Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHODS OF RENDERING LIGHT FIELD IMAGES FOR INTEGRAL-IMAGING-BASED LIGHT FILED DISPLAY
Document Type and Number:
WIPO Patent Application WO/2019/182592
Kind Code:
A1
Abstract:
Methods of rendering light field images for integral-imaging-based light field displays.

Inventors:
HUA, Hong (1630 E Universtiy Blvd, Tucson, AZ, 85721, US)
HUANG, Hekun (6580 N. Calle Sin Nombre, Tucson, AZ, 85718, US)
Application Number:
US2018/023682
Publication Date:
September 26, 2019
Filing Date:
March 22, 2018
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
ARIZONA BOARD OF REGENTS ON BEHALF OF THE UNIVERSITY OF ARIZONA (Tech Transfer Arizona, 220 West Sixth Street 4th Floo, Tucson AZ, 85701, US)
International Classes:
H04N13/00; G02B17/08; G02B27/01; G06T19/00
Foreign References:
US20150201176A12015-07-16
US20170078652A12017-03-16
US20050179868A12005-08-18
Other References:
HUANG ET AL.: "An integral-imaging-based head-mounted light field display using a tunable lens and aperture array", JOURNAL OF THE SOCIETY FOR INFORMATION DISPLAY 0 17 MAR 1, vol. 25, no. 3, 3 January 2017 (2017-01-03), pages 200 - 7, XP55636888
Attorney, Agent or Firm:
HAUN, Niels et al. (Dann, Dorfman Herrell and Skilliman, P.C.,1601 Market Street, Suite 240, Philadelphia PA, 19103, US)
Download PDF:
Claims:
CLAIMS

What is claimed is:

1. A method for rendering light field images of a 3D scene in an HMD using an integral- imaging-based light field display, comprising:

providing integral imaging (Ini) optics having a vari-focal element and a microdisplay in optical communication with the vari-focal element, the Ini optics having a central depth plane (CDP) associated therewith;

displaying image data on the microdisplay, the image data comprising elemental images each representing a different perspective of the 3D scene; and

setting the focal length of the vari-focal element to adjust the location of the CDP.

2. The method of claim 1, comprising sampling the 3D scene using a simulated virtual array of cameras so that each camera captures a respective portion of the 3D scene to create the elemental images.

3. The method of any one of the preceding claims, wherein the Ini optics is configured to create a virtual CDP that is the optical conjugate plane to the microdisplay in visual space, and wherein the 3D scene comprises a depth of interest (DOI) through which the 3D scene extends along a visual axis, the 3D scene having an average DOI, and comprising setting the focal length of the vari-focal element so that the location of the virtual CDP coincides with the average DOI of the 3D scene.

4. The method of any one of the preceding claims, wherein the Ini optics is configured to create a virtual CDP that is the optical conjugate plane to the microdisplay in visual space, and wherein the 3D scene comprises a depth of interest (DOI) through which the 3D scene extends along a visual axis, and comprising:

selecting a plurality of depths distributed along the visual axis within the DOI of the 3D scene; and

for each selected depth of the plurality of depths, setting the focal length of the vari-focal element so that the location of each virtual CDP coincides with the selected depth to create a plurality of virtual CDP’s each virtual CDP coinciding with a respective selected depth of the plurality of depths.

5. The method of claim 4, comprising for each selected depth of the plurality of depths, sequentially displaying on the microdisplay a portion of the 3D scene associated with each selected depth, and wherein the step of setting the focal length of the vari-focal element is synchronized to the timing of the sequential displaying on the microdisplay.

6. The method of any one of the preceding claims, wherein the Ini optics comprises a relay group, with the vari-focal element disposed therein, the relay group configured to receive light fields created by the microdisplay and to create an intermediate 3D scene on the optical axis of the selected 3D scene, the relay group configured to tune the position along the optical axis of the intermediate 3D scene.

7. The method of claim 6, wherein the microdisplay is configured to create light fields of the 3D scene at a selected position along an optical axis of the system, and the relay group disposed on the optical axis at a location so the selected position is an optical conjugate of the relay group.

8. The method of any one of claims 6-7, wherein the Ini optics comprises eyepiece optics for imaging the intermediate 3D scene from the relay group into an exit pupil of the system for viewing by a user of the head-mounted display system.

9. A method for rendering light field images of a 3D scene in an HMD using an integral- imaging-based light field display, comprising:

providing integral imaging (Ini) optics including a microdisplay, the Ini optics having a central depth plane (CDP) associated therewith;

sampling the 3D scene using a simulated virtual array of cameras so that each camera

captures a respective portion of the 3D scene to create a plurality of elemental images, the elemental images collectively comprising image data for display on the microdisplay; and displaying the image data on the microdisplay.

10. The method of claim 2 or 9, wherein the Ini optics includes a microlens array of lenslets, and wherein the step of sampling the 3D scene comprises positioning each virtual camera such that each virtual camera location corresponds to the intersection of the chief ray of a corresponding lenslet of the microlens array with an exit pupil of the Ini optics.

11. The method of claim 2 or 10, wherein each simulated virtual camera’s viewing axis matches the chief ray direction of a corresponding lenslet seen through the Ini optics.

12. The method of any one of claims 2, 10, or 11, wherein the step of sampling the 3D scene comprises providing a simulated virtual array of sensors, each sensor in optical

communication with a corresponding selected one of the virtual cameras to provide a simulated virtual camera-sensor pair, wherein the separation between each camera-sensor pair is such that the field of view of each camera-sensor pair matches the field-of-view of a corresponding lenslet of the microlens array.

Description:
METHODS OF RENDERING LIGHT FIELD IMAGES FOR

INTEGRAL-IMAGING-BASED LIGHT FIELD DISPLAY

GOVERNMENT LICENSE RIGHTS

This invention was made with government support under Grant No. 1422653 awarded by the

NSF. The government has certain rights in the invention.

FIELD OF THE INVENTION

The present invention relates generally to the field of head-mounted displays, and more particularly, but not exclusively to head-mounted displays based on integral imaging (Ini).

BACKGROUND

Head-mounted displays (HMD), also commonly known as near-to-eye displays (NED) or head-worn displays (HWD), have gained significant interest in recent years and stimulated tremendous efforts to push the technology forward for a broad range of consumer applications. For instance, a lightweight optical see-through HMD (OST-HMD), which enables optical superposition of digital information onto a user’s direct view of the physical world and maintains see-through vision to the real-world, is one of the key enabling technologies to augmented reality (AR) applications. A wide field-of-view (FOV), immersive HMD, which immerses a user in computer-generated virtual world or a high-resolution video capture of a remote real-world, is a key enabling technology to virtual reality (VR) applications. HMDs find a myriad of applications in gaming, simulation and training, defense, education, and other fields.

Despite the high promises and the tremendous progress made recently toward the development of both VR and AR displays, minimizing visual discomfort involved in wearing HMDs for an extended period remains an unresolved challenge. One of the key contributing factors to visual discomfort is the vergence-accommodation conflicts (VAC) due to the lack of the ability to render correct focus cues, including accommodation cue and retinal image blur effects. The VAC problem in HMDs stems from the fact that the image source is mostly a 2D flat surface located at a fixed distance from the eye. Figure 1 shows a schematic layout of a typical monocular HMD, which mainly includes a 2D microdisplay as the image source and an eyepiece that magnifies the image rendered on the microdisplay and forms a virtual image appearing at a fixed distance from the eye. An OST-HMD requires an optical combiner (e.g. beamsplitter) placed in front of the eye to combine the optical paths of the virtual display and real scene. The conventional HMDs, whether monocular or binocular, see-through or immersive, lack the ability to render correct focus cues for the digital information which may appear at other distances than that corresponding to the virtual image plane. As a result, conventional HMDs fail to stimulate natural eye accommodation response and retinal blurry effects. The problem of lacking correct focus cues in HMDs causes several visual cue conflicts. For instance, a conventional stereoscopic HMD stimulates the perception of 3D space and shapes from a pair of two-dimensional (2D) perspective images, one for each eye, with binocular disparities and other pictorial depth cues of a 3D scene seen from two slightly different viewing positions. Therefore, conventional stereoscopic HMDs force an unnatural decoupling of the accommodation and convergence cues. The cue for the accommodation depth is dictated by the depth of the 2D image plane while the convergence depth of the 3D scene is dictated by the binocular disparities rendered by the image pair. The retinal image blurring cues for virtual objects rendered by the display is mismatched from those created by the natural scene. Many studies have provided strong supportive evidence that these conflicting visual cues related to incorrectly rendered focus cues in conventional HMDs may contribute to various visual artifacts and degraded visual performance.

Several approaches proposed previously may overcome the drawbacks of conventional stereoscopic displays, including volumetric displays, super-multi-view auto-stereoscopic displays, Integral-Imaging-based displays, holographic displays, multi -focal -plane displays, and computational multi-layer displays. Due to their enormous hardware complexity, many of these different display methods are not suitable for implementation in HMD systems. On the other hand, the multi-focal-plane display, integral-imaging, and computational multi-layer approaches are commonly referred to be light field displays and are suitable for head-mounted applications. Their use in HMDs is referred to as head-mounted light field displays.

Head-mounted light field displays render a true 3D scene by sampling either the projections of the 3D scene at different depths or the directions of the light rays apparently emitted by the 3D scene and viewed from different eye positions. They are capable of rendering correct or nearly correct focus cues and addressing the vergence-accommodation mismatch problem in conventional VR and AR displays. For instance, an integral imaging (Ini) based display reconstructs the light fields of a 3D scene by angularly sampling the directions of the light rays apparently emitted by the 3D scene and viewed from different eye positions. As illustrated in Fig. 2, a simple Ini-based display typically includes a display panel and a 2D array which can be a microlens array (MLA) or pinhole array. The display renders a set of 2D elemental images, each of which represents a different perspective of a 3D scene. The conical ray bundles emitted by the corresponding pixels in the elemental images intersect and integrally create the perception of a 3D scene that appears to emit light and occupy the 3D space. The Ini-based display using 2D arrays allows the reconstruction of a 3D shape with full-parallax information in both horizontal and vertical directions, which is its main difference from the conventional auto- stereoscopic displays with only horizontal parallax using one-dimensional parallax barriers or cylindrical lenticular lenses. Since its publication by Lippmann in 1908, the Ini-based technique has been widely explored for both capturing the light fields of real scenes and for its use in eyewear-free auto-stereoscopic displays. It has been known for its limitations in low lateral and longitudinal resolutions, narrow depth of field (DOF), and narrow view angle. Compared with all other non-stereoscopic 3D display techniques, the simple optical architecture of an Ini technique makes it attractive to integrate with HMD optical system and create a wearable light field display.

However, like other integral-imaging based display and imaging technologies, the current Ini-based HMD method suffers from several major limitations: (1) narrow field of view (<30° diagonally); (2) low lateral resolution (about 10 arc minutes in the visual space); (3) low longitudinal resolution (about 0.5 diopters in the visual space); (4) narrow depth of field (DOF) (about 1 diopter for a 10-arc minute resolution criteria); (5) limited eyebox for crosstalk-free viewing(<5mm); and (6) limited resolution of viewing angle (>20 arc minutes per viewing). These limitations not only create significant barriers for adopting the technologies as high- performance solutions, but also potentially undermine the effectiveness of the technology for addressing the accommodation-convergence discrepancy problem.

Thus, the present disclosure details methods, design and embodiment of a high-performance head-mounted light field display based on integral imaging that overcomes some aspects of the performance limits of the state of the art summarized above.

SUMMARY

In one of its aspects the present invention provides methods associated with a high- performance HMD based on integral imaging that offers high lateral and longitudinal resolution, large depth of field, cross-talk free eyebox, and increased viewing angle resolution. In this regard, the present invention may provide a method for rendering light field images of a 3D scene in an HMD using an integral-imaging-based light field display, comprising: providing integral imaging (Ini) optics having a vari-focal element and a microdisplay disposed in optical communication with the vari-focal element, the Ini optics having a central depth plane (CDP) associated therewith; displaying image data on the microdisplay, the image data comprising elemental images each representing a different perspective of the 3D scene; and setting the focal length of the vari-focal element to adjust the location of the CDP. The method may include sampling the 3D scene using a simulated virtual array of cameras so that each camera captures a respective portion of the 3D scene to create a plurality of elemental images; the elemental images may collectively comprise image data for display on the microdisplay. The Ini optics may be configured to create a virtual CDP that is the optical conjugate plane to the microdisplay in visual space. The 3D scene may have a depth of interest (DOI) through which the 3D scene extends along a visual axis, and may have an average DOI. The method may include setting the focal length of the vari-focal element so that the location of the virtual CDP coincides with the average DOI of the 3D scene.

The method may also include selecting a plurality of depths distributed along the visual axis within the DOI of the 3D scene, and for each selected depth of the plurality of depths, setting the focal length of the vari-focal element so that the location of each virtual CDP coincides with the selected depth to create a plurality of virtual CDP’s each virtual CDP coinciding with a respective selected depth of the plurality of depths. For each selected depth of the plurality of depths, the method may sequentially display on the microdisplay a portion of the 3D scene associated with each selected depth, and the step of setting the focal length of the vari-focal element may be synchronized to the timing of the sequential displaying on the microdisplay. The Ini optics may include a relay group, with the vari-focal element disposed therein, the relay group may be configured to receive light fields created by the microdisplay and to create an intermediate 3D scene on the optical axis of the selected 3D scene. The relay group may be configured to tune the position along the optical axis of the intermediate 3D scene. The microdisplay may be configured to create light fields of the 3D scene at a selected position along an optical axis of the system, and the relay group may be disposed on the optical axis at a location so the selected position is an optical conjugate of the relay group. The Ini optics may also include eyepiece optics for imaging the intermediate 3D scene from the relay group into an exit pupil of the system for viewing by a user of the head-mounted display system.

In a further of its aspects the present invention may provide a method for rendering light field images of a 3D scene in an HMD using an integral-imaging-based light field display, comprising: providing integral imaging (Ini) optics including a microdisplay, the Ini optics having a central depth plane (CDP) associated therewith; sampling the 3D scene using a simulated virtual array of cameras so that each camera captures a respective portion of the 3D scene to create a plurality of elemental images, the elemental images collectively comprising image data for display on the microdisplay; and displaying the image data on the microdisplay. The Ini optics may include a microlens array of lenslets, and the step of sampling the 3D scene may include positioning each virtual camera such that each virtual camera location corresponds to the intersection of the chief ray of a corresponding lenslet of the microlens array with an exit pupil of the Ini optics. Each simulated virtual camera’s viewing axis may match the chief ray direction of a corresponding lenslet seen through the Ini optics. In addition, the step of sampling the 3D scene may include providing a simulated virtual array of sensors, with each sensor in optical communication with a corresponding selected one of the virtual cameras to provide a simulated virtual camera-sensor pair. The separation between each camera-sensor pair may be such that the field of view of each camera-sensor pair matches the field-of-view of a corresponding lenslet of the microlens array.

BRIEF DESCRIPTION OF THE DRAWINGS

The foregoing summary and the following detailed description of exemplary embodiments of the present invention may be further understood when read in conjunction with the appended drawings, in which:

Figure 1 schematically illustrates a conventional, monocular HMD in which an eyepiece magnifies the image rendered on a microdisplay and forms a virtual display appearing at a fixed, far distance from the eye;

Figure 2 schematically illustrates a near-eye light field display based on integral imaging;

Figure 3A schematically illustrates an exemplary configuration of a high-performance Inl- based head-mounted light field display in accordance with the present invention;

Figure 3B schematically illustrates an exemplary configuration of a micro-Inl unit in accordance with the present invention; Figures 4A-4D schematically illustrate an exemplary configuration of a micro-Inl unit in accordance with the present invention constructed to provide ray direction control by using: an aperture array (Fig. 4A), programmable spatial light modulator (Fig. 4B), a display source with controllable directional emissions engine (Fig. 4C); and a backlight source with a spatial light modulator as an exemplary controllable directional emissions engine (Fig. 4D);

Figure 5 schematically illustrates an exemplary configuration of a relay group in accordance with the present invention with a VFE (vari-focal element) placed at a position conjugate to the exit pupil of the eyepiece;

Figures 6A-6D schematically illustrate an exemplary configuration of an optical see-through Inl-HMD design in accordance with the present invention using a freeform waveguide prism where part of the vari-focal relay group is incorporated into the eyepiece, with Fig.6A showing the display path layout, Fig. 6B showing the see-through view layout, Fig. 6C showing a segmented rear surface of the waveguide prism for extended see-through view, and Fig. 6D showing a front view of the rear surface of the waveguide prism;

Figures 7A, 7B schematically illustrate an exemplary configuration of 2D optical layout of an Inl-HMD design configuration in accordance with the present invention, with Fig. 7A showing the light field display path and Fig. 7B the see-through path;

Figures 8A, 8B illustrate MTF (modulation transfer function) plots for the reconstruction central depth plane (CDP) depth of 3 diopters for fields on-axis (Fig. 8A) and for fields for the furthest MLA (micro lens array) element near the edge of the MLA (Fig. 8B);

Figures 9A, 9B illustrate MTF plots for the reconstruction CDP depth of 2 diopters for fields on-axis to the MLA (Fig. 9A) and fields for the furthest MLA element near the edge of the MLA (Fig. 9B);

Figures 10A, 10B illustrate MTF plots for the reconstruction CDP depth of 0 diopters for fields on-axis to the MLA (Fig. 10A) and for fields for the furthest MLA element near the edge of the MLA (Fig. 10B);

Figures 11 A, 11B illustrate MTF plots for the reconstruction points shifted away from CDP by 0.25 diopters for fields on-axis to the MLA (Fig. 11 A) and for fields for the furthest MLA element near the edge of the MLA (Fig. 11B); Figures 12A, 12B illustrate MTF plots for the reconstruction points shifted away from CDP by 0.5 diopters for fields on-axis to the MLA (Fig. 12A) and for fields for the furthest MLA element near the edge of the MLA (Fig. 12B);

Figures 13 A, 13B illustrate MTF plots for the reconstruction points shifted away from CDP by 0.75 diopters for fields on-axis to the MLA (Fig. 13A) and for fields for the furthest MLA element near the edge of the MLA (Fig. 13B);

Figures 14A, 14B illustrate MTF plots for the reconstruction points shifted away from CDP by 1 diopter for fields on-axis to the MLA (Fig. 14 A) and for fields for the furthest MLA element near the edge of the MLA (Fig. 14B);

Figure 15 illustrates the MTF for the see-through path FOV 65°x40°;

Figure 16 schematically illustrates a method for rendering the light field of 3D virtual scene in a fixed depth mode in accordance with the present invention;

Figure 17A illustrates an array of the elemental images (Els) on a microdisplay;

Figures 17B-17D illustrate captured images of both real and virtual targets through an Inl- HMD prototype fabricated in accordance with the present invention operated in a fixed -depth mode of the Els of Fig. 17A, with the camera focusing on 1 diopter (Fig. 17B), 0.5diopters (Fig. 17C), and 3 diopters (Fig. 17D), respectively;

Figure 18 schematically illustrates a method for rendering the light field of 3D virtual scene in a vari-depth mode in accordance with the present invention;

Figures 19A, 19B illustrate captured images of both real and virtual targets through an Inl- HMD prototype fabricated in accordance with the present invention operated in a vari-depth mode with the virtual CDP set at 3 diopters while the camera focusing on 3 diopters (Fig. 19 A) and 0.5 diopters (Fig. 19B), respectively;

Figure 20 schematically illustrates a method for rendering the light field of 3D virtual scene in a multi-depth mode; and

Figures 21 A, 21B illustrate captured images of both real and virtual targets through an Inl- HMD prototype fabricated in accordance with the present invention operated in a multi-depth mode with the virtual CDP set at 3 diopters while the camera focusing on 3 diopters (Fig. 21 A) and 0.5 diopters (Fig. 21B), respectively. PET ATT, ED DESCRIPTION

Referring now to the figures, wherein like elements are numbered alike throughout, as shown in Fig. 3 A, a HMD system 100 in accordance with the present invention may include three key subsystems: I) a microscopic Ini unit (micro-Inl) 130, II) a relay group 120 with a vari-focal element (VFE) 122 disposed therein for receiving the light fields from the Ini unit 130, and III) eyepiece optics 110 for receiving the tuned intermediate 3D scene from the relay group 120. As illustrated in Fig. 3B, the micro-Inl unit 130 can reproduce the full-parallax light fields of a 3D scene seen from a constrained viewing zone, where the full-parallax light fields offer the change of view perspectives of a 3D scene from both horizontal and vertical viewing directions. The constrained viewing zone optically corresponds to limiting the aperture of the micro-Inl unit 130, and the constrained viewing zone is optically conjugate to the exit pupil of the display system 100 where a viewer’s eye is placed to view the reconstructed 3D scene. The relay group 120 creates an intermediate image of the 3D scene reconstructed by the micro-Inl unit 130 with a tunable position of its central depth plane (CDP). Depending on the magnification power of the eyepiece 110, the position of the CDP may be tunable in the range from about 0.5mm to as large as hundreds of millimeters to create the perception of a 3D scene with a large depth range spanning from the optical infinity (0 diopter) to as close as 20cm (5 diopters). The relay group 120 may also facilitate the flip of the concavity of the reconstructed 3D scene AOB. The eyepiece optics 110 reimages the tunable 3D light fields into a viewer’s eye and enlarges the tunable depth range of the 3D light fields into a large depth volume spacing from meters far to as close as a few centimeters. A see-through unit (not shown), which may be optics with a beamsplitter function, may optically communicate with the eyepiece optics 110 to optically enable non-obtrusive view of a real-world scene if a see-through view is desired. The micro-Inl unit 130 of Fig. 3 A, as further illustrated in Fig. 3B, may include a high-resolution microdisplay and a micro-lens array (MLA) 132. The focal length of the lenslets 133 in the MLA 132 is denoted as f MLA and the gap between the microdisplay 134 and the MLA 132 is noted as g. A set of 2D elemental images, each representing a different perspective of a 3D scene AOB, may be displayed on the high-resolution microdisplay 134. Through the MLA 132, each elemental image works as a spatially-incoherent object and the conical ray bundles emitted by the pixels in the elemental images intersect and integrally create the perception of a 3D scene that appears to emit light and occupy the 3D space. The central depth plane (CDP) of the reconstructed miniature scene, with a depth range of zo, is located by the distance l cdp measured from the MLA 132. Such an Ini system 130 allows the reconstruction of a 3D surface shape AOB with parallax information in both horizontal and vertical directions. The light field of the reconstructed 3D scene (i.e., the curve AOB in Fig. 3B) may be optically coupled into eyepiece optics 110 via the relay group 120 for viewing by a user. In a resolution priority Ini system (f MLA ¹ g), the central depth plane CDP of the reconstructed 3D scene is optically conjugate to the microdisplay 134 and its location is given by

lcdp — d^MLA> (1)

Where M MLA is the magnification of the micro-Inl unit 130, which may be expressed by

As shown in Figs. 3A, 4A, optionally, an aperture array 136, including a group of ray- limiting apertures that matches the pitch of the MLA 132, may be inserted between the microdisplay 134 and MLA 132. The small aperture corresponding to each microlens 133 allows rays within the designed viewing window to propagate through the optics and reach the eyebox while blocking unwanted rays from reaching an adjacent microlens 133 or while blocking rays from neighboring elemental images to reach a microlens 133. For instance, the black zone between the aperture Al and A2 blocks the dashed rays originated from point Pl from reaching the MLA2 adjacent to the lenslet MLA1. These blocked rays are typically the main source of view cross-talk and ghost images observed in an Ini display system. The distance from the microdisplay 134 to the aperture array 136 is denoted as g a and the diameter of aperture opening is denoted as p a , which may be constrained by

Pei

(3)

Pei+Pmia

( Sa-max ~ Sa )

ei (4)

Sa-max

Where g a -max ar| d p a -max are the maximum allowable gap and aperture size, respectively, p ei is the dimension of the elemental image, and p mia is the pitch of the MLA 132.

One drawback in using an aperture array 136 with a fixed aperture size is that it can partially block rays for pixels located near the edge of each elemental images if the size of the elemental image changes. As illustrated in Figure 4A, a small part of the rays from point Pl which are supposed to propagate through lenslet MLA1 are blocked by the black zone between aperture Al and aperture A2, causing vignetting-like effects such that viewer may observe reduction of image brightness for points near the edge of each elemental images. Figure 4B shows an alternative configuration to that of Fig. 4A in which the aperture array 136 is replaced by a programmable spatial light modulator (SLM) 135 so that the size and shape of each aperture can be dynamically adapted to avoid partially blocking desired rays. Figure 4C shows another embodiment of a micro-Inl unit in accordance with the present invention in which the microdisplay 134 and aperture array 136 are replaced by a display source 131 with controllable directional emissions, where the light emission direction can be controlled precisely so that the rays from each pixel will only reach their corresponding MLA lenslet 133. Figure 4D demonstrates one possible configuration of such display source 131 where a spatial light modulator 135 is inserted between a backlight source 138 with non-direction emission and non-self-emissive microdisplay 137. The spatial light modulator 135 may be set to program and control the cone angle of the rays that illuminate the microdisplay 137 and reach the MLA 132.

A conventional Ini-based display system can typically suffer from a limited depth of field (DOF) due to the rapid degradation of spatial resolution as the depths of 3D reconstruction points shift away from that of the CDP. For instance, the 3D scene volume may need to be limited to less than 0.5 diopters in order to maintain a spatial resolution of 3 arc minutes or better in the visual space. In order to render a much larger 3D scene volume while maintaining a high spatial resolution, such as in the exemplary configuration of Fig. 3A, a relay group 120 with an electronically-controlled vari-focal element 122 sandwiched inside is inserted between the micro-Inl 130 and the eyepiece 110. Exemplary VFE’s 122 include liquid lenses, liquid crystal lenses, deformable mirrors, or any other tunable optical technology, such as electrically tunable optical technology. By dynamically controlling the optical power, cpR, of the relay group 120 by applying different voltages to the VFE 122, the relay group 120 forms an intermediate image A’O’B’ of the reconstructed miniature 3D scene created by the micro-Inl 130. The central depth position CDP of the relayed intermediate scene is tunable axially (along the optical axis) with respect to the eyepiece 110. As a result, the depth volume of the magnified 3D virtual scene by the eyepiece 110 can be shifted axially from very close (e.g. 5 diopters) to very far (e.g. 0 diopter) while maintaining high lateral and longitudinal resolutions.

Figure 5 schematically illustrates an exemplary configuration of the vari-focal relay group 120, such as the relay group 120 of Fig. 3A, including a front lens group“Front Relay” 126 adjacent to the micro-Inl unit 130, VFE optics 122 located in the middle functioning as the system stop, and rear lens group“Rear Relay” 124 adjacent to the eyepiece 110. The compound power, cpR, of the relay group 120 is given by

Where cpi, cp E, and qu are the optical power of the front lens group 126, VFE 122, and the rear lens group 124, respectively ti and t2 are the spaces between the front lens group 126 and VFE 122 and between the VFE 122 and the rear lens group 124. zo is the axial distance between the front lens group and the 3D scene reconstructed by the micro-Inl unit 130. The axial position of the relayed intermediate scene is given by

! _ _ 1 _

Z0 _ (ΐ- z0 f 1 )-[z 0 + (ΐ- z0 f 1 1 ]f nίq _ ( 5 )

[zo + (ΐ-zofi)ί 1 ]+{(ΐ-zofi)-[zo + (ΐ-zofi)ί 1 ]f nίq 2 y2

The lateral magnification of the vari-focal relay system is given by

MR =

(1— z 0 fi)— [z 0 tTl— ZoYi^iIfnίe— {[¾ + (!— z 0 <Pi)ti] + [(l— z 0 <Pi)— [z 0 + (l— Zo<Pi)ti]cp v fe]t 2 }<P2 ^

Assuming cp e is the optical power of the eyepiece 110 and ZRCDP is the distance from the relayed CDP to the eyepiece 110, the apparent CDP position of the reconstructed 3D virtual scene through the eyepiece 110 is given by

The lateral magnification of the entire system through the eyepiece 110 is given by

The field of view (FOV) of the entire system through the eyepiece 110 is given by, FOV =

Where t 3 is the spacing between the eyepiece 110 and rear relay lens 124; z xp is the spacing between the exit pupil and the eyepiece 110; h 0 is the image height of the reconstructed scene, and we further define u vfe = [(l - z xp cp e ) - (z xp + (l - z cr f q 3 )f 2 ] , and h vfe =

(z xp + (l - ¾ R f q 3 )f 2 )]Ϊ 2 ·

When the VFE 122 is set to be an optical conjugate to the exit pupil of the eyepiece 110 (i.e. h Vf e=0) where the entrance pupil of the eye is placed to view the display 134, we have h vfe =0 and the FOV is independent of the optical power of the VFE 122. The equation in Eq. (9) is simplified into:

uvfeti + [Uvfe u v feti<Pi]zo

As illustrated in Fig. 5, a preferred embodiment of the vari-focal relay group 120 is the placement of the VFE 122 at the back focal length of the front relay group 26 (i.e. ti=l/cpi) to make the VFE 122 an optical conjugate to the exit pupil of the eyepiece 110 (i.e. h vfe =0). With this preferred embodiment, the compound power, cpR, of the relay group 120 given by Eq. (4) is simplified into:

f k = fi - fif 2 ί 2 (11)

The lateral magnification of the vari-focal relay system given by Eq. (6) is simplified into

And so does the lateral magnification of the entire system given by Eq. (8).

When ti=l/(pi and h vfe =0,the FOV of the system is further simplified into

As demonstrated by Eqs. (10) through (13), the careful position of the VFE 122 in the preferred manner ensures that the compound optical power of the relay group 120 is maintained constant, independent of the optical power of the VFE 122 due to constant chief ray directions owing to the property of object-space tel ecentri city. As further demonstrated by Eq. (13), the subtended field angle of the display through the eyepiece 110 is further maintained constant, independent of the optical power of the VFE 122. Maintaining a constant optical power for the relay group 120 helps the virtually reconstructed 3D scene achieve constant field of view regardless of the focal depths of the CDP. Therefore a much larger volume of a 3D scene could be visually perceived without seams or artifacts in a gaze-contingent or time-multiplexing mode. It is worth noting that the lateral magnification of the relay group 120 given by Eq. (12) can be further maintained constant if ΐ 2 =1/f2 is satisfied, which makes the vari-focal relay group 120 a double-telecentric system.

The eyepiece 110 in Fig. 3A can take many different forms. For instance, to achieve a compact optical design of an optical see-through HMD, a wedge-shaped freeform prism can be adopted, through which the 3D scene reconstructed by the micro-Inl unit 130 and relay group 120 is magnified and viewed. To enable see-through capability for AR systems, a freeform corrector lens with one of the surfaces coated with beamsplitter coating can be attached to the freeform prism eyepiece to correct the viewing axis deviation and undesirable aberrations introduced by the freeform prism to the real-world scene.

In another aspect of the present invention, part of the relay group 120 may be incorporated into the eyepiece optics 110, such as freeform eyepiece, such that the tunable intermediate 3D scene is formed inside the freeform eyepiece. In such a context, the eyepiece may be a wedge- shaped freeform waveguide prism, for example. Figure 6A schematically illustrates the concept of a freeform waveguide-like prism 850 formed by multiple freeform optical surfaces. The exit pupil is located where the use’s eye is placed to view the magnified 3D scene. In the design, part of a traditional relay group 220 following the VFE 122 is incorporated into the prism 850 and fulfilled by the top portion 851 of the freeform waveguide prism 850 contained within the box labeled“Relay Group with VFE.” A light ray emitted from a 3D point (e.g. A) is first refracted by a closest optical element 126 of the relay group 220 and transmitted into the prism 850, followed by a reflection by one or multiple freeform surfaces to create an intermediate image (e.g. A’). The axial position of the intermediate image (e.g. A’) is tunable by the VFE 122. Multiple consecutive reflections by the subsequent surfaces and a final refraction through the exit surface 855allow the ray reaching the exit pupil of the system. Multiple bundles of rays from different elemental images may exist, but do so apparently from the same object point, each of which bundles represents a different view of the object, impinging on different locations of the exit pupil. These ray bundles integrally reconstruct a virtual 3D point (e.g.“A”) located in front of the eye. Rather than requiring multiple optical elements, the optical path is naturally folded within a multi-surface prism 850, which helps reduce the overall volume and weight of the optics substantially when compared with designs using rotationally symmetric elements. Compared with a design using a traditional wedge-shaped 3-surface prism, the waveguide-like eyepiece design incorporates part of the relay function, enabling a much more compact system than combining a standalone relay group 120 with a 3-surface prism. Besides the advantage of compactness, the waveguide-like multi-fold eyepiece design offers a much more favorable form factor, because it enables the ability to fold the remaining relay group and micro-Inl unit horizontally to the temple sides. The multiple folding not only yields a much more weight- balanced system, but also enables a substantially larger see-through FOV than using a wedge- shaped prism.

To enable see-through capability for AR systems, the bottom part 853 of the rear surface, marked as the eyepiece portion, of the prism 850 in Fig. 6A can be coated as a beamsplitting mirror, and a freeform corrector lens 840 including at least two freeform optical surfaces, may be attached to the rear surface of the prism 850 to correct the viewing axis deviation and undesirable aberrations introduced by the freeform prism 850 to the real-world scene. The see-through schematic layout is shown in Fig. 6B. The rays from the virtual light field are reflected by the rear surface of the prism 850 while the rays from a real-world scene are transmitted through the freeform corrector lens 840 and prism 850. The front surface of the freeform corrector lens 840 matches the shape of the rear surface of the prism 850. The back surface of the freeform corrector lens 840 may be optimized to minimize the shift and distortion introduced to the rays from a real-world scene when the lens is combined with the prism 850. The additional corrector lens“compensator” does not noticeably increase the footprint and weight of the overall system.

In another aspect of the present invention, the bottom part 853 of the rear surface, marked as the eyepiece portion, of the prism 850 in Fig. 6A may be divided into two segments, the segment 853-1 and the segment 853-2. As schematically illustrated in Fig. 6C, the segment of 853-1 may be a reflective or partial reflective surface which receives the light fields generated by the micro-

Inl unit. A beamsplitting mirror coating on the segment of 853-1 also allows the transmission of the light rays from a real-world scene. The segment 853-2 is a transmissive or semi-transmissive surface which only receives the light rays from a real-world scene, while it does not receive the light fields generated by the micro-Inl unit 130. Fig. 6D schematically illustrates a front view of the rear surface of the prism 850. The two surface segments, 853-1 and 853-2, intersect at an upper boundary of the aperture window required to receive the reconstructed 3D light fields by the micro-Inl unit 130, and they may be made by two separate freeform surfaces. The division of the bottom part of the rear surface 853 into two separate segments 853-1, 853-2 with different light paths provides the ability to substantially enlarge the FOV of the see-through view beyond the FOV of the display path without being subject to the constraints of the virtual display path. As shown in Fig. 6C, a freeform corrector lens 840 may be attached to the rear surface of the prism 850 to correct the viewing axis deviation and undesirable aberrations introduced by the freeform prism 850 to the real-world scene. The rays from the virtual light field are reflected by the segment 853-1 of the rear surface of the prism 850 while the rays from a real-world scene are transmitted through both the segments 853-1 and 853-2 of the prism 850 and the freeform corrector lens 840. The surface segment 853-2 may be optimized to minimize visual artifacts of see-through view when it is combined with the freeform corrector lens 840. The front surface of the freeform corrector lens 840 matches the shape of the surface segments 853-1 and 853-2 of the prism 850. The back surface of the freeform corrector lens 840 may be optimized to minimize the shift and distortion introduced to the rays from a real-world scene when the freeform corrector lens 840 is combined with the prism 850.

In accordance with yet another aspect of the present invention, Figure 7A schematically illustrates an optical design of a physical system 700 that embodies the conceptual system of Fig. 6A. Figure 7A illustrates the 2D optical layout of the light field display path, and Fig. 7B shows the optical layout of the see-through path. The optical system 700 of the light field display includes a micro-Inl unit, a relay group with VFE, and a freeform waveguide. A part of the relay group may be incorporated into the waveguide. The Micro-Inl unit may include a microdisplay SO, a pinhole array Sl, and a microlens array S2. The relay group may include four lenses, a commercially available VFE (Electrical Lens EL 10-30 by Optotune Inc.), and two freeform surfaces (Surface S19 and S20). The freeform waveguide prism 900 may be formed by multiple freeform optical surfaces which are labeled as S19, S20, S21, and S22, respectively. In the design, part of a traditional relay group following the VFE may be incorporated into the prism 900 and fulfilled by the Surface S19 and S20. A light ray emitted from a 3D point (e.g. A) is first refracted by the surface S19 of the prism 900, followed by a reflection by the surface S20 to create an intermediate image (e.g. A’). The axial position of the intermediate image (e.g. A’) is tunable by the VFE. Two more consecutive reflections by the surfaces S2T and S22-1 and a final refraction through the surface S21 allow the ray to reach the exit pupil of the system 700. There exist multiple bundles of rays from different elemental images but apparently from the same object point, each of which represents a different view of the object, impinging on different locations of the exit pupil. These ray bundles integrally reconstruct a virtual 3D point located in front of the eye. The rays reflected by the Surface S21’ of the waveguide are required to satisfy the condition of total internal reflection. The rear surfaces S22-1, S22-2 of the prism 900 may be coated with a mirror coating for building an immersive HMD system which blocks the view of the real-world scene. Alternatively, the surface S22-1 may be coated with a beamsplitting coating if optical see-through capability is desired using the auxiliary lens, as shown in Fig. 7B.

It should be noted that in the design disclosed hereby the Z-axis is along the viewing direction, the Y-axis is parallel to the horizontal direction aligning with interpupilary direction, and the X-axis is in the vertical direction aligning with the head orientation. As a result, the overall waveguide system is symmetric about the horizontal (YOZ) plane, and the optical surfaces (S19, S20, S21, and S22) are decentered along the horizontal Y-axis and rotated about the vertical X-axis. The optical path is folded in the horizontal YOZ plane. This arrangement allows the micro-Inl unit and the vari-focal relay group to be mounted on the temple side of the user’s head, resulting in a balanced and ergonomic system packaging.

Table 1 highlights some of the key performance specifications for the system 700 of Fig. 7A. The system 700 offers the ability to render the true 3D light field of a 3D scene which subtends a diagonal FOV of 35° and achieves an optical resolution as high as 2 arc minutes per pixel in the visual space. Furthermore, the system 700 offers a large depth range, tunable from 0 to 5 diopters, with a high longitudinal resolution of about 0.1 diopters for a monocular display. Moreover, the system 700 achieves a high view density of about 0.5/mm 2 , where the view density, s, is defined as the number of unique views per unit area on the exit pupil, given by:

N

s = t;—

Acr

where N is the total number of views and AXP is the area of the exit pupil of the display system. A view density of 0.5/mm 2 is equivalent to a viewing angle resolution of approximately 1 arc minute for objects at distance of 0.2 diopters. The exit pupil diameter for crosstalk-free viewing, also known as the eyebox of the display, is about 6mm. In this embodiment, the exit pupil diameter is limited by the aperture size of the commercial VFE and it can be increased if another larger-aperture VFE is adopted. Finally, the system offers a large see-through FOV, greater than 65° horizontally and 40° vertically. The microdisplay utilized in our prototype is a 0.7” organic light emitting display (OLED) with an 8pm color pixel and pixel resolution of 1920x1080 (ECX335A by Sony). The optics design itself, however, is able to support OLED panels of different dimensions or other type of microdisplays such as liquid crystal displays that have a color pixel size greater than 6 pm.

Table 1— First-order system specifications

An exemplary implementation of the system 700 of Fig. 7A is provided, Tables 2 through 5, in form of the optical surface data. Table 2 summarizes the basic parameters of the display path (units: mm). Tables 3 through 5 provide the optimized coefficients defining the non-spherical optical surfaces.

Table 2— Optical specifications of the Inl-HMD display path

A high resolution microdisplay with pixels as small as 6pm is adopted to achieve a high resolution virtual reconstructed 3D image. To achieve such high-resolution imaging for the micro-Inl unit, a microlens array (MLA) formed by aspherical surfaces may specifically be designed. Each of the aspherical surfaces of the MLA may be described as,

where z is the sag of the surface measured along the z-axis of a local x, y, z coordinate system, c is the vertex curvature, r is the radial distance, k is the conic constant, A through E are the 4th, 6th, 8th, lOth and l2th order deformation coefficients, respectively. The material of the MLA is PMMA. Table 3 provides the coefficients for the surfaces Sl and S2.

Table 3— Aspherical surface definitions for microlens array (MLA)

To enable enlarged see-through FOV, the freeform waveguide prism 900 may be formed by five freeform surfaces, labeled as surface S19, S20, S21/S2E, S22-1, and S22-2, respectively. The freeform corrector lens may be formed by two freeform surfaces, in which the front surface shares the same surface specifications as the surfaces S22-1 and S22-2 of the waveguide prism 900 and the rear surface is denoted as surface S23. The surface segment of S22-1 is a reflective or partial reflective surface which receives the light fields generated by the micro-Inl unit. A beamsplitting mirror coating on the segment of S22-1 also allows the transmission of the light rays from a real-world scene for see-through capability. The surface segment S22-2 is a transmissive or semi-transmissive surface which only receives the light rays from a real-world scene, while it does not receive the light fields generated by the micro-Inl unit.

The freeform surfaces, including S19, S20, S21/S2E, S22-1, and S23 may be described mathematically as

where z is the sag of the free-form surface measured along the z-axis of a local x, y, z coordinate system, c is the vertex curvature (CUY), r is the radial distance, k is the conic constant, and Q is the coefficient for x m y n . The material for both the waveguide prism and compensation lens is PMMA. Tables 4 through 8 provide the coefficients for the surfaces S19 through S21, S22-1, and S23, respectively, and Table 9 provides the surface references of each optical surface.

During the design process, the specifications for the Surface segment S22-1 were obtained after the optimization of the light field display path through the prism 900 composed of the micro-Inl unit, the relay lens group, and the surfaces S19. S20, S21/2F, and S22-1. The required aperture dimensions of Surfaces S20 and S22-1 were determined first for the light field display path. Then Surfaces S20, S21 and S22-1 were imported into 3D modeling software such as Solidworks ® from which the Surface S22-2 was created. The shape of the Surface S22-2 was created in the modeling software by satisfying the following requirements: (1) it intersects with Surface S22-1 along or above the upper boundary line of the required aperture for surface S22-1 defined by the display path; (2) along the intersection line between the surface S22-2 and S22-2, the surface slopes at the intersection points on the surface S22-2 approximately match, if not equal, with those corresponding points on the surface S22-1 to ensure the two surfaces to appear to be nearly continuous, which minimizes visual artifacts to the see-through view when it is combined with a matching freeform corrector lens; (3) the Surface S22-2 intersects with the surface S20 along or below the lower boundary line of the required aperture for surface S20, defined by the display path; and (4) the overall thickness between the surface S21 and S22-2 is minimized. Finally, a freeform shape of the Surface S22-2 is obtained in the 3D modeling software which is combined with the surfaces S19, S20, S21/2T, and S22-1 to create an enclosed freeform waveguide prism. Fig. 7B demonstrated a substantially enlarged see-through FOV through the method described above.

Table 4— Surface definition for freeform surface S19

Table 5— Surface definition for freeform surface S20

Table 6— Surface definition for freeform surface S21/S21’

Table 7— Surface definition for freeform surface S22-1

Table 8— Surface definition for freeform surface S23

Table 9— Definition of the local surface references in the global coordinate system

During the design process, three representative wavelengths, 465nm, 550nm, and 630nm were selected which correspond to the peak emission spectra of the blue, green and red emitters within the selected OLED microdisplay. A total of 21 lenslets in the MLA were sampled with each representing 9 element image points, which added up a total of 189 field samples. To evaluate the image quality, an ideal lens with the same power as the eyepiece is placed at the exit pupil of the system (viewing window), which resulted in a cut-off frequency of 20.83 lp/mm for the final image, limited by the pixel size of the microdisplay. The optical performance of the designed system was assessed at representative field angles for the three design wavelengths. By changing the power of the tunable lens VFE, the central depth plane could be shifted axially in a large range, for example, from 0 to 3 diopters, without noticeable degeneration of optical performance. Figures 8 through 10 plot the polychromatic modulation transfer function (MTF) for points reconstructed on the CDP set at the depth of 3, 1, and 0 diopters, respectively. For each CDP position, two sets of MTFs were plotted, one for fields corresponding to the on-axis MLA and one for fields correspond to the furthest MLA near the edge.

On the other hand, it is equally important to assess how the image quality of a 3D reconstruction point degrades when the reconstructed image is shifted away from the central depth plane for a specific tunable state. This can be evaluated by shifting the central depth plane a small amount of distance without changing the power of the tunable lens. Figures 11 through 14 plot the polychromatic MTF for reconstructed points shifted away from the CDP by 0.25, 0.5, 0.75, and 1 diopters, respectively. For each depth, two sets of MTFs were plotted, one for fields corresponding to the on-axis MLA and one for fields corresponding to the furthest MLA near the edge. Figure 15 plots the polychromatic MTF for the 65°x40° FOV. Across the entire the FOV, the see-through path achieved an average MTF value of over 50% at 30 cycles/degree frequency, corresponding to 20/20 normal vision, and nearly 20% at 60 cycles/degree frequency, corresponding to 20/10 vision or 0.5 arc minute of visual acuity.

A prototype system (“Inl-HMD prototype”) was constructed of the Inl-HMD 700 of Fig. 7A and Tables 1-9 and associated text.

In a further of its aspects, the present invention may provide methods for rendering light field images for an integral-imaging-based light field display. As one exemplary method, the flowchart of Fig. 16 illustrates rendering of a light field of a 3D virtual scene 1603, where the Inl-HMD optics 1600 creates a virtual central depth plane (CDP) 1601 at a fixed depth (ZCDP) from the VIEWER measured in diopters, referred to as a fixed-depth mode light field display. The virtual CDP 1601 is the optical conjugate plane of the microdisplay 1601 in the visual space. Usually the highest contrast and resolution of the 3D light field could be reconstructed for 3D objects located at the depth of the CDP 1609.

To render the light field of a 3D target scene 1603, the exemplary fixed-depth mode method of the present invention may start with determining the depth of the virtual CDP 1601 of the Inl- HMD optics 1600 with respect to the eye position of the VIEWER. A virtual camera array 1604 composed of I by J pinhole cameras may then be simulated. Each of the virtual cameras in the array 1604 may be positioned in the simulation in such a way that each location corresponds to the intersection of the chief ray direction of a corresponding lenslet of the microlens array (MLA) 1606 with the exit pupil of the Inl-HMD optics 1600, and each virtual camera’s viewing axis matches the chief ray direction of the corresponding lenslet seen through the Inl-HMD optics 1600. Corresponding to the simulated virtual camera array 1604 is a simulated virtual camera sensor array 1605 composed of I by J virtual sensors. Each of the virtual sensors may have a pixel resolution of K by L. The projection plane 1613 of the virtual cameras is set to coincide with the depth of the virtual CDP 1601 of the Inl-HMD optics 1600, and the separation between the simulated virtual camera array 1604 and the sensor array 1605, known as the camera equivalent focal length (EFL), f, is set such that the field of view (FOV) of each camera-sensor pair matches the FOV of each lenslet of the MLA 1606. A virtual 3D scene 1603 may be computed using the simulated virtual camera array 1604 as its reference. For the convenience of reference, hereafter the depths, Z, of 3D scene objects measured in diopters are referenced with respect to the VIEWER or equivalently to the simulated virtual camera array 1604. Each pair of the virtual cameras 1604 and sensors 1605 may correspond to a computed (rendered) 2D elemental image (El) of the 3D light field of the 3D scene, representing a slightly different perspective of the 3D scene seen by the simulated virtual cameras 1604. These Els may then be mosaicked to create a full-resolution light field image mosaic 1607 of I*K by J*L pixels for the microdisplay 1602. (It should be noted that element 1603, 1604, 1605, 1607 are non-physical elements that are computationally simulated to provide data to be delivered to the physical display 1602.) The full-resolution image 1607 may be displayed via the microdisplay 1602 of the Inl-HMD optics 1600. Through the Inl-HMD optics 1600, a reconstructed virtual 3D scene 1608 may be reconstructed for a VIEWER to view at the depth Z. For instance, in the present exemplary implementation, following the conventional rendering pipeline of 3D computer graphics (such as, F.S. Hill, r., Computer Graphics Using OpenGL, 2 nd Edition, Publisher: Prentice Hall, 1990), an array of 15x9 elemental images of a 3D target scene 1603 are simulated, each of which consists of 125x125 color pixels. These Els may be mosaicked to create the full- resolution image of 1920x1080 pixels for the microdisplay 1602.

Using the Inl-HMD prototype, a demonstration was performed by fixing the optical power of the tunable lens 122, S10-S16 so that the CDP 1609 of the display system 700, 1600 was set at a fixed distance of 1 diopter from the VIEWER, which simulates the display properties of a conventional Ini-based HMD. (For purposes of the instant fixed-depth mode method a tunable lens is not required, and so its optical power was fixed.) To demonstrate the optical performance of the light field optics 1600 in a fixed-depth CDP mode, the virtual 3D target scene 1603 having three depth planes located at 3, 1 and 0.5 diopters away from the viewer or the exit pupil of the Inl-HMD optics was created, Fig. 17A. On each depth plane three groups of Snellen letter E’s with different spatial resolutions (3, 6, and 10 arcmins for the individual strokes or gaps of the letters) and orientations (horizontal and vertical) as well as the depth indicators (‘3D’,‘1D’ and ‘0.5D’) were rendered. The images were rendered using the method described above in connection with Fig. 16. Figure 17A shows the exemplary mosaic 1607 of 1 lx 5 Els of the virtual 3D scene 1601 generated for the microdisplay 1602, where the virtual CDP 1601 was set at 1 diopter. For qualitative assessment of focus cues, three spoke resolution targets were physically placed at the corresponding depths of three depth planes of the virtual 3D scene 1603. A camera (not shown) with a 2/3” color sensor of 2448 by 2048 pixels and a l6mm lens was used in the place of the VIEWER. The camera system overall yielded a spatial resolution of 0.75 arcmin per pixel, which was substantially better than that of the display optics 1600. The entrance pupil diameter of camera lens was set to about 4mm such that it is similar to that of the human eye. Figure 17B shows the captured images of the reconstructed virtual 3D scene overlaying with the real-world targets where the camera was focusing on 1 diopter. It can be observed that only the targets, both the real (indicated by the arrow) and virtual (indicated by the box) ones, located at the same depth of the focus plane of the camera are correctly and clearly resolved, which suggests the ability of the Ini-based HMD 700, 1600 to render correct focus cues to the VIEWER. The ability to resolve the smallest Snellen letters on the top row of the 1 diopter targets further suggests the spatial resolution of the prototype matches with the designed nominal resolution of 3 arcmins. In this configuration of fixed lens focus, it can be further observed that the Els of the virtual targets at the depths (e.g. 3D and 0.5D) different from the focus plane of the camera do not converge properly, causing multiple copies of the letters being captured in Fig. 17B. These targets can properly converge when the camera focus is adjusted to focus on their corresponding depths, as demonstrated in Figs. 17C and 17D, which show the captured images of the same virtual and real-world scene with camera being focused at 0.5 and 3 diopters, respectively. The targets corresponding to the camera focus depth were marked by a box, respectively. However, alike a traditional Ini-based HMD, the image contrast and resolution of the targets reconstructed at the depth plane other than the CDP can only maintain in a relatively short, limited DOF and degrade severely beyond that, even though the Els of these targets converge correctly and located at the same depth as the focus plane of the camera. For instance, the captured images in Fig. 17C can still resolve the letters corresponding up to 6 arcmins while that in Fig. 17D can only resolve the letters corresponding to 10 arcmins and the Els start to converge improperly.

With the assistance of tunable lens 1811, 122 (Figs. 18, 7A) in accordance with the present invention, the depth of the CDP 1809 can be dynamically adjusted. This capability allows the system 1800 of the present invention to operate in two different modes: vari-depth mode (Figs. 18, 19A-19B) and time-multiplexed multi-depth mode (Figs. 20, 21A-21B). In the vari-depth mode, the depth of the CDP 1809 may be adaptively varied according to the average depth of the displayed contents or the depth of interest. In multi-depth mode, the power of the tunable lens

1810, 122 may be rapidly switched among several states corresponding to several discrete CDP depths, while in synchronization the light field rendering is updated at the same speed such that the contents of different depths are time-multiplexed and viewed as an extended volume if the switching occurs at flickering-free rate.

The method for rendering the light field of 3D virtual scene in a vari-depth mode is illustrated in the flowchart of Fig. 18. The vari-depth mode starts with determining the depth of interest, ZDOI, of a 3D target scene 1603 measured in diopters, which can be either determined by the point of the interest of VIEWER or specified by a computer algorithm. The point of interest of the VIEWER can be determined by an eyetracking device if available in the HMD system or other user input devices such as a computer mouse. Alternatively, instead of relying upon an eyetracking device or other input devices, a computer algorithm can specify the depth of interest of the target scene based on the average depth of the virtual 3D scene obtained from a depth map associated therewith or based on feature points of the virtual 3D scene detected by image processing algorithms. Once the depth of interest (DOI) of the scene 1603 is determined, a controller 1812, such as a PC, may apply an electrical control signal, V, to the VFE element 1811 of the vari -focal relay group 1810 which adaptively varies the distance, ZRCDP(V), between the relayed intermediate miniature 3D scene 1815 and the eyepiece 1817 of the Inl-HMD optics 1800 measured in diopters. Consequently, the depth, ZCDP(V), of the virtual CDP 1801 of the Inl-HMD optics 1800, which is measured in diopters, is adaptively set such that it coincides with the depth of interest of the target scene 1603. The simulated virtual camera array 1604 and the virtual camera sensor array 1605 are configured in a similar fashion to the fixed-depth one shown in Fig. 16 except that the camera projection plane 1813 coincides with the depth of interest of the 3D scene 1603. The rest of the rendering method remains the same as that discussed in connection with Fig. 16.

For the purpose of demonstrating the vari-depth mode, the optical power of the tunable lens 1811 was varied so that the CDP 1809 of the display optics 1800 was set to the depth of

3 diopters. The virtual camera and virtual sensor arrays 1604, 1605 were adapted to match the adjusted depth of the virtual CDP 1801 of the display optics 1800. The Els were then re rendered for targets at 3 and 0.5 diopters with the camera projection plane adjusted to match the depth of 3 diopters. Figures 19A, 19B show the captured images through the HMD with the camera (not shown) located at VIEWER focused at the depth of 3 and 0.5 diopters, respectively. By correctly adjusting the optical power of the tunable lens 1811 as well as regenerating the contents on the microdisplay 1602, the system 1800 was able to maintain the same level of the spatial resolution of 3 arcmins and image quality for the targets located at the depth of 3 diopters, Fig. 19A, as well as for the targets located at 1 diopter in Fig. 17B. The vari-depth mode, however, only achieves high-resolution display for targets near the specific depth dictated by the CDP of the display hardware. As shown in Fig. 19B, the targets at the depth of 0.5 diopters show more severely degraded resolution than in Fig. 17C due to its increased separation from the given CDP, even when the camera is focused at the depth of these 0.5-diopter targets.

In still a further of its aspects, a multi-depth mode method in accordance with the present invention for rendering the light field of a 3D virtual scene 2003 is illustrated in the flowchart of Fig. 20. In the multi-depth mode, we started with selecting multiple depths of interest, ZDOI (n) (n=l ...N), of a 3D target scene 2003 distributed along the visual axis measured in diopters, where ZDOI (1) may define the closest depth plane 2003-1 in diopters to the VIEWER and ZDOI (N) the furthest depth plane 2003-N. The placement of the multiple depths of interests may be constrained by multiple factors. The most important factors may be the angular resolution requirements, the depth of field requirements, the threshold tolerance to eye accommodation errors, and the longitudinal resolution requirements. Other factors that may affect the selection of the depths of interests include the depth range affordable by the vari -focal VFE 1811 and the depth distribution of the 3D scene 2003. The total number of depth planes, N, may be constrained by the hardware design. For instance, in a time-multiplexed implementation where the different depths of interests are rendered in a time-sequential fashion, the update frame rates of the VFE 1811, the microdisplay 1602, and the graphics hardware, may be expressed as

where f c is the threshold refresh rate required for flickering-free view, f E is the maximum response speed of the VFE 1811 to an electrical signal for optical power change, fdispiay is the maximum refresh rate of the microdisplay 1602, and f c is the maximum frame rate of the graphics rendering hardware. The number of depth planes can be increased if a spatial- multiplexing method can be implemented where the hardware can afford to render multiple depth planes concurrently. Once the placement and the number of the depths of interests are determined, the rest of the rendering method may be implemented as follows. For each of the selected depths of interests, ZDOI (n) (n=l ...N), a controller 1812 applies an electrical control signal, V(n), to the VFE element 1811 of the vari-focal relay group 1810, which adaptively varies the distance, ZRIM(VII), between the relayed intermediate miniature 3D scene 2105 and the eyepiece 1817 of the Inl-HMD optics 1800. Consequently, the depth of the virtual CDP 2001 of the Inl-HMD optics 1800, ZCDP(VII), is adaptively set such that it coincides with the given depths of interest, ZDOI (n) (n=l ...N). The simulated virtual camera array 1604 and the virtual camera sensor array 1605 may be configured in a similar fashion to that described in Fig. 18 such that the camera projection plane 1813 coincides with the depth of interest, ZDOI (n) (n=l ...N) 2003-1, 2003 -N, for example. To render the 2D elemental images of the 3D scene 2003 for the given depth of interest, a depth map of the 3D virtual scene 2003 is created to obtain depth information of the scene objects with respect to the VIEWER. Instead of rendering the 2D elemental images of the entire 3D scene 2003, we may only render the 2D elemental images located in the depth range defined by

Where Z D0/ (n— 1)— Z D0/ (in) and Z D0/ (n— 1)— Z D0/ (ni) define the dioptric spacings between the given depth of interests and its adjacent depth planes. When n=l, Z D0/ (n— 1) defines the nearest depth limit 2003-1 to be rendered by the displayl602, while when n=N, Z D0/ (n + 1) defines the furthest depth limit 2003-N to be rendered by the display 1602. The rendered 2D elemental images may be mosaiced together in the same way as in the fixed-depth or vari-depth modes to create the n th frame of full-resolution light field image which is then sent to the microdisplay 1602 for update. The same rendering method may repeat for the next depth of interest until all of the N depth planes are rendered. As stated earlier, all of the N depth planes may be rendered in a time-sequential fashion or in a concurrent manner or a hybrid of the two methods.

To demonstrate the multi-depth mode of Fig. 20, we decided to create an implementation of two time-multiplexed depth planes, one placed at 3 diopters and the other placed at 0.5 diopters. The optical power of the tunable lens VFE 1811 was electrically controlled by two different signals VI and V2 sequentially such that the virtual CDP 2001 of the display system 1800 was set to the depths of 3 and 0.5 diopters accordingly. At each of the two virtual CDP placements, we re-rendered the Els for the target scene 2003 which included two resolution targets placed at 3 and 0.5 diopters. For this simple case, the Els rendered for the 0.5 diopter CDP placement only rendered the target object placed at 0.5 diopters and similarly the Els rendered for the 3 diopter CDP placement only rendered the target object placed at 3 diopters. The separately-rendered Els were displayed in a time-multiplexing fashion at a frame rate of about 30Hz while in synchronization the CDP 2009 of the display 1602 was rapidly switched between the depths of 3 and 0.5 diopters. The refresh speed of 30Hz was due to the limit of the highest 60Hz refresh rate of the OLED microdisplay 1602. Figures 21A, 21B show the captured images through the HMD with the camera (not shown) placed at the location of the VIEWER and focused at the depths of 3 and 0.5 diopters, respectively. Along with the virtual display, two spoke resolution targets were physically placed at the corresponding depths of the letters. As shown in Fig. 21A, when the camera was focused at the near depth of 3 diopters, both of the virtual and real objects at the near depth (the letters and the spoke on the left) appears to be in sharp focus, while the far objects (the letters and the spoke on the right) show noticeable out-of-focus blurring as expected. Figure 21B demonstrates the case when the camera focus was switched to the far depth of 0.5 diopters. It can be clearly observed that both of the letters at far and near depths are comparably sharp at the corresponding focus of the camera. By driving the display in a dual-depth mode, the system achieved high-resolution displays of targets with a large depth separation of nearly 3 diopters while rendering focus cues comparable to their real counterparts.

The vari-depth and multi-depth modes of the Ini-based light field rendering methods of the present invention may share the feature that the depth of the CDP 1809, 2009 is either adaptively varied according to the depth of interest in the vari-depth mode or is rapidly switched among several discrete depths in the multi-depth mode. However, their visual effects and implications on focus cues are noticeably different. For instance, as demonstrated in Fig. 19, in the vari-depth mode of an Inl-HMD (Fig. 18), the contents away from the CDP 1809 are rendered with correct blurring cues, though in potentially degraded resolution, due to the nature of light field rendering, while in a conventional vari-focal HMD the contents away from its focal plane can be as high resolution as the contents on the focal depth unless artificially blurred but do not show proper focus cues due to its 2D rendering nature. In the multi-depth mode (Fig. 20), a significant advantage over the traditional multi-focal plane HMD approach is the requirement of much less number of depth switch to render correct focus cues in the same depth range, while depth blending is necessary in a multi-focal system to render focus cues for contents away from the physical focal planes. In the case of Ini-based light field rendering, covering a depth range of 3 diopters only requires 2 focal depth and the focus cues generated in this case are also more accurate and continuous.

These and other advantages of the present invention will be apparent to those skilled in the art from the foregoing specification. Accordingly, it will be recognized by those skilled in the art that changes or modifications may be made to the above-described embodiments without departing from the broad inventive concepts of the invention. It should therefore be understood that this invention is not limited to the particular embodiments described herein, but is intended to include all changes and modifications that are within the scope and spirit of the invention as set forth in the claims.