Method and device for determining parameters for spectacle fitting

Abstract

Methods for determining parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68), comprising: Acquiring (70; 80) depth information regarding a user's head (10), wherein the depth information comprises a distance (14) between the user's head (10) and a device used for acquiring, and Determining (72; 85) parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68) based on depth information, Taking (71; 81) a 2D image of the head, wherein the acquisition (71; 81) of the 2D image and the acquisition (70; 80) of the depth information are carried out via a common optical axis.

Description

[0001] The present application relates to methods and devices for determining parameters for fitting spectacles to a person's head, in particular the determination of centering parameters. Such centering parameters are used to correctly position spectacle lenses in a spectacle frame, i.e., to center them, so that the spectacle lenses are worn in the correct position relative to the person's eyes.

[0002] A generic device and a generic method are known, for example, from EP 1 844 363 B2.

[0003] The approach used in this publication employs a pair of image acquisition devices to generate stereo image data of a person's head or parts thereof. A three-dimensional model of the head is then calculated from this stereo image data. Based on this three-dimensional model, desired optical parameters can be determined. A corresponding procedure is also described. As an alternative described in this publication, a pattern projection can be used instead of the pair of image acquisition devices.

[0004] In the device described in EP 1 844 363 B2, the person being examined is positioned in front of the device, and the image acquisition devices then capture corresponding images of the head or parts thereof. Inaccuracies in the person's positioning can occur; that is, the person may be positioned differently from a desired target position, particularly a desired target distance. This can make it difficult to determine precise dimensions from the images captured by the image acquisition devices, which are needed to determine the parameters. Furthermore, the approach described in EP 1 844 363 B2 requires finding corresponding points or image areas within image pairs captured by the pair of image acquisition devices. Depending on the lighting conditions, it can be difficult to perform this for a sufficient number of image points.

[0005] Methods and devices are known from US Patent 2015 / 323310 A1 for determining pupillary distance and a scale using distance measurement. Determining parameters for spectacle fitting is not addressed in this document.

[0006] From the subsequently published DE 10 2005 001 874 A1, a method and a corresponding device for determining parameters for fitting spectacles are known, in which depth information regarding a user's head is acquired using a measuring device, wherein the measuring device determines the distance of at least one eye to an image acquisition device of the measuring device. A parameter of the usage position of spectacles or spectacle frames is then determined taking into account the determined distance.

[0007] From US 2010 / 0220285 A1, a method for determining parameters for spectacle fitting, in particular a pupillary distance, is known, in which a distance of a device used to a patient is measured and the pupillary distance is measured on the basis of the distance via a scale.

[0008] DE 689 28 825 T2 shows the reflection of a light source onto an optical axis of a camera.

[0009] A light field camera is disclosed in DE 10 2014 108 353 A1.

[0010] It is therefore an object of the present invention to provide methods and devices in which the accuracy of the determination of parameters for spectacle fitting is improved, particularly in the case of inaccurate positioning of a person to be examined as mentioned above.

[0011] For this purpose, a method according to claim 1 and a device according to claim 9 are provided. The dependent claims define further embodiments.

[0012] According to a first aspect, a method for determining parameters for spectacle fitting is provided, comprising: acquiring depth information regarding a user's head, wherein the depth information includes a distance between the user's head and a device used for acquiring it, and determining parameters for spectacle fitting based on the depth information. The depth information may also include multiple distances between the head and the device at different locations on the head, acquired simultaneously or at different times.

[0013] In addition, a 2D image (hereinafter also simply referred to as the image) of the head is captured, whereby the capture of the 2D image and the acquisition of the depth information take place via a common optical axis.

[0014] For the purposes of this application, "spectacle fitting" generally refers to the adjustment of spectacles to a specific person, particularly to that person's head. Such a fitting can begin, for example, with the selection of a specific type of spectacle frame, especially a particular brand. An expert, such as an optician, typically checks whether the frame fits the person's anatomy (e.g., frame size, bridge width, temple length). The fit of the frame is then checked (e.g., adjusting nose pads, temples to the facial geometry). Finally, various parameters are measured, such as pupillary distance, vertex distance, frame tilt, lens angle, and optical center height. This measurement of the aforementioned parameters is referred to as centration measurement or lens centration.Some of the measured parameters influence the lens power, such as the vertex distance. Other parameters determine how the lenses must be positioned or fitted into the frame, for example, the distance between the pupil centers when looking at infinity (pupillary distance) or the optical center height. Still other parameters can be used to calculate and manufacture lenses that are particularly well-tolerated and individually tailored to the wearer, such as the pantoscopic tilt of the frame and the lens angle.

[0015] Within the scope of the present invention, parameters for spectacle fitting are generally understood to mean information that is required or usable for the spectacle fitting described above. This includes, for example, dimensions relating to the person's head, particularly concerning the eye area, a type of spectacle frame and its dimensions, as well as the fit of the spectacle frame on the face.

[0016] The parameters for fitting spectacles can include, in particular, centration parameters, which can be used for the lens centration described above and which describe, for example, anatomical characteristics of the user (e.g., pupillary distance) and the position of the spectacles on the head. Examples of such centration parameters are described in DIN ISO 13666, edition 2013-10, namely, for example, monocular pupillary distance or pupillary distance. The monocular pupillary distance is the distance between the center of the pupil and the midline of the person's nose or the bridge of the spectacle frame when the eye is in its primary position. The primary position corresponds to the position of the eyes with the head and body held straight and the gaze directed straight ahead. The parameters for fitting spectacles can also include the dimensions of a spectacle frame.The parameters for fitting glasses can be determined based on a real frame worn by the person, or based on a virtual frame that is fitted to a model of the head created using depth information. In the case of using a virtual frame, the parameters for fitting glasses can also include the type of selected frame or parameters that describe the selected frame, such as its dimensions.

[0017] By determining a distance between the depth information acquisition device and the head of the person being examined, inaccurate positioning of the person being examined, i.e., positioning that deviates from a desired target position, can be compensated for when determining the parameters for spectacle fitting.

[0018] The method can further include determining a current head position (actual head position), particularly based on depth information, e.g., compared with a target head position. The head position can encompass the position of the head in space as well as its orientation (e.g., tilt, direction of gaze). The head position can, for example, include a lateral head rotation (i.e., around the person's body axis), a head position in the axial direction (relative to the measuring device), a lateral head tilt (i.e., towards a shoulder), or a head tilt along the body axis (i.e., forward or backward), wherein the head position can be determined relative to the device according to the invention, e.g., the depth information acquisition device. Head positions relative to the device according to the invention are relevant, for example, for the correct positioning of the person in front of the device according to the invention (e.g., within the measuring volume).Head tilt (along the body axis) has a significant influence on the measurement of the optical center height. Therefore, it is relevant for lens centration and can be used for subsequent correction of the optical center height. The parameters for spectacle fitting are determined based on the established actual head position. This allows, for example, deviations from the target head position and / or head position (e.g., in the axial direction) relative to a device being used to be taken into account when determining the parameters for spectacle fitting. As described above, the actual head position can include a lateral head tilt or a forward or backward tilt.This is particularly relevant when the person's natural, habitual head posture needs to be considered when determining the parameters for spectacle fitting, as the head tilt can influence the determination of certain parameters, such as the frame tilt or the optical center height. For this purpose, the person can be asked to assume their natural head posture, which is then recorded as the current head position. This recorded head position is then used to determine the spectacle fitting parameters. In this way, the spectacle fitting parameters can be determined to suit the natural head position.

[0019] For example, a rough 3D model of the head can be created from the depth information, revealing the user's head posture and position. This allows the user to receive feedback regarding head positioning; for instance, they can be instructed to reposition their head for a measurement if relevant parts of the head cannot be captured. Furthermore, it can be determined if the head posture differs from a previously established habitual posture or if the user's gaze direction is not the desired zero or main gaze direction for a measurement. This allows the user's head to be positioned as optimally as possible within the device's measurement range.

[0020] Determining the parameters for fitting spectacles can then be done based on the captured image of the head. The person can be wearing a frame without lenses, with support discs (simple plastic discs without optical effect), with lenses (lenses with corrective optics), or without a frame at all. Support discs are sometimes integrated into new frames sold in opticians' shops. In the latter case, the procedure can measure the three-dimensional topography of the face and determine anatomical parameters (e.g., pupillary distance) as the parameters that can then be used for fitting spectacles. By using three-dimensional topography, these parameters can be determined independently of head position.

[0021] By capturing the image, additional information besides depth information can be used to determine the parameters for fitting glasses, for example, dimensions of the person's head taken from the image.

[0022] The process can include scaling the captured image based on depth information and / or scaling image-based parameters for spectacle fitting based on depth information. Scaling the captured image based on depth information allows for a more precise determination of the spectacle fitting parameters, as the captured image can then accurately represent dimensions, and these dimensions can be extracted from the image.

[0023] The process can further include rectifying the captured image based on depth information. Rectification, in this context, means aligning and / or correcting the perceived distortion of the captured image, so that, for example, even with a tilted head position that leads to distortion in the image, correct parameters for eyeglass fitting can be extracted from the rectified image.

[0024] The acquisition of depth information and / or the capture of an image can be repeated multiple times, with the method further including a combination of the respective generated depth information and / or images, e.g., through temporal averaging. A suitable method for this is described, for example, in "Rapid Avatar Capture and Simulation using Commodity Depth Sensors"; A. Shapiro, A. Feng, R. Wang, Hao Li, M. Bolas, G. Medioni, E. Suma in Computer Animation and Virtual Worlds 2014, Proceedings of the 27th Conference on Computer Animation and Social Agents, 05 / 2014 - CASA 2014.

[0025] Such averaging can increase the accuracy of determining the parameters for spectacle fitting. However, multiple depth information points and / or multiple images can also be combined to increase accuracy in ways other than averaging. For example, this combining can involve temporal fitting of a function (or multiple functions; referred to as "fit"), such as fitting a polynomial function or other suitable function, or combined spatial and temporal fitting of the function, such as polynomial and spline functions. Temporal fitting involves adjusting the function across multiple depth information points or images acquired sequentially over time, while combined spatial and temporal fitting also includes spatial adjustment across the images or depth information (e.g., across different parts of the captured face).In this process, one or more corresponding functions are adapted to the depth information and / or images, for example by adjusting the function's coefficients (e.g., polynomial coefficients in a polynomial function), so that the function closely matches the depth information and / or images. Further processing steps can then be performed based on the function or functions.

[0026] Combining images or depth information, for example by averaging or adjusting a function, can be done using a rigid registration method (i.e., a method that uses only rotations and translations) to align different depth information and / or captured images. This aligns parts of the depth information or images relating to the same part of the head, such as measurement data. This is particularly necessary when the head moves between measurements. However, non-rigid methods (i.e., methods that also use other operations like distortion) can also be used for all or part of the depth information and / or images, for example...to take movements such as eyelid movements into account, for example the Dynamic Fusion method (described, for example, in “DynamicFusion: Reconstruction and tracking of non-rigid scenes in real-time”, Richard A. Newcombe, Dieter Fox, Steven M. Seitz; The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, pp. 343-352).

[0027] Non-rigid methods are preferably used only for areas of the face that are either not relevant for the subsequent determination of parameters or that can move quickly. The two types of registration methods mentioned above, used to align the various measurements (depth information and / or images), can be performed independently or in combination.

[0028] The process can further include discarding images and / or depth information that meet predetermined criteria. These criteria might include, for example, a head position unsuitable for measurement, such as one in which parts of interest like the eyes are not visible, or a closed eyelid. This allows images and / or depth information less suitable for determining the parameters for spectacle fitting (e.g., images with a closed eyelid) to be discarded.

[0029] The process can further include: creating a model of the head based on depth information, and virtually fitting glasses to the model, with the parameters for the glasses fitting being determined based on the virtual fitting. In such a case, it is preferred that the person is not wearing glasses when the depth information and, if applicable, the images are captured, so that a model of the head without glasses can be created more easily. Various eyeglass frames can then be virtually fitted to this model, i.e., also as models, for example, on a display. From the virtual fitting, the parameters for the glasses fitting can then be determined, e.g., the type of glasses to be used in reality, the dimensions of the glasses to be used in reality, and / or centration parameters for them.

[0030] The captured depth information, possibly combined with recorded images, can be displayed as a 3D model of the head on a screen. In other words, a 3D model of the head can be displayed in real time at a suitable refresh rate of the depth information capture device, effectively acting as a virtual mirror. This 3D model can then be combined with various virtual eyeglass frames to give the user a first impression of how different frames look. The user can then select one of the frames. The type of frame can also be used as a parameter for fitting the glasses. Furthermore, when fitting a virtual frame, it is possible to define parameters for fitting the glasses, which are then used to fit real glasses, and which describe the dimensions of the frame.This can be achieved, for example, by varying such parameters until an optimal fit is reached. Examples of such parameters include lens width, lens height, and bridge width of the spectacle frame. For instance, the bridge of the spectacles can be adjusted to the shape of the nose bridge in the 3D model.

[0031] Centering parameters can then be determined using such a selected virtual spectacle frame on the 3D model of the head. In other words, with embodiments of the present invention, centering parameters can be determined using real spectacle frames (which are worn by the person during the measurements) or with virtual spectacle frames (which are adapted to a 3D model of the head).

[0032] Furthermore, a computer program is provided containing program code which, when executed on a processor, causes the processor to execute one of the procedures described above and / or controls its execution. The computer program may be stored on a computer-readable medium.

[0033] According to a second aspect, a device for determining parameters for spectacle fitting is provided, comprising a depth information acquisition device for acquiring depth information regarding a user's head, wherein the depth information includes at least a distance of the head to the device, and an evaluation device which is configured to determine parameters for spectacle fitting based on the acquired depth information.

[0034] By determining a distance between the depth information acquisition device and the head of the person being examined, positions of the person being examined that deviate from a target position or preferred position can be compensated for when determining the parameters for spectacle fitting.

[0035] The depth information acquisition device can include a light field camera. Such a light field camera is also referred to as a plenoptic camera.

[0036] Light field cameras, similar in some ways to holographic imaging systems, capture the so-called light field of a scene. Unlike conventional cameras, they not only record intensity information but also additional information about the direction from which each light ray originates. The recorded signal (which, for example, would be captured by a conventional image sensor) thus contains both image information and depth information, derived from the intensity and direction of the light rays. This allows a captured object, such as a person's head, to be reconstructed three-dimensionally to a certain extent (depending on the specific implementation of the light field camera), and enables the measurement of distances between the head and the camera.With such a light field camera, the functionality of a depth information acquisition device and the functionality of a camera can be combined, for example. This allows for a correspondingly compact design.

[0037] In some embodiments of such light field cameras, a multi-microlens array is used, which is positioned in a defined plane in front of an image sensor. The individual lenses of the microlens array generate different image information on the image sensor. The light field can be reconstructed from the total image information on the image sensor.

[0038] The depth information acquisition device can be configured to capture a depth profile of a specific area of ​​interest on the head. A depth profile is defined as the distance of points within the area of ​​interest from a reference surface (e.g., defined by the depth information acquisition device) as a function of a position in a plane parallel to the reference surface, such as a position in a captured 2D image. Based on the depth profile, the evaluation unit can display a model of the area of ​​interest, essentially as a virtual mirror, which can, for example, enable the virtual fitting of eyeglasses. It is important to note that the user must either be wearing eyeglasses or not at the time the depth and image information is generated.For the virtual fitting of eyeglass frames, it is advantageous if the user does not wear eyeglass frames during the recording.

[0039] The depth information acquisition device can operate using infrared radiation. This means the measurement process does not disturb the user. For example, a camera-based infrared depth sensor can be used, in which an infrared pattern is generated by a projection system and the depth of objects in the scene is averaged by capturing the illuminated scene with an infrared camera, thus creating, for example, a depth profile of the head. Such depth sensors are commercially available. They allow for relatively high frame rates, such as 30 Hz or 60 Hz.

[0040] Instead of infrared patterns, patterns in the visible light spectrum can also be used. Such methods are also known as fringe projection. Various image patterns and / or image sequences are projected onto a surface (in this case, the head), and an image of the object is captured with a camera (which can be the aforementioned 2D camera or a different one). The camera is positioned at a defined angle to the projection system. Due to the three-dimensional surface of the object being measured, here the head, and the triangulation baseline (i.e., the distance between the projector and the camera), the patterns projected onto the camera's image sensor appear altered or deformed, from which the depth information and the topography of the illuminated surface can be determined. In particular, the distance of the head from the depth information acquisition device can also be determined in this way.However, the use of infrared light as described above is preferred, since in this case normal image capture is not disturbed and, for example, the user is not irritated by the light patterns.

[0041] The depth information acquisition device can include a time-of-flight sensor, which determines the distance over the time of flight of a signal, whereby the time of flight can be measured directly or in the form of a phase shift.

[0042] In such time-of-flight sensors, a signal is essentially sent from the depth information acquisition device to the head, and a signal reflected from the head is detected. The distance can then be determined from the time of flight of the signal to and from the head and the signal speed. This can be done, for example, using a scanning method at a large number of points. Instead of directly measuring the time of flight, a phase difference is often determined—especially when light pulses are used as the signal—between a modulation of the reflected beam and a corresponding modulation of the reference beam derived from the transmitted beam.

[0043] However, instead of light pulses, other types of signals, such as ultrasound signals, can also be used.

[0044] Time-of-flight sensors are particularly preferred, employing so-called time-of-flight cameras with laterally resolving depth sensors. Examples of suitable sensors for such time-of-flight cameras are photonic mixing device sensors (PMD sensors). These sensors use a modulated light signal, such as infrared light, to illuminate the head and detect the reflected light with the PMD sensor, which is also coupled to a modulation source. Here, the time of flight is measured indirectly via a phase shift.

[0045] Using appropriate time-of-flight sensors, objects such as the head can be scanned at a high frame rate, for example in the range of 30 to 60 Hz, and at high resolution, so that depth information can be provided at frame rates comparable to a video frame rate. This allows a complete depth profile of the head to be created.

[0046] These types of time-of-flight sensors make it possible to determine depth information with a high repetition rate and / or high accuracy. Robust depth information acquisition, largely independent of ambient lighting, etc., can also be achieved, since, for example, no corresponding points in stereo image pairs need to be found.

[0047] Another type of depth information acquisition device, which can be used in exemplary implementations, employs distance measurement via optical triangulation, for example, laser triangulation. The principle of optical triangulation is based on the generation of a light spot on the object to be measured (in this case, the head or a part thereof) using a laser, a light-emitting diode (LED), or another light source. This can be done in the visible or infrared range. The light spot is imaged by a camera, for example, a CCD camera (charge-coupled device), a CMOS camera, or a line scan camera. The light source and the camera are positioned at a defined angle to each other. Due to trigonometric relationships, the distance to the object to be measured can be determined from the displacement of the imaged light spot on the sensor, i.e., the position of the light spot on the sensor.In particular, the light spot in the image shifts in a direction from the light source to the camera as the distance to the object being measured increases, because due to the angle between the light source and the camera, the light also travels a greater distance towards the camera as the distance increases.

[0048] Such measurements can also be performed line by line. For this, a laser line is projected onto the head. Depth information along the laser line, particularly based on displacements perpendicular to its direction, can then be derived from a camera image of the laser line. This allows the topography of the head to be determined along the line. Using a scanning system in which the laser line moves across the head, the entire head, or a specific area of ​​interest (e.g., the eye area including glasses), can then be measured.

[0049] When using a laser light source, particular care must be taken to ensure that the eyes of the person being examined cannot be damaged by the laser. In particular, the laser intensity must be sufficiently low. Infrared light is preferable in this regard.

[0050] A camera for such optical distance measurement using triangulation can be separate from any additional 2D camera used for image acquisition. Alternatively, a single camera unit can be used, which, for example, when using visible light, can capture areas outside the aforementioned laser line or other light lines while simultaneously recording the laser line. Or, when using infrared light, it can be periodically switched between measuring depth information and capturing an image, for example, by means of a switchable filter. In other embodiments, no image is captured with a 2D camera at all; instead, a complete depth profile, and thus the topography of the head, is determined using a scanning system as described above, in which a laser line is moved over the head.

[0051] In another embodiment, depth information can also be determined using a stereo camera method via triangulation. In contrast to EP 1 844 363 B2 described at the beginning, here a pair of cameras is not only used to determine a 3D model, but a distance from the head to the stereo cameras is also explicitly determined. This is achieved using two cameras arranged at a predetermined angle to each other, or several cameras arranged at several predetermined angles to each other.

[0052] When using two cameras, for example in a stereo camera system, the depth of an object point is determined by calculating the parallax of that point in the two camera images, meaning a shift of object points between the camera images. The error of such a depth measurement is proportional to the square of the distance between the cameras and the head, and inversely proportional to the stereo baseline, i.e., the distance between the two cameras. Therefore, the stereo baseline must be sufficiently large to achieve adequate accuracy for typical operating distances and thus obtain the depth information. "Adequate accuracy" in this context means, in particular, that the parameters for fitting the glasses can be determined with the desired or required precision.

[0053] Since this approach requires identifying corresponding points or areas in the two images, the area of ​​the head under investigation must have sufficient structure (especially lines or edges) to allow such a correspondence to be identified using the appropriate method. Objects for which this typically works well include the eyes, eyebrows, nose, and other prominent facial features. In some implementations, a structure can also be projected onto the head, such as binary patterns, lines, dots, etc., to facilitate the identification of corresponding areas.

[0054] When implementing the depth information acquisition device, for example as a time-of-flight sensor or a light field camera, a more compact design is possible than with a stereo camera system, since no angle between two cameras is required. Therefore, time-of-flight sensors, light field cameras, and similar devices are preferred over stereo camera systems as depth information acquisition devices in cases where a compact design is crucial.

[0055] The present invention is not limited to the types of depth information acquisition devices described above. Depth information acquisition devices with other depth sensors can also be used, for example, as already mentioned, ultrasonic depth sensors, depth sensors that perform depth measurement using optical coherence tomography (OCT), confocal sensors, or chromatic confocal sensors. In general, any conventional depth sensor can be used that can acquire depth information with respect to the head, particularly with respect to the eye area and eyeglasses, with sufficient accuracy and without posing a danger to the user's eyes (for example, through excessive radiation). "Sufficiently accurate" means that the parameters for eyeglass fitting can ultimately be determined with the desired or required accuracy for eyeglass fitting.

[0056] In addition to the depth information acquisition device, the device also includes a 2D camera, i.e., a (conventional) camera for capturing two-dimensional images using an image sensor, for capturing an image of at least part of the head. This provides additional information for determining the parameters for fitting the glasses, for example, dimensions of the person's head taken from the image.

[0057] The evaluation unit can be set up to scale the images based on the depth information and / or to scale certain parameters based on the images for spectacle fitting based on the depth information, as already explained above.

[0058] The evaluation unit can also be set up to rectify the images based on the depth information, as explained above.

[0059] The device is configured such that the depth information acquisition device and the 2D camera detect the head via a common optical axis. For the depth information acquisition device, this optical axis corresponds to the axis along which the depth information is acquired, essentially a "viewing direction" of the device. If the depth information acquisition device uses an imaging optic, the optical axis corresponds to the optical axis of that imaging optic, typically a straight line connecting all the centers of curvature of the refracting or reflecting surfaces of the imaging optic. Light rays on the optical axis pass through the imaging optic without deflection. For the 2D camera, the optical axis corresponds to the optical axis of the 2D camera's lens, as explained above for an imaging optic.

[0060] In some embodiments, a beam splitter can be used to combine and separate the optical axes of the depth information acquisition device and the 2D camera. The beam splitter can be wavelength-selective. In such embodiments, for example, the 2D camera can capture visible light, while the depth information acquisition device operates based on infrared radiation. In other embodiments, the beam splitter is not wavelength-selective, and the 2D camera and the depth information acquisition device use at least part of the same spectrum. Such an arrangement allows for a more compact device than a stereo camera or similar, since no corresponding angle needs to be provided between image acquisition devices, for example.Furthermore, depth information and image information from the camera are captured from the same direction, thus eliminating the need for perspective correction of the depth information relative to the image data.

[0061] The evaluation device can be set up to determine the head position based on the depth information, whereby the parameters for spectacle fitting are determined based on the determined head position.

[0062] The device can be configured to repeatedly acquire depth information using the depth information acquisition device and / or capture images with the 2D camera, with the evaluation device then being configured to combine multiple acquired depth information and / or multiple captured images, e.g. in the form of averaging as described above.

[0063] The evaluation unit can also be configured to discard images and / or depth information that meet predetermined criteria.

[0064] The evaluation unit can be further configured to display a model of the head based on the depth information, e.g. on a display, and to enable a virtual fitting of glasses to the model, whereby the evaluation unit is configured to determine the parameters for the glasses fitting based on the virtual fitting.

[0065] The present invention will be explained in more detail below with reference to various exemplary embodiments and the accompanying drawings. These show: Fig. 1 a schematic representation of a device according to an exemplary embodiment, Fig. 2 a schematic representation of a device according to a further embodiment, Fig. 3 a schematic representation of a device according to a further embodiment, Fig. 4 A schematic representation to illustrate depth sensors that can be used in the exemplary embodiments, Fig. 5 a representation of a section through a depth profile, as can be generated in some embodiments, Fig. 6A-6F Illustrations to explain parameters for spectacle fitting, Fig. 7 a flowchart to illustrate a process according to an exemplary embodiment, Fig. 8 a flowchart to illustrate a method according to a further embodiment, and Fig. 9 and Fig. 10 illustrations to explain a rectification as it is carried out in some embodiments.

[0066] In the following, various embodiments of the present invention are explained in detail.

[0067] In Fig. Figure 1 is a device for determining parameters for fitting eyeglasses (i.e., parameters that can be used to fit eyeglasses to a person's head) according to an exemplary embodiment, shown schematically. The device of Fig. 1 comprises a depth information acquisition device 12 for determining depth information regarding a user's head 10, in particular for determining depth information regarding an eye area, that is, an area around the user's eyes. The user in Fig. 1 wears glasses 11, for which depth information can also be determined.

[0068] Depth information within the meaning of the present application comprises at least information regarding the distance between the user, in particular the user's head 10 and / or the glasses 11 worn by the user, and the depth information detection device 12. For example, the distance can be the length of the dashed line 14 of the Fig. Distances can be determined. However, distances can also be measured to other and / or multiple points on the head 10. In a preferred embodiment, a depth map is created, for example, by determining the distance described above for a plurality of points on the head 10, particularly the eye area, and / or a plurality of points on the glasses 11. The depth information provided by the depth information acquisition device 12 thus provides information about such a distance. This information can then be used to determine parameters for adjusting the glasses 11 to the user's head 10, for example, for centering lenses, regardless of the exact position of the head 10. Examples of suitable depth information acquisition devices and parameters to be determined will be explained in more detail later.

[0069] The device of Fig. The device 1 further comprises an evaluation unit 13, which receives the depth information from the depth information acquisition unit 12 and determines parameters for adjusting the glasses 11 based on this depth information. The evaluation unit 13 can be implemented, for example, as a suitably programmed computing device, such as a computer. However, hardwired hardware components such as application-specific integrated circuits (ASICs) can also be used. The evaluation unit 13 can include output devices such as a display, speakers, interfaces for outputting signals, and the like, in a conventional manner, to output the determined parameters for adjusting the glasses or to transmit them to other devices. Details of the determination of the parameters for adjusting the glasses will also be explained in more detail later.

[0070] It should be noted that the depth information acquisition device 12 and the evaluation device 13 of the device according to the invention can be arranged close together, for example in a common housing or in separate housings that are fixed in a spatial arrangement relative to each other. However, the evaluation device 13 can also be arranged spatially separate from the depth information acquisition device 12, and the depth information determined by the depth information acquisition device 12 can be transmitted to the evaluation device 13 in the conventional manner, wirelessly, via a wired connection, or via optical lines such as fiber optics. Such transmission is also possible, for example, via networks such as the Internet, so that essentially any distance between the depth information acquisition device 12 and the evaluation device 13 is possible.

[0071] Before various details of depth information acquisition devices, parameters for spectacle fitting, and their determination are explained in more detail, the following will now be discussed with reference to the Fig. 2 and Fig. 3 Variations and extensions of the exemplary embodiment of the Fig. 1 discussed. To avoid repetition, identical or corresponding elements in different figures in the following description bear the same reference symbols and are not explained in detail multiple times.

[0072] In the device of the Fig. 2 is the device of the Fig. 1 is extended by a camera 20, which captures a two-dimensional image, for example a black and white image or a color image, of the head 10 or a part thereof, for example the eye area. The camera 20 can be implemented in a conventional manner with a lens and an image sensor.

[0073] The image thus captured is also fed to the evaluation unit 13. In this case, the evaluation unit 13 additionally determines the parameters for adjusting the spectacles based on the captured image. The evaluation unit 13 can also control the camera 20 and the depth information acquisition unit 12 so that the image is captured simultaneously with the depth information.

[0074] For evaluation purposes, in one variant, for example, the image captured by camera 20 can be scaled based on the depth information acquired by the depth information acquisition device 12. Such scaling can be performed with a higher scaling factor if the depth information indicates that the head 10 is farther away from the depth information acquisition device 12, and with a lower scaling factor if the head is closer to the depth information acquisition device 12. The scaling factor can indicate either an increase or a decrease in size.

[0075] In this way, the image can be scaled in such a way that the dimensions of the head 10 can be extracted from the image with sufficient accuracy to correspond to the parameters to be determined for spectacle fitting, or on the basis of which such parameters for spectacle fitting can be determined. "Sufficiently accurate" means that the parameters for spectacle fitting can ultimately be determined with the desired or required accuracy for the spectacle fitting. Instead of scaling the image, another approach involves extracting corresponding dimensions from the captured image, and these extracted dimensions can then be scaled based on the depth information.

[0076] In Fig. Figure 2 shows the camera 20 separately from the depth information acquisition device 12. However, in some embodiments, the camera 20 can also serve simultaneously as part of the depth information acquisition device 12, for example in the case of fringe projection or laser triangulation, as will be explained in more detail later.

[0077] In the exemplary embodiment of the Fig. 2. The camera 20 and the depth information detection device 12 operate as shown at different angles and with different optical axes. In other embodiments, a setup is provided in which the depth information detection device 12 and the camera 20 view the head 10 coaxially. A corresponding embodiment is shown in the Fig. 3 shown.

[0078] In Fig. 3 are as in the exemplary embodiment of the Fig. 3 a depth information acquisition device 12 and a camera 20 are provided. In the embodiment of the Fig. 3. Additionally, a beam splitter 30 is provided, which combines the optical axis of the camera 20 with the optical axis 12 of the depth information acquisition device 12, so that the head 10 is viewed or measured on a single optical axis. This also enables a compact design. Furthermore, parallax errors and the like between the depth information and the image captured by the camera 20 are avoided or reduced. In the exemplary embodiment of the Fig. 2. However, such parallax errors and the like can be eliminated computationally by the evaluation unit 13.

[0079] The beam splitter 30 can be a wavelength-selective beam splitter. In such embodiments, for example, the camera 20 can capture visible light, while the depth information acquisition device 12 operates on the basis of infrared radiation. Alternatively, the beam splitter 30 is not wavelength-selective, and the camera 20 and the depth information acquisition device 12 use at least partially the same part of the spectrum.

[0080] Even if it is in Fig. 3 is not explicitly shown, the depth information provided by the depth information acquisition device 12 and one or more images provided by the camera 20 can also be processed by an evaluation unit such as the one for the evaluation unit 13 of the Fig. 1. The evaluation process is explained in order to obtain parameters for spectacle fitting.

[0081] Next, different types of depth information acquisition devices that can be used to obtain depth information will be explained in more detail.

[0082] For example, time-of-flight or phase-of-flight measurements can be used to obtain depth information. In such methods, as explained above, a signal is essentially sent from the depth information acquisition device to the head, and a signal reflected from the head is detected. With appropriate time-of-flight sensors, objects such as the head 10 can be scanned at a high frame rate, for example in the range of 30 to 60 Hz, and at high resolution, so that the depth information can be provided at frame rates comparable to a video frame rate. A complete depth profile of the head can then be created. Based on such depth profiles, as in the exemplary embodiment of the Fig. One parameter for adjusting glasses can also be determined without the use of an additional camera such as camera 20.

[0083] The Fig. Figure 4 schematically shows a depth information acquisition device, which can serve as a schematic representation for various usable types of depth information acquisition devices. In the case of a time-of-flight sensor, 40 denotes a signal source, for example a modulated light source, and 41 denotes a corresponding sensor for detecting the received light, for example a time-of-flight camera as described above.

[0084] As already explained, another type of depth information acquisition device, which can be used in exemplary implementations, employs distance measurement via optical triangulation, for example, laser triangulation. In the case of the Fig. In such a triangulation device, 4 can, for example, denote 40 a light source and 41 a camera, which are positioned at a defined angle to each other.

[0085] Another possibility for implementing the depth information acquisition device 12 is the use of a light field camera.

[0086] As described, another option for implementing a depth information system is a camera-based infrared depth sensor, in which an infrared pattern is generated by a projection system and the depth of objects in the scene is determined by capturing the illuminated scene with an infrared camera. As also described, a fringe projection using visible light can also be used.

[0087] In the case of the Fig. For example, element 40 would be the projection system, and element 41 the camera, which may be identical to or different from camera 20.

[0088] In another embodiment, the depth information can also be determined using a stereo camera method via triangulation, as described. In this method, two or more cameras are arranged at one or more predetermined angles. In the case of the Fig. In such a depth information acquisition device, for example, 40 is a first camera and 41 is a second camera.

[0089] By using depth sensors such as those described above, a profile of the head 10, and thus the three-dimensional surface of the head 10, can be modeled based solely on the depth information acquisition device, provided that appropriate high-resolution depth sensors are used. As an example, [reference to relevant figure] Fig. Figure 5 shows a schematic representation of a section 50 through such a 3D model, in which a section 51 through the glasses is also visible. The section 50 in the example of the Fig. Line 5 passes through the user's eye (52), meaning it does not run through the center of the subject's face, but rather through the center of the eye. Line 10 represents the profile of the head as shown in the Fig. 1-4 is shown, with this profile passing through the central axis of the head and thus differing from profile 50.

[0090] Based on the depth information, which includes the distance between the depth information acquisition device and the head 10, a true-to-scale three-dimensional profile of the head 10 can be created, with corresponding sections 50 at various points. From this, parameters for spectacle fitting, such as centration parameters, can be determined. This will be explained in more detail below.

[0091] First, with reference to the Fig. Sections 6A to 6F explain the various parameters that can be determined for fitting spectacles, in particular centration parameters. These parameters are shown below. Fig. 6A to 6F each show views of a pair of glasses, possibly together with a partial view of a head to illustrate various parameters.

[0092] The Fig. Figure 6A shows the monocular pupillary distance when looking at infinity. Arrow 60A indicates the monocular pupillary distance for the left eye, measured as the distance between the pupil and a central axis 61 of the head. Arrow 60B indicates the monocular pupillary distance for the right eye. The values ​​for the left and right eyes are different in most cases.

[0093] Fig. Figure 6B illustrates the optical center height, measured again while looking at infinity, fulfilling the eye rotation center requirement. The eye rotation center requirement means that the optical axis of the lens should pass through the eye's center of rotation. The optical axis runs through the optical center of the lens and is generally perpendicular to the lens. This can minimize unwanted prismatic effects when eye movements cause viewing through different parts of the lens. The optical center height indicates the distance between the pupil and a lower edge of the lens. Arrow 62A shows the optical center height for the right eye, and arrow 62B shows the optical center height for the left eye.

[0094] In Fig. 6C, where 63 is the vertex distance, is usually measured from the back of the spectacle lens to a vertex plane of the cornea. Fig. 6D denotes an angle of 64°, the forward tilt of the frame, essentially an inclination of the glasses towards the vertical. This depends, like the one in Fig. 6B shows that the viewing height also depends on the person's head position.

[0095] In Fig. 6E denotes an angle of 65°, the lens angle, an angle at which the spectacle lens is positioned compared to a "flat" spectacle. Fig. Finally, section 6F lists various dimensions of the spectacle frame itself. 67 indicates the lens width, and 68 the lens height. Lens width and height, along with pupillary distance and optical center height, are important pieces of information for determining the required lens diameter. 66 indicates the bridge width of the spectacle frame.

[0096] The Fig. Figures 6A to 6F show some parameters for spectacle fitting that can be determined using the devices shown. Further parameters for spectacle fitting can also be determined. For example, there may be additional parameters that can be determined for centration of lenses for near vision (such as reading glasses or work glasses) and for centration of progressive lenses. These include, for example, the "near pupillary distance," which is not measured when looking towards infinity as described above, but rather when looking at an object close to the head.

[0097] In addition to the in Fig. In addition to the parameters shown in section 6F for spectacle fitting, other geometric information about the spectacle frame can also be determined. This geometric information can be important for selecting suitable lenses, as it influences the lens thickness, for example, in lenses with a positive optical effect.

[0098] At least when centration single-vision lenses, the eye rotation center requirement is generally met. When centration vertically, that is, aligning the optical axis of the lens vertically, head posture must be taken into account, as the head tilt influences the measurement of the optical center height and also the frame's tilt. To account for this, for example, when using the device under discussion, the person's head posture can be adjusted so that the plane of the spectacle frame is perpendicular to the ground, which is also known as the zero gaze direction. In this state, the position of the pupils in the plane of the lens is then determined using the device. It is also possible to determine the optical center height with a natural head and body posture, whereby the person being examined usually looks down at a distance of 8 to 10 meters, which is also known as the primary gaze direction.As will be explained in more detail below, the head position can also be determined using the device according to the invention, and the person being examined can, if necessary, be instructed to change their head position.

[0099] The following describes the approaches for determining centering parameters based on depth information and, if necessary, based on a camera image (for example, taken with camera 20 of the preceding embodiments) with reference to the Fig. 7, Fig. 8, Fig. 9 to Fig. 10 explained in more detail.

[0100] The Fig. Figure 7 shows a flowchart that provides an overview of a procedure according to an exemplary embodiment for determining parameters for fitting spectacle glasses. Fig. Figure 8 then shows a detailed flowchart illustrating various possible details of such a procedure. While the procedures are depicted as a sequence of steps, the order shown is not to be interpreted as restrictive. In particular, some steps can also be performed in a different order, or one or more steps can be performed in parallel. For example, depth information and an image can be acquired simultaneously using a depth information acquisition device such as the depth information acquisition device 12 of the Fig. 2 and a camera like the camera 20 of the Fig. 2 or can be recorded with a light field camera as explained above.

[0101] In one step 70 in Fig. 7 Depth information is acquired, which includes the distance of a head to a device used for acquiring the depth information. This can be done with depth information acquisition devices 12 as discussed above.

[0102] In step 71, an image (also called an overview image) of at least one area of ​​interest, for example, the eye area or a head, is captured. Step 71 is optional, and in some embodiments, only the depth information is captured in step 70. In step 72, parameters used for spectacle fitting, in particular lens centration, are then determined. For example, one or more of the parameters referenced in the Fig. The parameters discussed in sections 6A to 6F will be determined.

[0103] In some implementations, a true-to-scale 3D model of the head with glasses is created from the captured depth information, and the parameters are then read from this 3D model. Additionally, if an overview image has been captured, it can be registered with the 3D model and used, for example, to add texture to the 3D model. This facilitates the recognition of individual features such as the eyes. Approaches to recognizing such features as the eyes, the eyeglass frames, and the like, which are needed to determine the parameters, are known in themselves, for example, from conventional photography, where various facial recognition algorithms are used.The distance between the head and the depth information acquisition device can be used to ensure a correct size of the 3D model even if the head is positioned differently than a specified target position.

[0104] In other embodiments, the depth information may also comprise only a single distance between the head and the device at a single point, and an overview image is additionally captured. Such depth information can be provided with a relatively simple depth sensor, which, for example, does not need to be a scanning sensor. In this case, the distance thus determined is used to scale the image with a scaling factor that depends on the distance and is essentially derived directly from the intercept theorem. In this way, the captured image can be scaled to the correct size, and dimensions such as those in the Fig. The dimensions shown in sections 6A to 6F can be taken from the scaled overview image. Depending on the shooting direction, only some parameters need to be determined for a single image. For example, parameters of the following can be determined from a frontal shot: Fig. 6A, Fig. 6B and Fig. 6F are determined, while for the parameters of the Fig. 6C and Fig. In 6D, an image must be taken from the side. Depending on the parameters to be determined, several images may be taken in such a case.

[0105] In some implementations, depth information compensates for the fact that the size of the subject's head, or relevant parts thereof, changes depending on the distance between the head and the camera. For example, the head appears larger when it is closer to the camera and smaller when it is farther away.

[0106] Now, with reference to the Fig. 8 different details and extensions of the procedure Fig. 7 discussed. While the Fig. 8 and the following description explain various possible extensions and details in combination, it should be noted that these can also be implemented independently of each other, i.e., they do not necessarily have to be implemented in combination with each other.

[0107] In step 80, depth information is repeatedly acquired, and optionally, image information is repeatedly acquired in step 81. Thus, steps 70 and 71 of the process are repeated. Fig. 7. Repeatedly. As already explained, some depth sensors offer high refresh rates in the range of 30 Hz or 60 Hz, i.e., in the range of typical video rates, and cameras can also record image information at such rates. However, lower refresh rates are also possible. These measurements at 80 and 81 can be analyzed in real time with appropriately equipped computers, at least with regard to some aspects.

[0108] As previously explained, a model of the head can be created from repeatedly captured depth information and displayed on a screen, enabling, for example, the fitting of a virtual eyeglass frame as described. The depth information can also be used, as described, for head alignment.

[0109] In step 82, suitable measurements can then be selected from those taken in steps 80 and 81. This allows unfavorable images or depth information to be excluded from the subsequent determination of the parameters for spectacle fitting. For example, the position of the pupil, which is important for the [missing information], can be determined. Fig. 6A and Fig. The parameter shown in Figure 6B can only be determined imprecisely when the eyelid is closed. Therefore, facial feature detection (also referred to as landmark detection in English-language literature) can be used to determine in which images or depth information one or both eyelids are closed, and the corresponding data can be discarded for subsequent processing steps. It should be noted that the acquisition of depth information and the recording of images at 80 and 81 can also be synchronized, so that, for example, each image is assigned a set of depth information. If, for instance, it is determined that the eyelid is closed in the image, the corresponding depth information can also be discarded immediately.Conversely, if, for example, the depth information reveals that the head was turned so far that a reasonable determination of desired parameters for spectacle fitting is difficult or impossible, the corresponding image can also be discarded.

[0110] Next, in step 83, a combination of multiple measurements and / or a registration of data relative to each other can be performed. Registration generally refers to a process in which several data sets are superimposed or aligned with one another. In so-called rigid registration methods, the topography of the individual measurements is not changed; that is, distances and angles are preserved. Instead, the individual measurements are only superimposed through rotation and translation.

[0111] In this way, several consecutive measurements (in steps 80 and 81 of the Fig. 8) can be combined after registration, which can improve accuracy, particularly with regard to depth information. The registration parameters can be determined using conventional optimization methods, such as the ICP (Iterative Closest Point) algorithm for textureless data, homographic estimation, or, for example, a DLT algorithm for image data (direct linear transformation; for example, for image data captured with a camera).

[0112] However, such rigid registration methods for combining facial features can be problematic for changing facial areas (e.g., moving eyelids), as this can lead to unfavorable registration, particularly for display purposes on a screen or the like. Therefore, in preferred embodiments, non-rigid methods are used for such image areas or areas of depth information where rapid changes can typically occur (especially eyelids, eye movements, and the like).

[0113] In step 84, the head position is determined during the measurements of steps 80 and 81. This head position can then be used to rectify captured images or, for example, to correct them in other ways when determining the parameters for spectacle fitting.

[0114] In one embodiment, the head of the person being recorded is registered using a standard head (i.e., a 3D model of a head of a defined standard shape) using registration methods such as those described above in section 83. The posture of this standard head is determined relative to a reference coordinate system of the device. This reference coordinate system is also defined as the world coordinate system. For this purpose, in one embodiment, the rigid transformation is determined that transforms the standard head, given in its own reference system, into a standard head in the world coordinate system, so that the best possible agreement with the measurement data of the depth information acquisition device is achieved.This transformation can be described, for example, using the following six parameters: translation in the x, y and z directions, and rotation about the x, y and z axes (for example, according to Euler's formula), where x, y and z describe axes of a world coordinate system.

[0115] For example, in Fig. Figure 9 shows a head 10 rotated by an angle of 90° around the z-axis (perpendicular to the image plane). 91 denotes the central axis of a straight head, and 92 the central axis of head 10. Here, the angle 90° would be determined by recording. Translations or rotations of head 10 around the x- and / or y-axis can be determined in a similar manner.

[0116] The head position described by these six parameters is then used in step 85 below to correct any tilt of the gaze towards the device. In this way, for example, a visual acuity point—that is, the intersection of the person's visual axis with a plane of the spectacle lens—can be corrected.

[0117] One way to apply a correction in advance is to rectify the camera image (for example, the images taken in step 81) in step 84 to convert the camera images to an image plane corresponding to the orientation of the face.

[0118] This is done with reference to the Fig. 10 explained. Fig. 10 is again the head 10 in the position of the Fig. Figure 9 is shown. 100 represents the optical center of a camera used (for example, camera 20 in the embodiments described above). 101 denotes an image plane of the central projection from center 100, for example, an image plane from which the image is captured by camera 20. Due to the tilt of the head 10, distortions occur, for example, between the left and right sides of the face. During rectification, the points of image plane 101 are essentially converted into an image plane 102 according to the drawn light rays, which corrects the distortion. The resulting corrected image is adapted to the orientation of the face, since image plane 102 corresponds to the orientation better than image plane 101. Parameters for adjusting the glasses can then be derived from such corrected (and, if necessary, scaled as discussed above) images.

[0119] In step 85, as mentioned above, an analysis of the depth information and / or images pre-processed in steps 82 to 84 is performed to determine the parameters for spectacle fitting. As previously explained, this can be done primarily using algorithms for detecting facial features or other characteristics (such as the spectacle frame). The corresponding spectacle fitting parameters can then be determined by measuring distances or angles within the data, for example, in a 3D model created based on the depth information and / or in image data scaled and / or rectified based on the depth information. In this way, the spectacle fitting parameters, especially the centration parameters, can be easily determined.

[0120] As already explained, in some implementation examples only some of the features described in Fig.The steps described in section 8 can be carried out. For example, steps 83 and 84 can also be performed independently, and can also be performed if only one depth information measurement is taken, or if only one depth information measurement and one image are captured. In other words, multiple image capture and averaging can also be omitted. Reference symbol list 10 heads 11 Glasses 12. In-depth information acquisition device 13 Evaluation unit 20 cameras 30 beam splitters 40 components 41 Component 50 Depth profile 51 Glasses 52nd eye 10 Head profile 60A,60B Pupil spacing 61 Center line 62A, 62B Viewing height 63 vertex distance 64 Forward tilt 65° mounting disc angle 66 bridge width 67 disc width 68 disc height 70-72; 80-85 Procedural steps 90 angle 91 Center axis 92 Center axis 100 Optical Camera Center 101 Image plane 102 Rectified image plane

Claims

[1] Method for determining parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68), comprising: Acquiring (70; 80) depth information regarding a user's head (10), wherein the depth information comprises a distance (14) between the user's head (10) and a device used for acquiring, and Determining (72; 85) parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68) based on depth information, Taking (71; 81) a 2D image of the head, wherein the acquisition (71; 81) of the 2D image and the acquisition (70; 80) of the depth information are carried out via a common optical axis. [2] Method according to claim 1, wherein the method further comprises determining a head position of the head (10), wherein the parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68) are determined on the basis of the determined head position. [3] Method according to claim 1 or 2, wherein the determination of the parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68) is based on the recorded 2D image. [4] Method according to claim 3, comprising scaling the 2D image based on the depth information and / or scaling parameters determined on the basis of the 2D image for spectacle fitting (60A, 60B, 62A, 62B, 63-68) based on the depth information. [5] Method according to claim 3 or 4, further comprising rectifying (84) the 2D image based on the depth information. [6] Method according to any one of claims 1 to 5, wherein the acquisition of the depth information (80) and / or the acquisition of the 2D image (81) is repeated several times, the method further comprising averaging over several acquired depth information and / or over several acquired images. [7] Method according to claim 6, further comprising a rejection (82) of 2D images and / or depth information which meet predetermined criteria. [8] Method according to any one of claims 1 to 7, further comprising: Representing a model of the head (10) based on depth information, and Virtual fitting of glasses to the model, where the parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68) are determined based on virtual fitting. [9] Device for determining parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68), comprising: a depth information acquisition device (12; 40, 41) for acquiring depth information regarding a user's head (10), wherein the depth information includes at least a distance (14) of the head (10) from the device, an evaluation unit (13) which is set up to determine the parameters for spectacle fitting (60A, 60B, 62A, 62B, 63-68) based on the acquired depth information, and a 2D camera (20) for taking an image of at least part of the head (10), wherein the device is arranged such that the depth information acquisition device (12) and the 2D camera (20) detect the head (10) via a common optical axis. [10] Device according to claim 9, wherein the depth information acquisition device (12) comprises a light field camera. [11] Device according to one of claims 9 or 10, wherein the depth information detection device (12; 40, 41) operates on the basis of infrared radiation. [12] Device according to one of claims 9 to 11, wherein the depth information acquisition device (12; 40, 41) is configured to acquire a depth profile of an area of ​​interest of the head (10), and wherein the evaluation device (13) is configured to display a three-dimensional model of the area of ​​interest. [13] Device according to one of claims 9 to 12, wherein the depth information acquisition device (12; 40, 41) comprises a device based on time-of-flight measurements and / or phase measurements and / or triangulation and / or pattern projection and / or stereo image acquisition.

Images