AUTOSTEREOSCOPE DISPLAY WITH P VIEWS FOR DISPLAYING AN AUTOSTEREOSCOPE PICTURE WITH I VIEWS AND DISPLAY DEVICE WITH SUCH A SCREEN

Description

1. Technical field of the invention

[0001] The invention relates to an autostereoscopic display screen for displaying an autostereoscopic image and a display device comprising such a display screen. The invention also relates to a method for displaying an autostereoscopic image on such a display screen. 2. Technological background

[0002] Autostereoscopy is a technique that allows the display of three-dimensional images without requiring the viewer to wear special glasses. This technique is known per se, in particular from patent documents WO2006 / 024764, WO2014 / 041504, WO2013 / 140363, WO2014 / 016768 in the name of the applicant.

[0003] DE 10 2004 059729 B3 describes a method for generating an autostereoscopic image from two images dedicated respectively to the right eye and the left eye of an observer. 2

[0004] WO 2008 / 011888 A1 describes a method for autostereoscopically displaying images on a screen in which image information of separate channels for a right eye and a left eye of a user is displayed in interlaced segments on the screen, and in which a lens array is arranged to deflect light emitted from the screen such that the image information of each channel is visible only to one eye of the user. 4

[0005] WO 2008 / 003373 A1 describes a method for displaying autostereoscopic images implementing steps of determining the intensity barycenters of the groups of sub-pixels as a function of the position of the observer, and of detecting a change in the position of the observer's head relative to the matrix screen and of modifying the display as a function of the detection of the change in position. 1

[0006] LIU JING ET AL: "Dynamic mapping for multiview autostereoscopic displays" (PROCEEDINGS OF SPIE, IEEE, US, vol. 9391, March 16, 2015) describes a method for displaying an autostereoscopic image.

[0007] Generally speaking, an autostereoscopic image consists of a plurality of nested elementary images, each corresponding to views of the same object or the same scene from different viewpoints. A selector device, typically consisting of a network of cylindrical lenticules or a parallax barrier, is arranged in front of the display screen so as to allow the projection of a pair of elementary images corresponding to two different viewpoints of the scene, respectively towards the two eyes of the observer, which creates an impression of relief in the brain of the observer.

[0008] A screen for displaying an autostereoscopic image with P viewpoints typically comprises a matrix of pixels arranged in rows and columns, each pixel being composed of a plurality of sub-pixels of different colors. The screen is further surmounted by a selector device (also referred to throughout the text as an optical device), such as an array of identical cylindrical lenticules, each having a focal length configured to be able to reflect to infinity the light rays coming from the screen. The pitch of the array of cylindrical lenticules is precisely calculated so that the observer sees, at a predetermined distance from the screen (called the flat tint distance), the images succeeding one another at regular intervals, thanks to the magnifying effect of the array of lenticules.

[0009] This magnifying glass effect results from the fact that a lens placed at the right distance (its focal length) magnifies the sub-pixel that is in alignment with its optical center and the pupil of the observer's eye. If the lens magnifies P times, the sub-pixel seen through the lens is perceived P times larger than it actually is and masks from the eye that receives the light through this lens the P-1 other sub-pixels that are not in the alignment described above.

[0010] The cylindrical lenticule network therefore forms a viewpoint selector device configured to allow the projection of a pair of differentiated viewpoints in the observation space of the screen, corresponding to two different viewpoints of the scene, respectively towards the two eyes of the observer.

[0011] When certain elements of the two images observed respectively by the left eye and the right eye of an observer do not present any disparity (defined as the difference, measured in pixels or sub-pixels, observable and / or measured between the left and right images), these elements overlap perfectly and are then perceived on the plane of the screen itself. On the other hand, when there is a disparity between the observed elements, the latter jump out of the screen or on the contrary are perceived in depth. By convention, depth corresponds to a positive disparity while jump out corresponds to a negative disparity. The shift in absolute value between the two images is the same in both cases but when the disparity is negative, the jump out elements of the left image are to the right of those of the right image and vice versa. It is this shift which produces a jump out effect (the two eyes converge in front of the screen).On the other hand, when the disparity is positive, the depth elements of the left image are to the left of those of the right image and vice versa (the two eyes converge behind the screen).

[0012] An object the size of a point is perceived further from the screen plane the greater the offset on the horizontal plane of this point between the left and right image.

[0013] Current autostereoscopic displays are limited in the depth and projection achievable without loss of quality. These displays allow the projected content to be placed in a viewing cube whose depth is close to the diagonal of the screen. The displays are generally used so that one-third of the viewing cube projects from the screen and the remaining two-thirds are perceived in depth, behind the screen plane.

[0014] For example, on a 24" screen, this means the standard viewing cube can have a total throw of about 60 cm, with a comfortable projection 20 cm in front of the screen and backgrounds 40 cm behind the screen.

[0015] These depth and burst limits are explained by physiological constraints of the observer, by technological constraints and by image processing constraints.

[0016] Physiological constraints result from the fact that each eye of the observer sees a different image in order to form a stereoscopic pair which it merges by converging its gaze where the disparity seems to place the objects in space.

[0017] The disparity between the left and right images induces three different convergence modes: when the observer perceives an object in depth behind the screen, the left eye looks more to the left and the right eye more to the right. Thus, the two optical axes intersect behind the screen, where the observed virtual object is located, when the observer perceives an object on the plane of the screen itself, the two eyes look at the same place on the screen, when the observer perceives an object bursting out in front of the screen, the left eye looks more to the right and the right eye more to the left. Thus, the two optical axes intersect in front of the screen, where the observed virtual object is located.

[0018] When objects are perceived in depth or in projection, the gaze converges at distances different from that of the screen plane. This necessarily leads to a dissociation between the different convergences described previously and accommodation, which must correspond to the effort of each eye to see the images clearly at the distance of the screen.

[0019] The further an object is from the screen plane, the greater the effort required to dissociate the convergence plane from the physical plane of the screen where the eyes accommodate. However, observers are all unequal in the ocular gymnastics required to achieve this dissociation. Only experience and training make it possible to effortlessly dissociate vergence and accommodation. Without practice, objects may appear blurred, or even doubled, when one converges far from the screen surface and binocular fusion becomes difficult. As in the physical world, the eyes seek clarity at the distance at which they converge and where the virtual objects are located, even though the two images never leave the physical surface of the screen.

[0020] The technological constraints result from the fact that it is technically complicated to display several images simultaneously on an autostereoscopic screen and to separate them perfectly for each eye.

[0021] Indeed, and as previously indicated, an autostereoscopic screen consists of a screen and an optical system, placed on its surface. By magnifying glass effect, this optical system makes it possible to subdivide the field of observation of the screen into contiguous zones, each specifically reserved for a different image. To see in relief, these images must at least form a stereoscopic pair. It is generally insufficient to display only a single pair from two images and the majority of the screens proposed by the applicant display a series of eight images simultaneously, stereoscopic two by two. The space formed by the eight contiguous zones occupied by each of these eight images is designated by the term lobe, which the optical system makes it possible to repeat periodically. The succession of viewpoints is always in the same order.This repetitive display allows the lobes to repeat without the lenticular array becoming too defocused, over an angle of approximately 100°.

[0022] Perfect partitioning of the information displayed in each contiguous area of ​​a lobe remains a challenge. Indeed, the light from the screen's backlight (better known as " backlight»), passes through many materials before reaching the observer's eye. A portion of the light that passes through each of the surfaces of these different materials is slightly deflected from its ideal trajectory. In addition, to limit the diffusion of the emitted light, the surfaces of the screen and the lenticular network must be perfectly polished and have no internal inhomogeneities such as frosted areas in the material itself, which are common for obtaining anti-reflective surfaces. Each micro-imperfection is likely to deflect the light and disrupt the refraction effect of the microlenses intended to separate the viewpoints. The shape of the pixels and the fact that the optical axis of the lenses can simultaneously intersect sub-pixels assigned to different viewpoints can prevent this partitioning.

[0023] Image processing constraints result from the fact that computer processing options are dependent on a number of parameters, including the number of viewpoints, their repetition, the width of the lobes, the observation distance, and the number of simultaneous observers. Sometimes the electronics of the screen or computer apply undetectable modifications to the content of the images to be displayed, which has the effect of modifying the strictly necessary location. For example, by modifying the intensity of an entire pixel, its three subpixels are affected at the same time, i.e., three viewpoints simultaneously (in the case where a single subpixel is used per image viewpoint).

[0024] The applicant has already proposed a system equipped with a device for tracking the observer's head (better known under the English name of "head tracking" device). tracking") to enable the real-time display of the viewpoints to be addressed to each eye of the observer. The speed and precision of tracking are all the more critical as the objects are far from the plane of the screen. A tiny movement of the head then results in a significant apparent displacement of the objects, in particular, when the objects are very deep, which is one of the objectives of the invention.

[0025] The inventors therefore sought a solution that would overcome at least some of the above-mentioned constraints. 3. Objectives of the invention

[0026] The invention aims to provide an autostereoscopic display screen which makes it possible to overcome the limits of projection and depth achievable by known autostereoscopic screens.

[0027] The invention aims in particular to provide a screen with almost perfect resolving power, which gives the observer the possibility of perceiving the volume in burst up to half the distance of the flat tint and that in depth up to several meters behind the plane of the screen, without there being any real limit, while maintaining optimal sharpness of the display for each eye, despite the significant disparity that it is necessary to include in the content to achieve this.

[0028] The invention aims in particular to provide a display screen which makes it possible to multiply the relief performance of known screens while maintaining, or even improving, the quality of the observed image.

[0029] The invention aims in particular to provide a screen for displaying an autostereoscopic image which has greater resolving power than known screens.

[0030] The invention also aims to provide a display screen which makes it possible to limit interference between the points of view visible to each eye of the observer.

[0031] The invention also aims to provide an autostereoscopic display device which comprises a display screen according to the invention.

[0032] The invention finally aims to provide a method for displaying an autostereoscopic image on an autostereoscopic display screen according to the invention. 4. Statement of the invention

[0033] To do this, the invention according to claim 1 relates to an autostereoscopic display screen with P viewpoints, called screen viewpoints, P being an integer greater than or equal to 6, intended to display an autostereoscopic image with I viewpoints, called image viewpoints, I being an integer greater than or equal to 2, less than or equal to P / 3, preferably equal to 2, said screen comprising: a matrix of pixels arranged in rows and columns, each pixel being composed of a plurality of sub-pixels of different colors, each intended to display a sub-pixel of one of the image viewpoints, a network of cylindrical lenticules or a parallax barrier, called an optical network, configured to allow the projection of pairs of predetermined screen viewpoints towards the eyes of an observer located at a nominal distance from the screen, called the flat tint distance. The screen according to the invention is characterized in that said optical network is configured so that the P viewpoints of the screen can be perceived successively in one direction, called the horizontal direction, by said observer at said flat tint distance, in spaces in front of the screen, called lobes, each covering I times the average interpupillary distance of an observer, and so that the screen viewpoints of each pair of viewpoints visible to the observer at the flat tint distance are separated by at least T screen viewpoints, called buffer viewpoints, T being greater than or equal to 2, such that it is possible to display successively for each lobe, each of the I image viewpoints, each repeated P / I times, and that each viewpoint visible to the observer at the flat tint distance can be framed by at least 2 buffer viewpoints each displaying this same image viewpoint.

[0034] A display screen according to the invention therefore has the particularity of being equipped with an optical network, such as a network of cylindrical lenticules (also designated by the terms optical component) configured to display P viewpoints, for example 10 viewpoints (in the case where P is equal to 10).

[0035] When the invention uses a single subpixel per viewpoint per lens and per horizontal line, each lens of the lenticular array covers P subpixels on each horizontal line. When the invention uses a single subpixel per viewpoint per lens on L horizontal lines, each lens of the lenticular array covers P / L subpixels on each horizontal line.

[0036] A screen according to the invention makes it possible to obtain an almost perfect separating power, thanks to the creation of buffer zones formed by duplication of the image viewpoints and isolating from each other the two viewpoints perceived by each of the two eyes. The I image viewpoints form 1-1 successive stereoscopic pairs two by two and these different viewpoints are separated from each other by buffer zones created by duplication of an image viewpoint. In other words, each of the I screen viewpoints visible to the observer at the flat tint distance and each displaying an image viewpoint is framed by at least 2 buffer viewpoints each displaying this same image viewpoint.

[0037] Furthermore, each lobe covering I times the average interpupillary distance of an observer, each of the I image viewpoints benefits from a space corresponding to an interpupillary distance to display the perceived viewpoint and its buffer viewpoints.

[0038] The invention thus goes against the prejudice of the person skilled in the art who generally considers that the improvement of an autostereoscopic system should be done by increasing the number of image viewpoints, to increase the lobe width, the parallax and the viewing distance in front of and behind the flat tint distance, so as to give the viewer greater freedom of positioning. This can be done at the expense of the perceived resolution if the screen resolution is not proportionally increased.

[0039] The invention does not seek to give the viewer greater latitude in positioning but to free up the volume he perceives. Rather, it tends to fix his position, so as to limit as much as possible the number of image viewpoints necessary, only two being sufficient, in such a way as to obtain an almost perfect resolving power and to make possible the perception of unusually deep volumes.

[0040] The invention makes it possible to create a monocular sub-lobe per image viewpoint, equal in width to the reference interpupillary distance (approximately 6.5 cm), and to produce in this monocular sub-lobe the means of displaying the same image several times, so that each instance of this image is visible over a width equal to the monocular sub-lobe divided by the number of times the image is duplicated.

[0041] In the absence of the invention and when the contents are pushed far in front of or behind the plane of the screen, the disparity is very significant and the problem of ghosts particularly acute. The image seen by each eye is likely to retain a remanence of the image intended for the other eye. Therefore, the left eye sees not only the left image intended for it, but also the ghost of the right image and conversely, the right eye sees not only the right image intended for it, but also the ghost of the left image. However, the ghost of the right image seen in the left image pairs stereoscopically with the right image itself, while the ghost of the left image seen in the right image pairs stereoscopically with the left image itself, which positions the ghosts on the plane of the screen itself, and not at the contours of the elements in depth.

[0042] With the invention, the points of view neighboring those observed by the observer's eyes come from the same image, which makes the crosstalk between them imperceptible.

[0043] Since the images dedicated to the observer's eyes are each replicated several times (at least 2 replicas to form at least 3 identical successive viewpoints, including the viewpoint observed by the observer's eye), each sub-pixel magnified by a lens for one eye is framed by sub-pixels that belong to the same image. Thus framed by its peers, a sub-pixel can no longer interfere with the one intended for the other eye. If it interferes with its immediate neighbors, this is not visible.

[0044] A display screen according to the invention therefore has the particularity of allowing the display of T buffer viewpoints between the viewpoints dedicated respectively to the left eye and the right eye of the observer.

[0045] A screen according to the invention not only allows one to benefit from almost perfect separating power, but also to present the advantages of traditional multi-point-of-view content, namely: the ability to move away from the flat tint distance through the recombination of viewpoints; the proportional increase in the lobe and therefore the space in which the viewer can place his eyes; the parallax effect obtained when moving horizontally along the screen; the guarantee that any remaining ghosts, in the case of very strong contrasts and extreme disparities for example, are perceived at the edge of the objects and not on the collimation plane of the images. This is an undeniable advantage for applications that require the implementation of renderings conducive to the perception of ghosts.

[0046] Preferably, the optical network of a screen according to the invention is formed of cylindrical lenses which each have a main axis which forms an angle α with the direction of the columns.

[0047] Various inclinations of the optical network (between the vertical network and the 45° network) can be used provided that a suitable viewpoint distribution grid is defined. For example, it is possible to use a network whose lenses are inclined at an angle α of 9.46°, 18.43°, 26.56°, or 33.69°, as shown in the figures discussed later.

[0048] According to a particularly advantageous variant of the invention, I is equal to 2 and said optical network is configured to be able to display in each lobe, M times the image viewpoint dedicated to the left eye of the observer and N times the image viewpoint dedicated to the right eye of the observer, M and N each being greater than or equal to 3 and their sum being equal to P, and forming T buffer viewpoints between the viewpoint dedicated to the left eye and the viewpoint dedicated to the right eye.

[0049] This advantageous variant concerns a screen intended to display an autostereoscopic image at 2 points of view intended respectively for the right eye and the left eye of the observer.

[0050] In this preferred case of displaying a 2-viewpoint autostereoscopic image (I=2), the screen is controlled to display only two different images intended respectively for the left eye and the right eye of the observer. In other words, each image is repeated several times for each eye, in this case M times for the left eye and N times for the right eye, M and N each being an integer greater than or equal to 3 and their sum being equal to P.

[0051] Furthermore, the optical parameters of the optical component are chosen so that the P viewpoints (M viewpoints for the left eye and N viewpoints for the right eye) cover a lobe which extends over a distance equal to twice the average interpupillary distance of an observer, that is to say in practice a distance of the order of 13 cm (the average interpupillary distance of an adult being 6.5 cm).

[0052] Thus, and in the case where P is equal to 10 and I is equal to 2, a change of screen viewpoint occurs, at the flat tint distance, every 1.3 cm.

[0053] The display screen according to this variant can be designed such that the number of screen viewpoints intended to display the image viewpoint dedicated to the left eye of an observer located at the flat tint distance is equal to or different from the number of screen viewpoints intended to display the image viewpoint dedicated to the right eye of the observer. In other words, the screen can be designed so as to obtain a symmetrical arrangement (M and N are then equal) or asymmetrical arrangement (M and N are then different from each other) of the viewpoints dedicated to each of the two eyes of the observer.

[0054] Preferably, M and N are equal and the screen allows a symmetrical display of the viewpoints dedicated respectively to the left eye and the right eye of the observer.

[0055] However, it is also possible to exploit an asymmetrical arrangement and minimize its drawbacks, by taking into account the direction of movement of the observer's head to position the greatest number of buffer viewpoints in the area of ​​the lobe towards which the eyes are directed, and by dynamically optimizing the allocation of viewpoints in a predictive manner, by operating permutations in parts of the lobe not perceived by the observer's eyes. This requires the use of a device for detecting the position of the observer's head, as discussed later in connection with the device according to the invention.

[0056] Preferably and according to this variant, the screen is further designed so that the sum of the number of viewpoints to the left of the left eye and the number of viewpoints to the right of the right eye is equal to the number of buffer viewpoints T separating those perceived by the two eyes of the observer.

[0057] This allows for the same number of successive T buffer viewpoints on each side of both eyes.

[0058] According to a particular embodiment of the variant of the invention with 2 image viewpoints (I equal to 2), P is equal to 10 so that the screen has 10 screen viewpoints and said optical network is configured so that the pairs visible to the observer at the flat tint distance are views 1-6, 2-7, 3-8, 4-9, 5-10 for screen viewpoints numbered successively from 1 to 10 so that each sub-pixel perceived by each eye of the observer can be surrounded on each side by 2 buffer viewpoints.

[0059] Thus, in the case where P is equal to 10 and M and N are each equal to 5, the points of view perceived by each eye are always separated by 4 other points of view, which serve as a buffer and avoid interference.

[0060] Thus, when the right eye perceives viewpoint No. 3, the left eye perceives viewpoint No. 8. This viewpoint No. 3 is framed on its left by viewpoints No. 1 and No. 2 and on its right by viewpoints No. 4 and No. 5, while viewpoint No. 8 is framed on its left by viewpoints No. 6 and No. 7 and on its right by viewpoints No. 9 and No. 10. In this case, the 2 viewpoints perceived by the observer are isolated from each other by 4 buffer viewpoints, which makes it possible to avoid interference. The resolving power is therefore optimal for the screen according to this preferred embodiment of the invention.

[0061] According to a particular embodiment of the variant of the invention with 2 image viewpoints (I equal to 2), M and N are odd so that each image viewpoint dedicated to each eye of the observer can be surrounded, in each lobe, by the same number of buffer viewpoints.

[0062] For example, in the case where M and N are each equal to 5 (forming a 10-viewpoint screen), the coding of the content of the autostereoscopic image is based on the encapsulation of a stereoscopic pair in a 10-viewpoint image format. Each subpixel perceived by an eye is framed by four buffer subpixels, 2 on its left and 2 on its right, which display the same image. Interference from neighboring subpixels with the perceived viewpoint is invisible since all display the same image.

[0063] The image is decoded using a lenticular array whose innovative architecture, with 10 viewpoints, allows for a change of viewpoint every 1.3 cm. The average interpupillary distance being 6.5 cm, it allows for 4 buffer viewpoints to be inserted between the left and right viewpoints, as required by the content. The lobe of this array is 13 cm at the flat tint distance (twice the interpupillary distance). This also allows for pupillary distances different from the chosen average value of 6.5 cm, resulting in a modification of the ideal symmetry.

[0064] The invention can be implemented using different pixel matrices. For example, it is possible to use a panel of pixels with horizontal sub-pixels (i.e. sub-pixels superimposed vertically) or a panel of pixels with vertical sub-pixels (i.e. sub-pixels juxtaposed horizontally), or even a panel with organic light-emitting diodes (better known by the English acronym OLED for " Organic Light-Emitting Diode ") or an active matrix with organic light-emitting diodes (better known by the English acronym AMOLED for " Active-Matrix Organic Light-Emitting Diode").

[0065] In the particular case of a slab formed of horizontal sub-pixels, each pixel of the pixel matrix is ​​composed of a plurality of sub-pixels of different colors superimposed on each other in the direction of the columns.

[0066] For a 2D display, the orientation of the subpixels in vertical or horizontal bands has no impact on the user, who does not perceive any difference in the display depending on the screen orientation. The information is always encoded at the pixel level and the arrangement of its substructure has no impact on the quality of the display.

[0067] On the other hand, for a 3D display, when a lenticular network acting by magnifying glass effect is glued to the surface of the screen, the orientation of the sub-pixels and the regularity of their geometry become decisive.

[0068] When using a screen in landscape mode with vertically aligned subpixels in portrait mode, or when using pixel panels with anisotropic pixel structure, the cylindrical microlenses of the lenticular array can no longer discriminate colors separately and must align entire pixels. A whole pixel then becomes the smallest entity that the lenses can allow to be observed individually by magnifying glass effect.

[0069] The invention also relates to a device for displaying an autostereoscopic image at I viewpoints, called image viewpoints, comprising: a display screen according to the invention, a display module configured to display for each lobe, each of the I image viewpoints, each repeated P / I times, each viewpoint visible to the observer at the flat tint distance being framed by at least 2 buffer viewpoints each displaying this same image viewpoint.

[0070] The technical advantages and effects of a display screen according to the invention apply mutatis mutandis to a display device according to the invention.

[0071] According to the invention, the display module is configured to assign to each sub-pixel of the pixel matrix of the display screen, a sub-pixel of one of the I images intended for the eyes of the observer. This display module is preferably formed by a computer device for controlling the pixel matrix.

[0072] Throughout the text, module means a software element, a subset of a software program, capable of being compiled separately, either for independent use or for assembly with other modules of a program, or a hardware element, or a combination of a hardware element and a software subprogram.

[0073] Preferably, the display module of a device according to the invention is a subset of a software program intended to be executed by a computer or by a microprocessor connected to the display screen.

[0074] According to a particularly advantageous variant, I is equal to 2 and said display module is configured to display in each lobe formed by said display screen, M times the image viewpoint dedicated to the left eye of the observer and N times the image viewpoint dedicated to the right eye of the observer and forming T buffer viewpoints, between the image viewpoint dedicated to the left eye and the image viewpoint dedicated to the right eye.

[0075] In the preferred case where I is equal to 2, these 2 viewpoints are intended respectively for the left and right eye of an observer. In the case where I is greater than 2, the 1-2 central viewpoints can be seen either by the left eye or by the right eye of an observer, depending on his position.

[0076] Advantageously and according to the invention, the device further comprises a device for detecting, at each instant, the position of the observer's head relative to said horizontal and / or vertical direction so as to allow said display module to slide the image viewpoints in each lobe so that said screen viewpoints visible to the observer are permanently framed by buffer viewpoints of the image viewpoint visible to this eye.

[0077] A display device according to this variant makes it possible to keep the observer permanently in the center of the lobe so that he perceives (in the case where I=2 and N=M=5 to form 10 screen viewpoints, P=10) viewpoints no. 3 and no. 8. A lobe of twice the interpupillary distance (i.e. of the order of 13 cm) offers a limited latitude of movement. Rather than forcing the observer to place himself in the center of the lobe, the invention makes it possible to dynamically position the lobe to center it on the position of the observer, thanks to a tracking device (also known under the English name of " tracking ”) of the observer's head.

[0078] The invention therefore makes it possible to identify the position of the observer's head relative to the horizontal and / or vertical direction of the screen (i.e. to detect the lateral movements of the observer's head relative to the screen in the case of horizontal detection) and to slide the viewpoints in the lobe so as to present viewpoints No. 3 and No. 8 to the observer. Insofar as the aim in this embodiment is only to locate the position of the observer's face relative to the horizontal and / or vertical direction, a simple 2D camera is sufficient to ensure the tracking of the observer's head.

[0079] According to an alternative embodiment of the invention, said detection device is configured to detect, at each instant, the position of the observer's head - in particular of each eye of the observer - relative to a predetermined point on the display screen, making it possible to define the X, Y, Z coordinates of each eye of the observer with a predetermined precision, in an orthonormal reference frame of which said predetermined point on the screen is the origin of said reference frame, said screen defining the X and Y axes of this reference frame and the direction perpendicular to the screen defining the Z axis of this reference frame.

[0080] According to this variant of the invention, the position of the observer's head is determined at each instant in three dimensions, which makes it possible to use a device according to this variant in augmented reality applications.

[0081] In particular, when the screen brightness allows it, it is possible to use the screen, not in direct vision, but in reflection on a semi-reflective mirror, so as to create an augmented reality experience of the "Pepper's ghost" type.

[0082] The 3D volume displayed on the screen then materializes in the room and offers the possibility of creating new immersive experiences, provided that care is taken not to create relief conflicts. In particular, care must be taken to ensure that a deep virtual object does not overlap with an element of the scene located in front of it.

[0083] The dissociation between vergence and accommodation then no longer poses a problem because the gaze relies on the physical world to converge on the virtual elements that are part of it.

[0084] When the viewer is not too close to the screen, the dioptric difference is small enough that the sensation of sharpness is good and visual fatigue is minimal.

[0085] To be credible, an augmented reality experience must, however, offer a perfect match between the physical and virtual worlds. Beyond the relief conflicts that must be absolutely avoided, virtual objects superimposed on the real world must obey the same rules of perspective and react to the observer's movements in the same way as the objects in the room. A virtual cube superimposed on a physical cube must behave in every way like the real object when the observer moves their head. The parallax between virtual objects perceived by the observer must also remain consistent with the position of their eyes. To achieve this, it is essential to measure in real time, with low latency and very precisely, the position of each of the observer's pupils in all three dimensions.

[0086] The Z measurement is particularly important in this context. Indeed, given the good resolving power of the screen according to the invention and the great depths that it can simulate, the disparity displayed produces a very different sensation of volume depending on the distance at which the content is collimated, but also depending on the distance at which the observer is from the screen. To generate images faithful to the physical world, the real-time image rendering engine must therefore know precisely the Z position of the observer's pupils, in order to adapt the stereoscopic base and the focal length accordingly, and that in X and Y to adjust the parallax.

[0087] The invention therefore makes it possible to know the relative position of the observer in relation to the environment in order to adapt the virtual scene accordingly. In particular, the invention according to this variant makes it possible to know the distance at which the pupils of the observer are from the screen, in order to adapt the rendering of the virtual scene and make it consistent with the physical world.

[0088] Such a device can, for example, be used in a flight simulator, construction machinery driving simulator, remote operation simulator, or any other simulator aimed at placing the user in hyper-realistic conditions, without requiring the wearing of an augmented reality headset.

[0089] In this particular application context, the room can be empty of objects and it is no longer essential to preserve the coherence between the physical world and the virtual world. The measurement of the Z position of the pupils nevertheless remains essential to adapt the stereoscopic base and the focal length, just as it is essential in X and Y to adapt the parallax.

[0090] The invention also relates to a method for displaying an autostereoscopic image with I viewpoints, I being an integer greater than or equal to 2, on a display screen with P viewpoints, called screen viewpoints, P being an integer greater than or equal to 6 and greater than or equal to 3I, comprising a matrix of pixels arranged in rows and columns, each pixel being composed of a plurality of sub-pixels of different colors, each intended to display a sub-pixel of one of the image viewpoints, and a network of cylindrical lenticules or a parallax barrier, called an optical network, configured to allow the projection of pairs of predetermined screen viewpoints towards the eyes of an observer located at a nominal distance from the screen, called the flat tint distance, and so that the P screen viewpoints can be perceived successively in one direction, called the horizontal direction, by said observer at said flat tint distance,in spaces in front of the screen, called lobes, each covering 1 times the average interpupillary distance of an observer.

[0091] The method according to the invention is characterized in that it comprises, for each lobe, the assignment to each adjacent screen viewpoint of each image viewpoint visible to the observer at the flat tint distance, of a copy of this image viewpoint such that, in each lobe, each of the I image viewpoints visible to the observer at the flat tint distance is framed by at least 2 buffer viewpoints each displaying this same image viewpoint, and in that it is made up of the succession of the I image viewpoints each repeated P / I times.

[0092] The technical advantages and effects of a display screen according to the invention apply mutatis mutandis to a display method according to the invention.

[0093] According to a particularly advantageous variant of the invention, I is equal to 2 and the method comprises, for each lobe: the assignment to the first M screen viewpoints of the image viewpoint dedicated to the observer's left eye, the assignment to the following N screen viewpoints of the image viewpoint dedicated to the observer's right eye,

[0094] M and N each being greater than or equal to 3 and their sum being equal to P.

[0095] The invention also relates to a display screen, a display device and a display method characterized in combination by all or part of the features mentioned above or below. 5. List of figures

[0096] Other aims, characteristics and advantages of the invention will appear on reading the following description given solely for non-limiting purposes and which refers to the appended figures in which: [ Fig. 1] is a schematic view of a display screen according to one embodiment of the invention, [ Fig. 2 ] is a schematic view of a display screen according to an embodiment of the invention with 10 viewpoints of a 2 viewpoint image screen comprising a pixel panel formed of vertical sub-pixels surmounted by an optical network having an inclination of 18.43° relative to the vertical (in other words, I=2; P=10; N=5; M=5; α=18.43°), [ Fig. 3 ] is a schematic view of the principle of arrangement of the left and right images on a screen according to one embodiment of the invention, [ Fig. 4 ] is a schematic view of a display screen according to an embodiment of the invention with 9 viewpoints, a screen of an image with 2 viewpoints comprising a pixel panel formed of vertical sub-pixels surmounted by an optical network having an inclination of 18.43° relative to the vertical (in other words, I=2; P=9; N=5; M=4; α=18.43°), [ Fig. 5] is a schematic view of a display screen according to an embodiment of the invention with 8 viewpoints, a screen of an image with 2 viewpoints comprising a pixel panel formed of vertical sub-pixels and surmounted by an optical network having an inclination of 18.43° relative to the vertical (in other words, I=2; P=8; N=4; M=4; α=18.43°), [ Fig. 6 ] is a schematic view of a display screen according to an embodiment of the invention with 10 viewpoints of a 2 viewpoint image screen comprising a pixel panel formed of vertical sub-pixels and surmounted by an optical network having an inclination of 9.46° relative to the vertical (in other words, I=2; P=10; N=5; M=5; α=9.46°), [ Fig. 7] is a schematic view of a display screen according to an embodiment of the invention with 10 viewpoints of a 2 viewpoint image screen comprising a pixel panel formed of vertical sub-pixels and surmounted by an optical network having an inclination of 26.56° relative to the vertical (in other words, I=2; P=10; N=5; M=5; α=26.56°), [ Fig. 8 ] is a schematic view of a display screen according to an embodiment of the invention with 10 viewpoints of a 2 viewpoint image screen comprising a pixel panel formed of vertical sub-pixels and surmounted by an optical network having an inclination of 33.69° relative to the vertical (in other words, I=2; P=10; N=5; M=5; α=33.69°), [ Fig. 9] is a schematic view of a display screen according to an embodiment of the invention with 10 viewpoints, a screen of an image with 2 viewpoints comprising a pixel panel formed of horizontal sub-pixels and surmounted by an optical network having an inclination of 18.43° relative to the vertical (in other words, I=2; P=10; N=5; M=5; α=18.43°), [ Fig. 10 ] is a schematic view of a display screen according to an embodiment of the invention with 6 viewpoints, a screen of an image with 2 viewpoints comprising a pixel panel formed of horizontal sub-pixels and surmounted by an optical network having an inclination of 26.56° relative to the vertical (in other words, I=2; P=6; N=3; M=3; α=26.56°), [ Fig. 11] is a schematic view of a display screen according to an embodiment of the invention with 10 viewpoints of a 2 viewpoint image screen comprising a pixel panel formed of horizontal sub-pixels and surmounted by an optical network having an inclination of 33.69° relative to the vertical (in other words, I=2; P=10; N=5; M=5; α=33.69°), [ Fig. 12 ] is a schematic view of a display screen according to an embodiment of the invention with 20 viewpoints of a 4 viewpoint image comprising a pixel panel formed of horizontal sub-pixels and surmounted by an optical network having an inclination of 9.46° relative to the vertical (in other words, I=4; P=20; α=9.46°), [ Fig. 13] is a schematic view of a display screen according to an embodiment of the invention with 20 viewpoints, a screen of an image with 4 viewpoints comprising a pixel panel formed of horizontal sub-pixels and surmounted by an optical network having an inclination of 18.43° relative to the vertical (in other words, I=4; P=20; α=18.43°), [ Fig. 14 ] is a schematic view of a display device according to an embodiment of the invention comprising a screen according to the invention, a tracking device and a display module. 6. Detailed description of an embodiment of the invention

[0097] In the figures, scales and proportions are not strictly respected, for the purposes of illustration and clarity.

[0098] There figure 1is a partial schematic view of an autostereoscopic display screen according to the invention comprising a matrix 10 of pixels (also referred to as a pixel panel) arranged in rows and columns surmounted by an optical network 11 of cylindrical lenticules. The optical network 11 is partially shown, but it is obvious that it extends in practice over the entire pixel panel, or where appropriate over the part of the screen dedicated to autostereoscopy.

[0099] The particularity of the screen according to the invention is that the optical network 11 is configured to allow the display of P viewpoints and that the solid angle in which these P viewpoints are spread is I times the interpupillary distance, where I represents the number of viewpoints of the autostereoscopic image to be displayed by this screen.

[0100] For example, for a 10-viewpoint (P=10) screen displaying a 2-viewpoint (I=2) image, the solid angle over which these 10 viewpoints are spread is 13 cm, i.e. twice the average interpupillary distance of 6.5 cm at the flat tint distance of the screen.

[0101] In other words, the specific screen proposed by the inventors is equipped with an optical network 11 intended for the display of P points of view. The innovation consists in displaying on this screen only I different images, each replicated a certain number of times. On either side of each sub-pixel magnified by a lens for one eye are a certain number of sub-pixels belonging to the same image. Thus framed by its peers, a sub-pixel can no longer interfere with the one intended for the other eye. If it interferes with its immediate neighbors, this is not seen.

[0102] In the specific case of a 10-viewpoint screen intended to display a 2-viewpoint autostereoscopic image, the manufacture of an optical network with 10 viewpoints, visible successively in a 13 cm lobe, requires a microlens focal length 5 times longer than that required for a 65 cm lobe (65 cm / 13 cm = 5) as already proposed by the applicant. Long focal length lenses have a proportionally larger radius of curvature and are easier to machine.

[0103] Machining a long focal length optical component allows for an almost perfect surface finish after machining and bonding the front face with optical silicone. Also, the imperfections inherent in prior art screens are greatly reduced with a screen according to the invention.

[0104] The person skilled in the art will be able to determine, depending on the targeted flat tint distance and the characteristics of the screen, the parameters of the optical network necessary to enable the display of the P points of view in lobes of 1x6.5 cm.

[0105] For example, for a 15.6-inch 4K panel, whose display area measures approximately 345 mm x 194 mm and each pixel 0.09 mm, the array pitch would be approximately 0.299 mm, the focal length of the lenticular array would depend on the usage conditions and could be 4 mm for a flat tint of 1.30 m (case of 10 viewpoints to display an image with 2 viewpoints) 6.1. Embodiment I=2; P=10; N=5; M=5; α=18.43°; vertical subpixels

[0106] There figure 2illustrates an embodiment of the invention representing a 10-viewpoint display for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has a tilt of 18.43° relative to the vertical (i.e., I=2; P=10; N=5; M=5; α=18.43°).

[0107] This screen is intended to display in each lobe, 5 times the image viewpoint dedicated to the observer's left eye and 5 times the image viewpoint dedicated to the observer's right eye. In this figure and all those referring to a 2-viewpoint autostereoscopic image, the subpixels assigned to the right eye are represented in gray and the subpixels assigned to the left eye are represented in white.

[0108] In addition, a black frame represents the smallest "modulo", that is, the smallest block of necessary and sufficient data, which allows the entire screen to be filled by periodic repetition, with or without vertical and / or horizontal phase shift. The "module" is most often written under a single lens, but this is not always the case, depending on the angle of the network and the number of viewpoints.

[0109] Each pixel in the pixel array is composed of subpixels of different colors, identified in the figures by the letters R, G, and B (for red, green, and blue). Each subpixel displays a subpixel from the left image perspective (white subpixel) or from the right image perspective (gray subpixel).

[0110] Each pixel is formed from an RGB triplet, represented schematically on the figure 3 by different levels of gray as well as by the letters R, G and B. On the figure 3, the subpixels dedicated to the left eye are topped with the letter G and the subpixels dedicated to the right eye are topped with the letter D.

[0111] The angle of the optical network being 18.43° in this embodiment, its tangent 1 / 3 and the pixels being square, at each new line of pixels on the screen, the optical axis of the lenses shifts by a third of a pixel. At 10 viewpoints the pitch of the network is approximately 10 / 3 of pixels.

[0112] Viewpoints 1 to 5 are dedicated to the right image and viewpoints 6 to 10 to the left image. Thanks to this particular arrangement, viewpoints 3 and 8 each have 2 buffer viewpoints on either side, which display the same image and with which there is no visible alteration of the perceived image.

[0113] Preferably, the screen is used with a user head tracking device to ensure that viewpoints 3 and 8 are the viewpoints addressed to the observer.

[0114] Thus, the points of view perceived by each eye (shown by circles on the figure 3 and by a transparency of the lens on the figure 2 ) are separated by 4 other viewpoints, which serve as a buffer for an interpupillary distance of 6.5 cm, or by 3 or 5 other viewpoints when the interpupillary distance is lower or higher. 6.2. Embodiment I=2; P=9; N=5; M=5; α=18.43°; vertical subpixels

[0115] There figure 4illustrates another embodiment of the invention representing a 9-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has an inclination of 18.43° relative to the vertical (i.e., I=2; P=9; N=5; M=4; α=18.43°).

[0116] In this embodiment, the distribution of viewpoints is not balanced since there is an odd number of screen viewpoints. In the example shown, we chose to assign 5 viewpoints to the left eye (represented by white sub-pixels on the figure 4) and 4 right-eye viewpoints (represented by gray subpixels on the figure 4). Of course, another distribution of viewpoints could have been chosen for a substantially identical result. Depending on the user's direction of movement, this distribution can be rebalanced. We favor the eye that moves toward the center of the lobe to give it more viewpoints between the two eyes.

[0117] In this embodiment, viewpoints 1 to 5 are dedicated to the right image and viewpoints 6 to 9 to the left image. Viewpoints 3 and 7 are the screen viewpoints projected towards the two eyes of the observer and the latter have 3 buffer viewpoints between them. 6.3. Embodiment I=2; P=8; N=5; M=5; α=18.43°; vertical subpixels

[0118] There Figure 5illustrates another embodiment of the invention representing an 8-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has a tilt of 18.43° relative to the vertical (i.e., I=2; P=8; N=4; M=4; α=18.43°).

[0119] In this embodiment, viewpoints 1 to 4 are dedicated to the right image (shown in gray) and viewpoints 5 to 8 to the left image (shown in white). Viewpoints 3 and 7 are the screen viewpoints projected towards the two eyes of the observer and the latter have 3 buffer viewpoints between them.

[0120] Of course, other configurations are possible than those shown, both in terms of the number of screen viewpoints and the inclination of the lenticular network. 6.4. Embodiment I=2; P=10; N=5; M=5; α=9.46°; vertical subpixels

[0121] There figure 6illustrates another embodiment of the invention representing a 10-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has a tilt of 9.46° relative to the vertical (i.e., I=2; P=10; N=5; M=5; α=9.46°).

[0122] It is distinguished from the embodiment of the figure 2 due to the angle of the optical network, which implies a grid for filling the viewpoints adapted to this specific arrangement, but also by a pitch of the lenses twice as fine. The viewpoints are in phase opposition, which has the consequence of spreading the "modulo" over 6 lines instead of 2. 6.5. Embodiment I=2; P=10; N=5; M=5; α=26.56°; vertical subpixels

[0123] There figure 7illustrates another embodiment of the invention representing a 10-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has a tilt of 26.56° relative to the vertical (i.e., I=2; P=10; N=5; M=5; α=26.56°).

[0124] It is distinguished from the embodiment of the figure 2 due to the angle of the optical network, which implies a viewpoint filling grid adapted to this specific arrangement, but the pitch of the lenses is the same. On the other hand, the "module" is spread here under three lenses instead of just one. The figure 7 therefore represents these three lenses. 6.6. Embodiment I=2; P=10; N=5; M=5; α=33.69°; vertical subpixels

[0125] There figure 8illustrates another embodiment of the invention representing a 10-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has a tilt of 33.69° relative to the vertical (i.e., I=2; P=10; N=5; M=5; α=33.69°).

[0126] Also, the only difference from the method of realization of the figure 4 is the angle of the optical network, which implies a viewpoint filling grid adapted to this specific arrangement. 6.7. Embodiment I=2; P=10; N=5; M=5; α=18.43°; horizontal subpixels

[0127] The invention also applies to pixel panels formed from horizontal sub-pixels.

[0128] So, the figure 9illustrates an embodiment of the invention representing a 10-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of horizontal subpixels and the optical array has an inclination of 18.43° relative to the vertical (i.e., I=2; P=10; N=5; M=5; α=18.43°).

[0129] In this screen, each pixel of the pixel matrix is ​​composed of a plurality of sub-pixels of different colors superimposed on each other in the direction of the columns and associated with an optical network formed of cylindrical lenses each having a principal axis which forms an angle of 18.43° with the direction of the columns.

[0130] The cylindrical microlenses of the lenticular array can no longer discriminate colors separately and simultaneously magnify the three RGB components of each pixel. The entire pixel then becomes the smallest entity that the lenses can allow to be observed individually by magnifying effect.

[0131] Furthermore, the three colored dots of a pixel (the subpixels) are assembled to approximately form a square. In other words, and according to this variant, the colored dots fit into three small juxtaposed and superimposed rectangles three times wider than they are high to form a square, the pixel.

[0132] We can clearly see on the figure 9 that the horizontal subpixels are grouped in blocks of three to each display the same viewpoint of the image dedicated either to the left eye (represented in white) or dedicated to the right eye (represented in gray).

[0133] As with the horizontal subpixel version, viewpoints 1 to 5 are dedicated to the right image and viewpoints 6 to 10 to the left image. 6.8. Embodiment I=2; P=6; N=3; M=3; α=26.56°; horizontal subpixels

[0134] There figure 10 illustrates an embodiment of the invention representing a 6-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of horizontal subpixels and the optical array has an inclination of 26.56° relative to the vertical (i.e., I=2; P=6; N=3; M=3; α=26.56°).

[0135] In this embodiment, each viewpoint addressed to each eye of the observer is framed by 2 buffer viewpoints on either side of this viewpoint. 6.9. Embodiment I=2; P=10; N=5; M=5; α=33.69°; horizontal subpixels

[0136] There figure 11illustrates an embodiment of the invention representing a 10-viewpoint screen for displaying a 2-viewpoint autostereoscopic image. The pixel panel is formed of horizontal subpixels and the optical array has an inclination of 33.69° relative to the vertical (i.e., I=2; P=10; N=5; M=5; α=33.69°).

[0137] Also, the only difference from the method of realization of the figure 9 is the angle of the optical network, which implies a viewpoint filling grid adapted to this specific arrangement.

[0138] Of course, other configurations are possible than those shown, both in terms of the number of screen viewpoints and the inclination of the lenticular network.

[0139] Note that the methods of realization of the figures 9 , 10 And 11can also be applied to an IPS panel (English acronym for “In-Plane Switching”), to an organic light-emitting diode panel (better known by the English acronym OLED for “ Organic Light-Emitting Diode ") or an active matrix with organic light-emitting diodes (better known by the English acronym AMOLED for " Active-Matrix Organic Light-Emitting Diode ”) .

[0140] Indeed, like a screen whose pixels are composed of a plurality of sub-pixels of different colors superimposed on each other in the direction of the columns, the panels with anisotropic pixels can benefit from the invention to obtain a quasi-perfect separating power, by preventing the optical axis of the lenses from intersecting contiguous pixels and generating ghost images. 6.10. Embodiment I=4; P=20; α=9.46°; vertical subpixels

[0141] There figure 12illustrates another embodiment of the invention representing a 20-viewpoint screen for displaying a 4-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has a tilt of 9.46° relative to the vertical (i.e., I=4; P=20; α=9.46°).

[0142] It is essentially distinguished from the mode of realization of the figure 6 due to the number of image viewpoints which is here equal to 4 instead of 2, which implies an adapted viewpoint filling grid. The lens pitch is in this case twice as large.

[0143] In this figure as well as in the figure 13 , the 4 image viewpoints are represented respectively by white subpixels, gray subpixels, vertically hatched subpixels, and black subpixels.

[0144] On the method of realization of the figure 12, screen viewpoints 1 to 5 display the first image viewpoint (gray subpixels), screen viewpoints 6 to 10 display the second image viewpoint (hatched subpixels), screen viewpoints 11 to 15 display the third image viewpoint (black subpixels), and screen viewpoints 16 to 20 display the fourth image viewpoint (white subpixels).

[0145] The lens shown in the figure allows to project towards the observer, according to his position, the stereoscopic pairs of viewpoints 3-8; 8-13 or 13-18, the viewpoints visible to the observer being separated from each other by 4 buffer viewpoints. Each viewpoint visible to the observer is well framed by at least 2 screen viewpoints which display the same image viewpoint.

[0146] The number of screen viewpoints allows, despite the number of image viewpoints, to maintain 2 buffer viewpoints on either side of each viewpoint addressed to the observer's eyes. In other words, each image viewpoint extends over 5 consecutive sub-pixels. As such, we can consider that M=5 and N=5 for each of the image viewpoints. 6.11. Embodiment I=4; P=20; N=5; M=5; α=18.43°; vertical subpixels

[0147] There figure 13 illustrates another embodiment of the invention representing a 20-viewpoint screen for displaying a 4-viewpoint autostereoscopic image. The pixel panel is formed of vertical subpixels and the optical array has a tilt of 18.43° relative to the vertical (i.e., I=4; P=20; N=5; M=5; α=18.43°).

[0148] It is essentially distinguished from the mode of realization of the figure 2due to the number of image viewpoints which is here equal to 4 instead of 2, which implies an adapted viewpoint filling grid. The lens pitch is in this case twice as large.

[0149] There figure 14 is a schematic view of a device for displaying a 2-point-of-view autostereoscopic image according to one embodiment of the invention.

[0150] Such a device is formed of a display screen 12 according to the invention, a display module 13 configured to display the right and left images according to the configuration described previously and a device 14 for detecting, at each instant, the position of the head of the observer 8 relative to the horizontal direction so as to allow the display module to slide the image viewpoints in each lobe so that said screen viewpoints visible to the observer are permanently framed by buffer viewpoints of the image viewpoint visible to this eye.

[0151] The tracking device 14 may be of any known type. For example, it may be the device marketed under the reference Intel RealSense ®< SR300 (SDK1 Gold Release). This is a time-of-flight RGBD camera, accompanied by an optimized face tracking library.

[0152] According to a variant, the detection device is configured to detect, at each instant, the position of the observer's head relative to a predetermined point on the display screen, making it possible to define the observer's X, Y, Z coordinates with a predetermined precision, in an orthonormal reference frame of which said predetermined point on the screen is the origin of said reference frame, the screen defining the X and Y axes of this reference frame and the direction perpendicular to the screen defining the Z axis of this reference frame.

[0153] As previously indicated, this variant makes it possible in particular to use a screen according to the invention in an augmented reality application, for example for simulation or head-up display applications in automobiles. Indeed, augmented reality requires significant contrasts. The black elements are transparent while the white elements are the most material. If one wishes to display a white square on a transparent background, it is essential to eliminate all possible causes of ghost images, which would interfere with the perceived images.

[0154] However, a device according to the invention which combines a screen according to the invention and a face tracking device makes it possible to obtain an almost perfect separating power and is therefore applicable to augmented reality applications.

[0155] The invention is not limited to the embodiments described. In particular, and as specified, different screens can be produced to obtain an autostereoscopic device with almost perfect resolving power. The invention makes it possible to revolutionize the experience of relief as we know it. The usual rules for composing relief scenes are obsolete and all stereoscopic content can be displayed with an unequaled relief perception.

[0156] Although the vergence / accommodation dissociation persists, it becomes easier to get used to it, in the absence of any optical interference in the stereoscopic pair, degradation of brightness and colors and without the bulk of 3D glasses or a headset.

[0157] Thanks to its unrivalled separating power and the exceptional amplitude of the relief that it allows to be exploited, the screens of the invention make it possible to respond to a very large number of demands for uses that are little or not met by existing devices.

[0158] The invention opens up countless market prospects for individual autostereoscopic applications in many technical fields including: medical imaging, surgical robotics, videoconferencing, automotive, education, slot machines, remote monitoring, remote operation, drone piloting, aeronautics, cartography, seismology and geological, mining and oil prospecting, scientific studies at all scales, from microscopes to satellites, simulation, production line for relief films, CAD, design, mechanics, architecture, virtual tours, museums, games, telephony / tablets, entertainment, etc.

[0159] A particularly interesting application case of augmented reality concerns the particular application context where the backdrop is not made up of a physical set but of the display on a screen or by means of a projector of a 2D image, synchronized with the scene and representing the backgrounds.

[0160] The autostereoscopic screen according to the invention makes it possible to display an autostereoscopic image of very large amplitude, for example 20 meters deep, which occupies all the physical space separating the screen according to the invention from the 2D background image. Beyond about twenty meters, the eyes perceive little disparity and monocular cues, such as perspective, become predominant. 2D backgrounds, if they integrate strong visual cues such as perspective, therefore perfectly complete the illusion and contribute to increasing the viewer's feeling of immersion.

[0161] Such an installation is mainly single-user (especially when using a 2-viewpoint autostereoscopic image), but could nevertheless, due to the exceptional filling quality of the two lobe sub-parts, be used by several users, positioned at the flat tint distance.

Claims

1. Autostereoscopic display screen with P points-of-view, called screen points-of-view, P being an integer greater than or equal to 6, intended to display an autostereoscopic image at I points-of-view, called image points-of-view, I being an integer greater than or equal to 2, less than or equal to P / 3, preferably equal to 2, said screen comprising: - a matrix of pixels arranged by rows and by columns, each pixel being composed of a plurality of sub-pixels of different colors, each intended to display a sub-pixel of one of the image points-of-view, - an array of cylindrical lenticules or a parallax barrier, called an optical array, configured to allow the projection of pairs of predetermined points-of-view of the screen towards the eyes of an observer located at a nominal distance of 1 screen, called flat tint distance, said screen being characterized in that said optical array is configured so that the P points-of-view of the screen can be perceived successively in a direction, called horizontal direction, by said observer at said flat tint distance, in spaces in front of the screen, called lobes, each covering I times the average inter-pupillary distance of an observer, and so that the screen points-of-view of each pair of visible points-of-view by the observer at the flat tint distance, are separated by at least T points-of-view, called buffer points-of-view, T being greater than or equal to 2, so that it is possible to display successively for each lobe, each of the I image points-of-view, each repeated P / I times, and that each point-of-view visible to the observer at the flat tint distance can be surrounded by at least 2 buffer points-of-view each displaying this same image viewpoint.

2. Display screen according to claim 1, characterized in that I is equal to 2 and in that the said optical array is configured to be able to display in each lobe, M times the image point-of-view dedicated to the left eye of the observer and N times the image point-of-view dedicated to the right eye of the observer, M and N each being greater than or equal to 3 and their sum being equal to P, and by forming T buffer points-of-view between the point-of-view dedicated to the left eye and the point-of-view dedicated to the right eye.

3. Display screen according to claim 2, characterized in that M and N are odd such that each image point-of-view dedicated to each eye of the observer can be surrounded, in each lobe, by the same number of buffer points-of-view.

4. Display screen according to one of claims 2 or 3, characterized in that M and N are equal so that the screen allows a symmetrical display of the points-of-view dedicated respectively to the left eye and to the observer's right eye.

5. Display screen according to one of claims 2 to 4, characterized in that P is equal to 10 so that the screen has 10 points-of-view and in that the said optical array is configured so that the pairs visible to the observer at the flat tint distance are views 1-6, 2-7, 3-8, 4-9, 5-10 for points-of-view numbered successively from 1 to 10 so that each sub-pixel perceived by each eye of the observer can be surrounded on each side by 2 buffer points-of-view.

6. Display screen according to one of claims 1 to 5, characterized in that the said optical array is formed of cylindrical lenses each having a main axis which forms an angle with the direction of the columns.

7. Display screen according to claim 6, characterized in that the said angle is between 0° and 45°, and preferably chosen from the group comprising the angles of 9.46°; 18.43°; 26.56°; and 33.69°.

8. Display screen according to one of claims 1 to 7, characterized in that each pixel of the said matrix of pixels is composed of a plurality of horizontal sub-pixels of different colors superimposed on each other in the direction columns.

9. Display screen according to one of claims 1 to 7, characterized in that each pixel of the said matrix of pixels is composed of a plurality of vertical sub-pixels of different colors juxtaposed to each other in the direction lines.

10. Device for displaying an autostereoscopic image at I points-of-view, called image point-of-view, comprising: - a display screen according to one of claims 1 to 9, - a display module configured to display for each lobe each of the I image points-of-view, each repeated P / I times, each point-of-view visible to the observer at the flat tint distance being framed at least 2 buffer points-of-view each displaying this same image point-of-view.

11. Display device according to claim 10, characterized in that I is equal to 2 and in that said display module is configured to display in each lobe formed by said display screen, M times the image point-of-view dedicated to the left eye of the observer and N times the point-of-view image dedicated to the right eye of the observer and forming T points-of-view, called buffer points-of-view, between the image point-of-view dedicated to the left eye and the image point-of-view dedicated to the right eye.

12. Display device according to claim 11, characterized in that it further comprises a device for detecting, at any instant, the position of the observer's head with respect to the said horizontal and / or vertical direction so as to allow said display module to slide the image points-of-view in each lobe such that said screen points-of-view visible to the observer are permanently framed by buffer points-of-view from the image point-of-view visible by that eye.

13. Display device according to claim 12, characterized in that the said detection device is configured to detect, at each instant, the position of the observer's head with respect to a predetermined point of the display screen, making it possible to define the coordinates X, Y, Z of the observer with a predetermined precision, in an orthonormal frame of reference of which said predetermined point of the screen is the origin of said frame, said screen defining the axes X and Y of this frame and the direction perpendicular to the screen defining the Z axis of this frame.

14. Method for displaying an autostereoscopic image at I points-of-view, I being an integer greater than or equal to 2, on a display screen at P points-of-view, called screen points-of-view, P being an integer greater than or equal to 6 and greater than or equal to 3I, comprising a matrix of pixels arranged by rows and by columns, each pixel being composed of a plurality of sub-pixels of different colors, each intended to display a sub-pixel of one of the image points-of-view, and an array of cylindrical lenticules or a parallax barrier, said optical array, configured to allow the projection of pairs of points-of-view of the predetermined screen towards the eyes of an observer located at a nominal distance from the screen, called the flat tint distance, and so that the P points-of-view of the screen can be perceived successively in one direction, called the horizontal direction, by said observer at said flat tint distance, in spaces in front of the screen, said lobes, each covering each I times the average inter-pupillary distance of an observer, characterized in that it comprises, for each lobe, the allocation to each adjacent screen point-of-view of each image point-of-view visible by the observer at the flat tint distance, of a copy of this image point-of-view such that, in each lobe, each of the I image points-of-view visible to the observer at the tint distance plate is flanked by at least 2 buffer points-of-view each displaying this same image points-of-view, and in that it is made up of the succession of I image points-of-view each repeated P / I times.

15. Display method according to claim 14, characterized in that I is equal to 2 and in that it comprises, for each lobe: - the assignment to the first M screen points-of-view of the image point-of-view dedicated to the left eye of the observer, - the assignment to the following N screen points-of-view of the image point-of-view dedicated to the right eye of the observer, M and N each being greater than or equal to 3 and their sum being equal to P.