JP2952687B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and apparatus for performing processing such as image smoothing in MR and CT, and more particularly, to a method for continuous processing in one axial direction such as a human body. This is useful for subjects that have a structure that

[Prior Art] Conventionally, in an image in MR or CT, for example, a 3 × 3 or 5 × 5 spatial filter is applied to each pixel in one image in order to improve the S / N ratio of the pixel and increase the contrast resolution. To smooth the image.

[Problems to be Solved by the Invention] However, in the conventional method, since the data (luminance) of one pixel is reflected by the data of a plurality of pixels in the vicinity thereof, the contrast resolution of an image can be increased. On the other hand, there is a problem that the spatial resolution is deteriorated.

In particular, when obtaining a cross-sectional image perpendicular to the body axis of the human body, the human body has a continuous structure in the body axis direction, but various tissues such as bones, internal organs, and muscles in the direction perpendicular to the body axis. Are distributed, and when smoothing is performed in a cross-sectional image perpendicular to the body axis by the above-described conventional method, data of different tissues is reflected on data of each pixel. Therefore, in this case, there is a problem that not only the spatial resolution deteriorates, but also the reliability of the image decreases.

SUMMARY OF THE INVENTION An object of the present invention is to provide an image processing apparatus capable of increasing the contrast resolution while minimizing the deterioration of the spatial resolution.

[Means for Solving the Problems] The image processing method of the present invention processes a plurality of slice images, and performs filtering between corresponding pixels of a target slice and auxiliary slices at least on both sides thereof. And

Further, the image processing apparatus of the present invention includes a first filtering unit that performs filtering between pixels in a target slice, and a second filtering unit that performs filtering between corresponding pixels of the target slice and auxiliary slices at least on both sides thereof. And characterized in that:

[Operation] In the image processing method and apparatus of the present invention, filtering is performed between pixels on the target slice and pixels corresponding to the pixels of the target slice on the auxiliary slices set on both sides of the target slice. The pixel data at the corresponding position on a different slice is reflected on the upper pixel data. Therefore, when filtering is performed between neighboring pixels on the target slice, deterioration of the spatial resolution is extremely small.

On the other hand, the image processing method and apparatus of the present invention reflect the data of the corresponding pixel on the auxiliary slice instead of the neighboring pixel data on the target slice, so that the contrast resolution is similarly improved.

Further, the image processing apparatus of the present invention includes the first filtering means for performing filtering between pixels on the target slice and the second filtering means for performing filtering between pixels on the target slice and the auxiliary slice. Depending on the structure of the target and the direction of the target slice,
The two filtering means can be operated by switching them appropriately.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. It should be noted that the present invention is not limited by this.

FIG. 5 is a block diagram showing an MRI apparatus 1 according to one embodiment of the present invention.

The computer 2 controls the entire operation based on instructions from the console 13.

The sequence controller 3 operates the magnetic field drive circuit 4 based on the stored sequence, and generates a static magnetic field and a gradient magnetic field using the static magnetic field coil and the gradient magnetic field coil of the magnet assembly 5. Further, it controls the gate modulation circuit 7, modulates the RF signal generated by the RF oscillation circuit 6, and applies the RF signal from the RF power amplifier 8 to the transmission coil of the magnet assembly 5.

NMR obtained with the receiving coil of the magnet assembly 5
The signal is input to the phase detector 10 via the preamplifier 9 and further to the computer 2 via the AD converter 11.

The computer 2 reconstructs an image based on the data of the NMR signal obtained from the AD converter 11 and displays the image on the display device 12.

The image processing method of the present invention is performed according to a procedure stored in the computer 2. Further, the image processing apparatus of the present invention is configured by the computer 2.

FIG. 1 is a conceptual diagram showing the operation of a computer 2 as an image processing device according to the present invention.

First smoothing means 21, an original slice G intended to smoothing process within each object slice, as in FIG. 1, five Taking raw slice G five smoothing slice H ₁ is obtained Can be Second smoothing means 22 and the target slice, for example, those smoothing processing between each two auxiliary slice on both sides, the five of one from the original slice G smoothed slices H ₂ obtained . The two smoothing means 21 and 22 can be arbitrarily switched by the switching means 23 according to the select signal.

First smoothing means 21, for example, as shown in FIG. 2 (a), 8 pixels in the case of smoothing the image of the object slice S _0, around the central pixel P ₀ on purpose slices S ₀ data to d, the f to i, is reflected in the data e of the central pixel P _0.
For example, when the 3 × 3 noise removal operator OP shown in FIG. 2B is used, each data a to i is weighted as shown in FIG. 2B, and the weighted result is averaged. since the smoothing data of the center pixel P ₀ is obtained. Therefore, the processing is performed for all the pixels on the object slice S _0, it is possible to perform a smoothing of the image of interest slices S _0.

The operator OP, since the weighting of twice the 8 pieces of pixel data in the vicinity thereof in the data of the central pixel P _0, also simultaneously removing noise and smoothing.

Data E after smoothing of the central pixel P ₀ in this case is determined as E = (1/10) × (a + b + c + d + 2e + f + g + h + i).

FIG. 3 is a conceptual diagram showing the processing of the second smoothing means 22.

In the second smoothing unit 22, the data of the pixel P ₀ on the object slice S _0, the auxiliary slice S ₁ of the two set on both sides of the object slice S _0, S _2, S _-1, the S _-2 Smoothing is performed by reflecting pixel data. That is, the data γ pixel P ₀ on the object slice S _0, the auxiliary slice S ₁ in the corresponding ie mutually overlapping position to the pixel P _0, S _2, S _-1, the pixels on S _-2 P
₁ , an appropriate operator is applied to the data β, α, δ, ε of P ₂ , P ₋₁ , and P ₋₂ .

By this operator, each data γ, β, α, δ, ε
Predetermined weighting is performed, image smoothing purposes slices S ₀ is performed.

The operators used here include, for example, −3/35, 12/35, 17/3 for each of the data α, β, γ, δ, and ε, respectively.
One that weights 5,12 / 35, -3 / 35 is conceivable. This operator performs fitting using a quadratic curve to the pixel P ₀ of the target slice S ₀ .

At this time, the data Γ obtained by smoothing the data γ of the pixel P ₀ is obtained by Γ = (1/35) × (−3α + 12β + 17γ + 12δ−3ε).

If the operator displays the component, (−3 / 35,12 / 3
5, 17/35, 12/35, −3/35) (hereinafter referred to as “weight vector”). On the other hand, each data α, β, γ, δ, ε
"Pixel vector" with components as (α, β, γ, δ, ε)
In consideration of the above, the processing of weighting the data α, β, γ, δ, and ε by the operator illustrated above and averaging the results is performed by calculating the inner product of the “weight vector” and the “pixel vector”. Equivalent to.

The smoothing process by the second smoothing means 22 is performed according to the flowchart of FIG.

In the first step, a target slice is selected at a position where a cross-sectional image of a subject such as a human body is required.

In the second step, it is determined whether or not there are auxiliary slices on both sides of the target slice. If it is determined that there is an auxiliary slice, the process proceeds to the third step, and if it is determined that there is no auxiliary slice, the process is stopped.

If you want to stop and not perform the processing after step 3,
The first smoothing means 21 performs smoothing. When proceeding to the third step, smoothing is performed by the second smoothing means 22 as follows.

In the third step, a pixel at a certain position on the target slice and a corresponding pixel on the auxiliary slice used for smoothing are extracted.

In the fourth step, the inner product of the “weight vector” and the “pixel vector” as illustrated above is calculated.

In a fifth step, the calculation result is stored at a position corresponding to a pixel position of the target slice in the output image.

In the sixth step, it is determined whether or not the above processing has been completed for all pixels on the target slice. If it is determined that the processing has not been completed, the process proceeds to step 7, the position of the pixel is changed in the same target slice, and then the process jumps to step 3 to repeat the above processing.

If it is determined that the processing has been completed for all the pixels, the process proceeds to step 8, selects another target slice, and then jumps to step 2. Then, the above flow is repeated for the next target slice. Thus, processing is performed for all necessary slices.

FIG. 6 is a schematic explanatory view showing a torso of a human body. A spine b and a blood vessel v penetrate inside the torso B in the body axis direction. Accordingly, the body B has a structure that is continuous in the body axis direction, that is, a structure in which a cross-sectional structure perpendicular to the body axis does not change significantly along the body axis.

Therefore, when it is desired to obtain a cross-sectional image perpendicular to the body axis, the target slice S ₀ and the auxiliary slice
S ₁ , S ₂ , S ₋₁ , S ₋₂ may be set as shown in FIG. 6, and filtering such as smoothing may be performed between pixels on these slices. This not only results in very good contrast resolution and spatial resolution, but also in each auxiliary slice.
Since the pixel data at the corresponding positions on S ₁ , S ₂ , S ₋₁ , and S ₋₂ is obtained with almost the same organization as the pixel data on the target slice S ₀ , a highly reliable pixel is obtained. be able to.

Moreover, according to this method, the S / N ratio can be improved without deteriorating the spatial resolution in the body axis direction as in the case where the thickness of the target slice is simply increased.

In the above embodiment, the number of slices is five in total and fitting is performed with a quadratic curve. However, the number of slices may be increased to increase the degree of the curve to be fitted, or the number of data may be increased. Of course. This further smoothes the image.

Alternatively, the number of slices may be set to four and fitting may be performed using a quadratic curve. When the number of slices is 5, an image cannot be obtained at a portion corresponding to the thickness of two slices near the edge of the subject, but when the number of slices is 4, two images are provided on one side of the target slice. Since one auxiliary slice comes on the opposite side, the bias is generated in the smoothed data,
By making one auxiliary slice come to the edge of the subject, there is an advantage that an image can be obtained up to the edge of the subject by two slices as compared with the case of five.

In the above embodiment, smoothing and noise removal are mentioned as filtering, but it goes without saying that other processing is also possible.

According to the image processing apparatus of the present invention, an image can be obtained by arbitrarily switching between the first filtering means and the second filtering means. Accordingly, there is an effect that a clearer image can be obtained.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing the operation of the image processing apparatus of the present invention,
FIG. 2A is an explanatory diagram showing a center pixel and a neighboring pixel of a target slice, FIG. 2B is an explanatory diagram of a noise removal operator, and FIG. 3 is smoothing between pixels of the target slice and the auxiliary slice. FIG. 4 is a flowchart showing the operation of the image processing apparatus of the present invention, FIG. 5 is a block diagram of an MRI apparatus having the image processing apparatus of the present invention, and FIG. 6 is the inside of the torso of a human body. FIG. 3 is a schematic explanatory view showing a structure. (Explanation of Signs) 21 First Smoothing Means 22 Second Smoothing Means 23 Switching Means S ₀ … Target Slices S ₁ , S ₂ , S ₋₁ , S ₋₂ … Auxiliary Slice P ₀ , P ₁ , P ₂ , P _-1 , P _-2 …… Pixel OP …… An operator for noise removal.

──────────────────────────────────────────────────続 き Continued on the front page (58) Fields surveyed (Int.Cl. ⁶ , DB name) G06T 1/00 G06T 5/20 H04N 1/40 A61B 6/00

Claims (1)

(57) [Claims] 1. A smoothing process is performed on one slice image of a plurality of slice images by using a value of a predetermined pixel in the one slice image and a value of a pixel located near the predetermined pixel. A first smoothing means for performing, for the plurality of slice images, a value of a predetermined pixel in one slice image and a pixel of a slice image located in the vicinity of the one slice image, at a position of the predetermined pixel. A second smoothing unit that performs a smoothing process by performing an operation using a value of a pixel at a corresponding position; and an image that is smoothed by the first smoothing unit or smoothed by the second smoothing unit. An image processing apparatus comprising: a selection unit that selects a converted image.