EP0612024B1 - Method for generating tomographic images and device for carrying out the method - Google Patents

Description

The invention relates to a method for generating layer images of a Examining area from a variety of radiation source positions for the purpose of generating separate individual images with X-rays is enforced, with image values being saved for each individual image correspond to the absorption in its pixels and where slice images are generated in the layer image values from the image values of the pixels of the individual images are derived from a layer pixel geometrically assigned.

A finite - preferably - is used as the pixel and in the following referred to as the square area of a single image, while as Layer image point, a corresponding area is designated in the layer image.

Such a method and such an arrangement are known, for example from U.S. Patent 3,499,146. The individual images are created using a X-ray emitter, which is successively in different Radiation source positions is brought. The radiation relief is from one Image converter, e.g. by an image intensifier, whose Output screen is scanned using a television camera. The the video signal thus generated is digitized. The emerging digital Data words correspond to the image values of the single image. You will be in one preferably stored digital storage device. The layer image values for the different layer pixels are derived from the image values of Pixels derived, the geometrical layer pixel in question are assigned, that is to say from those pixels which are generated when the Layer image on the connecting line of the relevant layer pixel with the different radiation source positions.

It has been shown that areas within the examination area that lie outside the layer, the details of which are reproduced sharply, in experience a reversal of contrast or a pseudo resolution of the slice image can, if the radiation source positions lie on a straight line or on an arc around a horizontal axis.

It is known that such effects also in conventional systems for linear tomography can occur with continuous Displacement between a film and an X-ray tube Layer recording is generated. It is known (see Harding et al in Phys. Med. Biol., 1975, vol. 20, no. 1, 144-149) that these effects can eliminate that during recording the for Triangular X-ray tube current increases and decreases. At a Improvement of this method (Phys. Med. Biol., 1977, Vol. 22, No. 4, Pages 747 - 759) is replaced by a triangular modulation bell-shaped modulation related, being a better one Compromise between the suppression of undesirable effects and the Decrease in blurring of details outside the shift results.

Both methods require an intervention in the X-ray generator, which is the X-ray tube feeds, and assume that the current during a Layer recording can be regulated quickly enough.

The object of the present invention is a method of the beginning mentioned type so that without intervention in the X-ray generator the contrast reversal or the pseudo resolution in slice images can be at least partially eliminated, and an arrangement for To indicate implementation of the procedure.

This object is achieved in that the derivative of Slice image values are carried out by a weighted summation, in which the Weight with which the image values of the individual images are included in the summation, the smaller the distance between the radiation source position at the Generation of the relevant image from a medium one Radiation source position was. An arrangement to carry out the The method is characterized in that the image processing unit is a Computing unit that uses the image values to generate slice images one from the radiation source position when generating the associated one Single image dependent weighting factor multiplied and the weighted Image values of the image elements geometrically assigned to a layer pixel summed up.

In the invention according to the present claims, the undesirable effects mentioned are thereby removed that the single pictures, from which a layer (with freely selectable Location) is reconstructed with different weights in the reconstruction come in. An intervention in the X-ray generator is not necessary.

As with the known method mentioned above, the Reduction in contrast reversal or pseudo-resolution effects, that objects or details in the examination area outside of the the slice image is sharply depicted, less in the slice image blurred, which can also affect the diagnosis. On the other hand, the contrast reversal and pseudo-resolution effects in different recordings occur more or less strongly. The Weighting of the individual images during the generation of a slice image but also in those cases where the effects mentioned are eliminated a reduction in blurring would not have been required.

Therefore, it is provided in a preferred embodiment of the invention that the Image processing unit contains a memory in which several of each other different sets of weighting factors are stored and that Selection means for - preferably interactive - selection of a set of Weighting factors are provided. The user can then each Choose set of weighting factors that make the most favorable compromise between reducing the undesirable effects on the one hand and the Blurring of details outside of the layer level on the other hand.

Fig. 1

Teile eines Schichtaufnahmegerätes, mit dem die Erfindung ausführbar ist,

Fig. 2

die geometrischen Verhältnisse bei einer Schichtaufnahme,

Fig. 3

ein Blockschaltbild einer Einheit zur Durchführung des Verfahrens,

Fig. 4

die Abhängigkeit der Gewichtungsfaktoren von der Strahlenquellenposition bei verschiedenen Gewichtungsfaktorsätzen,

Fig. 5a bis 5e

die mit den verschiedenen Gewichtungsfaktorsätzen einhergehenden Modulatiosübertragungsfunktionen.

The invention is explained below with reference to the drawing. Show it:

Fig. 1

Parts of a layer recording device with which the invention can be carried out,

Fig. 2

the geometrical relationships in a slice image,

Fig. 3

2 shows a block diagram of a unit for carrying out the method,

Fig. 4

the dependence of the weighting factors on the radiation source position for different weighting factor sets,

5a to 5e

the modulation transfer functions associated with the different weighting factor sets.

The layer recording device shown partly and purely schematically in FIG. 1 comprises an x-ray emitter 1 which has an x-ray beam 2 on it Input screen of an X-ray image intensifier 3 is emitted. Like through the two Arrows 200 and 300 indicated, the X-ray emitter 1 and X-ray image intensifier 3 along parallel paths in opposite directions to one another a first solid line shown end position in a second Movable end position, which is indicated by dashed lines. In each of the two radiation source positions as well as in one - preferably odd - Number of intermediate radiation source (and image intensifiers) Positions, X-rays are switched on briefly and a single image generated. The shift to the different radiation source positions takes place in such a way that the central rays of the beams 2 in all Cut positions in the same point 4. This can be done be ensured that X-ray tube 1 and image intensifier 3 by a Rod to be coupled to each other by one through point 4 horizontal axis perpendicular to the plane of the drawing is pivotable. The one containing point 4 and parallel to the directions of movement Level EF is called the Fulcrum level.

The output image of the X-ray image intensifier 3 is obtained from a video camera 5 converted into an electrical signal by an analog-digital converter 6 digitized and stored in a memory 7, which is sufficient Has capacity to the digitized video signals of all frames to record. Each frame can e.g. 256 x 256 or 512 x 512 pixels include. The number of frames must be large enough; for one Swivel angle from -20 ° to + 20 ° (opposite the vertical) has one between 30 and 50 number of single images has been shown to be sufficient. Thus, in the memory 7 for each pixel of each frame digital form stores an image value that indicates the absorption of the X-ray radiation in the examination area that represents point 4 surrounds.

Fig. 2 shows the - for clarification compared to Fig. 1 - changed geometric relationships when recording the individual images. Three radiation source positions are shown, the radiation source position P _{0 being} in the middle of the range of motion and the radiation source positions P _-N and P _{+ N being} at the ends of the range of motion. The individual images B ₀ , B _-N and B _{+ N are} generated in the three positions. For the sake of simplicity, it is assumed that each individual image contains only five pixels which (at their center) are connected to the associated radiation source position by straight lines.

For the generation of a slice image sharply depicting a specific slice, the image values for each slice pixel must be superimposed by those pixels that are geometrically assigned to this slice pixel. For example, for a slice image of the Fulcrum plane EF, the slice image value for the slice pixel S _{01 must be determined} from the slice image value of the pixel on the far left in each of the individual images. The other slice image values for the fulcrum level can be determined analogously. In general, it applies to this layer image plane that a layer image point with the coordinates x and y is assigned a layer image value S ₀ (x, y) from the individual images the image values B _n (x, y) of the image points with the coordinates x, y. The value n ranges from -N to + N.

A different combination of image values from the other image values must be used for the other layers. For example, for the left layer pixel S _{11 of} the layer S ₁ lying below the fulcrum plane (in which the straight lines leading from the radiation source positions to the pixels of the individual images intersect), the image value of the left pixel in the individual image B _-N , the image value of the second image point from the left in the individual image B ₀ and the image value of the central image point in the individual image B _{+ N} are used. It generally applies to this layer that the layer image value S ₁ (x, y) is composed of the image values B _n (x + n, y) for a pixel with the coordinates x, y. The coordinates x, y count from the straight line G which connects the central radiation source position P ₀ to the center of the single image B ₀ generated from this position. It is further assumed that the x-direction coincides with the direction of movement of the radiation sources and y represents the horizontal direction perpendicular thereto (perpendicular to that to the plane of the drawing in FIGS. 1 and 2).

Similarly, one finds for the layer S _-1 (this is the first layer above the Fulcrum plane EF) that its layer image values S _-1 (x, y) are derived from the pixels B _n (xn, y). In general, for pixels in a layer S _i (where i is a positive or negative integer which indicates the position of this layer from the layer in the Fulcrum plane), the layer image values S _i (x, y) result from one another Derive image values B _n (xi • n, y).

3 shows a schematic block diagram, with the aid of which images of a layer S _i can be generated from the stored individual images. The images B _n stored therein may preferably be stored in a standardized form in such a way that all individual images may be identical for a uniformly thick, homogeneous object in the examination area. A layer image S _{i is} generated from the individual images stored in the image memory 7 by an image processing unit 8 and displayed on a suitable display unit 9, for example a video monitor. The user can on the one hand specify the position i of the layer S _{i to} a suitable setting unit 10 and on the other hand set the extent to which he wants to eliminate contrast reversal and resolution effects.

The image processing unit 8 contains a computing and control unit 80, which determines the address for the calculation of layer image values of layer image points with the coordinates x, y in a layer i, under which the image value B _n (x + i • n, y ) is saved. This image value is called up from the memory 7 and fed to a multiplier 81. In this multiplication stage, the image value is multiplied by a weighting factor g _j (n). This weighting factor is the same for all pixels of the single image B _n . The weighting factor is taken from a memory 82.

The image processing unit 8 further contains a layer image memory 83, in which - after the reconstruction of a layer image has been completed - the layer image values S _i (x, y) are stored. The computing and control unit 80 forms the address x, y of these slice image values and calls up the value stored under this address. At the beginning of the reconstruction, this value is zero. It is added to the value calculated by the multiplier 81 with an adder 84, and the sum thus formed is stored in the memory 83 at the previously called address. The arithmetic and control unit 80 then calculates the addresses in the memories 7 and 83 for a different value of x until all x values of a layer image line have been processed. This is then repeated for another value of y until all contributions of a single image B _n to the slice image have been determined in this way. The procedure described is then repeated for all individual images, after which the slice image values S _i (x, y) are available in the memory 83. The slice image thus calculated is displayed on a monitor 9.

According to the schematically represented block diagram of FIG. 3, the Multiplication and addition by hardware units 81 and 84 carried out. However, it is clear to the person skilled in the art that the Image processing unit can contain a microcomputer that the Address calculations, multiplications and additions in software performs.

If the weighting factors g _j (n) all had the value 1, then the described image processing method would lead to the same images that result from the method mentioned at the beginning, in which the image values of the various individual images geometrically assigned to a layer pixel are added to one another. This results in a modulation transfer function for objects which are located in the examination area at a defined distance from the sharply imaged layer S _i , which is shown in FIG. 5e. It can be seen that spatial frequencies below the value 1 are transmitted the better the closer the spatial frequency is to zero. In the area between the spatial frequencies 1 and 2, the curve of FIG. 5e has negative values, that is to say that the contrast is reversed (areas of strong absorptions are therefore imaged as if only a small absorption would take place there and vice versa). In the spatial frequency range between 2 and 3 there is again a normal contrast - although not as great as in the range between 0 and 1. Since the contrast is already zero at the local frequency 1, this course of the modulation transfer function leads to one in the spatial frequency range between two and three Pseudo resolution. In the spatial frequency range above three, the modulation transfer function shows further negative and positive vibrations, the amplitude of which is however becoming smaller and smaller.

Fig. 5e applies in principle to all layers outside of the Reconstruction of sharply rendered slice. The limit of real Resolution shifts all the more to low values of the spatial frequency towards, the greater the distance of the respective structures from the sharp pictured layer is in the examination area. If you do that Takes into account that one varies the unit of the spatial frequency, 5e applies to all of these layers.

If there are any significant structures in the layers that are not sharply imaged that fall within the spatial frequency range between 1 and 2 or 2 and 3 or an area lying above them, then the image reversal or pseudo-resolution effects result in the diagnostic evaluation of the layer image can complicate. In order to suppress these effects, the weighting factors provided for the various images must depend on the radiation source position when the individual image in question is generated, so that the greater the distance between the radiation source position and the mean radiation source position (P ₀ ), the smaller the weighting factor .

Various dependencies are possible here. However, a particularly favorable compromise between the suppression of the effects mentioned on the one hand and the blurring of details outside the layer on the other hand arises if the following equation is fulfilled: G j (n) = 1 + a j1 • b 1 + a j2 • b 2 The index j is intended to indicate that there are several sets of weighting factors (with their own factors a _j1 and a _j2 ), one of which can be selected for a _{slice image} . The following applies to the factors b _k (with k = 1.2) b k = √2 • cos (π • k • n / N) N denotes the number of the radiation source position to the left or right of the center and n indicates the number of radiation source positions (seen from the center) in which the radiation source was when the single image b _n was generated. Further applies a j1 = 0.693; a j2 = 0.0267 (j = 1)

If the weighting factors g _j (n) calculated in this way are plotted as a function of the value n / N and these factors are connected to one another with an envelope, the curve (a) in FIG. 4 results. FIG. 5a represents the modulation transfer function for a plane lying outside the sharply depicted layer plane if the layer image is created with this set of weighting factors. It can be seen on the one hand that there is no reversal of contrast and no pseudo resolution, because the contrast decreases monotonically as a function of the spatial frequency. However, it can also be seen that only structures above a spatial frequency of 2 no longer produce any contrast in the sharply imaged layer, that is to say the blurring effect is only about half as great as in the modulation transfer function according to FIG. 5e, where the spatial frequency 1 already Contrast disappears.

For these reasons, it only makes sense to weight the individual images with the weighting factor set according to equations (1) to (3) if relatively strong contrast reversal or pseudo-resolution effects become noticeable in a slice image. If these effects are less pronounced, it may make sense to use a set of weighting factors that has a lower dependence on n than the weighting factor set shown in FIG. 4 by the envelope (a). Such a weighting factor set can be calculated using equations (1) and (2) if a _j1 and a _{j2 are} selected in accordance with a j1 = 0.5358; a j2 = - 0.0316 (j = 2) The curve (b) in FIG. 4a shows the associated envelope and FIG. 5b the resulting modulation transfer function. It can be seen that the blurring effect is more favorable than in FIG. 5a, but that slight phase reversal or pseudo-resolution effects can possibly occur.

Another set of weighting factors is obtained using equation (1) and equation (2) if one _{chooses the} following values for a _j1 and a _j2 : a j1 = 0.3248; a j2 = - 0.0494 (j = 3) The curve (c) in FIG. 4 represents the associated envelope and FIG. 5c the modulation transfer function. The blurring effect is stronger than, for example, in FIG. 5b, but phase reversal and pseudo-resolution effects can also be stronger - but not as strong as in FIG. 5e, which shows an even stronger blurring effect.

FIG. 5d shows a modulation transfer function in which the disturbing or desired effects are more pronounced than in FIG. 5c, but not yet as strong as in FIG. 5e. The associated envelope is represented by curve (d) in FIG. 4 and the associated weighting factors are calculated from equations (1), (2) with the values a j1 = 0.0814; a j2 = - 0.0182 (j = 4)

In the case of a tomographic examination, 2N + 1 individual images are first produced and then the user specifies the parameter i via the input unit c, which parameter specifies which layer the layer image to be generated is to be depicted sharply. A certain weighting factor set g _j (n) can initially be provided by the setting unit 10 by presetting j, for example that according to curve (a) in FIG. 4 or according to equations (1) to (3). This weighting factor set, like the other possible weighting factor sets, is stored in the memory 82, and the weighting factor g _j (n) that belongs to the respective individual image B _n is called up by the computing and control unit.

After a layer image has been generated in this way, the user can cause a stronger blurring effect by having another Weighting factor set, e.g. 4 (b) with the Modulation transfer function according to Fig. 5b (j = 2). If there too No disruptive effects can occur by choosing another one Weighting factor set the blurring can be further increased until a reasonable compromise between blurring on the one hand and annoying Effects on the other hand is achieved. If necessary, it can be useful yet another set of. not explained in FIGS. 4 and 5 To provide weighting factors, in which the weighting factors with increasing quotient n / N increase. But you will only use it make, even if all the weighting factors are the same Effects do not appear in the slice image; the blurring is then still more pronounced than in Fig. 5e. - When creating layer images of others Layers of the examination area are moved accordingly.

In the arrangement explained in Figures 1 and 2 move X-ray tube 1 and image recording device 3 in opposite directions on straight lines to each other. The invention is also applicable to arrangements in which x-ray emitter and image recording device in opposite directions to each other moved on circular paths around a horizontal axis in the Fulcrum plane become. Also a combination (circular path of the spotlight and straight line Path of the image recording device) is possible. However, this changes the magnification factor, so that the images are initially geometrical Undergo transformation before they are added.

In the exemplary embodiment explained above, the radiation source positions were symmetrical with respect to point 4 (FIG. 1) in the Fulcrum plane, their distance from one another was the same, and their number was odd. Although these requirements are expedient, they can be omitted individually or as a whole. The "mean radiation source position" in the sense of the invention is then the geometric center between the two outer positions; this position does not have to be identical to one of the radiation source positions for the individual images. To calculate the weighting factor sets according to equation (1) ff, the expression n / N in equation (2) must then be replaced by x / x _o , where x denotes the distance of the respective radiation source position from the center and x _o the center distance of the outer two radiation source positions .

Claims (3)

1. A method of generating slice mages of an examination zone which is irradiated by X-rays from a plurality of radiation source positions in order to generate separate single images, for each single image there being stored image values which correspond to the absorption in its image points, slice images being generated in that slice image values are derived from the image values of the image points of the single images which are geometrically associated with a slice image point, characterized in that the slice image values are derived by a weighted summing operation where the weight with which to the image values of the single images enter the summing operation is smaller as the distance between the radiation source position and a central radiation source position was greater upon generation of the relevant single image.
2. A device for carrying out the method claimed in Claim 1, comprising at least one X-ray source for irradiating an examination zone from a plurality of radiation source positions which are situated in one plane, an image converter for converting the single images picked up in the radiation source positions into image values, and an image storage device for storing the image values of the single images, an image processing unit for generating slice images from the image values of the single images, characterized in that the image processing unit comprises an arithmetic unit which, in order to generate slice images, multiplies the image values by a weighting factor which is dependent on the radiation source position upon generation of the associated single image and which sums the image values, thus weighted, of the image points geometrically associated with a slice image point.
3. A device as claimed in Claim 2, characterized in that the image processing unit comprises a memory in which several, different sets of weighting factors are stored, there being provided selection means for the preferably interactive selection of a set of weighting factors.

Images