DE10311172A1 - Image segmenting of objects, especially of human/animal organs and/or results of pathological changes, involves generating object/part profiles from three-dimensional image information in image planes - Google Patents

Abstract

The method involves using three-dimensional image information of at least one object and/or object part, generating first (1) and second (2 or 3) profiles of the object/part from the three-dimensional image information in mutually perpendicular first (Z) and second (X or Y) image planes and automatically generating profiles in further planes parallel to the first plane using points in the second image plane. An independent claim is also included for the following: (a) a computer program for implementing the inventive method (b) a data medium for loading into a computer for implementing the inventive method (c) a computer program product for implementing the inventive method (d) and an arrangement for image segmenting of objects, especially human or animal organs and/or results of pathological changes.

Description

The invention relates to the image segmentation of Objects, especially for segmenting human or animal Organs and / or results of pathological changes, being local three-dimensional image information of at least one object and / or of an object part can be used.

Prostate cancer is a common diagnosis in men at the age of 50 or older. Prostate cancer is the second most common form of cancer worldwide Leads to patient death.

Symptoms due to prostate cancer are usually not evident until the carcinoma has already led to considerable prostate growth or until metastases have already developed. In contrast, tumors with a size of 1 to 1.5 cm ^{3 are} usually still limited to the diseased organ and can be treated successfully. It is therefore important to identify growth early and reliably. It is also important during treatment to specifically destroy the cancer cells and maintain the healthy cell material.

The use of ultrasound for Generation of three-dimensional image information, in particular in the diagnosis of prostate cancer, for example, in the US patents 5,562,095, 5,454,371 and 5,842,473. The present However, the invention is not limited to the use of ultrasound. alternative or additionally to ultrasound images the image data are obtained in any other suitable way, z. B. by means of computer tomography, magnetic resonance imaging, positron emission tomography (PET) and / or classic X-ray methods.

The above, related Statements made with the prostate also affect other organs to. There is also outside cancer diagnosis and therapy areas of application in which similar diagnostic approaches be followed, such as determining the volume of a heart muscle. Here, too, three-dimensional image information is generated and is precise and strive for quick segmentation of the object.

WO 01/89386 A1 describes the automatic, model-based segmentation of prostate, rectum and urethra in Ultrasound images. Furthermore, in the introduction to this document described a segmentation method in which in a variety of ultrasound images in mutually parallel planes, each manually a contour of the prostate is lined. This enables the determination of the total volume the prostate.

EP 1 076 318 A1 describes the marginal segmentation of the prostate from two-dimensional and three-dimensional ultrasound images. After manually positioning points on the outer contour of the prostate, further segmentation is automatically carried out using a model. The purely manual contouring is described in the scripture as time-consuming, exhausting and therefore unsuitable during an operation.

Previously known automatic, model-based segmentation methods have the disadvantage that there is a considerable percentage of copies of the respective organ or object in which the model fails. For example, the automatic, model-based segmentation according to the EP 1 076 318 A1 an organ with an unusual bulge cannot be contoured.

It is a task of the present Invention, a method and an apparatus for image segmentation of objects, especially for segmenting human or animal organs and / or results of pathological changes in the human or animal body, to indicate a precise and fast segmentation of an object and / or an object part enable.

Objects or object parts should also be included atypical and / or unusual Shapes can be segmented.

• – örtlich dreidimensionale Bildinformationen zumindest eines Objektes und/oder eines Objektteils verwendet werden,
• – aus den dreidimensionalen Bildinformationen in einer ersten Bildebene ein erster Umriss des Objektes und/oder des Objektteils erzeugt wird,
• – aus den dreidimensionalen Bildinformationen in einer zweiten Bildebene, die quer zu der ersten Bildebene verläuft, ein zweiter Umriss des Objektes und/oder des Objektteils erzeugt wird,
• – automatisch in einer Mehrzahl weiterer Bildebenen, die parallel zueinander und parallel zu der ersten Bildebene verlaufen, jeweils ein weiterer Umriss des Objektes und/oder des Objektteils erzeugt wird, wobei in der zweiten Bildebene liegende Punkte der weiteren Umrisse durch den zweiten Umriss vorgegeben sind.

A method for image segmentation of objects, in particular for the segmentation of human or animal organs and / or results of pathological changes, is proposed, wherein

• Locally three-dimensional image information of at least one object and / or part of an object is used,
• A first outline of the object and / or the object part is generated from the three-dimensional image information in a first image plane,
• A second outline of the object and / or the object part is generated from the three-dimensional image information in a second image plane, which runs transversely to the first image plane,
• - A further outline of the object and / or the object part is generated automatically in a plurality of further image planes that run parallel to one another and parallel to the first image plane, points of the further outlines lying in the second image plane being predetermined by the second outline.

Segmentation means any marking, ruling, contouring or other generation of an outline of the object and / or the object part. The outline need not be a complete outline and not a real, truthful outline. Rather, it may also be desirable to define an image subarea to eliminate irrelevant image information by specifying the outline. B. contains the object and / or another object. Therefore, the term object is not necessarily synonymous with an organ or other delimited real object. For example using the method, a user segment an object which has a tumor and an area surrounding the tumor. Even if it is so in many practical cases, the outline does not have to correspond to an outer contour of an object and / or partial object. Furthermore, the outline can also correspond wholly or partially to an inner contour of an object. It is also possible to segment not just one object, but a plurality of objects. For example, the relative spatial position of the objects can be preserved and / or displayed.

The segmentation becomes advantageous for image data from human or animal organs and / or of results of space-consuming processes in human or animal body applied. Examples are: prostate, urethra, brain, eye, lungs, Heart, stomach, liver, kidney, a cyst and / or a pathological change (especially tumor), or a part of each. This enumeration is not concluding to understand.

By creating the first outline and the generation of the second outline from the image information the segmentation accelerated considerably because of the wider outlines not individually and not in the planes parallel to the first image plane must be generated one after the other. Furthermore, the generation offers in different, cross (in particular mutually perpendicular planes have the advantage that the image information in certain circumstances can be displayed with better contrast than with the individual Segmentation in a large number of parallel image planes the case is. Is the image information in the case of the method according to the invention in one of the two mutually opposed levels of worse Quality, can switch to a parallel image plane and in this Image plane of the outline are generated. This happens e.g. B. thereby that the common intersection line of a user in the first image plane the first and second image plane or the common intersection the first to third image level.

The first outline does not have to be before the second outline. Rather, the first outline can be whole or partially after the second (and or after that described below third outline) are generated and / or changed.

It is included with the ending, that the first and / or the second outline automatically or semi-automatically is generated using a model. Preferably however these two outlines as well as the optional third outline on the even closer "manual" is generated. "manual" is not literal in to understand the sense that someone's hands are used Need to become. Rather you can also or exclusively other means are used, such as. B. a speech recognition device, with which a person controls or specifies the generation of the outlines.

This embodiment of the invention is particularly beneficial as it takes advantage of manual generation combines the advantages of automatic segmentation. The Can outline namely from a user with large reliability and precision generated, which ensures is that unusual too Object shapes are handled correctly. Still, a small one is enough Number of outlines (in the simplest case the first and the second Outline) that are created manually. A variety of others Outlines can then be created automatically from the manually generated outlines be generated. In particular, part of the three-dimensional Image information is displayed using image display means and submits a user of the method, d. H. for example the one mentioned Person, based on the displayed image information, the first and / or the second outline.

In a particularly preferred further training is transformed from the three-dimensional image information into a third Image plane which is transverse (in particular perpendicular) to the first and transverse (in particular perpendicular) to the second image plane third outline of the object and / or the object part generated, wherein points of the further outlines lying in the third image plane the third outline are given. Thus, as a rule, at least four points of further outlines through the second and third Outline given.

It is also possible, in at least one image plane, that runs parallel to the second or third image plane, one additional second or additional to generate the third outline from the image data, through which in turn usually two additional Points of the further outlines can be specified.

It is further preferred that from the three-dimensional image information at least in a fourth Image plane that runs parallel to the first image plane additional Outline of the object and / or the object part is generated, in particular manually. Especially for objects whose shape is transverse in one direction changed significantly to the first image level is the additional one Outline beneficial for a precise Segmentation.

An advantageous way the other outline is to provide information about the Shape and / or size of the first Outline and / or, if available, information about the shape and / or size of the additional Outline to use.

In one configuration, temporary outlines are generated in the further image planes using information about the shape and / or size of the first outline and or, if available, using information about the shape and / or size of the additional outline, and are generated using informatio generates the further outlines from the temporary outlines using the size of the second outline and / or, if available, using information about the size of the third outline. The temporary outlines do not have to be fully defined or even displayed graphically before the other outlines are created. Rather, z. B. first coordinates of points of one of the further contours are calculated before the associated second coordinates of the corresponding temporary contour are calculated from the first contour. At least one of the temporary outlines and the first outline can have a congruent shape.

Is next to the first outline too the at least one additional Outline present in a parallel plane can be done as follows in particular the temporary outlines between the first Image plane (the plane of the first outline) and the fourth image plane (the level of the additional Outline}: At least one becomes one of the others Outline temporary outline, that in a plane between the first and the fourth image plane lies, by interpolation from the information on the Shape and / or size of the first Outline and from the information about the shape and / or size of the additional outline generated.

When creating the further outlines from the temporary Outlined is z. B. a ratio an expansion of the temporary Outline to an extent of the second outline and / or, if available, the third outline along a common line of intersection of the respective levels of the two outlines are used. In this way can they Dimensions of the other outlines are and are precisely scaled at least in one direction (the one defined by the cutting line Direction) can be clearly determined.

The extension also includes a computer program that when running on a computer, the process in at least one of the configurations described above.

In particular, the computer program has program code means on, which are configured, the method in at least one of the Perform configurations described above when the program is on running on a computer becomes. The program code means can be stored on a computer-readable data carrier.

The invention further comprises a disk, on which a data structure is stored that after a load in a main memory of a computer the method in at least executes one of the configurations described above.

The extension also includes a computer program product with program code means stored on a machine-readable medium, which are configured, the method in at least one of the previously described configurations to perform when the program running on a computer becomes.

Taking a computer program product will understood the program as a tradable product. It can be in any Form, for example on paper, a computer-readable disk or about a data transmission network be distributed.

• – Bildeingabemitteln zum Eingeben örtlich dreidimensionaler Bildinformationen zumindest eines Objektes und/oder eines Objektteils,
• – Erzeugungsmitteln zur Erzeugung eines ersten Umrisses des Objektes und/oder des Objektteils in einer ersten Bildebene aus den dreidimensionalen Bildinformationen,
• – Erzeugungsmitteln zur Erzeugung zumindest eines zweiten Umrisses des Objektes und/oder des Objektteils in einer zweiten Bildebene, die quer zu der ersten Bildebene verläuft, aus den dreidimensionalen Bildinformationen,
• – automatischen Erzeugungsmitteln, die ausgestaltet sind, in einer Mehrzahl weiterer Bildebenen, die parallel zueinander und parallel zu der ersten Bildebene verlaufen, jeweils einen weiteren Umriss des Objektes und/oder des Objektteils zu erzeugen, wobei in der zweiten Bildebene liegende Punkte der weiteren Umrisse durch den zweiten Umriss vorgegeben sind.

Furthermore, a device for image segmentation of objects, in particular for segmenting a human prostate, heart and / or another human organ, is proposed

• Image input means for inputting locally three-dimensional image information of at least one object and / or part of an object,
• Generating means for generating a first outline of the object and / or the object part in a first image plane from the three-dimensional image information,
• Generating means for generating at least a second outline of the object and / or the object part in a second image plane, which runs transversely to the first image plane, from the three-dimensional image information,
• Automatic generating means, which are designed to generate a further outline of the object and / or the object part in a plurality of further image planes that run parallel to one another and parallel to the first image plane, with points of the further outlines lying in the second image plane the second outline are given.

The device is particular a computer system, part of a computer system and / or cooperates a computer system together. The computer system can include a storage device for storing a computer program on a storage medium and can be a processing unit for loading the computer program the storage device and for executing the computer program have, the computer program being at least configured, control the generation of the further outlines and / or the further ones To create outlines. Furthermore, the device can have one or more Have means that are designed, individual or any reasonable combinations of the process steps described above perform.

Based on the attached drawing, preferred embodiment (Best mode) of the present invention explained. The ending, however, is not limited to this embodiment. Same Reference symbols in the figures denote identical or functionally identical Elements. The individual figures of the drawing show:

1 three image planes perpendicular to each other in pairs with an object outline in a first of the image planes;

2 the representation according to 1 , where additional outline points are defined in the second and third image plane;

3 a screen representation of the three image planes in one of four screen fields, with a fourth screen field in which a segmented object can be represented spatially;

4 the screen display according to 3 , with points of an outline being highlighted in the first image plane;

5 the screen display according to 3 and 4 during the process of creating a second outline in the second image plane;

6 the screen display according to 3 to 5 during the process of creating a third outline in the third image plane;

7 the screen display according to 6 after creation of the third outline;

8th the screen display according to 7 , wherein in the second and in the third image plane intersection lines of the object defined by the second or third outline are shown with planes running parallel to the first image plane and wherein a spatial representation with corresponding outline lines is recognizable in the fourth screen field;

9 the screen display according to 8th , with additional outlines of another object being shown;

10 a flowchart with steps of an embodiment of the method according to the invention;

11 the scaling of an outline in a plane parallel to the first image plane; and

12 a device for image segmentation of objects.

In the 1 The three image planes shown perpendicular to each other in pairs are designated with the letters Z (first image plane), X (second image plane) and Y (third image plane). The letters Z, X, Y each correspond to the coordinate z, x, y of a three-dimensional, Cartesian coordinate system, which has a constant value in the plane. The directions of the axes of the coordinates x, y, z are also in 1 shown, but not the origin of the coordinate system. Furthermore, in 1 the intersection lines of the three image planes Z, X, Y and the common intersection point S are represented by broken lines. A line of an outline 1 , which corresponds to an outer contour of a human prostate, is located in the image plane Z.

In 2 are points of outline 1 , which lie on the intersection lines with the image planes X, Y, are particularly emphasized and are designated with the letters A, B (intersection line with the image plane Y) or with the letters C, D (intersection line with the image plane X). Furthermore, an additional point is drawn in the image plane X (point F) and in the image plane Y (point E). The points C, D, F represent the beginning of a second outline that is generated in the image plane X. The points A, B, E represent the beginning of a third outline that is generated in the image plane Y.

In 3 is a similar outline 1 like in the 1 and 2 shown, in the first screen field at the bottom left of the figure. The first outline 1 is also known as the basic outline or the basic contour. Furthermore, in 3 which are already based on 2 Points A, B, C, D described are shown in the respective image planes, ie the points A, B in the first image plane Z (first screen field, bottom left) and in the third image plane Y (third screen field, top left} and the points C , D in the first image plane Z and in the second image plane X (second screen field, bottom right), as in 1 the common intersection S of the three image planes Z, X, Y is shown. The directions of the axes of the coordinates x, y, z are also shown. A spatial representation of the segmented object can be seen in the fourth screen field (top right). In the in 3 state shown, the segmented object is only that of the outline 1 enclosed area.

In practical applications of the method according to the invention or when operating the device according to the invention, the associated image information (input data of the segmentation) is also shown in the first to third screen fields. In order to better recognize the outlines, the image information is in the 3 to 9 omitted.

The first screen field (bottom left) in 4 has a variety of dots on the outline 1 on, two of which are exemplified by the letter P. The sections of the outline 1 that connect two of the points P with each other are straight lines in the exemplary embodiment described here. With the outline 1 - and the other outlines that will be described - are irregular polygons.

3 and 4 correspond to an early stage of the segmentation process. In 10 a flow diagram of the segmentation is shown. In step S1, a user, in particular due to the in the 3 to 9 Image information displayed in the first screen field (ie in the basic level) is a basic outline (the outline 1 ) defined, e.g. B. by clicking the in 4 points P shown using a computer mouse.

The outline 1 doesn't have to be in 3 and 4 shown convex shape. Rather, he can e.g. B. have at least one indentation or have any other shape. In the exemplary embodiment of the method, which will be explained in more detail, the outline defines 1 however, in any case a self-contained outline and thus a coherent bordered area. The second and (optional) third outline, there is still another limitation in the embodiment. In order to be able to clearly interpret the second and third contours and to use them for scaling the further contours that lie in further image planes running parallel to the first image plane, the second and third contours may only have a maximum of two points on each possible intersection with one of the have other image layers. This means that the second and the third outline in this direction (the direction of the cutting lines) must not have a concave shape. However, this condition does not apply to the direction perpendicular to it.

Thus, one of the two or three outlines can have any shape (in the terminology used above, the "first outline"). In general, the outline does not have to 1 his. The second and, if present, the third outline are subject to the restrictions mentioned in the exemplary embodiments.

In step S2, the segmentation device determines the (length) extension of the basic outline in at least one image plane running transversely to the basic plane, namely at least in the second image plane and / or the third image plane. In the in 3 and 4 illustrated embodiment, the points A, B, C, D.

Now in step S3, further outline points are entered in the second and / or the third image plane, ie in particular the generation of the second and / or third outline, as is also based on 5 to 7 is explained in more detail by way of example.

Step S3 is followed by step S4 the generation of temporary Outlined in the other image planes, which are parallel to the first Image plane run. The temporary Outlines are made from the basic outline and, if available, from at least an additional Outline generated from the three-dimensional image information was generated and in a plane parallel to the base plane lies. If no such additional If there is an outline, the temporary outlines are preferably the same in terms of their shape and size the basic outline. she are copied and parallel-shifted, i.e. H. in particular through uniform change one of the three Cartesian coordinates generated from the basic outline. If at least one additional Outline exists, the temporary outlines in the area between the base plane and the plane of the additional outline by interpolation from the basic outline and the additional Outline won. Other additional Outlines are also preferred by copying and parallel shifting from the closest outline generated, which is obtained from the three-dimensional image information has been. This is either the basic outline or an additional one Outline.

Step S4 is finally followed by step S5, in which the further outlines are generated by scaling the temporary outlines, information from the second and / or the third outline being used. This is further exemplified (with reference to 11 ) received.

This completes the segmentation and it stands by the basic outline, if necessary by the additional ones Outlines, through the second outline and possibly the third outline as well object or sub-object defined by the further outlines Available. The segmented object can be used for further interpretation, for example and / or evaluation of the image information can be used. To the For example, the volume of the object can be determined and / or radiation treatment can be planned and / or simulated.

As mentioned above, the 5 to 7 detailed an exemplary embodiment. 5 shows the same outline in the first screen field 1 as the 3 and 4 , Furthermore shows 5 , in the second screen field (bottom right} the emergence of a contour or an outline 2 , which is generated, for example, by clicking on the points shown, in particular the points P, H, L, M. Again, the outline is 2 z. B. a polygon. As with the other contours, however, other, in particular curved, connecting lines of the points can alternatively be used and / or models with predetermined shape properties can be used.

The outline 2 points in the in 5 shown stage a total of thirteen points shown, of which the points C, D (as already explained} by the outline 1 are given and the point G has been obtained by interpolation from the neighboring points shown. The other points were obtained by evaluating the three-dimensional image information, in particular in the second image plane. The points G, H lie on the common intersection line of the second and the third image plane. Therefore, these points are also drawn in the third screen field (top left). The outline 3 The third screen field therefore already has four defined points A, B, G, H.

For the same reason, changes to the outline can be made 3 again to changes to the outline 2 to lead. Namely, becomes the outline of the outline 3 changed in such a way that the position of one of the points G, H changes, a corresponding correction must be made to the outline 2 be performed.

In 5 , second screen field (bottom right) there is also a point N which could be defined instead of the point H and which is inadmissible as described above, because otherwise in the vertical direction in the second screen field sometimes more than two points of the outline 2 could be with the same z coordinates. The extent or length of the outline 2 in the z direction would therefore not be clearly defined everywhere.

In 6 , second screen field (bottom right) is the finished outline 2 recognizable. Furthermore, in the third screen field (top left) is the already partially generated outline 3 shown, in particular in the same way as the outline 2 is generated. 7 also shows the finished outline 3 ,

Now the temporary outlines described above are generated in a plurality of parallel planes to the first image plane and the so-called further outlines are generated therefrom. In 8th are the corresponding cutting lines 11 to 26 of the parallel planes with those through the second outline 2 and through the third outline 3 defined areas. Furthermore, the generated further outlines are shown in a spatial representation in the fourth screen field, the ones in front of the first image plane (with the outline 1 ) lying further outlines with the reference numerals 29 to 36 are designated.

On the scaling of one of the temporary outlines (outline 8th ) and thus the generation of one of the other outlines (outline 9 will now refer to 11 discussed in more detail. The outline 8th was, in particular as described above, from the basic outline (e.g. outline 1 ) and / or generated from an additional outline. In the 11 a line is drawn in which corresponds, for example, to the direction of the coordinate y. The line intersects the outline 8th at points U, V. Furthermore, the line is the intersection of the image plane of 11 and the image plane of the second outline 2 or the third outline 3 , for example the second outline 2 , If the line also intersects this outline in two points (which can usually be achieved by a suitable choice of the position of the first to third image planes), the scaling factor for adapting the size of the outline is thus 8th in the direction of the line. The scaling factor is equal to the ratio of the distance between the points U, V to the distance between the two points of intersection of the line with the outline mentioned. With this scaling factor (in the direction of the line} the entire temporary outline 8th scaled. In particular, the point U is mapped to the point U 'and the point V is mapped to the point V' in this way. Taking all three spatial coordinates into account, the points U ', V' are identical to the two intersections of the line with the outline mentioned (in the example, the outline 2 }.

In the same way, scaling can be carried out in a direction perpendicular to the direction of the line, in particular using the information from the outline 3 , In the example, this corresponds to the scaling of the x coordinate of all outline points. 11 shows the result of scaling in the y and x direction.

9 shows that a second object can be segmented in a corresponding manner, while the information from the segmentation of the first object is retained.

In particular, the first object a human prostate and the second object is the associated urethra. When segmenting the second object, it is preferred that the common intersection of the first, second and third image plane (and thus the position of the three image planes) in a suitable manner is positioned.

In the 12 The preferred embodiment of a device according to the invention shown has input means 52 . 53 for entering three-dimensional image data into a computer 55 on. The input means 52 have, for example, an ultrasound head for generating and recording ultrasound images. The input means 53 that with the input means 52 connected, are designed to convert raw data into a suitable data format that can be processed by a computer. The input means 53 are z. B. formed as a adapted to the data format of the raw data graphics card from a recording of the computer 55 is recorded and to a data bus of the computer 55 connected.

The input means 52 . 53 are using the computer 55 connected, which has at least one storage device for storing the image data. Furthermore, the device has generating means 54 . 55 on, for generating a first outline of an object to be segmented and / or contoured and / or an object part and for generating at least a second outline of the object and / or the object part in a second image plane which runs transversely to the first image plane. The first and the at least one second outline are obtained from the three-dimensional image data. For this purpose, the generating means 54 in a certain embodiment, a device for marking pixels on, e.g. B. a computer mouse, a trackball and / or a touch-sensitive screen (touchscreen). In particular, one or more screens and / or other image display means can be used to display the three-dimensional object data with the computer 55 be combined. The computer 55 generated using that from the generating means 54 information received the outline.

Furthermore, with the computer 55 An institution 57 connected, for evaluating the result of the segmentation or contouring, for example a device for calculating the volume of an area defined by the segmentation or contouring. Such a device is, for example, software in the computer 55 realized.

In the context of the invention, how it through the attached claims numerous modifications and further training are defined of the described embodiments realizable.

Claims (15)

Method for image segmentation of objects, in particular for segmentation of human or animal organs and / or results of pathological changes, whereby - locally three-dimensional image information of at least one object and / or part of an object is used, - from the three-dimensional image information in a first image plane ( Z) a first outline ( 1 ) of the object and / or the object part is generated, - from the three-dimensional image information in a second image plane (X or Y), which runs transversely to the first image plane (Z), a second outline ( 2 or 3 ) of the object and / or the part of the object is generated - automatically in a plurality of further image planes that run parallel to one another and parallel to the first image plane (Z), each another outline ( 29 to 36 ) of the object and / or the object part is generated, with points of the further outlines lying in the second image plane (X or Y) ( 29 to 36 ) through the second outline ( 2 or 3 ) are specified. The method of claim 1, wherein a third outline (3) from the three-dimensional image information in a third image plane (Y), which runs transversely to the first (Z) and transversely to the second (X) image plane 3 ) of the object and / or of the object part is generated and wherein points of the further outlines lying in the third image plane (Y) ( 29 to 36 } through the third outline ( 3 ) are specified. Method according to one of the preceding claims, wherein from the three-dimensional image information at least in a fourth Image plane that runs parallel to the first image plane additional Outline of the object and / or the object part is generated. Method according to one of the preceding claims, wherein using information about the shape and / or size of the first outline ( 1 ) and / or, if available, the further outlines are generated using information about the shape and / or size of the additional outline. Method according to the preceding claim, wherein from the information about the shape and / or size of the first outline ( 1 ) and or, if available, using information about the shape and / or size of the additional outline, temporary outline ( 8th ) are generated in the further image planes and using information about the size of the second outline ( 2 or 3 ) and or, if available, using information about the size of the third outline from the temporary outline ( 8th } the further outlines are created. Method according to the preceding claim and according to claim 3, wherein at least one of the further outlines associated temporary outline, which lies in a plane between the first (Z) and the fourth image plane, by interpolation from the information about the shape and / or size of the first outline ( 1 ) and is generated from the information about the shape and / or size of the additional outline. Method according to one of the two preceding claims, wherein at least one of the temporary outlines ( 8th ) and the first outline ( 1 ) have a congruent shape. Method according to one of claims 5 to 7, wherein when generating the further outlines from the temporary outlines ( 8th ) a ratio of an extent of the temporary outline to an extent of the second outline and / or, if present, of the third outline is used along a common intersection line of the respective planes of the two outlines. Method according to one of the preceding claims, wherein a part of the three-dimensional image information is displayed using image display means and wherein a user of the method the first ( 1 ) and / or the second ( 2 or 3 ) Outline. Computer program, characterized in that it when running on a computer, the method according to one of the preceding Claims. Computer program with program code means, thereby characterized in that the program code means are designed a method with the features according to any one of claims 1 to 9 perform, if the program is running on a computer. Computer program with program code means according to the preceding Claim that are stored on a computer-readable data carrier. disk, on which a data structure is stored that after a load in a main memory of a computer, the method according to one of the Expectations 1 to 9 executes. Computer program product with program code means stored on a machine-readable carrier, characterized in that the program code means are designed, a Ver drive with the features according to any one of claims 1 to 9 to perform when the program is executed on a computer. Contraption ( 51 ) for image segmentation of objects, in particular for the segmentation of human or animal organs and / or results of pathological changes, with - image input means ( 52 . 53 ) for entering locally three-dimensional image information of at least one object and / or part of an object, - generating means ( 54 . 55 } to create a first outline ( 1 ) of the object and / or of the object part in a first image plane (Z) from the three-dimensional image information, - generating means ( 54 . 55 ) to generate at least a second outline ( 2 or 3 ) of the object and / or the object part in a second image plane (X or Y), which runs transversely to the first image plane (Z), from the three-dimensional image information, - automatic generation means ( 55 ), which are designed in each case a further outline in a plurality of further image planes that run parallel to one another and parallel to the first image plane (Z) ( 29 to 36 ) of the object and / or the object part, with points of the further outlines (X or Y) lying in the second image plane ( 29 to 36 ) through the second outline ( 2 or 3 ) are specified.