WO2009121504A1 - Method for rendering the inside surface of a hollow organ - Google Patents

Abstract

The invention relates to a method for rendering the inside surface (1) of a substantially spherical hollow organ using an endoscope (15, 15'), comprising the following steps: producing overlapping partial images (2 - 5) of the inside surface (1), dividing the inside surface (1) into regions (6-8), associating the partial images (2-5) with the regions, producing flat projection surfaces (6'-10'; 40, 41; 51 - 53) that are disposed at an angle to each other and are associated with a region (6-8) of the inside surface (1) and are disposed as parallel thereto as possible, projecting the partial images (2-5) of each region (6-8) onto the projection surface (6' - 8') associated with the respective region, assembling the partial images (2-5), rendering the images obtained on the projection surfaces (6' - 10') in a rendition plane (11).

Description

 Process for displaying the inner surface of a hollow organ

The invention relates to a method for displaying the inner surface of a hollow organ by means of endoscopic partial images.

Hollow organs in the human body are frequent targets of endoscopic examinations. It is a very important task of the examining surgeon to get an overview of the entire inner surface of the hollow organ, so as not to miss anything. This overview must be as complete as possible, otherwise the danger, e.g. to overlook a dangerous tumor.

By nature, endoscopes always only survey very small parts of the inner surface and thus provide very small partial images in different orientation and direction. If you look at these parts one after another, you can quickly lose your orientation and do not remember what you have already seen.

There are some approaches to solving this problem with tubular hollow organs, in particular the human intestine. Examples of this can be found in EP 1832223 Al and EP 1862101 Al. Hollow-shaped hollow organs can basically be regarded as cylindrical surfaces. The juxtaposition of images on a cylinder surface is simple and clear. These problems are relatively easy to solve. However, in substantially spherical hollow organs, such as in particular the human bladder, the areas to be displayed are curved in two directions. Problems arise here, as had generations of cartographers in the representation of the globe on a flat map. In the present case, in addition to endoscopic image acquisition only individual, relatively small-scale partial images exist, which in themselves have only slight problems with the curvature, but when composing over larger areas of curvature considerable problems.

The object of the present invention is to present the inner surface of a hollow organ in a particularly clear manner.

This object is achieved with the features of claim 1.

According to the invention, first overlapping partial images are produced. This enables the composition of the images in a known manner by means of recognition and corresponding alignment and equalization of matching structures in the overlapping regions.

The composition of overlapping partial images, which are normally recorded in different directions and thus lie in image planes which are at an angle to each other, requires bringing the images to be composed into a common plane, which requires distorting transformations or projections. This in turn leads to problems with strongly curved inner surfaces. The further successive images change their angle, the larger the distortions. Finally, if, for example, the inner surface of a hollow organ is rotated by 180 °, for example, distortions occur which render the images unrecognizable. The invention avoids this by not all projecting the partial images in a common plane, but on different projection surfaces which are at an angle to each other and are as parallel as possible to areas of the inner surface. The projection onto an at least approximately parallel projection surface leads to only slight distortions. The partial images can therefore be assembled on the projection surfaces into well-understandable images. According to the invention the projection surfaces are then displayed in a common display plane, for example on the usual monitor. In the display plane you can see the images of the projection surfaces, which are easy to understand and belong to certain areas of the inner surface and are therefore easy to assign. On the presentation level, for example, one can provide the individual projection surfaces with written instructions, such as "left area", "right area" or the like.

It succeeds in this way with very simple means using conventional image processing procedures to represent the inner surface of a hollow organ in a very clear manner. In particular, this method does not fail in strongly curved inner surfaces of hollow organs, which inevitably result from their closed mold.

Advantageously, the features of claim 2 are provided. In this embodiment of the display method, the projection surfaces adjoin each other with straight edges. The images merge without major distortions, so that in a very clear way, the transition from one surface to the other surface can be understood.

Advantageously, the features of claim 3 are provided. With individually different winkers between the projection surfaces, these can be better adapted, namely as parallel as possible to areas of the inner surface of the hollow organ. Advantageously use right angle between the projection But there is a better intuitive understanding of the images. If the individual projection areas at right angles represent easily comprehensible viewing directions, such as "left" and "straight ahead", this significantly increases the clarity and the fast orientation in the evaluation of the image.

Advantageously, the features of claim 4 are provided. The partial images can be assembled before or even after their projection onto the projection surfaces. The first alternative is advantageous since simple calculation models can be used. In addition, the previous composition can be helpful to determine their assignment to the areas of chaotic partial images, which can be simpler on the basis of composite images.

Advantageously, the features of claim 5 are provided. In the case of endoscopic imaging of hollow organs, the half of the inner surface lying in the direction of insertion of the endoscope into the hollow organ is usually considered. If this is projected in regions on the surfaces of a half-cube, adjoining surface areas arise on the display plane, which in an extremely clear manner point in the middle to the area lying straight ahead, to the right and left of which the right and left areas are adjacent, and above and below the corresponding upper and lower areas of the inner surface. Such a presentation is intuitively understood even by inexperienced people.

Using a half-cube to simplify simulation of a hemisphere is known per se in a completely different field, namely that of light and shadow generation on object surfaces when rendering digital images. The essay "Radiosity" by Hugo Elias in

http://freespace.virgin.net/hugo.elias/radiosity/radiosity.htm describes this. The half cube is used here as a substitute for a hemisphere in the representation of the incident on a surface segment environment. The semi-cube surfaces are each obtained as whole images.

In the drawings, the invention is shown for example and schematically. Show it:

1 is a schematic representation of a hemispherical half of a hollow organ with semi-cube-shaped projection surfaces,

2 shows a central section through the illustration of FIG. 1,

3 shows the common representation of the projection surfaces of FIG. 1 on a monitor, FIG.

4 is a representation corresponding to FIG. 1, but with only two projection surfaces,

5 shows a representation corresponding to FIG. 1, but with three projection surfaces intersecting in a point, FIG.

Fig. 6 is a representation corresponding to FIG. 2, but with oblique winkers between the projection surfaces and

7 is a representation corresponding to FIG. 3 of the projection surfaces of FIG. 6. FIG.

Fig. 1 shows a hemispherical inner surface 1, which represents in schematic form the half of a human bladder opposite the access through the urethra. The inner surface of the human bladder becomes very common examined endoscopically. Partial images of the inner surface 1 are then obtained.

Fig. 2 shows again in section the inner surface 1 with two overlapping partial images 2, 3, which were created at assumed ideal recording technique from the center of the hemisphere 1 from different angles and thus with their image planes at an angle shown to each other. All partial images must each overlap at least one other partial image. Based on pattern matching in overlapping areas, partial images can then be adapted and put together.

The entire hemisphere surface 1 to be displayed is detected with a multiplicity of partial images, which are all at different angles to one another. As shown in FIG. 2, the two images 2, 3 are not very different in angle. However, a remote field 4 is orthogonal to field 3. This results in significant problems.

You can look at the individual fields individually, but then loses the overview very quickly, even if the fields very neatly, like tiles are numbered consecutively. Therefore, the partial images are to be assembled into a clear overall picture.

Since no hollow spherical presentation media are available, but these, such as the usual computer monitor, are level, the partial images for obtaining an overview image not only have to be assembled, but also brought into a common plane. It should be noted that the images, if they are meaningfully assembled on the basis of overlapping patterns, a distorting projection is required, with the first the pattern can be brought to cover. With small angle differences, as between the sub-images 2, 3, they can be brought into a common plane, without larger Strains occur. The partial images 2, 3 on the one hand and the partial image 4 on the other hand, however, are too distorted when displayed in a common plane to be still recognizable.

As FIG. 2 shows, the inner surface 1 is therefore subdivided into the three regions 6, 7, 8, that is to say a left region 6, a central region 7 and a right region 8, with boundaries shown in FIG Horizontal lines lie.

The partial images from each of the regions 6, 7, 8 are projected onto different projection surfaces. This can be seen in FIG. 2. Thus, the left region 6 of the inner surface 1 is projected onto the projection surface 6 ', the middle region 7 onto the projection surface T and the right region 8 onto the projection surface 8'. The projection surfaces 6 ', T and 8' are perpendicular to each other.

In the projection, therefore, the partial images 2, 3 are projected onto the projection surface 6 ', the partial image 4 onto the projection surface 8 ¹ and a partial image 5 lying in the central region 7 onto the projection surface T.

In Fig. 2 is also shown how the frames are obtained, by means of endoscopes. An endoscope 15 is shown, which captures the partial image 3 with an oblique-looking objective. An endoscope 15 'with a straight looking lens is shown in the detection of the partial image 5.

The endoscope can record individual partial images or, in particular, deliver a stream of partial images as a conventional video endoscope.

2 shows that the regions 6, 7 and 8 of the hemispherical surface 1 are assigned to their associated projection surfaces 6 ', T and 8' in such a way that the directions in each case approximately coincide between the region and the projection surface, ie at least approximately parallelism exists. This ensures that In the projection of the partial images 2 - 5 of the respective areas on the associated projection surfaces only relatively small distortions occur, the images so well understood.

Fig. 1 shows a three-dimensional representation of the inner surface 1, which corresponds to the inner surface of a half hollow organ, in particular the human bladder. The projection surfaces around the inner surface 1 are arranged at right angles to one another, specifically the projection surfaces 6 ', T and 8' from FIG. 2 and additionally an upper projection surface 9 'and a lower projection surface 10'.

After projection and assembly of the partial images, the images resulting from projection and composition on the projection surfaces are displayed together on a display plane 11, which in the example of FIG. 3 is the screen plane of a conventional monitor 12. In FIG. 3, all projection surfaces are shown in one plane. In this overall picture you can find your way around very clearly and intuitively. It can be seen immediately that T is an area straight ahead, while 6 'and 8' are right and left and 9 'and 10' up and down. It is therefore intuitive to say that a detail lying on the surface 9 'lies in the bubble at the top. It is also important that the overall picture shown in Fig. 3 corresponds to the complete area of the illustrated hemisphere.

In the exemplary embodiment illustrated in FIGS. 1 to 3, a projection of the partial images 2 to 5 onto the projected partial images 2 'to 5' is required, specifically in the sense of a projection from the sphere onto the planar projection surfaces. In this case, on the one hand, the projection is required and, on the other hand, the composing of the partial images on the basis of their overlapping regions. It may be advantageous to first assemble the images, for example, in a plane in any direction, for example, arbitrarily directed to the first image, with which you begin the composition. Then that will turn into that level resulting overall image converted to the projection surface, which belongs to the area of the inner surface 1, from which the partial images originate.

In the embodiment of FIGS. 1-3, the projection surfaces are arranged in the form of a half cube or cuboid, that is to say correspond to a hemisphere to be displayed. However, projection surfaces can also be arranged completely differently.

Fig. 4 again shows the same hemispherical inner surface 1, but with only two projection surfaces 40 and 41, which intersect in a section line behind the inner surface 1. The angle between the surfaces 40 and 41 can be largely arbitrary and in particular can not be a right angle. Then two halves of the inner surface 1 can each be projected onto the projection surface 40 and the projection surface 41. As shown on the illustration plane 11 of FIG. 3, two surfaces adjoining one another correspond to the left half and the right half of the hemisphere surface 1. Even such a representation is very clear, but in some areas, especially in the upper and lower areas relatively distorted.

FIG. 5 shows a variant of FIG. 4, in which not two projection surfaces but three projection surfaces 51, 52 and 53 forming a tetrahedron with a tip 50 are used to again represent the same hemisphere surface 1. With three projection surfaces instead of two projection surfaces according to FIG. 4, the projection surfaces can be better adapted to the inner surface 1, so that the distortions become smaller. But in such a presentation, however, the intuitive clarity will suffer a little.

FIGS. 6 and 7 show, in accordance with FIGS. 2 and 3, a variant shown greatly simplified, in which the angles between the projection surfaces 6 ¹ , T and 8 'are greater than 90 °. As a result, the projection surfaces, as shown in FIG. 6, can be better adapted to the associated regions 6, 7 and 8 of the hemisphere 1. They are much more parallel than in FIG. 2, so that the distortions of the partial images are smaller in the projection. The overview image resulting according to FIG. 7 essentially corresponds in its intuitive interpretability to that of FIG. 3.

Claims

Method of depicting the inner surface of a hollow organ PATTERN CLAIMS: 1. A method for displaying the inner surface (1) of a substantially spherical hollow organ with an endoscope (15, 15 '), comprising the following steps: Producing overlapping partial images (2 - 5) of the inner surface (1), Dividing the inner surface (1) into areas (6 - 8), Assigning the sub-images (2 - 5) to the areas Producing planar, angled projection surfaces (6'-K) '; 40, 41; 51-53), which are each assigned to a region (6-8) of the inner surface (1) and are as parallel to this as possible, Projecting the sub-images (2 - 5) of each area (6-8) on the projection area (6 '- 8') assigned to the respective area, Composing the partial images (2 - 5), Representing the on the projection surfaces (6 ¹ - 10 ') resulting images in a display plane (11). 2. The method according to claim 1, characterized in that the projection surfaces (6'-1O ¹ ; 40, 41; 51, 52, 53) border each other at edges. 3. The method according to claim 1, characterized in that the projection surfaces (6 '- 10') are at right angles to each other. 4. The method according to claim 1, characterized in that the partial images (2-5) are assembled before their projection. 5. The method according to claims 2 and 3, characterized in that for the representation of the hemispherical half (1) of an inner surface of the projection surfaces (6 '- 10') form a half-cube.