Conference PaperPDF Available

Super stereoscopy 3D glasses for more realistic 3D vision

Authors:

Abstract and Figures

This paper introduces a new major twist on stereoscopic displays, where users suffer less from the accommodation-vergence con-flict with the help of improved monocular parallax. Our method provides two different views to each eye by using special aper-tures equipped with color filters. The design can be embedded into conventional stereoscopic glasses or special contact lenses. Subjective tests verified that the accommodation-vergence conflict is avoided to a large degree. The technique is also applicable to multi-view 3DTV displays in general.
Content may be subject to copyright.
SUPER STEREOSCOPY 3D GLASSES FOR MORE REALISTIC 3D VISION
Kaan Aks¸it, Amir Hossein Ghanbari Niaki, Osman Eldes, Hakan Urey
Koc¸ University, Department of Electrical Engineering, Istanbul, 34450, Turkey
ABSTRACT
This paper introduces a new major twist on stereoscopic displays,
where users suffer less from the accommodation-vergence con-
flict with the help of improved monocular parallax. Our method
provides two different views to each eye by using special aper-
tures equipped with color filters. The design can be embedded
into conventional stereoscopic glasses or special contact lenses.
Subjective tests verified that the accommodation-vergence conflict
is avoided to a large degree. The technique is also applicable to
multi-view 3DTV displays in general.
Index Terms Super Stereoscopy, Accommodation-vergence
conflict, Monocular parallax, Rotating Screen, Pinhole Glasses
1. INTRODUCTION
The main drawback of conventional stereoscopic and autostereo-
scopic displays is absence of monocular parallax. Next gener-
ation autostereoscopic displays [1] can be expected to solve the
accommodation-vergence conflict [2] where the viewer’s focus is
at the screen instead of the spatial position of three-dimensional
(3D) content in the real world. Autostereoscopic displays with a
solution to the both problems are classified as Super Multi View
(SMV) displays [3], [4]. SMV display provides high number of
images with narrow interval of parallax to overcome the men-
tioned problems [2], [3], [5]. Therefore, such designs require high
density of views and pixels with great level of optical complexity.
This paper introduces a new technique for glasses type stereo-
scopic displays, where users’ accommodation vergence conflict
problem can be reduced. Our solution proposes using special aper-
tures in front of the user’s eye. Unlike conventional stereoscopy,
our initial implementation provides four different perspectives by
employing pinhole size Anaglyph type wavelength bandpass fil-
ters to satisfies the SMV region criterion [3] providing monocular
parallax. This new method is named as Super Stereoscopy 3D
(SS3D). Through subjective tests, we have shown depth percep-
tion and eyes vergence can be enhanced with our method. Due to
the nature of apertures, the visible disadvantage of our method is
a decrease in intensity perceived by the viewers.
2. SUPER STEREOSCOPY TECHNIQUE
Conventional SMV displays produce narrow viewing slits by hav-
ing a narrower angular spread than a single eye [3], [5]. SMV
region criterion is met when at least more than one perspective is
fed to a single eye which is known as monocular parallax. Same
assumption is valid for the case of stereoscopic displays. SMV re-
gion criterion can be expressed as a mathematical formula through
equation 1 representing angular parallax interval, in which deye is
the eye pupil size varying between 2-8 mm [6] and dvrepresents
the distance between the viewer and the three dimensional (3D)
object. Assuming a 5 mm eye pupil with an 3D object at 570 mm
Figure 1. Sketch showing how a voxel is perceived with (a) a conventional
stereoscopic glass (b) our proposed SS3D design
from the viewer, an angler parallax interval of θ0.5degree
is required using equation 1.
θdeye
dv
(1)
Figure 1 provides a comparison in between our method and
the conventional stereoscopy, in which dIP D represents interpupil-
lary distance (60-70 mm). A careful reader can easily notice the
dramatic decrease in perceived voxel volume with the help of
SS3D glasses. There are three different common types of stereo-
scopic glasses to separate left and right eye, which are passive
polarization glasses, anaglyph glasses, and shutter glasses. One
can use one of these techniques as a basis to separate the eyes
from each other, but combination of these methods with the pro-
posed technique can provide number of views per eye. Remaining
accommodation-vergence conflict problem in conventional stereo-
scopic glasses can be easily enhanced by embedding an aperture
in front of the wavelength band-pass filters. Thus, a reader eas-
ily realizes other combinations can also be beneficial to provide
978-1-4799-4758-4/14/$31.00 c
2014 IEEE
natural 3D perception through monocular parallax.
The voxel shapes and sizes differ from each other when con-
ventional system is compared with our proposal. Using ray-optics,
a conventional stereoscopic display’s voxel size can be formalized
in terms of width wcand height hcshown in Figure 1. As an
examplary, deye = 6 mm,dI P D = 65 mm,ds= 500 mm,
dv= 500 mm,doff = 65 mm will lead to wc= 3 mm, and
hc= 42 mm. On the other hand for this example the values
are wb0.3mm and hb0.3mm which shows a substantial
decrease in voxel dimensions.
3. ROTATING AUTOSTEREOSCOPIC DISPLAY
We tested the SS3D concept using autostereoscopic display sys-
tem developed in our lab. As illustrated in Figure 2, the system
contains a transfer screen which is composed of a retro-reflective
surface and a 1D light diffuser, two pico-projectors, a camera for
head-tracker, and a servo motor to rotate the transfer screen. The
transfer screen is a retro-reflective linear light diffusing screen.
Each of two pico-projectors projects one perspective of stereo im-
age onto the transfer screen. The projected images from the pico-
projectors are retro-reflected back to the projectors and create two
viewing zones which overlap with MEMS scanner of picoprojec-
tors, as seen in Figure 2. 1D light diffuser in the transfer screen
diffuses retro-reflected light with a larger diffusing angle in verti-
cal direction than in horizontal direction. Due to the vertical dif-
fusion of light, viewing zones are vertical slits centered with the
MEMS scanner of pico-projectors. If two projectors are placed
horizontally apart from each other by interpupillary distance, then
viewing slits are also apart from each other by the same distance.
By looking through these viewing slits, a viewer sees the left im-
age by his left eye and the right image by his right eye, and per-
ceives 3D image. The 1D Light diffuser allows viewer to move in
vertical direction without losing 3D image perception. However,
if the viewer moves in horizontal direction, he loses 3D percep-
tion. In order to overcome this problem, we added pupil tracker
to change the slits’ angle in order to preserve the viewer’s 3D vi-
sion using OpenCV through python coding language. Detailed
information about the screen can be found in our group’s related
journal article [7].
Figure 2. The rotating screen used during the test, which shows (a) a
user wearing the special glasses, (b) pico-projectors (c) in-house built au-
tostereoscopic display [7] with the 1D diffuser attached on it, and (d) the
user-screen distance
Figure 3. Photograph of the SS glasses used in autostereoscopic displays
with (A) slit and (B) pinhole apertures.
4. EXPERIMENTAL TESTS
The proposed method was tested with our in-house made aper-
ture equipped glasses, with color filters detached from conven-
tional anaglyph type glasses. Instead of using regular stereoscopic
glasses to separate the left and right eye’s content, we used our in-
house built screen [7] which has the capability of separating left
and right image information with no difficulty. Figure 3 shows
the image of glasses built in-house for the setup in which the slit
version has 10 slits per eye with a slit width of 1 mm and center to
center distance of 1.8 mm, and the pinhole version has 6×10 pin
holes with a diameter of 0.8 mm and a center to center distance of
1 mm.
Subjective test was conducted to investigate the proposed tech-
nique. A photograph of the test bench is showed in Figure 2 which
the distance between the user and the screen is fixed at 100 cm
through the tests. The subjects where chosen randomly among
the university students and staffs. All the participants were young
and were not using correction for their vision. In each case the
users were first shown same standard images in order to see if the
subject were experienced enough to distinguish a 3D image, to
calibrate the white balance, and to apply the required alignment
due to differences in their IPDs. All the images shown in the test
were static 3D images.
Viewers were asked to estimate the distance of 3D objects in
the shown content. All images contained a reference point at in-
finite distance, which is used to register stereo pair correctly. Ad-
ditionally, each image contained three different objects at differ-
ent distances which two of them (the right pumpkin and the car-
toon character) were fixed and one of them (the big pumpkin) was
shown in different distances. Note that all of the images were cre-
ated in-house using open-source 3D models in Blender rendering
software. Figure 4 shows the estimated distances by testers with
different contents. Only 13% of subjects could perceive the most
extreme content as 3D before using SS3D glasses, but with the
Figure 4. (a) Estimated distance by testers with different contents. (b)
Right eye content of the object (pumpkin) at 36 cm (c) Right eye content
of the object (pumpkin) at 48 cm. The image of the cartoon character
appears on the screen at 1 m from the user
help of our purposed 3D glasses this number increased to 100%.
Major observed drawback of SS3D glasses during the test was
found to be a decrease in brightness of the screen by testers. An-
other possible problem was resolution due to pinhole diffraction
nature. As the distance between the viewer and the display in-
creases, the pinhole or the alternative slit structure limits the small-
est resolvable spot size to about 0.3 mm (at 500 mm subject dis-
tance). This is consistent with the geometrical calculations given
in section 2.
5. CONCLUSION
A new method entitled as Super Stereoscopy Technique was pro-
posed to enhance the capabilities of stereoscopic displays, and
multi-view autostereoscopic displays. The method provides a so-
lution via improved monocular parallax to a common problem in
stereoscopy domain, accommodation-vergence conflict, and shrinks
dramatically the voxel dimensions via changing its shape. The
method was demonstrated by using an in-house built multi-view
autostereoscopic display.
Our implementation provides two different perspectives to each
eye by employing wavelength band-pass filters used in Anaglyph
type glasses. Through subjective tests, the method was found
to improve the depth perception, and to solve accommodation-
vergence problem. By means of our SS3D glasses, 100% of sub-
jects who were not able to perceive the most extreme contents
were seeing 3D easily. Major drawback of the method is a de-
crease in perceived light intensity by users. This technique is gen-
eral and can be used in 3D cinemas and 3DTV glasses as well as
wearable displays.
6. ACKNOWLEDGEMENT
Financial support from ERC Wear3D Project No: EC.00062 and
T¨
UB˙
ITAK Project No: 111E183 are gratefully acknowledged.
7. REFERENCES
[1] E. Erden H. Urey, K. V. Chellappan and P. Surman, “State of
the art in stereoscopic and autostereoscopic displays,” Pro-
ceedings of the IEEE, vol. 99, pp. 540–555, 2011.
[2] Y. Takaki and N. Nago, “Multi-projection of lenticular dis-
plays to construct a 256-view super multi-view display, Op-
tics Express, vol. 18, pp. 8824–8835, 2010.
[3] H. Yoshikawa Y. Kajiki and T. Honda, “Hologram-like video
images by 45-view stereoscopic display,” in in Proc. SPIE,
1997, pp. 154–166.
[4] K. Tanaka Y. Takaki and J. Nakamura, “Super multi-view
display with a lower resolution flat-panel display,” Optics Ex-
press, vol. 19, pp. 4129–4139, 2011.
[5] J Nakamura, K Tanaka, and Y Takaki, Accommodation re-
sponses to horizontal-parallax-only super-multiview display,”
in IS&T/SPIE Electronic Imaging. International Society for
Optics and Photonics, 2013, pp. 864818–864818.
[6] SG De Groot and JW Gebhard, “Pupil size as determined
by adapting luminance,” JOSA, vol. 42, no. 7, pp. 492–495,
1952.
[7] Osman Eldes, Kaan Aks¸it, and Hakan Urey, “Multi-view au-
tostereoscopic projection display using rotating screen,” Op-
tics Express, vol. 21, no. 23, pp. 29043–29054, 2013.
Thesis
Full-text available
This thesis presents four novel 3D display methods and a number of new applications using commercial laser scanning pico projectors that have been developed with major contributions from our laboratory during the last decade. The new methods developed in this thesis are (i) mixed-polarization stereoscopic display using single-projector and passive polarized glasses, (ii) a new type of glasses that can relieve the accommodation and vergence conflict associated with all the stereoscopic displays, (iii) rotating screen based auto-stereoscopic display for a single-viewer using two pico-projectors, and (iv) multi-user multi-view auto-stereoscopic display using an array of pico-projectors. Furthermore, a head-mounted pico-projector is also demonstrated as a new augmented- reality application aimed at motion-capture actors. The first method introduces a new twist on glasses based stereoscopic projection displays as it utilizes polarization and color multiplexing simultaneously and avoids weaknesses associated with previous methods. This new method is named Mixed Polarization 3D. Color imbalance artifacts associated with anaglyph glasses are avoided by alternating the colors presented to each eye. In mixed polarization 3D, flicker is not observed even at 60 Hz since both eyes receive at least one color frame in every single frame. The second method, named Super Stereoscopic 3D Display (SS3D), introduces a major improvement on stereoscopic 3D displays for avoiding accommodation-vergence conflict. Our method provides at least two views to a single eye by using special apertures equipped with selective filters in front of the users’ eye. Our designs can be embedded into conventional stereoscopic glasses or can be developed as special contact lenses. The third method introduces a single viewer multi-view autostereoscopic projection display, which is invented in our laboratory by a co-worker. The technique uses two mobile projectors, a rotating retro-reflective diffuser screen, and a head-tracking camera. As two dynamic viewing slits are created at the viewers position, the slits can track the position of the eyes by rotating the screen. The last method introduces a new modular multi-user multi-view autostereoscopic display architecture based on an array of pico-projectors and a vertical diffuser screen. A single projector provides a portion of the each perspective observed through the vertical diffuser. Unlike similar projector-array based displays, the screen assembly complexity and the size of the projection display hardware have been decreased dramatically. Our proposal has the capability to provide horizontal expansion in the screen size, and an increase in the number of different perspectives as well. Lastly, an augmented reality (AR) system is proposed and developed for supporting motion capture actors. This system allows seeing and exploring the digital environment without occluding the actors visual field. An in-house prototype is built by combining a retro-reflective screen that covers the walls and a headband consisting of a laser scanning projector with a smartphone. Built-in sensors on the smartphone provide navigation capabilities in the digital world.
Article
Full-text available
A new technique for multi-view autostereoscopic projection display is proposed, and demonstrated. The technique uses two mobile projectors, a rotating retro-reflective diffuser screen, and a head-tracking camera. As two dynamic viewing slits are created at the viewer’s position, the slits can track the position of the eyes by rotating the screen. The display allows a viewer to move approximately 700 mm along the horizontal axis, and 500 mm along the vertical axis with an average crosstalk below 5 %. Two screen prototypes with different diffusers have been tried, and they provide luminance levels of 60 Cd/m², and 160 Cd/m² within the viewing field.
Article
Full-text available
This overview covers most of the 3-D displays that are in use today and presents recent developments and advances in this field. ABSTRACT | Underlying principles of stereoscopic direct-view displays, binocular head-mounted displays, and autostereo-scopic direct-view displays are explained and some early work as well as the state of the art in those technologies are re-viewed. Stereoscopic displays require eyewear and can be categorized based on the multiplexing scheme as: 1) color multiplexed (old technology but there are some recent devel-opments; low-quality due to color reproduction and crosstalk issues; simple and does not require additional electronics hardware); 2) polarization multiplexed (requires polarized light output and polarization-based passive eyewear; high-resolution and high-quality displays available); and 3) time multiplexed (requires faster display hardware and active glasses synchronized with the display; high-resolution com-mercial products available). Binocular head-mounted displays can readily provide 3-D, virtual images, immersive experience, and more possibilities for interactive displays. However, the bulk of the optics, matching of the left and right ocular images and obtaining a large field of view make the designs quite challenging. Some of the recent developments using uncon-ventional optical relays allow for thin form factors and open up new possibilities. Autostereoscopic displays are very attractive as they do not require any eyewear. There are many possibi-lities in this category including: two-view (the simplest implementations are with a parallax barrier or a lenticular screen), multiview, head tracked (requires active optics to redirect the rays to a moving viewer), and super multiview (potentially can solve the accommodation–convergence mis-match problem). Earlier 3-D booms did not last long mainly due to the unavailability of enabling technologies and the content. Current developments in the hardware technologies provide a renewed interest in 3-D displays both from the consumers and the display manufacturers, which is evidenced by the recent commercial products and new research results in this field.
Article
Full-text available
A lenticular-type super multi-view (SMV) display normally requires an ultra high-resolution flat-panel display. To reduce this resolution requirement, two or more views are generated around each eye with an interval smaller than the pupil diameter. Cylindrical lenses constituting a lenticular lens project a group of pixels of the flat-panel display to generate a group of viewing zones. Pixel groups generating left and right viewing zones through the same cylindrical lens are partitioned to separate the two zones. The left and right pixel groups for different cylindrical lenses are interlaced horizontally. A prototype SMV display is demonstrated.
Conference Paper
Hologram-like video images have been demonstrated using a stereoscopic display developed by a member of the 3-D project at TAO. Images are made of 45-view stereoscopic images at resolution of 400 by 400 pixels -- each pixel has 45 horizontal parallaxes -- and refreshed at 30 Hz. Stimulation of conventional stereoscopic video display causes a physiological problem, called accommodation-vergence conflict, though it provides multi-view images. Holography is a usual approach to avoid this conflict but it has great difficulty of providing video images. A new stereoscopic approach is described in this paper, called super multi-view (SMV) region. SMV region means the following condition; there are more than two parallax-images passing through a pupil of viewer's each eye. At the SMV region, if viewer accommodates one's focus to the spatial position of 3-D image -- perceived by binocular disparity -- then the parallax-images passed through a pupil are focused to the same position on a retina. Therefore, there is an ability of fusing accommodation and vergence due to the effect of monocular and binocular parallax. To satisfy the SMV condition, the display must provide numerous view images with narrow interval of parallax. It is difficult to satisfy the SMV condition using conventional displaying technique, but the FLA concept (proposed at Practical Holography X, EI '96) realized providing 45-view images with 0.5 degree interval. The design of this display is also described in this paper. SMV technique is just a quasi-holography because reproduced images have natural stimulus for monocular and binocular vision.
Article
Super multi-view (SMV) displays have been developed to solve the vergence-accommodation conflict that causes visual fatigue with conventional three-dimensional (3D) displays. An ideal SMV display should generate dense viewpoints in both the horizontal and vertical directions. However, the SMV displays developed so far generate dense viewpoints only in the horizontal direction because of the system complexity required for full-parallax systems. Therefore, the accommodation responses to SMV displays were measured mostly for 3D images that contain distinctive vertical edges. In this study, we investigated the influences of the edge properties contained in 3D images upon the accommodation responses evoked by horizontal-parallax-only (HPO) SMV displays. We used the recently developed reduced-view SMV display, whose interval of viewpoints was 2.6 mm, for the accommodation measurements. Two test images were used: a "Maltese cross", which contains several sharp edges, and a standard test image of "Lenna", which contains various types of edges. We found that the HPO-SMV display still evoked the accommodation responses when "Lenna" was displayed. There were two types of accommodation responses for "Lenna"; the amounts of the evoked accommodation were smaller than or similar to those for the "Maltese cross", depending on where the viewers gazed.
Article
A new super multi-view (SMV) display system that enables the number of views to be increased is proposed. All three-dimensional (3D) images generated by multiple multi-view flat-panel displays are superimposed on a common screen using a multi-projection system. The viewing zones of the flat-panel 3D display are produced in the pupils of the projection lenses and then imaged to the observation space by a screen lens. Sixteen flat-panel 3D displays having 16 views were used to construct a SMV display having 256 views. The 3D resolution was 256 x 192. The screen size was 10.3 inches. The horizontal interval of the viewing zones was 1.3 mm.
Article
Available measurements by different investigators on the relationship between pupil diameter and luminance of large adapting fields were compiled and weighted for the number of observers used. The equation for the curve that best fits eight sets of data is logd=0.8558-0.000401 (logB+8.1)3, where d is the diameter of the natural pupil in millimeters and B is the luminance of the adapting field in millilamberts. This is the best estimate that can be made to date of the pupil size—luminance relationship. To make the data of more general use they were corrected for the Stiles-Crawford effect and expressed so that retinal illuminance, adjusted for average pupil size, may be obtained directly from knowledge of the luminance of the stimulus field. The expression for this is logE(trolands) =logB+1.8614-0.000986 (logB+6.5)3, where E is the retinal illuminance in trolands and B is luminance in millilamberts. It is unfortunate that no extensive measurements have been made in this important area.