Method and configuration method for simulating a second camera objective lens and movie camera

Abstract

A method for simulating a second camera objective lens and a corresponding configuration method and a movie camera, wherein the configuration method comprises: determining a plurality of first images recorded using a first camera objective lens; determining a plurality of second images recorded using a second camera objective lens, wherein the image content of the second images corresponds to the image content of the first images; and configuring a simulation method for simulating the second camera objective lens based on the plurality of first images and the plurality of second images.

Description

Technical Field

[0001] The present application relates to a method and a device for simulating a camera objective (camera lens), and in particular to a method and a device for adapting the imaging characteristics of a camera objective, and to a method and a device for configuring such an adaptation method and device. Background Art

[0002] An electronic movie camera generally includes an integrated optical system (e.g., a camera objective lens, which may be referred to as an objective lens or lens) or a lens mount for mounting an interchangeable objective lens (interchangeable lens); further includes an electronic image sensor for generating an image signal from incident light; and an image signal processing device for generating a digital image data stream based on the image signal. The image data stream may include or consist of multiple image data sets, which may be referred to as images.

[0003] In professional cinematography, cameras with interchangeable objectives are mainly used. For example, the focal length of the individual objectives is different, so that different recording situations can be taken into account due to the change in focal length. Furthermore, it is known to adapt the individual objectives of this set to one another so that the image impression (for example color reproduction or resolution performance) remains unchanged when changing from one focal length to another.

[0004] It is well known that every real object has physical limitations in its imaging properties. Depending on how the objective is calculated, designed and manufactured, more or less visible imaging errors such as color errors, distortions or scattered light reflections occur. On the one hand, attempts are made to minimize these errors; on the other hand, minimal residual errors are also considered part of the imaging properties of the objective. These properties may also be intentional in terms of the aesthetic effect of the recorded image. The aesthetic effect in this regard is not only related to the sharply imaged areas of the image, but the way in which the objective represents the blurred areas of the image may also contribute to the overall effect.

[0005] Until now, the optimization of objectives has focused mainly on quantitatively measurable parameters such as contrast transfer (modulation transfer function, MTF), distortion, chromatic aberrations, etc. Known simulation programs can calculate these parameters for a predetermined objective design based on the laws of optics to support the optimization process.

[0006] Specific information about the objective design can be used to improve the correction image processing. In particular in the case of interchangeable objectives, it may be advantageous if the image processing can be adapted individually. For example, the relevant information can be stored in the respective objective and then transferred to the image processing software when the objective is connected to the camera. Furthermore, it is possible to identify specific targets by the image processing software and to retrieve information relevant for the correction from the corresponding database.

[0007] The camera objective used has a decisive influence on the appearance of the image recorded by a movie camera, i.e. the subjective perception of the imaging characteristics of the objective, which can also be called the "look" of the objective.

[0008] Such a scene may be produced by the specific characteristics and settings of the lens used, which may be, for example, characteristic of a certain objective lens manufacturer and / or a certain time period.Such a characteristic scene is recognizable to a person skilled in the art.

[0009] Users are always looking for new perspectives with objectives. In this respect, developing objectives with a specific perspective is very complex and / or expensive. The use of filters or the adjustment of individual optical parameters can only approximate to a certain extent the subtle effects of a real objective that results in a specific perspective.

[0010] When very specific imaging characteristics are to be achieved in the design of objectives (e.g. by minimizing residual errors, adapting an existing objective, or producing a specific aesthetic impression), all known methods rely on modeling the optical conditions according to known physical laws.

[0011] In order to achieve accurate modeling of imaging behavior, a very complex mathematical representation of the physical conditions is required. Effects such as the representation of scattered light or blurred areas can be very difficult to model because they depend on a large number of parameters.

[0012] Therefore, the correction of the objective lens based on the subsequent image processing is limited in its implementation. In particular, in embodiments where the effect is to be produced in real time or directly in the camera, the limits of the available processing power can be reached quickly. Only simplified processing procedures can then be used, which in turn can only simulate the desired optical effect to a limited extent.

[0013] Therefore, physical models of the objective can only describe the subjective properties of the objective to a limited extent or may be too complex to be useful in real time. Summary of the invention

[0014] An object of the present application is to provide a method and a device which are capable of generating an image having a specific objective view without directly using the corresponding objective.

[0015] This object is achieved by a simulation method for configuring a second camera objective having the following characteristics, a method for simulating a second camera objective having the following characteristics, a computer program product having the following characteristics, and a movie camera having the following characteristics. The description or the drawings describe preferred embodiments of these methods, computer program products, and movie cameras.

[0016] This method for configuring a simulation method for simulating a second camera objective lens includes at least the following steps: a first image determination step: determining a plurality of first images recorded using the first camera objective lens; a second image determination step: determining a plurality of second images recorded using the second camera objective lens, wherein the image content of the second images corresponds to the image content of the first images; and a simulation method configuration step: configuring a simulation method for simulating a second camera objective lens based on the plurality of first images and the plurality of second images.

[0017] In other words, this method configures a simulation method that simulates the view of a second image based on an image recorded by a first objective lens (without changing the actual content of the recorded image, i.e., the content displayed on the recorded image). This method configures a simulation method based on a plurality of first images recorded by a first objective lens. In this respect, a second image is associated with each first image, so that the first image and the associated second image have the same image content (but they are recorded by different objective lenses).

[0018] In this respect, the "simulation" of the second camera objective can be understood as the reproduction of an image obtained using the second camera objective. The second camera objective can also be designated as the objective to be simulated. The result of the simulation can be called a simulated image.

[0019] It can be seen that the first image and the second image are provided to the method in such a way as to configure (also known as: train) the simulation method, i.e. the first image is provided as input to the simulation method, the output of the simulation method (i.e. the simulated image) is provided, wherein the simulation method is evaluated by the second image. Ideally, the simulated image is similar or identical to the second image, at least in terms of the image impression or effect or subjective impression (i.e. the scene). The plurality of first images and the plurality of second images may be referred to as a training data set. The training data set may include any desired real scene or transmitted light image template. The training data set may include a plurality of images, wherein the settings of the plurality of images, such as resolution, contrast range, color or depth level, etc., are different.

[0020] In this respect, the simulation method can be configured iteratively, for example, by feeding back an evaluation of the simulated image. The feedback can include an evaluation of an image impression or effect or subjective impression of similarity or identity between the first image and the second image.

[0021] A simulation method or an apparatus for performing a simulation method may also be considered a generator. A method or an apparatus for evaluating a simulated image may be considered a discriminator.

[0022] According to an embodiment, the step of configuring the simulation method comprises determining parameters representative of the second camera objective.Thus, different parameters (or different sets of parameters) may characterize different second camera objectives.

[0023] According to an embodiment, a plurality of first images are recorded under predetermined conditions, and a plurality of second images are recorded under predetermined conditions, wherein the predetermined conditions include at least one predefined lighting situation; and / or at least one predefined camera objective lens aperture setting; and / or at least one predefined camera objective lens focus setting; and / or at least one predefined camera objective lens focal length setting; and / or at least one camera sensor sensitivity setting; and / or at least one camera sensor white balance setting. Since the first image and the second image correspond to the same predetermined conditions, such as the same aperture setting of the objective lens used to generate the respective images and the same focus setting, the simulation method can be configured particularly accurately.

[0024] Thus, each first image has a corresponding second image recorded under the same conditions. These corresponding images can be configured with the simulation method. The simulation method can be explicitly configured so that the first images are processed so that they correspond to the second images, in particular in terms of the scene. The simulation method configured in this way can then also generate simulated images for images recorded by the first objective but not included in the training data set (and therefore simulate the second objective), i.e. images that do not correspond to the images recorded by the second objective.

[0025] According to an embodiment, the method comprises a step of generating a plurality of first images and a step of generating a plurality of second images, wherein the plurality of first images and the plurality of second images are recorded simultaneously (ie in pairs) by two different image sensors.

[0026] According to one embodiment, in order to simultaneously record a plurality of first images and a plurality of second images, the first camera objective images the corresponding image motif onto a first image sensor via a beam splitter (e.g. a partially transmissive mirror), and the second camera objective images the corresponding image motif onto (e.g. by transmission or reflection) a second image sensor via a beam splitter. Due to the beam splitter, the optical axis of the first objective and the optical axis of the second objective may be superimposed on each other. In this regard, the first image sensor and the second image sensor may be of the same type (i.e. identical in design) and / or may be arranged identically, so that the difference between the first image and the second image is essentially only due to the use of the first objective for the first image and the use of the second objective for the second image.

[0027] In various embodiments, a plurality of first images and a plurality of second images can be recorded one after another by the same image sensor. This is particularly possible for static image content, ie, image content that does not change between the time of recording a first image and the time of recording a second image corresponding to the first image.

[0028] Alternatively, the first image and / or the second image may also be based on existing (e.g. past) image material. Typically, the first image and / or the second image may also be artificially generated or subjected to post-digital processing.

[0029] According to one embodiment, a plurality of second images and corresponding templates or image motifs may be provided as a training data set, which makes it possible to generate a plurality of first images decoupled in space and time from the second images in order to configure a simulation method for simulating a second objective. Thus, a training data set may be provided, which may in particular correspond to a history of images generated by the second camera objective. Due to the presence of such a training data set, it is not necessary to directly use the second camera objective to configure the simulation method. It may be provided that the simulation method is configured in such a way that a plurality of first images are indeed generated by different (first) camera objectives, however, it is preferred to use the same image sensor or an image sensor of the same type or an image sensor configured in the same way as the original generation of the plurality of second images. The plurality of first images may be generated based on a template of the training data set.

[0030] According to one embodiment, the simulation method uses a first artificial neural network. According to one embodiment, the first artificial neural network has a deconvolutional neural network. Such a deconvolutional neural network may include at least one deconvolutional layer. The deconvolutional neural network may also include multiple deconvolutional layers. The deconvolutional neural network may also include other layers (not deconvolutional layers) in addition to the at least one deconvolutional layer.

[0031] In this regard, artificial neural networks of predetermined structures may be used, in particular deconvolutional neural networks, wherein parameters describe the connections between individual neurons within the artificial neural network.

[0032] The configuration of the simulation method, ie the configuration of the artificial neural network, can then include the determination of these parameters. In particular, the parameters can be determined so that the simulated image, ie the image generated by the simulation method from the first image, coincides as closely as possible with the second image.

[0033] According to one embodiment, the simulation method can be configured, in particular, by optimizing an objective function.

[0034] According to one embodiment, the target function is obtained according to the extent to which a first image (of the plurality of first images) differs from a corresponding second image (of the plurality of second images) processed by the simulation method. In this respect, the difference between the two images can be determined pixel by pixel (i.e. by image element), or pixel by pixel group (i.e. by a group of a plurality of image elements, in particular adjacent image elements), or in another way. For example, according to one embodiment, the target function is obtained according to the impression that each first image, after being processed by the simulation method, produces on a (human) observer, which has been recorded by the second objective lens.

[0035] According to an embodiment, instead of a human observer, a second artificial neural network may be used, which is trained to recognize whether an image is recorded with the aid of a second camera objective. According to an embodiment, based on the second artificial neural network (which may be an example of a discriminator), an objective function is obtained. According to an embodiment, the second artificial neural network has a convolutional neural network. Such a convolutional neural network may include at least one convolutional layer. The convolutional neural network may also include a plurality of convolutional layers. In addition to the at least one convolutional layer, the convolutional neural network may also include other layers (not convolutional layers).

[0036] Therefore, in cooperation with the first artificial neural network, a competition may arise between the first artificial neural network and the second artificial neural network, wherein the first artificial neural network attempts to generate a simulated image that is as indistinguishable as possible from the image recorded through the second objective lens, and the second artificial neural network attempts to identify any image that is not recorded with the aid of the second objective lens (that is, any simulated image) as such an image. Therefore, according to one embodiment, the simulation method can be configured based on a zero-sum game between the first artificial neural network and the second artificial neural network. The principle of zero-sum games is well known to technicians in technical fields such as game theory. The goal of training the first artificial neural network (i.e., the simulation method) is that the first artificial neural network generates a simulated image so that the second artificial neural network regards the simulated image as originating from the second objective lens.

[0037] According to an embodiment, the method further comprises the step of alternately optimizing the first artificial neural network and the second artificial neural network.

[0038] In this respect, the input data for the optimization can be used multiple times (ie in multiple optimization steps, for example for a simulation method and for configuring a simulation method).

[0039] The input data for the optimization generator (ie the simulation method) can be a first image recorded by a first objective and a second image (of the same content) recorded by a second objective.

[0040] The input data for optimizing the generator may also include historical images or historical film material recorded by a second objective (which is no longer available). Since only the second objective may have been used in the process of generating the old film material, it is necessary to use a method of configuring the simulation method to provide the corresponding image (or corresponding film material) recorded by the first objective. This can be done, for example, by manual processing so that the first image gives the impression of being recorded by the first objective.

[0041] The input data for optimizing the discriminator (i.e. the evaluation unit) might be as follows:

[0042] - images actually recorded by the second objective and information about these images being recorded by the second objective (the discriminator should recognize these images as originating from the second objective, i.e. as images recorded by the second objective);

[0043] - simulated images that actually originate from the first objective (but are simulations of the view of the second objective by simulation methods) and information that these images were not recorded by the second objective (the discriminator should recognize these images as not originating from the second objective); and

[0044] - images recorded by the first objective and information that these images were not recorded by the second objective (the discriminator should recognize these images as not originating from the second objective).

[0045] The present application also relates to a method for simulating a second camera objective, the method comprising the steps of: determining an image recorded using a first camera objective; determining parameters characterizing the second camera objective; and applying processing rules to the determined image using the parameters to simulate the second camera objective.

[0046] In this sense, the simulation method can use the parameters to generate an image based on a first image recorded by a first objective, which image has the same content as the first image, but has a view of a second objective. In this respect, by carrying out the described method for configuring the simulation method, different parameter sets can be prepared for different second camera objectives, wherein each parameter set characterizes one of a plurality of second camera objectives. Thus, starting from a first image, a plurality of different second camera objectives can be simulated (by selecting the corresponding parameter sets).

[0047] The simulation method may have a fixed structure, ie it may comprise fixed rules on how to process the image by means of parameters in order to obtain a simulated image in this way. In this respect, the parameters parameterize these rules.

[0048] The simulation method can be performed in the camera (on the fly, essentially directly after recording the images by means of the real objective) or in the camera, i.e. after storing the image data set, in particular at a location independent of the camera. For example, a plurality of images or one or more moving image sequences (each image sequence comprising a plurality of images) can be recorded and stored by means of the first camera objective, and then the simulated images or moving image sequences can be generated based on these images or moving image sequences.

[0049] According to an embodiment of the simulation method, the processing algorithm comprises an artificial neural network, for example a deconvolutional neural network.

[0050] According to an embodiment, the simulation method is configured according to the configuration method as described above or according to the embodiments described with respect thereto.

[0051] According to an embodiment of the simulation method, the configuration of the method includes determining parameters. For example, the parameters can be stored in a movie camera or in a database and can be read from the movie camera or the database. The database can also be stored on a decentralized server (e.g., cloud storage).

[0052] A plurality of parameter sets may be stored in the database. Each set of parameters may characterize a specific objective lens to be simulated. Each set of parameters may also characterize a specific first objective lens (as the objective lens actually used) through which the image of the second objective lens to be simulated is recorded.

[0053] According to an embodiment of the simulation method, the parameters are provided in encrypted form, wherein the step of determining the parameters comprises decrypting the parameters based on a user key (which may also be designated as a usage key). The user key may indicate which of the plurality of parameter sets a user may access (i.e. which simulation of the plurality of objective lenses to be simulated and / or which of the plurality of objective lenses to be actually used). Thus, different user authorizations may be defined.

[0054] The user key may be applicable to the processed image, the first objective and / or the second objective and the simulation method (ie transformation) relative to the first objective and the second objective (ie relative to a pair of objectives). The use of the key may be limited in time.

[0055] According to an embodiment of the simulation method, a plurality of parameter groups are provided, which correspond to a plurality of different second camera objective lenses; and / or a plurality of camera objective lens aperture settings; and / or a plurality of different first camera objective lenses; and / or a plurality of camera objective lens focus settings; and / or a plurality of camera objective lens focal length settings; and / or a plurality of camera sensor sensitivity settings; and / or a plurality of camera sensor white balance settings. The method further comprises the following steps: selecting a second camera objective lens, and / or a first camera objective lens, and / or a camera objective lens aperture setting, and / or a camera objective lens focus setting, and / or a camera objective lens focal length setting, and / or a camera sensor sensitivity setting, and / or a camera sensor white balance setting; determining a parameter group, which includes parameters characterizing the selected second camera objective lens, and / or the selected first camera objective lens, and / or the selected camera objective lens aperture setting, and / or the selected camera objective lens focal length setting, and / or the selected camera sensor sensitivity setting, and / or the camera sensor white balance setting; determining a parameter group, which includes parameters characterizing the selected second camera objective lens, and / or the selected first camera objective lens, and / or the selected camera objective lens aperture setting, and / or the selected camera objective lens focal length setting, and / or the selected ... and applying the processing rules to the determined image using the parameters of the determined parameter group to simulate the second camera objective lens (i.e., using the first camera objective lens, and / or the selected camera objective lens aperture setting, and / or the selected camera objective lens focus setting, and / or the selected camera objective lens focal length setting, and / or the selected camera sensor sensitivity setting, and / or the selected camera sensor white balance setting to simulate the selected second camera objective lens).

[0056] For a pair of objectives (i.e. a specific first objective and a specific second objective), there may be multiple parameter sets for the simulation method, for example for different settings of the objectives (e.g. focus or iris aperture). When the second objective is used, its settings can be detected and the data can be used to select a suitable parameter set for the transformation.

[0057] In this regard, one of the plurality of second camera objectives may be selected based on a user input. The corresponding camera objective settings may be selected based on the user input, or the selection may be made automatically (e.g. based on an electronic identification, such as based on metadata included in the image data set (recorded by the first objective), such as, for example, LDS (Lens Data System) data). The electronic identification may be forwarded when the first image is recorded in order to select the correct parameters for the simulation method.

[0058] If the parameters (or parameter sets) are available for different first objective lenses, the selection of the parameters relevant to the first objective lens actually used may occur as a result of user input or may be performed automatically, for example by means of LDS (Lens Data System) data comprising information about the first objective lens model, for example based on metadata comprised by the image dataset (of the images recorded by the first objective lens).

[0059] According to one embodiment, in order to determine the image, the first camera objective images the image motif onto the image sensor, and the image sensor generates a corresponding image signal. The image data set includes the image signal in digital form and can also be further processed (e.g. in a linear or logarithmic manner). The image data set can also additionally include metadata, such as a time signal, a sound signal and / or a camera objective setting. The image data set can also be referred to as an image for short.

[0060] The application also relates to a computer program product comprising instructions which, when executed by a computer, cause said computer to perform one of the above methods (configuration method, simulation method).An electronic computing component is considered herein to be a computer.

[0061] The application also relates to a film camera for recording a time-ordered sequence of frames, the film camera having a control unit for processing the frames according to one of the above-described simulation methods, for example in real time.

[0062] The control unit may, for example, include one or more of the following: an integrated circuit (IC); a microprocessor; a central processing unit (CPU); a graphics processing unit (GPU); an application specific integrated circuit (ASIC); or a field programmable gate array (FPGA). The control unit may be located within the housing of the camera. Alternatively, components of the control unit may be located outside the housing of the camera and / or form a separate computing device that is permanently or removably connected to the camera.

[0063] According to an embodiment, the control unit is used to form an artificial neural network. As described above in conjunction with the simulation method, the artificial neural network can be parameterized by parameters.

[0064] According to an embodiment, the control unit has a decryption module for decrypting the parameters.

[0065] According to one embodiment, the movie camera has a housing that accommodates an image sensor for generating an image signal and has an interchangeable objective lens mount on which a first camera objective can be selectively fixed. In this case, the parameters can characterize not only the second camera objective (so that different second camera objectives can be simulated) but also the first camera objective (so that the second camera objective can be simulated based on different first objectives). Obviously, parameters for different associations of the first camera objective (i.e. the camera objective that actually records the image) with the second camera objective (i.e. the camera objective to be simulated) can thus be provided.

[0066] In the method according to the present application, the transformation of the image cannot be based on the physical parameters of the second objective, but on the image impression; thus, the necessity of complex modeling based on optical laws can be avoided. After training once, the less computationally intensive transformation can subsequently be used to align the recordings (e.g., the trained artificial neural network of the generator).

[0067] Furthermore, no physical description of the objective properties is required. Therefore, even very subjective imaging features can be detected.

[0068] This method allows the detection of any desired new objective lens even without detailed technical data. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The present application will be described below with reference to the embodiments and in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar components are marked with the same reference numerals. The accompanying drawings include:

[0070] Figure 1 is a schematic diagram of a system having a simulator;

[0071] Figure 2 is a schematic diagram of a system for configuring a simulator;

[0072] Figure 3 is a system for providing images with the aid of different objective lenses;

[0073] Figure 4 is a schematic diagram of a system for configuring a simulator;

[0074] Figure 5 is a schematic diagram of the expected working mode of the evaluation unit;

[0075] Figure 6 is a schematic diagram of a system for configuring an evaluation unit. DETAILED DESCRIPTION

[0076] Figure 1 Schematically shown is a system 100 with a simulator 104 which, starting from a first image 102 recorded by a first objective (i.e. using a first objective), generates a simulated image 106 which gives a human observer the impression that it has been recorded by a second objective different from the first objective. In this respect, the image content of the simulated image 106 is identical to the image content of the first image 102. However, the simulated image 106 differs from the first image 102 in the characteristic impression of the respective objective.

[0077] In the images shown in the various figures, in order to make the figures clear and legible, the upper right corner of the image indicates whether the image is an image recorded through the first objective lens (not processed by the simulation method) (upper right corner label "1"), or an image recorded through the second objective lens (upper right corner label "2"), or a simulated image recorded by the first objective lens and intended to simulate the image impression of the second objective lens (upper right corner label "1→2").

[0078] The simulator 104 may be configured (ie, set; ie, trained) by machine learning. For example, the simulator 104 may be an (artificial) neural network.

[0079] Figure 2 The system 200 for configuring the simulator 104 is schematically shown. To this end, a second image 202 recorded by a second objective lens can be provided. In fact, the content displayed by the second image 202 is the same as that displayed by the first image 102 (i.e., for example, the same scene or the same theme). Therefore, the difference between the first image 102 and the second image 202 is only that, for example, the following combination Figure 3 As further described, the first image is recorded by a first objective lens and the second image 202 is recorded by a second objective lens.

[0080] The simulated image 106 represents a simulation of the second objective lens based on the first image 102 recorded by the first objective lens. The simulated image 106 and the second image 202 actually recorded by the second objective lens can be compared using an evaluation unit 204. The comparison result obtained by executing the evaluation unit 204 can be used to configure (as shown by arrow 206) the simulator 104. During the configuration, the parameters of the simulator 104 can be updated (e.g. slightly modified) according to each pair of images formed by the first image 102 and the associated second image 202. The parameters of the simulator 104 can, for example, include the links of the individual neurons in the artificial neural network.

[0081] exist Figure 2 In an embodiment not shown in FIG. 1 , the first image 102 and the second image 202 may be directly transmitted to a configuration unit, which configures the simulator 104 based on the first image 102 and the second image 202. Therefore, it is not absolutely necessary to compare the simulated image 106 with the second image 202.

[0082] although Figure 2Only one first image 102 and one second image 202 associated therewith are shown, but a plurality of first images 102 and a plurality of second images 202 are generally required to configure the simulator 104, wherein one of the plurality of second images 202 is associated with each of the plurality of first images 102. Therefore, the simulator 104 can be clearly configured based on possible inputs (i.e., the first images 102) in normal operation, and the simulator 104 should generate the expected output in response to the inputs (i.e., the second images 202).

[0083] Figure 3 A system 300 is schematically shown of images 102, 202 provided by different objective lenses 306, 308. The template 302 is the content of the recorded image 102, 202, which can also be represented as an image motif, and can be imaged by a beam splitter 304 and a first objective lens 305 to obtain a first image 102. Moreover, the second image 202 can be obtained by imaging the template 302 by the beam splitter 304 and the second objective lens 308. The first objective lens 306 and the second objective lens 308 can have the same focal length so that the template 302 appears at least substantially the same in the first image 102 and the second image 202 (in other words: so that the image content of the first image 102 at least substantially corresponds to the image content of the second image 202). Therefore, the first image 102 and the second image 202 differ only in the imaging characteristics (i.e. the view) of the used objective lenses (the first objective lens 306 and the second objective lens 308). In order to eliminate or minimize the influence of the image detection unit (such as the image detection unit of the image sensor, or the image processing / pre-processing occurring in the camera) on the images 102, 202, the first objective lens 306 and the second objective lens 308 can work under the same camera hardware (especially the same image sensor).

[0084] The use of a beam splitter 304 to generate two images 102, 202 having (at least substantially) the same content enables the generation of the images to occur simultaneously, which is particularly advantageous for moving content 302. However, in particular for static content 302, the second image 202 may also be recorded directly using the second objective 308 (i.e. without the use of the beam splitter 304), and after or before this, the first image 102 may be recorded using the first objective 306 (without the use of the beam splitter 304).

[0085] Figure 4The system 400 for configuring the simulator 104 is schematically shown. As indicated by arrow 402, the training goal of the simulator 104 is that the image recorded by the first objective lens 306 cannot be distinguished from the image from the second objective lens 308 after being transformed into a simulated image by the simulator 104. The training goal of the evaluation unit 204 is to train it, and after the training, the image recorded by the first objective lens 306 can still be distinguished from the image from the second objective lens 308 after being transformed into a simulated image by the simulator 104.

[0086] Using a (fixed or determined) first objective lens 306, multiple trainings can be performed with different second objective lenses 308, so that simulation methods with different configurations (e.g., with different parameters or different sets of parameters) can be generated. Thus, different parameters can be stored (and associated with the (fixed or determined) first objective lens 306 and the corresponding second objective lens 308), and a certain set of parameters can be used for the simulation method as needed. Thus, an image from the first objective lens 306 can be made to look like an image from a certain second objective lens 308.

[0087] Using the (fixed or determined) second objective lens 308, multiple trainings can be performed with different first objective lenses 306, so that simulation methods with different configurations (e.g., with different parameters or different sets of parameters) can be generated. Thus, different parameters can be stored (and associated with the (fixed or determined) second objective lens 308 and the corresponding first objective lens 306), and a certain set of parameters can be used for the simulation method as needed. Thus, images from different first objective lenses 306 can be made to look like images from the second objective lens 308.

[0088] Figure 5 An example 500 of a desired operating mode of the evaluation unit 204 is shown. The evaluation unit 204 processes the image as input and outputs information about whether the image was recorded by the second objective lens. Figure 5In the example shown, the evaluation unit 204 outputs "yes" (indicated by a check mark, i.e., √) or "no" (indicated by an X) to indicate information about whether the image is recorded by the second objective lens. However, the information about whether the image is recorded by the second objective lens can also be generally expressed in a probability manner, rather than in a "yes" or "no" (i.e., a binary manner). This probability manner can be, for example, expressed in a percentage, such as a real number or a natural number between 0 and 1 or between 0% and 100%, or in a "fuzzy" tone such as "very sure it is", "probably", "cannot be sure", "probably not", "very sure it is not", etc. For another example, the information about whether the image is recorded by the second objective lens can be expressed in a manner of whether there is a gap in pixels between the image and the image actually recorded by the second objective lens. Alternatively, the information about whether the image is recorded by the second objective lens may include information about how similar a human observer judges the image to the image actually recorded by the second objective lens.

[0089] For the second image 202 recorded by the second objective lens as input, the information output by the evaluation unit 204 on whether the second image 202 is recorded by the second objective lens should show that the second image 202 is recorded by the second objective lens, and is indicated by a check mark 502. For the image 102 recorded by the first objective lens as input, the information output by the evaluation unit 204 on whether the first image 102 is recorded by the second objective lens should show that the first image 102 is not recorded by the second objective lens, and is indicated by a mark 504, i.e., X. For the image generated by the simulator 104 (i.e., the simulated image 106), the information output by the evaluation unit 204 on whether the image 106 generated by the simulator 104 is recorded by the second objective lens should show that the image 106 generated by the simulator 104 is not recorded by the second objective lens, and is indicated by a mark 506, i.e., X.

[0090] As further described above, the simulator 104 can be configured (i.e., trained). The evaluation unit 204 can also be configured (i.e., trained). In this regard, the goal is to train the simulator 104 so that the images it generates cannot be distinguished from the images recorded by the second objective lens; while the goal of training the evaluation unit 204 is to enable it to recognize any images that were not recorded by the second objective lens. In this regard, the evaluation unit 204 should also be able to recognize images that were generated by the simulator 104 but were not recorded by the second objective lens. Therefore, in this sense, the simulator 104 and the evaluation unit 204 compete with each other, so that the configuration of the simulator 104 and the configuration of the evaluation unit 204 can be regarded as a zero-sum game. The "better" the simulator 104 is, the harder it is for the evaluation unit 204 to recognize that the simulated image is not an image actually recorded by the second objective lens.

[0091] The evaluation unit 204 may be configured by machine learning. For example, the evaluation unit 204 may be an (artificial) neural network.

[0092] Figure 6 Schematically shown is a system 600 for configuring the evaluation unit 204. The data input for configuring the evaluation unit 204 may be an image 602 and its associated information about whether the image 602 was recorded by the second objective lens. The number of images 602 and their associated information 604 may also be specified as a training data set.

[0093] Each image 602 is fed to the evaluation unit 204, which outputs information 606 about whether the image 602 was recorded by the second objective lens (after evaluation by the evaluation unit 204). The comparison unit 608 processes (e.g. compares) the information 606 output by the evaluation unit 204 with the information 604 characterizing that the image 602 is actually an image 602 recorded by the second objective lens. Based on this processing, the evaluation unit 204 is configured (as indicated by arrow 610). In this regard, the configuration may include determining the parameters used by the evaluation unit 204. Depending on each image 602 to be processed and its associated information 604, the parameters may be updated (e.g. slightly modified) using, for example, a gradient method. The configuration may clearly include setting the parameters of the evaluation unit 204 so that for as many input images 602 as possible, the input information 604 and the information 604 determined by the evaluation unit 204 may be as consistent as possible.

[0094] Reference numerals list

[0095] 100 Systems with simulators

[0096] 102 First Image

[0097] 104 Simulator

[0098] 106 Simulated images

[0099] 200 System for configuring the simulator

[0100] 202 Second Image

[0101] 204 Evaluation Units

[0102] 206 Arrow indicating configuration

[0103] 300 Systems providing images

[0104] 302 Template

[0105] 304 Splitter

[0106] 306 First Objective

[0107] 308 Second objective lens

[0108] 400 System for configuring the simulator

[0109] 402 Training objectives of the evaluation unit

[0110] Example of the expected operating mode of the 500 evaluation unit

[0111] 502 Checkmark

[0112] 504 X

[0113] 506 X

[0114] 600 System for configuring evaluation units

[0115] 602 images

[0116] 604 Information related to the image

[0117] 606 Information determined as to whether the image is recorded by the second objective lens

[0118] 608 Comparison Unit

[0119] 610 Arrow indicating configuration

Claims

1. A configuration method for simulating a second camera objective lens (308), characterized in that: The following steps are involved: A first image determination step: determining a plurality of first images (102) recorded using a first camera objective lens (306); A second image determination step: determining a plurality of second images (202) recorded using a second camera objective lens (308), wherein image contents of the second images correspond to image contents of the first images (102); Configuration simulation method step: Based on the multiple first images (102) and the multiple second images (202), configure a simulation method for simulating a second camera objective (308), wherein the first image (102) is provided as an input of the simulation method, and the simulated image output of the simulation method in response to the input is evaluated through the second image (202).

2. The configuration method according to claim 1, characterized in that: The step of configuring the simulation method comprises determining parameters characterizing the second camera objective (308).

3. The configuration method according to claim 1, characterized in that: Recording the plurality of first images under predetermined conditions (102); Recording the plurality of second images under predetermined conditions (202); The predetermined condition includes at least one of the following: at least one predefined lighting situation; at least one predefined camera objective aperture setting; at least one predefined camera objective focus setting; at least one predefined camera objective lens focal length setting; At least one camera sensor sensitivity setting; or At least one camera sensor white balance setting.

4. The configuration method according to claim 1, characterized in that: The method comprises the steps of generating the plurality of first images (102) and generating the plurality of second images (202), wherein the plurality of first images (102) and the plurality of second images (202) are recorded simultaneously by two different image sensors.

5. The configuration method according to claim 4, characterized in that: In order to simultaneously record multiple first images (102) and multiple second images (202), the first camera objective (306) images the corresponding image motifs onto the first image sensor via the beam splitter (304), and the second camera objective (308) images the corresponding image motifs onto the second image sensor via the beam splitter (304).

6. The configuration method according to claim 1, characterized in that: The simulation method uses a first artificial neural network.

7. The configuration method according to claim 6, characterized in that: The first artificial neural network has a deconvolutional neural network.

8. The configuration method according to claim 1, characterized in that: The simulation method is configured by optimizing an objective function.

9. The configuration method according to claim 8, characterized in that: The objective function is obtained according to the degree to which the first image (102) differs from the corresponding second image (202) processed by the simulation method.

10. The configuration method according to claim 8, characterized in that: The target function is obtained based on the impression that each first image (102) produces on an observer after being processed by the simulation method and has been recorded by the second camera objective (308).

11. The configuration method according to claim 8, characterized in that: Based on the second artificial neural network, the objective function is obtained.

12. The configuration method according to claim 11, characterized in that: The second artificial neural network has a convolutional neural network.

13. The configuration method according to claim 11, characterized in that: The simulation method uses a first artificial neural network; the simulation method is configured based on a zero-sum game between the first artificial neural network and the second artificial neural network.

14. The configuration method according to claim 11, characterized in that: The simulation method uses a first artificial neural network, and the method further includes: alternately optimizing the first artificial neural network and the second artificial neural network.

15. A method for simulating a second camera objective (308), characterized in that According to the configuration method of claim 1, the method for simulating the second camera objective lens comprises the following steps: determining an image (102) recorded using a first camera objective (306); Determining parameters characterizing the second camera objective (308) according to the configuration method of claim 1; Processing rules are applied to the determined image (102) using the parameters to simulate the second camera objective (308).

16. The method according to claim 15, characterized in that The processing algorithm has an artificial neural network.

17. The method according to claim 16, characterized in that The artificial neural network has a deconvolutional neural network.

18. The method according to claim 15, characterized in that The parameters are provided in encrypted form, and the step of determining the parameters characterizing the second camera objective (308) comprises decrypting the parameters based on a user key.

19. The method according to claim 15, characterized in that A plurality of parameter groups corresponding to at least one of the following are provided: a plurality of different second camera objective lenses; a plurality of different first camera objective lenses; Multiple camera objective aperture settings; Multiple camera objective focus settings; Multiple camera lens focal length settings; Multiple camera sensor sensitivity settings; or Multiple camera sensor white balance settings; The method further comprises the following steps: Select at least one of the following: Second camera objective lens, First camera objective lens, Camera objective lens aperture setting, Camera objective lens focus setting, Camera objective lens focal length setting, Camera sensor sensitivity settings, or Camera sensor white balance settings; A parameter set is determined that includes parameters characterizing at least one of: The selected second camera objective, the selected first camera objective, The selected camera objective aperture setting, The selected camera objective focus setting, The selected camera objective lens focal length setting, The selected camera sensor sensitivity setting, or The selected camera sensor white balance setting; The processing rules are applied to the recorded images using the parameters in the determined parameter set to simulate a second camera objective.

20. The method according to claim 15, characterized in that To determine the image, the first camera objective images the image motif onto an image sensor which generates a corresponding image signal.

21. A computer program product, characterized in that The method comprises instructions which, when executed by a computer, cause the computer to execute the configuration method according to claim 1 .

22. A movie camera for recording a sequence of time frames and having a control unit for processing the frames according to the method of claim 15.

23. The movie camera according to claim 22, characterized in that The control unit is used to form an artificial neural network.

24. The movie camera according to claim 22, characterized in that The control unit has a decryption module for decrypting parameters.

25. The movie camera according to claim 22, characterized in that The movie camera has a housing that accommodates an image sensor for generating an image signal, and the housing has an interchangeable objective lens mount on which a first camera objective lens can be selectively fixed.

Images