GB1603568A

GB1603568A - Optically forming a selected sectional-layer-image from a composite of superposed shadowgraph images of a three-dimensional object

Info

Publication number: GB1603568A
Application number: GB1674578A
Authority: GB
Original assignee: Philips Gloeilampenfabrieken NV
Current assignee: Koninklijke Philips NV
Priority date: 1977-04-30
Filing date: 1978-04-27
Publication date: 1981-11-25
Also published as: FR2389154B1; DE2719386C3; DE2719386A1; CA1101705A; FR2389154A1; DE2719386B2; JPS54893A

Description

(54) OPTICALLY FORMING A SELECTED SECTIONAL LAYER-IMAGE FROM A COMPOSITE OF SUPERPOSED SHADOWGRAPH IMAGES OF A THREE-DIMENSIONAL OBJECT (71) We, N.V. PHILIPS' GLOEILAM PENFABRIEKEN, a limited liability Company, organised and established under the laws of the Kingdom of the Netherlands, of Emmasingel 29, Eindhoven, the Netherlands, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: The invention relates to a method of reconstructing an image of a selectable transverse section of a three-dimensionsal object, e.g. part of a human body, from a composite image comprising overlapping two-dimensional shadowgraph images of said three-dimensional object formed in a recording plane by a plurality of irradiating point sources from a corresponding plurality of different perspective directions. The shadowgraph images making up the composite image are all recorded on the same recording material, eg photographically as radiograms, and the composite image can be regarded as being source-position encoded since the relative positions of the overlapping images depend on the relative positions of the irradiating point sources.

It is known that three-dimensional objects can be source-position encoded by exposing the object to incoherent light or to X-rays from different positions and by photographically recording the shadow images on one and the same film. Thus a sourceposition enclosed composite image is obtained, from which the information about the object cannot be retrieved directly.

Discrete planar sections of said object cannot be visualized until after a second step, decoding of the source-position encoded composite image.

Methods of decoding such a composite image have been proposed. In accordance with U.K. Patent Specification Number 1,505,999 decoding is effected by replicating and shifting the composite image in an incoherent optical Fourier arrangement, the source distribution being stored in a Fourier hologram. Because of a limited aperture diameter of the Fourier-transform lens the primary composite image must first be reduced photographically.

In a German Published Patent Application P 24 32 595.9 the composite image is replicated by a so-called point hologram, whose points corresponds to the source distribution, and the sectional layer is reproduced by changing the scale of the composite image by means of a zoom lens. This method imposes stringent requirements (high speed, distortion free imaging) on the zoom lens, when a large-format composite image is to be processed.

It is an object of the invention to provide a method of the kind mentioned in the preamble, which also enables large-format composite images to be processed without lenses.

According to the invention there is provided a method of reconstructing an image of a selectable transverse section of a three-dimensional object from a composite image comprising overlapping twodimensional shadowgraph images of said three-dimensional object formed in a recording plane by a plurality of irradiating point sources from a corresponding plurality of different perspective directions, said method comprising viewing the illuminated composite image through an imagemultiplying ocular device arranged to generate a plurality of virtual images of said composite image respectively angularly displaced relatively to one another so as to correspond to the relative angular dispositions of the set of perspective directions from which the object was illuminated by the point sources in order to form said composite image, and varying the size of said composite image or varying the dis tance between said composite image and said image-multiplying ocular device in order to obtain clear virtual images of various transverse sections of the object.

An image-multiplying device is to be understood herein to refer to a device which generates a plurality of relatively displaced but otherwise substantially identical versions of an image.

The decoded virtual image of a sectional layer can be observed by the naked eye through the image-multiplying ocular device. It is particularly advantageous to use an amplitude or phase point hologram as the image multiplying device. An arrangement with birefringent prisms is also suitable as an image multiplying device.

The novel method may be employed in a particularly advantageous manner in medical X-ray imaging of fast-moving threedimensional objects, such as a pulsating heart. Shadowgraph images of the moving object are then simultaneously flash projected onto a single film by a plurality of X-ray tubes from different positions and subsequently, after development of the film, it can be decoded at will with respect to any sectional layer by a method embodying the invention. Object sections can be decoded directly from a large-format primary radiographic composite image; a time-consuming and expensive reduction of the composite radiograph is no longer necessary. Moreover, a method of observing the decoded layers without the use of lenses in accordance with the invention is compatible with the now generally used method of observing X-ray images on a light box.

Embodiments of the invention will now be described by way of example, with reference to the accompanying drawings, of which: Figure I schematically illustrates the principle of recording the point image function of a multiple radiation source assembly, Figure 2 illustrates the multiple projection of a three-dimensional object, Figure 3 illustrates an optical arrangement for recording a point hologram to form an image-multiplying device, Figure 4 shows an embodiment of an observational arrangement for decoding selectable discrete sectional layers of a three-dimensional object from a sourceposition encoded composite image, Figure 5 shows a further embodiment of an observational arrangement for decoding selectable discrete sectional layers from source-position encoded composite images which have been recorded on cine film, Figure 6a-d schematically illustrate the decoding process applied to a composite image with respect to two recording source positions and two planes, and Figure 7 shows an embodiment employing an arrangement of birefringent prisms.

In Figure 1 a point image field representing the distribution of individual sources RQ1 and RQ2, for simplicity of illustration formed by two X-ray tubes, is recorded by means of a pinhole camera diaphragm LKI in the plane P. Each source forms a corresponding shadowgraph image of a point Oi via the pinhole aperture in the diaphragm LK1 as a respective image point I and II.

This point image field recorded in the plane P, contains information about the positions of the sources RQ1 and RQ2 relative to the point 1. If then, instead of the pinhole camera diaphragm LK1, a threedimensional object 0 were to be included in the irradiation path, each point of the object will cast a shadow image at two corresponding points, for example a point 2 in the plane LK2 will relate to a point III and to a point (IV) (which latter will not necessarily coincide with the point I as shown in Figure 1). Thus, two radiographic images would be superimposed to form a source-position encoded composite image of the threedimensional object 0.

In Figure 2 an extension of the recording principle to several radiation sources RQ is illustrated by an example in which geometrical figures F1 and F2 are located in two spaced sectional planes of the object 0. B represents the source-position encoded composite image of these geometrical figures.

Figure 3 shows how by means of the point image recording in the plane P (Figure 1) and representing the source distribution RQ1 and RQ2 of Figure 1, an imagemultiplying ocular device in the form of a point hologram H, is recorded photographically. A plane coherent light wave front LS1 from a laser illuminates the planar image P which is arranged so as to provide, in consequence, the effect of two point light sources via transparent point regions at the respective image points I and II, which relative to each other correspond to have the positions of the corresponding sources relative to the point Oi shown in Figure 1.

By means of a reference wave LS2 the point hologram H of the points I and II is recorded. This hologram stores the information about the source directions represented by the points I and II, i.e.information about the angle a between respective pairs of source directions. In this way the directions relating to all the emitting points of the radiation source assembly RQ of Figure 2 can be stored in a point hologram H.

During the synthesising construction of a sectional image using the point hologram H as an image-multiplying ocular device (Figures 4 and 5), each image point viewed therethrough is replicated by the number of points stored in the point hologram. When the composite image B formed as shown in Figure 2 is thus decoded. each image point in the image B is correspondingly replicated by means of the point hologram H through which it is viewed. This property of a point hologram. together with a change in the size or the relative scale of the composite image B by means of a zoom lens (eg V in Figure 5) or a change in the distance (eg I in Figure 4) between the point hologram H and the composite image B viewed therethrough, is employed for selectively decoding the various individual layers F1 and F2 of the three-dimensional object 0 illustrated in Figure 2.

Figure 4 shows an embodiment of the optical observational arrangement for decoding the source-position encoded composite images (B in Figure 2). A light source LQ (for example a Hg high-pressure lamp) or a plurality of light sources illuminate the radiographic composite image B. Behind the composite image B in the observational direction, a frosted-glass screen M is placed, so that a luminous image form of reproduction is obtained. The complete arrangement comprising the light source LQ, frosted glass screen M and radiographic composite image B may be formed by a conventional light-box viewing apparatus L. An image multiplying ocular device formed by the point hologram H produced by the arrange ment of Figure 3, and which is disposed at some distance in front of the composite image B in the viewing direction, causes an observer to observe a plurality of replicas of said image along directions corresponding to the stored relative dispositions of the source images I and II to form the image Bl,ll. At the distance 11 (source-position encoded composite image B - point holo gram H) a sectional layer image S1 of the encoded three-dimensional object, for ex ample an image of the sectional example layer F2 (triangle) of the object 0 of Figure 2. is reinforced in the composite image superposition region of Bl Il. The sectional layer image Si is observed as a virtual image by a naked eye A through the point holo gram H. By changing the distance 1 (compo site image B - point hologram H) the various layers successive transverse sections Sn of the three-dimensional object, can be de coded consecutively. In order to ensure that the point of observation for the observer (the eye A) does not change, the position of the point hologram H is maintained station ary and only the light box L together with the source-position encoded composite im age is moved in the direction of the arrow only. At the distance 12 a different sectional layer Sl of the three dimensional object is decoded, for example the sectional layer F, (circle) of the object 0 of Figure 2. The point hologram H is combined with a narrow-band interference filter J, so as to pass substantially only one of the various spectral lines emitted by the lamp which is used. It will be apparent that such filtering may alternatively be effected adjacent the light source LQ.

In Figure 5 the a succession of sourceposition encoded radiographic composite images B are first photographically recorded on a cine film F which is then threaded into a projector PJ and back-projected onto a translucent projection screen W by said projector PJ (light source LQ, filter J, film transport FT and zoom lens V). An imagemultiplying ocular device comprising the point hologram H of Figure 3, replicates the projected composite image B in accordance with the stored directional distribution of sources (I and II) so that an image superposition region B 1.11 is obtained provided in which displaced virtual image replicas of the composite image B' are superimposed with respect to an observer A. For a specific size of the projected composite image B' a sectional layer image Svl of the sourceposition encoded three-dimensional object, for example the sectional layer F2 of the object 0 shown in Figure 2, is obtained in the superposition region B'1,11. By means of the zoom lens V (arrow) of the projector PJ, the size of the projected composite image B' (arrow) can be varied, so that respective images of other sectional layers Sv of the three-dimensional object can be observed, for example the sectional layer F1 of the object 0 shown in Figure 2. By cinematographic projection of a succession of sourceposition encoded composite images B on a film F it is possible to provide motionpicture images of the sectional layers.

The decoding operation will now be described in more detail in a simplified manner with reference to Figure 6. for only two radiation sources RQI and RQIl and only two object planes, a circle K and a triangle D.

Figure 6a schematically represents a composite image B of the geometrical figures F1 and F2 of Figure 2. For reasons of clarity the images K1 K11 and Dl, Dull, of the objects are shown separately.

Figure 6c represents the replication of the composite image B of Figure 6a with respect to a reference point PR by means of the point hologram H of Figure 3 using the stored points I and II. If the distance a between the circles Kl and K11 in the composite image B (Figure 6a) is equal to the distance a' between the corresponding source position related points I and II stored in the hologram H (Figure 6c), decoding of the sectional plane containing the circle is effected because the circles Kl and K11 will then be constructively superimposed to form an image Kl ll, as shown in Figure 6c.

The remaining images which are shown dashed in Figure 6c, are those components of the replicated composite images derived from Figure 6a, which are not superimposed constructively, but which only disturb the constructively superimposed image K1,11.

For the triangles D1 and D11 decoding is effected for example by a reduction V of the size of the composite image B of Figure 6a by means of the zoom lens V of Figure 5, to provide the reduced composite image B shown in Figure 6b.

In Figure 6b the distance a between the triangles D1 and D11 in the reduced image, is the same as the distance a' apart of the points I and II stored in the point hologram H (Figure 6d), so that these are constructively superimposed to form an image DI II in Figure 6d. Reduction of the composite image B shown in Figure 6a to that shown in Figure 6b thus results in an adaptation of the image overlap required to synthesise the image of a corresponding object section by image reinforcement to the fixed angular overlap provided by the hologram H, and this is achieved by means of the zoom lens V of Figure 5.

The previously decoded circles Kl and KII are now no longer constructively superimposed and are also shown dotted in a similar manner to that of the other replicated secondary images. If the object 0 is recorded using more than two sources, for example n sources, the constructive image will accordingly be "boosted" n-times relative to the secondary images which cannot be superimposed constructively when a nonredundant source distribution is employed, as discussed in UK Patent Number 1,505,999. Such distributions are more generally described by M.J.E. Golay in J. Opt.

Soc. Amer. 61 272.

Figure 7 shows a possible arrangement of birefringent prisms (for example calcite prisms), which may be used as the imagemultiplying ocular device.

The source position encoded composite image 1 in the plane 11 with the individual overlapping radiographic images 2, 3, 4 and 5, is imaged on the retina by eye 8 through an arrangement of for example two birefringent calcite prisms 6 and 7. During observation through the two prisms a virtual image 9 is formed. By a suitable choice of the prisms the individual component radiographic images 2, 3, 4 and 5 of the replicated composite images 1, are superimposed in the virtual image in such a way that a clear image of a sectional layer 12 of the object is obtained. By changing the distance 10 the three-dimensional object can be decoded so that images of a continuous succession of sectional layers are formed.

WHAT WE CLAIM IS: 1. A method of reconstructing an image of a selectable transverse section of a three-dimensional object from a composite image comprising overlapping twodimensional shadowgraph images of said three-dimensional object formed in a recording plane by a plurality of irradiating point sources from a corresponding plurality of different perspective directions, said method comprising viewing the illuminated composite image through an imagemultiplying ocular device arranged to generate a plurality of virtual images of said composite image respectively angularly displaced relatively to one another so as to correspond to the relative angular dispositions of the set of perspective directions from which the object was illuminated by the point sources in order to form said composite image, and varying the size of said composite image or varying the distance between said composite image and said image-multiplying ocular device in order to obtain clear virtual images of various transverse sections of the object.

2. A method as claimed in Claim 1, wherein said image-multiplying ocular device is an amplitude or a phase pointhologram.

3. A method as claimed in Claim 1, wherein said image-multiplying ocular device is constituted by an arrangement of birefringent prisms.

4. A method as claimed in any one of Claims 1, 2 and 3, wherein a succession of composite images respectively formed using different source distributions (point codes) are decoded in rapid succession to form corresponding images of the same transverse section and said corresponding images are additively combined.

5. A method as claimed in any one of the preceding claims, wherein the composite image is illuminated by a light source which emits light radiation restricted to a narrow frequency band.

6. A method as claimed in Claim 5, wherein the light source includes a lamp emitting a line spectrum, and a single spectral line is selected for illuminating the composite image, by means of a filter.

7. A method as claimed in any one of Claims 1 to 4, wherein the imagemultiplying ocular device is combined with a narrow band optical filter.

8. A method as claimed in any one of the preceding Claims, wherein a diffusing screen is disposed behind the composite image.

9. A method as claimed in any one of Claims 1 to 7, wherein the superposition image is displayed on a light box.

10. A method as claimed in Claim 8 or Claim 9, wherein the composite image together with the diffusing screen or light box are arranged to be displaceable along

**WARNING** end of DESC field may overlap start of CLMS **.

Claims

**WARNING** start of CLMS field may overlap end of DESC **. The remaining images which are shown dashed in Figure 6c, are those components of the replicated composite images derived from Figure 6a, which are not superimposed constructively, but which only disturb the constructively superimposed image K1,11. For the triangles D1 and D11 decoding is effected for example by a reduction V of the size of the composite image B of Figure 6a by means of the zoom lens V of Figure 5, to provide the reduced composite image B shown in Figure 6b. In Figure 6b the distance a between the triangles D1 and D11 in the reduced image, is the same as the distance a' apart of the points I and II stored in the point hologram H (Figure 6d), so that these are constructively superimposed to form an image DI II in Figure 6d. Reduction of the composite image B shown in Figure 6a to that shown in Figure 6b thus results in an adaptation of the image overlap required to synthesise the image of a corresponding object section by image reinforcement to the fixed angular overlap provided by the hologram H, and this is achieved by means of the zoom lens V of Figure 5. The previously decoded circles Kl and KII are now no longer constructively superimposed and are also shown dotted in a similar manner to that of the other replicated secondary images. If the object 0 is recorded using more than two sources, for example n sources, the constructive image will accordingly be "boosted" n-times relative to the secondary images which cannot be superimposed constructively when a nonredundant source distribution is employed, as discussed in UK Patent Number 1,505,999. Such distributions are more generally described by M.J.E. Golay in J. Opt. Soc. Amer. 61 272. Figure 7 shows a possible arrangement of birefringent prisms (for example calcite prisms), which may be used as the imagemultiplying ocular device. The source position encoded composite image 1 in the plane 11 with the individual overlapping radiographic images 2, 3, 4 and 5, is imaged on the retina by eye 8 through an arrangement of for example two birefringent calcite prisms 6 and 7. During observation through the two prisms a virtual image 9 is formed. By a suitable choice of the prisms the individual component radiographic images 2, 3, 4 and 5 of the replicated composite images 1, are superimposed in the virtual image in such a way that a clear image of a sectional layer 12 of the object is obtained. By changing the distance 10 the three-dimensional object can be decoded so that images of a continuous succession of sectional layers are formed. WHAT WE CLAIM IS:

1. A method of reconstructing an image of a selectable transverse section of a three-dimensional object from a composite image comprising overlapping twodimensional shadowgraph images of said three-dimensional object formed in a recording plane by a plurality of irradiating point sources from a corresponding plurality of different perspective directions, said method comprising viewing the illuminated composite image through an imagemultiplying ocular device arranged to generate a plurality of virtual images of said composite image respectively angularly displaced relatively to one another so as to correspond to the relative angular dispositions of the set of perspective directions from which the object was illuminated by the point sources in order to form said composite image, and varying the size of said composite image or varying the distance between said composite image and said image-multiplying ocular device in order to obtain clear virtual images of various transverse sections of the object.

the optical axis of the arrangement.

11. A method as claimed in any one of Claims 1 to 7, wherein the composite image is projected for observation onto a transparent projection screen.

12. A method as claimed in Claim 11, wherein the size of the projected composite image is varied by means of a zoom lens.

13. A method as claimed in Claim 11 or Claim 12, wherein a sequence of composite images formed at regular intervals are projected cinematographically.

14. A method as claimed in any one of the preceding claims, wherein the virtual image of the selected transverse section formed by the image-multiplying ocular device, is observed via a telescope lens.

15. A method as claimed in any one of Claims 1 to 13, wherein the virtual image of the selected transverse section is recorded via the image-multiplying ocular device by means of a TV-camera and is displayed on a monitor.

16. A method of reconstructing an image of a selectable transverse section of a three-dimensional object from a composite image comprising overlapping twodimensional shadowgraph images of said three-dimensional object formed in a recording plane by a plurality of irradiating point sources from a corresponding plurality of different perspective directions, substantially as herein described with reference to Figures 4, 5 and 7 of the accompanying drawings.