JPH0454455B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention is an X-ray computed tomography (CT)
Tomography) relates to an X-ray CT image processing system that uses only reconstructed images to quickly remove bone beam hardening artifacts that occur in images.

(Prior Art) Bone beam hardening correction is conventionally described in the literature of Joseph et al. (Joseph-Spital-A Method for
Correcting Bone Induced Artifacts in
Computed Tomography Scanners, Journal of
Computer Assisted Tomography Vol2 p
100-108, Jan. 1978), the amount of correction is calculated from the path length of the soft tissue part and bone (separated by CT value) included in the scan of the reconstructed image. However, based on this, the actual scan data was corrected and reconstructed again.

However, this method has the following problems.

Since two reconstructions and an image scan are required, the amount of calculation is enormous, and high-speed processing cannot be expected.

Post-processing is not easy because actual scan data is required.

(Object of the Invention) The present invention has been made in view of the above points, and its purpose is to perform bone beam hardening correction quickly and easily from only image data without impairing the effect of removing artifacts. The object of the present invention is to provide an X-ray CT image processing device that can obtain the desired results.

(Structure of the Invention) A first invention to achieve such an object includes a first image storage device capable of inputting and outputting a reconstructed image of X-ray CT to an external device, and a first image storage device capable of inputting and outputting a reconstructed image of an X-ray CT to an external device. The image provided by the device is
A binarization device that binarizes based on the CT value, and
a second image storage device for storing value images; a weighting function storage device storing weighting functions; a mask addition adder for adding the weighting functions using the binary image as a mask; a third image storage device that stores the addition result obtained by the adder; and a third image storage device that performs addition between the images of the first image storage device and the third image storage device, and stores the addition result in the first image storage device. The system is equipped with an image adder that provides a corrected image by calculating the sum of two-dimensional weighting functions spread around this point for all pixels representing the bone component of the reconstructed image, and uses this corrected image as the original reconstructed image. The second invention is characterized in that the bone beam hardening correction is performed by adding it to the image, and the second invention is capable of inputting and outputting the reconstructed image of the X-ray CT to an external device. a first image storage device, an image reduction device that reduces the image provided from the first image storage device both vertically and horizontally, a fourth image storage device that stores this reduced image, and this fourth image storage device 2 based on the predetermined CT value of the image given by
A binarization device for converting into a value, a second image storage device for storing the binary image, a weighting function storage device storing a weighting function, and a second image storage device for storing the weighting function using the binary image as a mask. A mask addition adder performs the addition, a third image storage device stores the addition results obtained by the mask addition adder, and the image provided from the third image storage device is interpolated to restore the original image both horizontally and vertically. an image enlarging device for enlarging the size of the image; and an image adder for performing addition between the images of the first image storage device and the image enlarging device and providing the addition result to the first image storage device; The binary image representing the bone and the weighting function are reduced to 1/N vertically and horizontally, and then the process of adding the weighting function two-dimensionally using the binary image as a mask is repeated for all pixels representing the bone, and the resulting added image is interpolated. This is characterized in that the beam hardening correction for the bones is performed by enlarging the image N times vertically and horizontally using the reconstructed image, and then adding it to the original reconstructed image.

(Example) Hereinafter, an example of the present invention will be described in detail using the drawings. FIG. 1 is a block diagram of essential parts showing an embodiment of an X-ray CT image processing apparatus according to the first invention. In this figure, 1 is a first image storage device that stores a reconstructed image, 2 is a binarization device, 3 is a second image storage device, 4 is a weighting function storage device, 5 is a mask addition adder, and 6 is a third image storage device, and 7 is an image adder.

The first image storage device 1 is an external device (not shown)
It is now possible to input and output images to and from. The binarization device 2 converts the image in the first image storage device 1 into 2 images using a preset CT value as a threshold value.
The second image storage device 3 stores this binary image. The mask addition adder 5 receives the outputs from the weighting function storage device 4 and the second image storage device 3, and spreads the weighting function two-dimensionally by aligning the center of the weighting function with a pixel position with a high CT value representing the upper bone of the binary image. The weighting function is added and stored in the third image storage device 6. The image adder 7 is the third image storage device 6
and the image in the first image storage device 1, and the added image is stored in the first image storage device 1.

In such a configuration, a reconstructed X-ray CT image input from an external device to the first image storage device 1 is compared with a preset threshold in the binarization device 2 and binarized. . This binary image is
The images are sequentially stored in the image storage device 3. The mask addition adder 5 receives 2 images from this second image storage device 3.
After receiving the value image, the center of the weight function from the weight function storage device 4 is aligned with the pixel position of the high CT value representing the bone on this binary image, and a weight function spread two-dimensionally is added to the third image storage device 6. Remember. Note that the initial storage value of the third image storage device 6 is set to zero. This addition process is repeated for all pixels having CT values greater than or equal to the threshold value.

The image thus created is added to the initially input reconstructed image (stored in the first image storage device 1) by the image adder 7. This makes it possible to remove bone beam hardening artifacts, and the image is written anew to the first image storage device 1 and output to an external device as required.

FIG. 2 is an explanatory diagram showing another embodiment of the present invention. The difference from FIG. 1 is that the third image storage device 6 and image adder 7 are omitted. In FIG. 2, instead of calculating the sum of weighting functions and adding them later between two images, the weighting functions are added directly to the corresponding positions of the reconstructed image, and the result is stored in the first image storage device 1. I try to write to . With such a configuration as well, bone beam hardening artifacts can be removed in the same manner as described above.

Although the same weighting function is added to pixels with a CT value or higher, it is also possible to add a weighting function modulated by the CT value to pixels with a CT value or higher. be.

Further, the mask addition adder 5 is connected to the second image storage device 3.
2 of the binary image and the weighting function of the weighting function storage device 4
It is also possible to replace it with dimensional convolution.

In the embodiment of the first invention described above, two-dimensional weighting functions are added for the number of pixels representing bones, which may result in a considerable amount of calculation. Therefore, if the binary image weighting function representing the bone is reduced to 1/N vertically and horizontally, then the weighting function is added two-dimensionally using the binary image as a mask, and the addition is repeated for all pixels representing the bone. The amount can be reduced, and high-speed processing can be achieved accordingly.

FIG. 3 is a block diagram showing an embodiment of the X-ray CT image processing apparatus according to the second invention, which achieves such high-speed processing. In FIG. 3, parts equivalent to those in FIG. 1 are given the same reference numerals. Components different from those in FIG. 1 are an image reduction device 31, a fourth image storage device 32, an image enlargement device 33, and a fifth image storage device 34.

The image reduction device 31 receives the image in the first image storage device 1 and reduces it to 1/N in both the vertical and horizontal directions, and the result is stored in the fourth image storage device 32. The output of the fourth image storage device 32 is led to the binarization device 2.

The image enlarging device 33 multiplies the reduced image stored in the third image storage device 6 by N times and enlarges the reduced image to the original size. The enlarged image is stored in the fifth image storage device 34.
The image is once stored in the image adder 7 and then sent to the image adder 7.

The operation in such a configuration will be explained next.
A reconstructed X-ray CT image input from an external device to the first image storage device 1 is reduced to 1/N in both the vertical and horizontal directions by the image reduction device 31, and is stored in the fourth image storage device 32. Note that N is usually an integer.

The reduced image in the storage device 32 is a preset CT image.
The image is binarized by the binarizer 2 according to the value and stored in the second image storage device 3.

It is assumed that the weighting functions in the weighting function storage device 4 have also been reduced in advance to 1/N in the vertical and horizontal directions, corresponding to the above-mentioned reduction rate of 1/N. In the mask adding adder 5, the weighting function of the weighting function storage device 4 is centered on each pixel representing the bone of the binary image (storage device 3) and
It is added and stored in the image storage device 6 of. Note that the initial storage value of the third image storage device 6 is set to zero. The result of performing this addition process on all pixels representing the bone is a corrected image for the reduced image, so this is multiplied by N in the image enlarging device 33 and restored to its original size. For this expansion, linear interpolation is used to reduce calculation time, and two-dimensional interpolation is performed in two stages for each dimension. The enlarged image is stored in the fifth image storage device 34, and this enlarged corrected image is two-dimensionally added to the input image on the first image storage device 1 in the image adder 7, and the result is the first image. The data is output on the storage device 1.

FIG. 4 is an explanatory diagram illustrating enlarged interpolation when N=4. Based on the known value (○ mark), first calculate the value at the × mark from the values of the ○ marks sandwiching it using the following formula.

f^(i)=f(1)+(i/4)・{f(2)−f(1)} Here, i=1, 2, 3 Next, change the value at * mark to the left and right sides of it. Calculate using the following formula from the value of the × mark.

f(i)=f^(1)+i/4){f^(2)f^(1)} Here, i=1, 2, 3 Interpolation is performed in this manner. The same applies to cases other than N=4.

Note that the present invention is not limited to the embodiment, and the following method may be used.

The image reduction device 31 does not actually reduce the image, but is integrated with the binarization device 2 to sample and binarize every N pixels both vertically and horizontally.

The image enlargement device 33 is the fifth image storage device 34
The image data is directly added to the first image storage device without passing through the image storage device.

The mask addition adder 5 is replaced with a two-dimensional convolution device.

Image reduction device 31, fourth image storage device 32
Without using the second image storage device 3, the adder 5 is turned on and off according to the value of every N samples output by the binarization device 2.

(Effects of the Invention) As explained above, according to the present invention, by appropriately setting the weighting function, it is possible to perform bone and beam hardening correction equivalent to the conventional method of Joseph et al. It is possible to obtain an image with good sharpness and no cutting caused by bones.

In addition, it is a complex and computationally intensive process that involves repeating the scan on the image, examining the components along the path, correcting the actual scan data, and reconstructing it again. In comparison, it is extremely practical because the algorithm is simple and the amount of calculations is small, such as by simply repeating simple calculations between images based on reconstructed images.

For example, use image 320x320 pixels, use weight function
When using 61 x 61 pixels, the conventional method of Joseph et al. took about 15 hours, but according to the present invention, it took about 50 hours.
The processing time is only 1 minute.

Furthermore, according to the present invention as shown in FIG. 3, the image is first reduced to 1/N ² and the reduced image is binarized and the weighting function is added, so the weighting function is also original. Since a reduced value of 1/N ² is used, the number of additions is reduced from M ² ×k to (M/N) ² ×k/N ² and 1/N ⁴ . Here, M ² is the area of the weighting function expressed in the number of pixels, and k is the number of pixels representing the bone.

According to computer simulation, N=
4. When M=61, the image quality deterioration is less than ±1CT value,
The calculation time was 1/203 times that of the case without reduction.

This method is the result of skillfully exploiting the phenomenon that beam-hardened images of bones contain only low spatial frequency components.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram showing one embodiment of the X-ray CT image processing device according to the first invention, FIG. 2 is an explanatory diagram showing another embodiment of the first invention, and FIG. FIG. 4 is a block diagram showing one embodiment of the invention, and is an explanatory diagram for explaining enlargement interpolation. DESCRIPTION OF SYMBOLS 1... First image storage device, 2... Binarization device, 3... Second image storage device, 4... Weighting function storage device, 5... Mask addition adder, 6... Third Image storage device, 7... Image adder, 31... Image reduction device, 32... Fourth image storage device, 33...
...image enlarging device, 34...fifth image storage device.

Claims (1)

[Claims] 1. A first image storage device capable of inputting and outputting a reconstructed X-ray CT image to/from an external device, and a predetermined CT A binarization device that binarizes based on the value,
a second image storage device for storing this binary image; a weighting function storage device storing weighting functions; a mask addition adder for adding the weighting functions using the binary image as a mask; A third image storage device stores the addition result obtained by the mask addition adder, performs addition between the images in the first image storage device and the third image storage device, and stores the addition result in the first image. The system is equipped with an image adder that supplies the image to the storage device, and calculates the sum of the two-dimensional weighting functions spread around this point using the design element representing the bone component of the reconstructed image as a corrected image, and uses this corrected image as the original. An X-ray CT image processing apparatus characterized in that beam hardening correction of bones is performed by adding the information to a reconstructed image. 2. A first image storage device that can input and output reconstructed X-ray CT images to and from an external device; an image reduction device that reduces the image provided from the first image storage device both vertically and horizontally; a fourth image storage device that stores reduced images; and a predetermined CT
a binarization device that binarizes based on the value; a second image storage device for storing this binary image;
a weighting function storage device that stores weighting functions; a mask addition adder that adds the weighting functions using the binary image as a mask; and a third image storage device that stores the addition results obtained by the mask addition adder. , this third
an image enlarging device that performs interpolation on the image given from the image storage device and enlarges it to the original size both vertically and horizontally; and an image adder for supplying the image to the first image storage device, and after reducing the binary image representing the bone and the weighting function to 1/N vertically and horizontally, adding the weighting function two-dimensionally using the binary image as a mask. The process was repeated for all pixels representing the bone, and the resulting added image was enlarged vertically and horizontally by N times using interpolation, and then added to the original reconstructed image to perform beam hardening correction for the bone. Characteristic X-ray CT image processing device.