is performed on three-dimensionally neighboring pixels, and an image obtained by medical image diagnosis or the like which lacks a partly necessary image is obtained. On the other hand, interpolation processing can be performed, and it is possible to reproduce a more realistic image.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is an explanatory diagram of an image processing apparatus according to an embodiment of the present invention. FIG. 2 is an explanatory diagram of a filter according to a first embodiment of the present invention. FIG. FIG. 4 is an explanatory diagram of a filter in the embodiment. FIG. 4 is an explanatory diagram of an expanded filter in the embodiment. FIG. 5 is an explanatory diagram of a rotation angle of the filter in the embodiment. [Description of Signs] 1 Three-dimensional pixel data to be input 2 First selector 4 Second selector 6 Three-dimensional image output means 7 Center pixel 100 Center 101 of pixels Concentric circle indicating that the distance between pixels is 1 102 Concentric circle 104 indicating that the distance of the pixel is 2 Interpolation straight line 105 Intersection point at a pixel distance of 2 Intersection point 107 at a pixel distance of 1 Subpixel at a pixel distance of 2 (4 ×
4) A sub-pixel (4 ×
4) Search line segment indicating the angle obtained by rotating the 110 line segment 1 111 Search line segment indicating the angle obtained by rotating the line segment 2 112 Search line segment indicating the angle obtained by rotating the line segment 3 113 line segment Search line segment indicating the angle at which was rotated 4

Continuation of the front page (56) References JP-A-4-285545 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06T 1/00 315 A61B 5/055 A61B 6/03 360 G06T 7/60 250 JICST file (JOIS)

Claims (1)

(57) [Claims 1] A lattice point is formed in a three-dimensional space region including x, y, and z coordinate axes orthogonal to each other, and a space of physical properties of an object existing in the region. The feature amount observed at each grid point corresponding to the distribution is set as input three-dimensional pixel data, and processing is performed on the pixel data at all grid points based on the three-dimensional pixel data to output an image. An image processing apparatus, wherein a line segment having a fixed length with the grid point as an end point is rotated by a certain angle on each of three planes including the end point perpendicular to the x, y, and z coordinate axes, and the position is set at each position. A first selection unit that selects a grid point sequence representing the line segment thus selected, and outputs pixel data of a grid point included in the grid point sequence corresponding to the line segment determination information for determining a plurality of line segments; Pixel data included in the grid point sequence input from the first selection means. Integrating the the pixel data values of the line segment,
A first integrating means for outputting together with the line segment discriminating information, and a pixel data value of the plurality of line segments output from the first integrating means, wherein the comparison result is larger than a predetermined threshold value In the case, a pixel data value of the line segment corresponding to the line segment discrimination information is output from a larger one.
A direction component of the end point is calculated from the plurality of line segment discrimination information input from the second selection unit, and a plurality of pixels of the line segment input from the second selection unit A second method of integrating data values and outputting as pixel information at the end points
Accumulating means, and accumulating the direction component and the pixel information input from the second selecting means as grid point information of the grid points, based on the information after all grid point information is obtained An image processing apparatus comprising: image output means for outputting an image.