CN101366638B - Method for improving image quality - Google Patents

Abstract

The invention provides a method for improving the quality of an image, which comprises the following steps: (1) whether each pixel on the image is an edge point is judged; (2) if the pixel is the edge point, an image value of the pixel is still taken as a new image value of the pixel; and (3) if the pixel is a non-edge point, more than two symmetrical windows which take the pixel as the center are selected first, then mathematical expectations and deviations of the pixels forming the symmetrical windows are respectively calculated, and finally the mathematical expectation of the symmetrical window with smallest deviation is selected as the new image value of the pixel. The method smoothes noises contained in CT positioning images, particularly various random noises and noises caused by metal parts and other embedded parts inside human bodies, and can not produce large impact on outlines of positioning images, thereby being more convenient for the analysis, identification and final determination of FOV.

Description

A kind of method of improving picture quality

Technical field

The present invention relates to a kind of method of improving picture quality, especially a kind of location picture at computed tomography (Computed Tomography is hereinafter to be referred as CT) improves the method for picture quality on (Topogram also claims the Topo picture).

Background technology

Generally before CT equipment begins series scanning or helical scanning, need sweep width of cloth location picture to patient, to look like to determine sweep limits and the location of carrying out image reconstruction (ImageReconstruction) by this location.Utilize x-ray source that thereby patient's the position that is examined is scanned and obtained data for projection and carry out image reconstruction and produce medical image then.

After obtaining CT location picture, the scope of scanning can be set in the above and rebuild the visual field (Field ofView is hereinafter to be referred as FOV), the FOV on the picture of location represents with a rectangle usually, under some special situations, also can adopt parallelogram to represent.Wherein, (be orthogonal following) is used for determining sweep limits along the limit of patient body short transverse, and utilize this sweep limits, interested check point is included in scanning and the reconstructed image, and other positions of patient avoid scanning; And being used for determining FOV along the limit (being limit, the orthogonal left and right sides) of patient body width, this FOV is used for the specified image indication range, promptly is examined display position and the size of position in image.CT equipment will carry out series scanning or helical scanning subsequently and carry out image reconstruction according to the scope that above-mentioned rectangle is set.The existing multiple at present method of determining and be provided with FOV in the location as the upper edge body's border automatically.

In when scanning location picture, may comprise some noises in the picture of resulting location, for example various random noises, at metal parts and the caused noise of other embedded components and the other kinds noise of inside of human body.The existence of these noises will bring adverse effect to analysis, identification and last the determining of FOV.

Summary of the invention

Main purpose of the present invention provides a kind of method of improving picture quality, can smoothly locate the noise that comprises in the picture on the one hand by this method, can not produce big influence on the other hand, thereby be convenient to determine the reconstruction visual field of location picture more the profile of location picture.

For achieving the above object, technical scheme of the present invention specifically is achieved in that a kind of method of improving picture quality, and the method includes the steps of:

(1), judges whether described pixel is marginal point for each pixel on the image;

(2) if described pixel is a marginal point, then still with the image value of described pixel new image value as described pixel;

(3) if described pixel is non-marginal point, then at first selecting two or more is the symmetrical window at center with described pixel; Calculate the mathematic expectaion and the deviation of the pixel of forming described symmetrical window then respectively; The mathematic expectaion of symmetrical window of selecting described deviation minimum at last is as the new image value of described pixel.

Wherein, described every kind is that the symmetrical window at center comprises at least two symmetrical subwindows with described pixel.Preferably, described every kind is that the symmetrical window at center comprises 4 symmetrical subwindows with described pixel.

Wherein, described symmetrical subwindow be respectively with described pixel be summit or edge point the tetragon subwindow, be the pentagon subwindow of summit or edge point, be at least two kinds of subwindows in the hexagon subwindow of summit or edge point with described pixel with described pixel.

Wherein, when calculating the mathematic expectaion of pixel of every kind of described symmetrical window and deviation, be mathematic expectaion and the deviation of calculating the pixel on the contour line of forming every kind of described symmetrical window; Or the mathematic expectaion and the deviation of the pixel on the contour line of all symmetrical subwindows of every kind of described symmetrical window of calculating composition.

As a preferred version of the present invention, described tetragon subwindow is for being the square subwindow on summit with described pixel; Described pentagon subwindow is respectively along 0 °, 90 °, 180 °, 270 ° symmetric pentagon subwindows of direction of this pixel; Described hexagon subwindow is respectively along 45 °, 135 °, 225 °, 315 ° symmetric hexagon subwindows of direction of this pixel.

As another preferred version of the present invention, described tetragon subwindow is for being the square subwindow on summit with described pixel; Described pentagon subwindow is respectively along 45 °, 135 °, 225 °, 315 ° symmetric pentagon subwindows of direction of this pixel; Described hexagon subwindow is respectively along 0 °, 90 °, 180 °, 270 ° symmetric hexagon subwindows of direction of this pixel.

Wherein, the image value of described the above pixel of image is the CT value of described pixel.

After having adopted said method,, especially various random noises, carried out smoothly at the metal parts and the caused noise of other embedded components of inside of human body to the noise that comprises in the CT location picture.This method can not produce big influence to the profile of location picture simultaneously, thereby is convenient to analysis, identification and last the determining of FOV more.

Description of drawings

Below in conjunction with accompanying drawing the specific embodiment of the present invention is described in further detail, wherein:

Fig. 1 is the method flow diagram of the improvement location picture element amount that adopts among the present invention.

Fig. 2 is the embodiment of the invention one, wherein, the location that Fig. 2 a indicates to calculate is non-marginal point P (3,3) as pixel, and Fig. 2 b, Fig. 2 c, Fig. 2 d adopt tetragon, pentagon, hexagon subwindow to form three kinds of symmetrical windows to carry out calculation process explanation sketch map.

Fig. 3 is the embodiment of the invention two, wherein, the location that Fig. 3 a indicates to calculate is non-marginal point P (3 as pixel, 3), Fig. 3 b, Fig. 3 c, Fig. 3 d be respectively adopt with Fig. 2 in different tetragon, pentagon, hexagon subwindow form three kinds of symmetrical windows and carry out calculation process sketch map is described.

The specific embodiment

Fig. 1 is the method flow diagram of the improvement location picture element amount that adopts among the present invention.In general, width of cloth CT location picture is made up of 512*512 pixel usually, and each pixel has an image value, is called the CT value.According to the inventive method, judge at first whether each pixel on the image is marginal point.Here so-called marginal point, the symmetrical window that promptly can not adopt the present invention to mention below calculates the pixel of its mathematic expectaion.Usually all belong to marginal point at preceding two row of image, last two row, preceding two row, last two pixels that list.

If CT location belongs to above-mentioned said marginal point as a last pixel, then still with the CT value of this pixel new CT value as described pixel;

If the CT location does not belong to above-mentioned said marginal point as a last pixel, then at first selecting two or more is the symmetrical window at center with described pixel, calculates the mathematic expectaion and the deviation of the pixel of forming described symmetrical window then respectively; The mathematic expectaion of symmetrical window of selecting described deviation minimum at last is as the new image value of described pixel.

Every kind is that the symmetrical window at center comprises at least two symmetrical subwindows with described pixel, comprises 4 symmetrical subwindows in the embodiment of the invention.

The symmetry subwindow be respectively with described pixel be summit or edge point the tetragon subwindow, be the pentagon subwindow of summit or edge point, be at least two kinds of subwindows in the hexagon subwindow of summit or edge point with described pixel with described pixel.

Wherein, when calculating the mathematic expectaion of pixel of every kind of described symmetrical window and deviation, be mathematic expectaion and the deviation of calculating the pixel on the contour line of forming every kind of described symmetrical window; Or the mathematic expectaion and the deviation of the pixel on the contour line of all symmetrical subwindows of every kind of described symmetrical window of calculating composition.

Described tetragon subwindow is for being the square subwindow on summit with described pixel; Described pentagon subwindow is respectively along 0 °, 90 °, 180 °, 270 ° symmetric pentagon subwindows of direction of this pixel; Described hexagon subwindow is respectively along 45 °, 135 °, 225 °, 315 ° symmetric hexagon subwindows of direction of this pixel.

Described tetragon subwindow is for being the square subwindow on summit with described pixel; Described pentagon subwindow is respectively along 45 °, 135 °, 225 °, 315 ° symmetric pentagon subwindows of direction of this pixel; Described hexagon subwindow is respectively along 0 °, 90 °, 180 °, 270 ° symmetric hexagon subwindows of direction of this pixel.

Through after the aforementioned calculation, the mathematic expectaion of the above pixel of direction of selection deviation minimum is as the new CT value of described pixel.

Fig. 2 is the embodiment of the invention one, wherein, the location that Fig. 2 a indicates to calculate is non-marginal point P (3,3) as pixel, and Fig. 2 b, Fig. 2 c, Fig. 2 d adopt tetragon, pentagon, hexagon subwindow to form three kinds of symmetrical windows to carry out calculation process explanation sketch map.In Fig. 2, at first select 3 kinds to be the symmetrical window at center with described pixel P (3,3).Every kind of symmetrical window comprises 4 symmetrical subwindows.First kind of symmetrical window comprises 4 square subwindows, wherein each square subwindow is all with pixel P (3,3) be a summit, second kind of symmetrical window comprises 4 with pixel P (3,3) be the pentagon subwindow on summit, these 4 pentagon subwindows are respectively along this pixel P (3,3) 0 °, 90 °, 180 °, 270 ° of direction symmetries, the third symmetrical window comprises 4 with pixel P (3,3) be the hexagon subwindow on summit, these 4 hexagon subwindows are respectively along 45 ° of this pixel P (3,3), 135 °, 225 °, 315 ° of direction symmetries.

In the present embodiment, when the mathematic expectaion of the pixel of first kind of symmetrical window of calculating and deviation, be mathematic expectaion and the deviation of calculating the pixel on the contour line of forming first kind of symmetrical window; And when calculating the mathematic expectaion of pixel of second kind of symmetrical window and deviation, be mathematic expectaion and the deviation of calculating the pixel on the contour line of all symmetrical subwindows of forming second kind of symmetrical window.Calculate the mathematic expectaion and the deviation of the pixel of the third symmetrical window, the method that adopts is identical with the computational methods of second kind of symmetrical window, i.e. the mathematic expectaion and the deviation of the pixel on the contour line of all symmetrical subwindows of the third symmetrical window of calculating composition.

The coordinates table of pixel is shown among Fig. 2:

P(1，1)P(1，2)P(1，3)P(1，4)P(1，5)

P(2，1)P(2，2)P(2，3)P(2，4)P(2，5)

P(3，1)P(3，2)P(3，3)P(3，4)P(3，5)

P(4，1)P(4，2)P(4，3)P(4，4)P(4，5)

P(5，1)P(5，2)P(5，3)P(5，4)P(5，5)

In Fig. 2 b, when the mathematic expectaion of the pixel of first kind of symmetrical window of calculating and deviation, with the pixel P (3 that will calculate, 3) be the center, along on 0 ° of direction to 360 ° direction of this pixel with the mathematic expectaion and the deviation of the CT value of the pixel calculating pixel point P (3,3) on the contour line of this symmetry window:

${\mathrm{Ave}}_{\mathrm{ir}}=\frac{1}{n}\underset{k=1\·\·\·n}{\underset{i=1}{\mathrm{\Σ}}}{\mathrm{rect}}_{\mathrm{ik}}$

(formula 1)

${\mathrm{Diff}}_{\mathrm{ir}}=\frac{1}{n-1}\underset{k=1\·\·\·n}{\underset{i=1}{\mathrm{\Σ}}}|{\mathrm{Ave}}_{\mathrm{ir}}-{\mathrm{rect}}_{\mathrm{ik}}|$

(formula 2)

In formula 1 and formula 2, rect represents rectangular window, and subscript r also represents rectangular window, and i represents orthogonal number, i=1 herein, and n represents to form orthogonal pixel number, n=8 herein.

According to above-mentioned formula 1 as can be known, the mathematic expectaion of calculating intermediary image vegetarian refreshments P (3,3) is promptly calculated the mathematic expectaion of the CT value of 8 pixels on the contour line of forming this symmetry window, that is to say calculating pixel point P (2,2), P (2,3), P (2,4), P (3,2), P (3,4), P (4,2), the mathematic expectaion of the CT value of P (4,3), P (4,4).Obtain intermediary image vegetarian refreshments P (3, after the mathematic expectaion of CT value 3), can calculate CT value poor of each pixel on described mathematic expectaion and the contour line of forming symmetrical window according to formula 2, absolute value to described difference is sued for peace, average (divided by (n-1)) again, can obtain deviation with the CT value of rectangular window calculating.

And the mathematic expectaion of pixel of calculating second kind and the third symmetrical window is during with deviation, is mathematic expectaion and the deviation of calculating the pixel on the contour line of all symmetrical subwindows of forming second kind of symmetrical window and the third symmetrical window respectively.

In Fig. 2 c, when the mathematic expectaion of the pixel of second kind of symmetrical window of calculating and deviation, be the center with the pixel P (3,3) that will calculate, the mathematic expectaion and the deviation of the pixel on the contour line of all symmetrical subwindows of this symmetry window of calculating composition.More particularly, promptly calculate along the mathematic expectaion and the deviation of the pixel on all four pentagons symmetry subwindow contour lines of 0 °, 90 °, 180 °, the 270 ° direction of this pixel P (3,3).Dotted line among Fig. 2 c is represented 0 °, 90 °, 180 °, 270 ° directions along this pixel P (3,3), and the solid line that an overstriking is arranged respectively on each direction is represented corresponding pentagon subwindow.The computational mathematics expectation is as follows with the formula of deviation:

${\mathrm{Ave}}_{\mathrm{ip}}=\frac{1}{n}\underset{k=1\·\·\·n}{\underset{i=1\·\·\·4}{\mathrm{\Σ}}}{\mathrm{pentagon}}_{\mathrm{ik}}$

(formula 3)

${\mathrm{Diff}}_{\mathrm{ip}}=\frac{1}{n-1}\underset{k=1\·\·\·n}{\underset{i=1\·\·\·4}{\mathrm{\Σ}}}|{\mathrm{Ave}}_{\mathrm{ip}}-{\mathrm{pentagon}}_{\mathrm{ik}}|$

(formula 4)

In formula 3 and formula 4, pentagon represents the pentagon subwindow, and subscript p also represents the pentagon subwindow, and i represents the number of pentagon subwindow, i=4 herein, and n represents to form the pixel number of i pentagon subwindow, herein n=6.

According to above-mentioned formula 3 as can be known, the mathematic expectaion of calculating intermediary image vegetarian refreshments P (3,3) is promptly calculated the mathematic expectaion of the CT value of all pixels of forming four pentagon subwindows.That is to say, when i=1, be to being positioned at the pentagon of 0 ° of direction of pixel P (3,3), this moment n=1,2 ... 6, forming pentagonal pixel is P (3,3), P (2,4), P (2,5), P (3,5), P (4,4), P (4,5).When i=2, be to being positioned at the pentagon of 90 ° of directions of pixel P (3,3), this moment n=1,2 ... 6, forming pentagonal pixel is pixel P (3,3), P (2,2), P (1,2), P (1,3), P (1,4), P (2,4).When i=3, be to being positioned at the pentagon of 180 ° of directions of pixel P (3,3), this moment n=1,2 ... 6, forming pentagonal pixel is pixel P (3,3), P (4,2), P (4,1), P (3,1), P (2,1), P (2,2).When i=4, be to being positioned at the pentagon of 270 ° of directions of pixel P (3,3), this moment n=1,2 ... 6, forming pentagonal pixel is pixel P (3,3), P (4,2), P (5,2), P (5,3), P (5,4), P (4,4).Calculate above-mentioned 24 pixels the CT value and, average (divided by n) again, can obtain along the mathematic expectaion of the pixel on all four pentagons symmetry subwindow contour lines of 0 °, 90 °, 180 °, the 270 ° direction of this pixel P (3,3).

Obtain pixel P (3, after the mathematic expectaion of CT value 3), can at first calculate CT value poor of described mathematic expectaion and each pixel of forming the pentagon subwindow according to formula 4, absolute value to described difference is sued for peace, average (divided by (n-1)) again, can obtain along the deviation of the pixel on all four pentagons symmetry subwindow contour lines of 0 °, 90 °, 180 °, the 270 ° direction of this pixel P (3,3).

In Fig. 2 d, when calculating the mathematic expectaion of pixel of the third symmetrical window and deviation, be the center with the pixel P (3,3) that will calculate, calculate the mathematic expectaion and the deviation of the pixel on the contour line of all symmetrical subwindows of forming this symmetry window.More particularly, promptly calculate along the mathematic expectaion and the deviation of the pixel on all four hexagonal symmetry subwindow contour lines of 45 °, 135 °, 225 °, the 315 ° directions of this pixel P (3,3).Dotted line among Fig. 2 d is represented 45 °, 135 °, 225 °, 315 ° directions along this pixel P (3,3), and the solid line that an overstriking is arranged respectively on each direction is represented corresponding hexagon subwindow.The computational mathematics expectation is as follows with the formula of deviation:

${\mathrm{Ave}}_{\mathrm{ih}}=\frac{1}{n}\underset{k=1\·\·\·n}{\underset{i=1\·\·\·4}{\mathrm{\Σ}}}{\mathrm{hexagon}}_{\mathrm{ik}}$

(formula 5)

${\mathrm{Diff}}_{\mathrm{ih}}=\frac{1}{n-1}\underset{k=1\·\·\·n}{\underset{i=1\·\·\·4}{\mathrm{\Σ}}}|{\mathrm{Ave}}_{\mathrm{ih}}-{\mathrm{hexagon}}_{\mathrm{ik}}|$

(formula 6)

In formula 5 and formula 6, hexagon represents the hexagon subwindow, and subscript h also represents the hexagon subwindow, and i represents the number of hexagon subwindow, i=4 herein, and n represents to form the pixel number of i hexagon subwindow, herein n=6.

According to above-mentioned formula 5 as can be known, the mathematic expectaion of calculating intermediary image vegetarian refreshments P (3,3) is promptly calculated the mathematic expectaion of the CT value of all pixels of forming four hexagon subwindows.That is to say, when i=1, be for the hexagon that is positioned at 45 ° of directions of pixel P (3,3), this moment n=1,2 ... 6, forming hexagonal pixel is P (3,3), P (3,4), P (2,5), P (1,5), P (1,4), P (2,3).When i=2, be for the hexagon that is positioned at 135 ° of directions of pixel P (3,3), this moment n=1,2 ... 6, forming hexagonal pixel is P (3,3), P (2,3), P (1,2), P (1,1), P (2,1), P (3,2).When i=3, be for the hexagon that is positioned at 225 ° of directions of pixel P (3,3), this moment n=1,2 ... 6, forming hexagonal pixel is P (3,3), P (3,2), P (4,1), P (5,1), P (5,2), P (4,3).When i=4, be for the hexagon that is positioned at 315 ° of directions of pixel P (3,3), this moment n=1,2 ... 6, forming hexagonal pixel is P (3,3), P (4,3), P (5,4), P (5,5), P (4,5), P (3,4).Calculate above-mentioned 24 pixels the CT value and, average (divided by n) again, can obtain along the mathematic expectaion of the pixel on all four hexagonal symmetry subwindow contour lines of 45 °, 135 °, 225 °, the 315 ° directions of this pixel P (3,3).

Obtain pixel P (3, after the mathematic expectaion of CT value 3), can at first calculate CT value poor of described mathematic expectaion and each pixel of forming the hexagon subwindow according to formula 6, absolute value to described difference is sued for peace, average (divided by (n-1)) again, can obtain along the deviation of the pixel on all four hexagonal symmetry subwindow contour lines of 45 °, 135 °, 225 °, the 315 ° directions of this pixel P (3,3).

Through after the aforementioned calculation, the deviation of the CT value of the pixel of three symmetrical windows of composition of calculating is compared.Select the new CT value of the mathematic expectaion of the symmetrical window of deviation minimum wherein as pixel P (3,3).

Fig. 3 is the embodiment of the invention two, wherein, the location that Fig. 3 a indicates to calculate is non-marginal point P (3 as pixel, 3), Fig. 3 b, Fig. 3 c, Fig. 3 d be respectively adopt with Fig. 2 in different tetragon, pentagon, hexagon subwindow form three kinds of symmetrical windows and carry out calculation process sketch map is described.With identical among Fig. 2 be at first also to select 3 kinds to be the symmetrical window at center with described pixel P (3,3).Every kind of symmetrical window also comprises 4 symmetrical subwindows.

Identical with Fig. 2, first kind of symmetrical window comprises 4 with pixel P (3,3) be the square subwindow on summit, second kind of symmetrical window comprises 4 with pixel P (3,3) comprise that for the pentagon subwindow on summit, the third symmetrical window 4 is the hexagon subwindow on summit with pixel P (3,3).With different among Fig. 2 be that described pentagon subwindow is 4 pentagon subwindows along 45 °, 135 °, 225 °, the 315 ° directions of this pixel P (3,3); Described hexagonal window openings is 4 hexagon subwindows along 0 °, 90 °, 180 °, the 270 ° direction of this pixel P (3,3).

Its computational process is then similar with embodiment one.For example, when the mathematic expectaion of the pixel of first kind of symmetrical window of calculating and deviation, be mathematic expectaion and the deviation of calculating the pixel on the contour line of forming this symmetry window; And the mathematic expectaion of pixel of calculating second kind and the third symmetrical window is during with deviation, is to calculate mathematic expectaion and the deviation of forming second kind of pixel on the contour line with all symmetrical subwindows of the third symmetrical window respectively.

Except that the cited situation of the foregoing description one and embodiment two, when the mathematic expectaion of calculating these three kinds of symmetrical windows and deviation, the pixel that can also calculate respectively on the contour line of forming three kinds of symmetrical windows comes computational mathematics expectation and deviation.Perhaps, the pixel of calculate forming respectively on the contour line of all symmetrical subwindows of three kinds of symmetrical windows comes computational mathematics expectation and deviation.

After having adopted said method, the CT location all has been updated to new CT value as the CT value of all last pixels.Location picture with new CT value is to wherein noise, especially various random noises, carried out smoothly at the metal parts and the caused noise of other embedded components of inside of human body.If desired, can carry out once again as last pixel location according to the method described above or repeatedly calculating with new CT value.This method can not produce big influence to the profile of location picture, thereby is convenient to analysis, identification and last the determining of FOV more.

Claims (6)

1. method of improving picture quality is characterized in that the method includes the steps of:

(1), judges whether described pixel is marginal point for each pixel on the image;

(2) if described pixel is a marginal point, then still with the image value of described pixel new image value as described pixel;

(3) if described pixel is non-marginal point, then at first selecting two or more is the symmetrical window at center with described pixel; Calculate the mathematic expectaion and the deviation of the pixel of forming described symmetrical window then respectively; The mathematic expectaion of symmetrical window of selecting described deviation minimum at last is as the new image value of described pixel;

Wherein, described every kind is that the symmetrical window at center comprises at least two symmetrical subwindows with described pixel; When calculating the mathematic expectaion of pixel of every kind of described symmetrical window and deviation, be mathematic expectaion and the deviation of calculating the pixel on the contour line of forming every kind of described symmetrical window; Or the mathematic expectaion and the deviation of the pixel on the contour line of all symmetrical subwindows of every kind of described symmetrical window of calculating composition.

2. the method for improving picture quality according to claim 1 is characterized in that, described every kind is that the symmetrical window at center comprises 4 symmetrical subwindows with described pixel.

3. the method for improving picture quality according to claim 1, it is characterized in that, described symmetrical subwindow be respectively with described pixel be summit or edge point the tetragon subwindow, be the pentagon subwindow of summit or edge point, be at least two kinds of subwindows in the hexagon subwindow of summit or edge point with described pixel with described pixel.

4. the method for improving picture quality according to claim 3 is characterized in that, described tetragon subwindow is for being the square subwindow on summit with described pixel; Described pentagon subwindow is respectively along 0 °, 90 °, 180 °, 270 ° symmetric pentagon subwindows of direction of this pixel; Described hexagon subwindow is respectively along 45 °, 135 °, 225 °, 315 ° symmetric hexagon subwindows of direction of this pixel.

5. the method for improving picture quality according to claim 3 is characterized in that, described tetragon subwindow is for being the square subwindow on summit with described pixel; Described pentagon subwindow is respectively along 45 °, 135 °, 225 °, 315 ° symmetric pentagon subwindows of direction of this pixel; Described hexagon subwindow is respectively along 0 °, 90 °, 180 °, 270 ° symmetric hexagon subwindows of direction of this pixel.

6. the method for improving picture quality according to claim 1 is characterized in that, the image value of described the above pixel of image is the CT value of described pixel.

Images