EP1665815A2

EP1665815A2 - Method for creating a stereoscopic image master for imaging methods with three-dimensional depth rendition and device for displaying a stereoscopic image master

Info

Publication number: EP1665815A2
Application number: EP04764457A
Authority: EP
Inventors: Armin Grasnick
Original assignee: Individual
Current assignee: Individual
Priority date: 2003-09-15
Filing date: 2004-08-25
Publication date: 2006-06-07
Also published as: WO2005029871A3; WO2005029871A2; JP2007506167A; US20070159476A1

Abstract

The invention relates to a method for creating stereoscopic image masters from two-dimensional image data, in particular from image data emanating from image sequences, video films or similar. According to said method, a virtual three-dimensional image structure (307), based on an assumed three-dimensional gradation of image depth is created on the basis of the image information of reproduced objects (303, 304, 305, 306) that has been determined from monocular original-image data, whereby said original-image data is adapted to the virtual three-dimensional image structure (307) to create a virtual three-dimensional image master and a series of individual images that reproduce the virtual three-dimensional image model are obtained from the latter. The virtual individual images are merged in a combination step to form a stereoscopic image master for carrying out an imaging method with supplementary depth rendition.

Description

Description: The invention relates to a method for creating a spatial image template according to the preamble of claim 1 and a device for displaying a spatial image template according to the preamble of claim 16.

Three-dimensional objects are only imaged two-dimensionally with monocular recording devices. The reason for this is that these objects are recorded from a single observation location and from only one observation angle. With such a recording method, the spatial object is projected onto a film, a photovoltaic receiver, in particular a CCD array or another light-sensitive surface. A spatial impression of the depicted object only arises when the object is recorded from at least two different observation points or at least two different viewing angles and presented to an observer in such a way that the two two-dimensional monocular images are perceived separately by both eyes and in their physiological Perceptual apparatus can be put together. For this purpose, the monocular individual images are combined to form a spatial image template which, using a suitable imaging method, leads to a spatial image impression on the viewer. Such methods are also referred to as "anaglyph technique".

A spatial image template that can be used for such a method can be present or created in various ways. The simplest example is the well-known stereo slide viewer, in which the viewer looks at an image motif taken from a different angle with one eye each. In a second possibility, the image produced from the first viewing angle is colored with a first color and the other image photographed from the second viewing angle with a second color. Both images are printed on top of one another or projected on top of one another with a displacement corresponding to the natural difference in the angle of view of the human eyes or the difference in the angle of view of the camera system, in which case the viewer uses two-color glasses to view the image. The other viewing angle component is filtered out through the correspondingly colored spectacle lens. Each eye of the viewer thus receives one corresponding to the Different image from different angles, whereby the viewer creates a spatial impression of the original image. Such a method is advantageous if data from a stereo camera are to be transmitted and displayed in real time and with little expenditure on equipment. In addition, simulated spatial images are also represented with such a method for generating a spatial image template, the viewer being able to get a better impression of complicated spatial structures, for example complicated simulated molecular structures and the like.

Furthermore, mechanisms of the physiological perceptual apparatus with a more subtle effect can be used to generate the spatial image template. It is known, for example, that two images perceived shortly within the reaction time are brought together to form a subjective overall impression. If you send two pieces of image information in succession as a combined spatial image template, each consisting of photographs taken from the first or the second viewing angle, these are combined in the viewer's perception to form a subjective overall spatial impression using shutter glasses ,

However, all of the methods mentioned have in common that at least one binocular image of the spatial image motif must be available in advance. This means that at least two pictures taken from different viewing angles must be available from the start or (as for example in drawings) must be created from the beginning. Images or films, video sequences and similar images that were created monocular from the start and therefore only carry monocular image information can therefore not be used for a spatial representation of the object. A photo taken with a monocular camera, for example, is a two-dimensional projection with no spatial depth. The information about the depth of the room has been irretrievably lost due to the monocular image and has to be interpreted by the viewer based on experience in the picture. Understandably, however, there is no real spatial image with depth effect. This is disadvantageous in that in a whole series of such two-dimensional monocularly produced recordings a considerable part of the original effect and the information of the image motif is lost. The viewer has to think it over or try to explain it to other viewers, whereby of course the original impression of the spatiality cannot be regained with any spatial mapping method of the above examples.

The object is therefore to specify a method for creating spatial image templates from two-dimensional image data, in particular image data from image sequences, video films and the like, in which a spatial image template for an imaging method with spatial depth effect is generated from a two-dimensional image.

This object is achieved with a method according to the features of claim 1, the subclaims containing at least features of the invention.

In the following description, the term "archetype" is understood to mean the initially given monocularly produced two-dimensional image. It is immediately clear that an application of the method according to the invention described below can also be applied to sequences of such archetypes and therefore also readily for moving images , in particular video or film recordings, can be used, provided that these consist of a series of successive images or can be transferred to them.

According to the invention, a virtual three-dimensional image structure based on presumption-based image depth grading is generated on the basis of image information of images of objects determined from monocular master image data. The original image data are adapted to the virtual three-dimensional image framework to generate a virtual three-dimensional image model. The data of the virtual three-dimensional image model are used as a template for creating the spatial image template for the imaging method with spatial depth effect. According to the invention, the objects depicted on the two-dimensional image are thus initially determined. Then these objects are each assigned a guess about their depth. A virtual three-dimensional model is created, the original image data of the two-dimensional image being adapted to this virtual three-dimensional model. This virtual three-dimensional model now forms a virtual object, the data of which represent the starting point for generating the spatial image template.

A method for edge detection of the depicted objects with generation of an edge-marked image is carried out on the monocular master image data in order to determine the image information. In the presumption-based image depth grading, original image areas are assigned to different virtual depth planes, in particular a background and / or foreground, on the basis of a determined richness of edges.

Here, the knowledge is used that objects that are rich in detail and thus have many edges generally belong to a different image depth and thus depth level than objects that are poor in detail and therefore also poor in edges. The step of edge detection accordingly sorts out components of the original image that are assumed to be in the background of the image and separates them from those that are suspected in the foreground or a further depth plane.

In a further procedure for determining the image information, a method for determining the color information of given archetype areas is carried out. In the presumption-based image depth grading, at least first identified color information is assigned to a first virtual depth level and second color information to a second virtual depth level.

The factual experience that certain colors or color combinations with certain image motifs preferentially occur in a different depth plane than other colors or color combinations is used here. Examples are blue as a typical background color for landscapes on the one hand and red or green as typical foreground colors of the motif shown on the other. The method for edge detection and the method for determining the color information can be used either individually or in combination with one another, with a combined application of edge detection and determination of the color information allowing further differentiation options of the original image data, in particular a more precise definition of further depth planes.

In an expedient embodiment, a softening method is used on the edge-marked image to reinforce and unify an edge-rich archetype area. On the one hand, this compensates for possible errors in edge detection and, on the other hand, strengthens adjacent, not randomly specified structures. Optionally, the values of the edge-marked image can also be corrected for tonal values.

On the basis of the soft-drawn and / or additionally tonally corrected edge-marked image, an assignment of a relevant image section to a depth plane is carried out based on the tonal value of a pixel. The structures of the edge-marked image, which is softened and optionally tonally corrected, are now assigned to individual defined depth levels depending on their tonal value. The edge-marked, soft-drawn and optionally tone-corrected image therefore forms the basis for a clear assignment of the individual image structures to the depth levels, such as the defined virtual background, a virtual image level or a virtual foreground.

When a fixed point definition is carried out, the color and / or tonal values are limited to a predetermined value. This defines a virtual pivot point for the individual views to be created later. The selected color and / or tonal value forms a reference value, which is assigned to a virtual image plane and thus separates a virtual deep background from a foreground that virtually protrudes from the image plane.

A virtual depth level can be assigned in different ways. The method steps already described expediently suggest an assignment of a depth level to a respectively predetermined color and / or brightness value of an image pixel. Objects with image pixels, which thus have the same color and / or brightness values, are thus assigned to a depth level. Alternatively, there is also the possibility of assigning arbitrarily defined image sections, in particular an image edge and / or the image center, to a virtual depth plane. In this way, in particular, a virtual “curvature”, “bracing”, “tilting” and the like three-dimensional image effects are achieved.

To generate the virtual three-dimensional image model, the virtual three-dimensional image structure is generated as a virtual network structure deformed in accordance with the virtual depth planes, and the two-dimensional original image is adapted as texture to the deformed network structure in a mapping process. The network structure forms a kind of virtual three-dimensional "die" or "profile shape", while the two-dimensional archetype represents a kind of "elastic cloth" that is stretched over the die and pressed into the die in a kind of virtual "deep-drawing process". The result is a virtual three-dimensional image model with the image information of the two-dimensional archetype and the "virtual thermoforming structure" of the virtual three-dimensional matrix additionally impressed on the archetype.

Virtual binocular or multi-ocular views can be derived from this three-dimensional image model. This is done by generating a series of virtual individual images representing the views of the virtual three-dimensional image model from the series of virtual three-dimensional image models, in which the image sections of the original image corresponding to a defined depth plane are shifted and / or displaced in accordance with the virtual observation angle be distorted. The virtual three-dimensional image model thus serves as a virtual spatial object that is viewed virtually binocularly or multiocularly, whereby virtual views are obtained that differ according to the observation angles.

These virtual individual images are combined according to an algorithm suitable for the imaging method with additional spatial effect for generating a spatial image template. The virtual single images are treated like real binocular or multi-ocular single images, which are now suitably prepared and combined for a three-dimensional display process. This means that virtually binocular or multi-ocular image information is available that can be used for any spatial imaging method.

To create the spatial image template, in one embodiment of the method, image processing of individual image areas of the original image, in particular scaling and / or rotating and / or mirroring, is carried out and the spatial image template generated in this way is displayed by means of a monofocal lenticular screen lying above it.

The image structures, which are assigned to certain depth levels in the virtual three-dimensional image model, are changed in such a way that, when the spatial image template generated in this way is displayed, they offer a sufficient accommodation stimulus for the viewing human eye. The image structures highlighted in this way are perceived by means of the optical image through the lenticular screen either in front of or behind the given image plane and thus lead to a spatial impression when viewing the image. This method only requires a relatively simple spatial image template in connection with a simple implementation of the imaging method with spatial depth effect.

The two-dimensional archetype can also be displayed directly without image processing using the monofocal lenticular grid. The two-dimensional archetype can thus be used immediately as a spatial image template for display by the monofocal lenticular screen. Such a procedure is particularly expedient if simple image structures against a homogeneously structured background, in particular characters against a uniformly designed text background, are to be displayed with a depth effect. The accommodation stimulus achieved by the imaging effect of the monofocal lenticular grid for the viewing eye then causes a depth effect, the original image itself not having to be prepared beforehand for such a display.

A device for displaying a spatial image template is characterized by a two-dimensional spatial image template and a monofocal lens grid arranged above the spatial image template. The monofocal lenticular grid depicts areas of the spatial image template and creates a corresponding accommodation stimulus in the eye looking at it. For this purpose, the two-dimensional image template is expediently made from a mosaic of image sections assigned to the raster structure of the lenticular screen, wherein essentially one image section each is an imaging object for essentially one associated lens element of the monofocal lenticular screen. The two-dimensional image template is therefore divided into a total of individual image areas, which are each indicated by a lens element.

In principle, two embodiments of the image template and in particular the image areas are possible with this device. In a first embodiment, the image sections are essentially unchanged image components of the two-dimensional image template of the original image. This means that in this embodiment the essentially unchanged two-dimensional image forms the spatial image template for the lenticular screen. In this embodiment, apart from changes in size or scaling of the entire image, image processing of individual image areas is therefore dispensed with.

In a further embodiment, the image sections are scaled and / or mirrored and / or rotated to compensate for the imaging effects of the lenticular screen. This results in an improved image quality, although the effort for creating the spatial image template increases.

The two-dimensional image template is in particular an image generated on a display, while the lenticular screen is attached to the surface of the display. The lenticular screen is thus attached to a pre-existing display, for example a tube or flat screen, at a suitable location and is thus located above the image shown on the display. This arrangement can be implemented in a very simple manner.

In a first embodiment, the lenticular grid is designed as a grid-like Fresnel lens arrangement that adheres to the display surface. The use of Fresnel lenses ensures a flat and simple design of the lenticular grid, the groove structures typical of Fresnel lenses being able to be worked into a transparent plastic material, in particular a plastic film, in the manner known from the prior art. In a second embodiment, the lenticular grid is designed as a grid-like, in particular flexible, zone plate arrangement adhering to the display surface. A zone plate is a concentric system of light and dark rings, which cause light to bundle the light that passes through it, thereby creating an imaging effect. Such an embodiment can be produced by printing on a transparent, flexible film in a low-cost and inexpensive manner.

In a third embodiment, the lenticular grid is also possible as an arrangement of conventionally shaped convex lenses, although the thickness of the entire arrangement and thus also its material consumption increases.

The method and the device are to be explained in more detail below on the basis of exemplary embodiments. The attached figures serve for clarification. The same reference numerals are used for process steps and process components having the same or the same effect. It shows:

1 shows a first part of an exemplary schematic program flow chart of the method,

2 shows an exemplary flow chart of an edge detection,

3 shows a second part of an exemplary schematic program flow chart of the method,

4 shows an exemplary selection menu for executing edge detection,

5a shows an exemplary archetype,

5b shows an exemplary edge-marked image as a result of an edge detection carried out on the exemplary original image from FIG. 5a,

6a shows an exemplary result of a soft focus carried out on the edge-marked image according to FIG. 5b, 6b shows an exemplary result of a tonal value correction carried out on the edge-marked image according to FIG. 6a,

7a shows an exemplary selection menu for a fixed point definition,

7b shows an exemplary fixed point-defined image,

8a shows a schematic example of graphic objects in a schematic two-dimensional original image,

8b shows a schematic example of a depth level assignment of the graphic objects from FIG. 8a and a generation of a virtual three-dimensional image model along exemplary sections along the lines A-A and B-B from FIG. 8a,

9a shows a schematic example of a virtual binocular view and projection of the virtual three-dimensional image model along the line A-A from FIGS. 8a and 8b,

9b shows a schematic example of an exemplary virtual binocular view and projection of the virtual three-dimensional image model along the line B-B from FIGS. 8a and 8b,

10 shows a schematic example of virtual single image generation from an exemplary viewing angle along the exemplary line A-A according to FIG. 9a,

11a shows a series of exemplary virtual single images from different viewing angles using the original image from FIG. 5a,

11b shows an exemplary combination of the virtual single images shown in FIG. 11a in an exemplary spatial image template for an imaging method with additional depth effect, 12a, b are exemplary representations of a two-dimensional image template and a monofocal lenticular screen located above it,

13a-c show exemplary representations of an image of a two-dimensional image section by means of the monofocal lenticular screen of the previous images.

Figures 1 and 3 show in two parts an exemplary schematic flow chart of the method. Figure 2 illustrates an edge detection method in a more detailed flow chart. FIGS. 4 to 11b show exemplary results and further details of the method according to the invention explained in the flowcharts.

The starting point of the method is a set of original image data 10 of a predefined two-dimensional, expediently digitized, original image. If the original image is a single image as part of an image sequence or a digitized film, the following description assumes that all further individual images of the image sequence can be processed in a manner corresponding to the single image. Therefore, the method described below by way of example can also be used for image sequences, films and the like.

It is expediently assumed that the original image data 10 are present in an image file, a digital storage device or a comparable storage unit. These data can be generated by the usual means for generating digitized image data, in particular by a known scanning process, digital photography, digitized video information and the like, other known image acquisition methods. In particular, this also includes image data obtained from video or film sequences using so-called frame grabbers. In principle, all known image formats, in particular the BMP, JPEG, PNG, TGA, TIFF or EPS format, can be used as data formats in all the respective versions. Although reference is made to figures in the exemplary embodiments described below, which are depicted in black and white or in gray values for reasons of illustration, the original image data can easily contain color information. The master image data 10 are loaded into a working memory in a read-in step 20 for executing the method. In a process step of an adaptation 30, the original image data are first adapted for an optimal process execution. Adjusting 30 the image properties includes at least changing the image size and color model of the image. Smaller images are generally preferred if the computing time of the method is to be minimized. However, a change in image size can also be a possible source of error for the method according to the invention. In principle, all color models currently in use, in particular RGB and CMYK or grayscale models, but also lab, index or duplex models can be used as the color model to be adapted, depending on requirements.

For further processing, the adapted image data are temporarily stored in a step 40 for repeated access. The temporarily stored image data 50 essentially form the basis of all subsequent data operations.

Depending on the assumed spatial image structure, i.e. the assumed gradation of the depth planes in the original image, an access to the temporarily stored image data 50 now either results in a change in the color channel / color distribution 60 or a conversion of the image data into a gray value graded image by means of a gray value gradation 70 The gray value gradation 70 is particularly advantageous when it can be assumed that depth information is predominantly to be assigned to the object contours shown on the image. In this case, all color information of the image is equally relevant for a deep interpretation of the original image and can therefore be converted into gray values in the same way. A modification of the color channel or the color distribution in the image data is expedient if it can be assumed that essentially one color channel is the carrier of the interpreted depth information and should therefore be particularly emphasized or taken into account for the subsequent processing. In the exemplary description of the method sequence given here, for reasons of better representation, in particular with regard to the figures, it is assumed that the temporarily stored image data 50 are converted into gray values independently of their color values, the color information remaining unchanged. An edge detection 80 follows in the further course of the method. The assumption is that the depth planes interpreted into the two-dimensional archetype are primarily defined by the objects present in the image motif. For example, it can be assumed that richly structured objects, which are strongly characterized by contours and thus edge-like structures, are predominantly present in the foreground of the image and that low-contour, washed-out and therefore low-edge objects form the background of the image. The edge detection method 80 is carried out in order to uniquely identify the different areas of the original image which, because of their structure, belong to different depth planes and to differentiate them from one another as clearly as possible.

FIG. 2 shows a schematic and exemplary flow chart for edge detection. FIG. 4 shows in connection therewith an exemplary input menu 89 for determining the changes to be made to the brightness values of a central pixel and a defined pixel environment. The image pixels 81 defined by their gray value are continuously processed in a loop process. First, a pixel 82 is selected and its brightness value 83 is read in. This brightness value is multiplied by a positive value that is as large as possible (in the example shown here by the arbitrary value +10), which produces a very bright image pixel 85. In contrast, a brightness value of a pixel to the right of it is multiplied by a strong negative value (in the example shown here by -10) in a step 86, whereby a very dark pixel 87 is generated. The next pixel is then read in in a step 88. The result of the edge detection is an edge-marked image. Where the original image shows a great deal of structure and therefore a wealth of edges, the edge-marked image data now given consist of a structure of very light or very dark pixels, while low-structure and therefore low-contour and low-edge image areas have a uniform dark color. The structure of the alternating very bright and very dark pixels and the object marked with them therefore has a higher average brightness value than an area of consistently dark pixels.

Structures lying next to one another are subsequently reinforced by means of a method step 90 referred to as “soft focus” Brightness values of a certain selected pixel amount in the edge-marked image are averaged according to a certain algorithm and assigned to the pixels of the selected amount. A Gaussian soft focus method has proven particularly useful here. In the softened, edge-marked image, the object structures stand out as a lighter pixel set from the rest of the image and enable identification of a uniform object.

If necessary, a tonal value correction of the edge-marked blurred image can then be carried out in a step 100. The tonal values of the image points are preferably corrected in such a way that there is as clear a contrast as possible between the object structure and the rest defined as the background of the image.

The next method step is designated by a fixed point determination 110. In this step, the color and / or gray values of the edge-marked, blurred image are limited to a certain value in such a way that the virtual pivot point of the virtual individual views to be generated is defined. In other words, the fixed point definition 110 defines the objects or structures which are supposed to be virtually located in front of or behind the image surface and are to be subsequently imaged in their depth effect.

Furthermore, further fixed point options can optionally be taken into account in a method step 120. For example, a first assumption can be made that larger blue areas predominantly form a background (blue sky, water, etc.), while smaller, sharply delineated objects that stand out in terms of color form the foreground of the image. Certain color values can also be assigned certain virtual depth levels from the start. For example, color values that correspond to a face color can be assigned to a virtual depth level that corresponds to an average image depth. Defined image sections, such as, for example, the image edge or the center of the image, can likewise be assigned to certain depth levels, for example the foreground or the background, in which case a “pre-tensioning” or “arching” of the spatial image produced later can be generated. The gradations of the virtual depth planes generated in this way produce a virtual three-dimensional image structure which serves as a distortion mask or "displacement map" and can be visualized in the form of a grayscale mask. This virtual three-dimensional image structure is stored in a step 130 for further use.

The virtual three-dimensional image framework serves as a distortion mask and virtual form for generating a virtual three-dimensional image model. In a method step 150, which is referred to in FIG. 3 as "Displace", the original image is placed as a texture over the virtual image frame and distorted in such a way that the corresponding original image sections are "deep-drawn" on the virtual depth planes, i.e. be assigned to these depth levels. From this series of virtual three-dimensional image models, virtual individual images 160 of the virtual three-dimensional image model are now generated from a number of different virtual viewing angles by virtual projection of the image data of the virtual three-dimensional image model in accordance with known perspective mapping laws.

In a combination step 170, the virtual individual images are combined in accordance with an algorithm defined for the imaging method with an additional depth effect in such a way that finally image data 180 are available for the three-dimensional imaging of the initial original image.

Some image processing is explained in more detail below using examples. 5a shows a generally colored, two-dimensional archetype 200. As can be seen in the picture, a number of plants are obviously in the foreground of the image motif, while indistinct port facilities, buildings and a largely structureless beach are evident in the background are. Furthermore, the practically infinite background in the original image 200 is formed by a soft sky. As can be seen from FIG. 5a, the plants arranged in the obvious foreground are distinguished by a considerable richness of detail in comparison with the background, which among other things shows up in a large number of “edges”, for example in the area of the leaves or the flowers In comparison, the background has little or no edges, so it makes sense to use the density of the edges in the original image 200 as an indicator of the spatial position of the objects shown. FIG. 5b shows an edge-marked image 210 obtained from the original image 200 after a gray scale conversion and an optional size correction. Where the structurally rich plants from the original image 200 shown in FIG. 5a are located, the edge-marked image 210 shows a large number of bright pixels Edges that lead to a higher average image brightness, especially in the right image area. In contrast, both the sky and the beach area from the original image 200 are low-edged and therefore predominantly dark in the edge-marked image 210, while the buildings recognizable in the original image 200 produce a few edge structures in the form of isolated bright pixels.

FIGS. 6a and 6b show a soft-drawn edge-marked image 220 and a soft-drawn edge-marked and additionally tone-corrected image 230. It can be seen on the soft-drawn image 220 that the right image part differs from the left image part by a higher image brightness value. This difference is shown even more clearly in the tone-corrected image 230 in FIG. 6b. The richness of edges in the area of the plants of the archetype, in other words the richness of structure of the assumed foreground, is clearly shown in pictures 220 and 230 as a bright area. A slightly lighter stripe can be clearly seen in the tone-corrected image 230 on the left half of the image, which is, however, significantly darker than the image area of the plants. This stripe corresponds to the depicted buildings from the original image 200 from FIG. 5a. The significantly darker brightness value refers to the structure of the depicted buildings with less edges and thus to their arrangement in the assumed image background.

In the soft-drawn and tonally corrected image, the sky and the beach from the original image 200 form a uniformly dark surface. Although the beach is to be counted as the middle foreground of the picture rather than the background formed by the sky, it is not possible to clearly determine the middle foreground position from the edge-marked, soft-drawn and tonal-corrected picture alone. Here it is advisable to assign the beach to a virtual middle depth level based on the yellow or brown color value, which in this example clearly differs from the color value of the blue sky. This is possible with the fixed point definition 110 already mentioned above. 7a shows an exemplary menu 239 in this regard, FIG. 7b the image 240 corresponding to the menu. A histogram 241 shows a series of color channels, which in the exemplary embodiment shown in FIG. 7a are a series of gray values. The corresponding gray values are shown in a gray value bar 242. In the left part of the histogram 241 or the gray value bar 242 there are the dark brightness values, in the right part the light brightness values. The size of the histogram bar shows the frequency distribution of the corresponding gray values. It can be seen that the bright area of the soft-drawn or tonally corrected image 220 or 230 is shown in a wide maximum of the histogram 241, while the dark areas in the images 220 and 230 become a maximum in the left part of the histogram 241 in the dark Brightness values. Certain brightness values can be selected by means of indicator pointers 243. With the help of the buttons 245, brightness values of selected pixels can be read out directly from the image 241 into the histogram 241 and transmitted.

In the example shown here it turns out that the area corresponding to the beach from the original image 200 differs in its brightness value from the image section corresponding to the sky from the original image 210. This can be selected as a virtual image plane using a selection indicator 244 and forms a possible fixed point for virtual individual views of the virtual three-dimensional image model to be generated later.

FIGS. 8a and 8b use a very highly schematic example to show a depth level assignment and a construction of a virtual three-dimensional image frame. FIG. 8a shows a highly schematic, monocularly generated two-dimensional archetype 301, the individual objects of which in their spatial position in the image have already been identified by the contour recognition methods described above. The schematic master image 301 shown by way of example in FIG. 8a has a first object 303, a second object 304 and a third object 305, which are arranged in front of an area 306 identified as background and stand out from the latter.

The previously described methods for contour marking, for the definition of fixed points and further assumptions about the image depth make it more exemplary 8a seems to make sense for the schematic original image 301 in FIG. 8a to arrange the first object 303 in the depth plane in the foreground, while the objects 304 and 305 are further assigned to presumed depth planes in the image background. The area 306 forms an image background which is virtually at infinity. 8b shows the virtual image structure 307 generated from the assignments of the objects from FIG. 8a to the corresponding depth planes in a section along the lines A-A or the line B-B from FIG. 8a. The cuts along the section lines A - A or B - B thus result in a virtual “height profile” of the virtual image frame. As can be seen from FIG. 8b, the object 303 is arranged in this virtual “height profile” on the top depth level, while the object 304 is assigned to a depth plane below. The object 305 forms a further depth plane in the virtual image structure in FIG. 8b. For reasons of illustration, the virtual depth plane of the image background 306 is arranged relatively close to the depth planes of the other objects 303, 304 and 305 in FIG. 8b. An expedient virtual image structure expediently has to have depth level gradations which correspond to the presumed actual spatial position of the objects. Accordingly, the virtual depth plane of the image background should expediently be arranged such that its distance from the other defined depth planes of the virtual image framework corresponds to a multiple of the distances between the other. If, for example, the distances between the virtual depth planes of objects 303 and 304 or between the virtual depth planes of objects 304 and 305 are defined in the range of a few meters, the appropriate virtual distance between the depth plane of object 305 and the depth plane of background 306 must be a Realistic image structure expediently assume a size that is in the kilometer range, since experience has shown that objects in the background are depicted practically unchanged with small differences in the viewing angle.

The two-dimensional archetype is adapted to the virtual frame. In the schematic example shown in FIGS. 8a and 8b, this is done in such a way that the image data of the original image 301, in particular the image information of the individual pixels, are virtually assigned to the individual depth planes in the virtual image frame. A virtual three-dimensional image model is created which, in the example shown here, is comparable to an arrangement of “backdrops” in which the object 303 is virtually in the foreground at the level of a first virtual depth plane and the further “scenes” of objects 304 and 305 located on the levels of the corresponding other depth levels are “hidden”.

In addition to this, smooth transitions between the individual virtual depth levels can be achieved by refining the grid of the gradations of the virtual distances between the individual depth levels and by carrying out further gradations at the individual depth levels. On the other hand, it is also possible to appropriately deform the edges of the depth planes or the objects located on the depth planes in such a way that they merge into one another. For example, in the case of the schematic object 303 in FIG. 8 a, it would be possible to virtually deform the edge of its depth plane in such a way that the object 303 receives a virtual spherical curvature. In a corresponding manner, the depth planes of the virtual image framework can be deformed in such a way that the image information of the two-dimensional archetype mapped thereon can in principle have any shape or distortion within the scope of an expedient or desired virtual three-dimensional image model, which either largely corresponds to a real three-dimensional body shape or also can be enriched with any artistic effects.

FIGS. 9a and 9b show a virtual three-dimensional image model 807 generated from the virtual image framework from FIG. 8b in section along the lines A - A and B - B from FIG. 8a and their virtual recordings from two viewing angles 351 and 352 The configuration shown in FIGS. 9a and 9b corresponds to a binocular view of a three-dimensional object, which is carried out virtually within the method. In accordance with the laws of perspective, the virtual three-dimensional objects 303, 304, 305 and 306 appear to be shifted differently from one another from the relevant virtual viewing angles 351 and 352. This perspective shift is the basis for the binocular or multi-ocular viewing of spatial objects and is virtually modeled within the scope of the method according to the invention.

10 shows an example of a virtually generated displacement under the influence of the virtual viewing projection from the virtual viewing angle 352 am

Example of the section of the virtual three-dimensional image model from FIG. 9a. To calculate the virtual perspective shift of the virtual objects of the Different three-dimensional image models can be used. In the method shown by way of example in FIG. 10, the principle of centric stretching is applied, the objects 303, 304 and 306 virtually sighted from the viewing angle 352 being projected onto a virtual projection plane 308 and thereby experiencing a change in size. The projection plane can be virtually in front of the virtual three-dimensional image model as well as behind the three-dimensional image model. A position of the virtual projection plane within the virtual three-dimensional image model, for example in one of the screen planes determined during the fixed point determination, is also possible and even the most expedient, since such a projection best reflects binocular viewing conditions. In the projection shown in FIG. 10, the viewing angle 352 simultaneously forms a projection center, the virtual three-dimensional image model being viewed as it were in a virtual “incident light method” in which the radiation source virtually coincides with the camera.

Other virtual projection techniques can also be used or are expedient. Thus, the projection center can be arranged virtually behind the background of the virtual three-dimensional image model and project the corresponding objects of the virtual depth planes as a “silhouette” onto an appropriately positioned projection plane that is viewed from a viewing angle. In the case of such a virtual projection, they appear in the virtual foreground located objects are enlarged compared to the objects virtually behind them, which can create an additional spatial effect.

Furthermore, it is possible to provide several virtual projection centers in connection with several virtual projection levels in any suitable combination. For example, the virtual background can be projected onto a first projection level from a projection center that is located very far behind the virtual three-dimensional image model, while an arrangement of many, very densely graduated objects in the virtual foreground is projected through a second projection center that does not enlarge them Objects, but only a virtual displacement of these objects. The choice of the virtual projection mechanisms or the number of viewing angles depends on the specific individual case, in particular on the image motif of the two-dimensional archetype, on the depth ratios interpreted into the archetype, on the desired and / or suppressed image effects and, last but not least, on the computing effort deemed appropriate and on the ultimately used spatial imaging method for which the spatial image template is to be generated. In principle, however, any number of perspective single images with any number of arbitrarily arranged virtual projection centers, virtual projection planes, viewing angles, etc. can be generated from the virtual three-dimensional image model, the very simple exemplary embodiment shown in FIG. 10 only showing an unrepresentative, but only exemplary embodiment ,

FIG. 11 a shows a series of individual images 208 a to 208 d generated virtually according to one of the previously described projection methods from a virtual three-dimensional image model of the original image 200 shown as an example in FIG. 5 a. Although the virtual individual images 208a to 208d are shown in black and white in this exemplary embodiment, they are generally colored. In the individual images 208a, 208b, 208c and 208d, a different deformation of this image section can be seen above all by comparing the flower structure shown in the upper part of the image. This arises from the virtual projection of the virtual three-dimensional image model for the respective, in this exemplary embodiment four virtual viewing angles.

11 shows a spatial image template 209 combined from the virtual individual images 208a, 208b, 208c and 208d for an imaging method with spatial impression in conjunction with an enlarged image section 211 of the upper middle image part from the spatial image template 209. The spatial image template 209 is made up of the individual images 208a-d combined with a depth effect for an imaging process used in each case.

Exemplary two-dimensional image templates and their images are described below with reference to FIGS. 12a and 12b, or FIGS. 13a to 13c, using a monofocal lenticular screen. 12a shows an exemplary two-dimensional original image 200, which is divided into a series of image sections 361. The size of the individual image sections is in principle arbitrary and is essentially determined by the average size of the smallest closed image objects and the individual pixels. If it is assumed that clearly recognizable image structures lie in the foreground of the image, these must expediently be detected essentially as a unit by the image sections, so that they can be distinguished from other structures and offer the viewing eye a sufficient accommodation stimulus. This means that with an increasingly small grid for an increasing number of details, accommodation stimuli can be created, which lead to a deep impression on the viewer, provided that the individual image points, ie the image pixels, are not emphasized.

FIG. 12a shows a matrix-shaped grid of essentially square image sections. Another division of the two-dimensional archetype 200 is, however, readily possible. Among other things, it is useful to have circular image sections, not shown here, in a hexagonal arrangement. The hexagonal arrangement of circular image sections offers the advantage that a given image section has six immediate neighbors compared to the matrix-shaped image division, and thus there is a more homogeneous transition for the accommodating eye from a first image section to the immediate surroundings of the image.

The image sections 361 can contain preprocessed, in particular scaled, rotated or also mirrored image data with respect to several axes, which are carried out in advance with a view to compensating for the imaging effect of the lenticular screen. In this case, the image sections form a mosaic that is actually present on the two-dimensional image template. 12a also shows that some image sections 361a contain predominantly structurally poor image information, while some other image sections 361b are particularly structurally rich.

In the example shown in FIG. 12a, however, the raster of the image sections is initially not actually present in the original image itself and only appears through the lens raster placed above it. An exemplary arrangement for this is in one Side view shown in Fig. 12b. The two-dimensional image template appears on a display surface 370, for example the fluorescent surface of a picture tube or the liquid-crystalline matrix of a flat screen, and is viewed through a display surface 375. The monofocal lenticular grid 360 is arranged on the display surface 375 and can be designed, for example, as a transparent film which contains a number of matrix-like or hexagonally arranged Fresnel lenses or zone plates. The film itself adheres firmly to the display surface through adhesive, electrostatic forces or a transparent adhesive film. Each lens element 365 of the lenticular screen forms an image section 361 located underneath it in such a way that it appears in front of or behind the image plane of the display 370 as a result of the enlargement thereby caused. Therefore, the lens elements 365 are designed such that the display surface is either just before or behind the individual focal points of the lenticular screen.

This is shown in more detail in FIGS. 13a to 13c. The figures show an exemplary image detail 200a from the two-dimensional original image 200 shown in FIG. 12a with the changes in the image detail caused by the lens grid 360 or the local lens elements.

The image section 200a is formed by an unchanged part of the two-dimensional original image 200 from FIG. 12a, which is shown on the display 370. 13b, the image section 200a is divided by an arrangement of four exemplary image sections 361. The two left-hand image sections 361a each contain more structureless and diffuse background information, while the right-hand image sections 361 show structure-rich content that is obviously located in the foreground of the image.

As illustrated by way of example in FIG. 13c, each of these image sections is imaged enlarged by a lens element 365. In the exemplary illustration shown in FIG. 13c, the magnification factor when using a lens element with a focusing effect 365 is approximately 1: 2. In this exemplary representation, the left image parts 361a, which contain a diffuse structure-free image background, also result in a slight accommodation stimulus when they are enlarged by the lenticular grid due to their structurelessness, while the two right image sections 361b from FIG. 13c contain structures which accommodate the Have your eyes on the image content displayed in this way. As a result, the image contents of the right image sections 361b from FIG. 13c appear significantly closer to the viewer than the contents of the left image sections 361a. With a suitable size of the individual image sections 361, the gaps generated during the imaging by the lenticular grid are compensated for and integrated by the mode of action of the physiological visual perception apparatus.

In the exemplary embodiment from FIGS. 13a to 13c, the imaging of the image sections 361 leads to a horizontally and vertically mirrored representation. There are basically two ways to counter this effect. In a first procedure, the individual image sections of the two-dimensional original image are prepared in the sequence of the image processing method mentioned above, in particular scaled or mirrored horizontally or vertically, so that their mapping leads back to the original starting image. The strength of the preparatory scalings or reflections is derived on the basis of the magnification factor of the lenticular grid or the position of the objects to be represented, which is derived from the virtual three-dimensional image model, and is carried out in advance on the image sections.

In a second possibility, which can be used in particular for simple image motifs, such as characters or simple geometric structures on a uniform image background, the number, arrangement and size of the lens elements in the lenticular grid are selected such that the imaging factors for the entire image are insignificant. Above all, this embodiment offers the advantage that image-intensive image preparations are sometimes omitted and the spatial image template can be easily recognized without a lenticular screen. The image 200 acts as a normal two-dimensional image without a monofocal lenticular screen, while it appears staggered in depth as a result of the use of the lenticular screen, the depth effect being able to be brought about simply by attaching the lenticular screen, that is to say with very simple means.

LIST OF REFERENCE NUMBERS

10 archetype data 20 import of archetype data Adaptation of the original image data Buffering of the adapted original image data between stored image data optional color channel / color distribution change conversion into gray values edge recognition method data of the image pixel selection of the image pixel reading in of the brightness value of the image pixel increasing the brightness value image pixel with increased brightness value lowering the brightness value image pixel with reduced next brightness value Go to image menu Edge detection soft focus procedure optional: tonal value correction fixed point definition optional: setting further fixed point options saving the grayscale mask created grayscale mask distorting the original image texture, creating the virtual three-dimensional image model, generating virtual single images virtual single images combination of the virtual single images image data for spatial imaging method exemplary two-dimensional original image a single image first virtual ldc third virtual single image d fourth virtual single image combined room image template a enlarged section of a combined room image template exemplary edge-marked image 220 exemplary edge-marked, blurred image

230 exemplary tonal value-corrected soft-drawn image

239 Fixed point definition menu

240 fixed point defined image 241 histogram

242 gray scale bar

243 indicator pointers

244 selection indicator

245 Direct selection for brightness values 301 Original image, schematic

303 first object

304 second object

305 third object

306 assumed background 307 virtual image framework with virtual depth planes

308 virtual single image

351 first virtual viewing point with first viewing angle

352 second virtual viewing point with second viewing angle 360 monofocal lens grid 361 image section

361a structurally poor image sections

361b structurally rich image sections

365 lens element

370 display 375 display surface

Claims

claims

1. Method for creating and displaying a spatial image template for imaging methods with spatial depth effects from two-dimensional image data, in particular image data from images, image sequences, video films and similar two-dimensional original images, characterized in that - on the basis of image information determined from monocular original image data (10) a virtual three-dimensional image frame (307) based on a presumption-based spatial image depth gradation is generated, - the original image data are adapted to the virtual three-dimensional image frame (307) for generating a virtual three-dimensional image model (150), - the data of the virtual three-dimensional image model as a template for creating the spatial image template (209, 209a) can be used.

2. The method according to claim 1, characterized in that on the monocular master image data (10) for determining the image information, a method for edge detection (80) of the depicted objects is carried out with the generation of an edge-marked image (210), different edges being determined on the basis of a determined edge wealth Original image areas can be assigned to different virtual depth levels, in particular a background and / or a foreground.

3. The method according to claim 1, characterized in that on the original image data (10) for determining the image information, a method for determining the color information of given original image areas is carried out, wherein in the presumption-based image depth grading at least a first identified color information of a first virtual depth level and a second color information be assigned to a second virtual depth level.

4. The method according to any one of the preceding claims, characterized in that the method for edge detection (80) and the method for determining the color information are carried out individually and independently of one another or in combination.

5. The method according to any one of claims 1 to 3, characterized in that a softening process (90, 220) is used on the edge-marked image (210) to reinforce and standardize an edge-rich original image area.

6. The method according to any one of claims 1 to 5, characterized in that a tonal value correction (100) of the edge-marked image (210) is optionally carried out.

7. The method according to any one of claims 1 to 6, characterized in that on the basis of the soft-drawn and / or additionally tone-corrected edge-marked image (210, 220) an assignment based on the tonal value of a pixel of a relevant image section to a virtual depth plane (303, 304, 305, 306, 307) is carried out.

8. The method according to any one of the preceding claims, characterized in that in the case of a fixed point definition (110) the coloring and / or tonal values are limited to a predetermined value and a virtual pivot point of the virtual individual views generated later is defined.

9. The method according to any one of the preceding claims, characterized in that a fixed predetermined virtual depth level (303, 304, 305, 306, 307) is optionally assigned to a predetermined color and / or brightness value of an image pixel.

10. The method according to any one of the preceding claims, characterized in that fixed image sections, in particular the image edge and / or the image center, is assigned a predetermined virtual depth level.

11. The method according to any one of the preceding claims, characterized in that for generating the virtual three-dimensional image model, the virtual three-dimensional image structure (307) as a virtual network structure deformed in accordance with the virtual depth planes (303, 304, 305, 306, 307) and the two-dimensional one Original image is adapted as texture to the deformed network structure in a mapping process.

12. The method according to claim 1 and 11, characterized in that of the virtual three-dimensional image model from a number of virtual observation angles (351, 352) a number of virtual individual images (208a, 208b, 208c, 208d) representing the views of the virtual three-dimensional image model , 308) are generated, in which the image sections of the original image (200, 301) corresponding to a defined depth plane are shifted and / or distorted according to the virtual viewing angle.

13. The method according to any one of the preceding claims, characterized in that the virtual individual images (208a, 208b, 208c, 208d, 308) are combined according to an algorithm suitable for the imaging method with additional spatial effect for generating a spatial image template (209, 209a).

14. The method according to claim 1, characterized in that to create the spatial image template (209, 209a) image processing of individual image areas of the original image, in particular scaling and / or rotating and / or mirroring, and the spatial image template generated thereby by means of a monofocal layer above it Lenticular grid (360) is displayed.

15. The method according to claim 14, characterized in that the two-dimensional original image (200) is displayed without image processing by the monofocal lens grid (360), the two-dimensional original image (200) forming the spatial image template for display by the monofocal lens grid.

16. Device for displaying a spatial image template, characterized by a two-dimensional original image (200) as a two-dimensional image template and a monofocal lenticular screen (360) extended over the image template.

17. The apparatus as claimed in claim 16, characterized in that the two-dimensional image template is made from a mosaic of image sections (361, 361a, 361b) associated with the raster structure of the lenticular screen (360), wherein essentially one image section is an image object for essentially one Lens element (365) of the monofocal lenticular grid.

18. Device according to one of claims 16 or 17, characterized in that in a first embodiment the image sections (361, 361a, 361b) are essentially unchanged image components of the two-dimensional image template (200).

19. Device according to one of claims 16 or 17, characterized in that in a further embodiment, the image sections (361, 361a, 361b) are scaled and / or mirrored and / or rotated to compensate for the imaging effects of the lens grid (360).

20. Device according to one of claims 16 to 19, characterized in that the two-dimensional image template (200) is an image generated on a display (370) and the lenticular screen (360) is attached to the surface (375) of the display.

21. Device according to one of claims 16 to 20, characterized in that the lens grid (360) is designed as a grid-like Fresnel lens arrangement that adheres to the display surface.

22. Device according to one of claims 16 to 20, characterized in that the lens grid (360) is designed as a grid-like zone plate arrangement adhering to the display surface.

23. Device according to one of claims 16 to 20, characterized in that the lenticular screen (360) is designed as a grid-like, conventional convex lens arrangement adhering to the display surface.