Eye posture and screen alignment with simulated see-through head-mounted displays by Gibaldi A, Liu Y, Kaspiris-Rousellis C, Mahadevan SM, Read JCA. Vlaskamp BNS, Maus GW, GibaldiLiuKaspirisRousellisMahedevanReadVlaskampMaus2025.pdf (1.6 MiB) - When rendering the visual scene for near-eye head-mounted displays, accurate knowledge of the geometry of the displays, scene objects, and eyes is required for the correct generation of the binocular images. Despite possible design and calibration efforts, these quantities are subject to positional and measurement errors, resulting in some misalignment of the images projected to each eye. Previous research investigated the effects in virtual reality (VR) setups that triggered such symptoms as eye strain and nausea. This work aimed at investigating the effects of binocular vertical misalignment (BVM) in see-through augmented reality (AR). In such devices, two conflicting environments coexist. One environment corresponds to the real world, which lies in the background and forms geometrically aligned images on the retinas. The other environment corresponds to the augmented content, which stands out as foreground and might be subject to misalignment. We simulated a see-through AR environment using a standard three-dimensional (3D) stereoscopic display to have full control and high accuracy of the real and augmented contents. Participants were involved in a visual search task that forced them to alternatively interact with the real and the augmented contents while being exposed to different amounts of BVM. The measured eye posture indicated that the compensation for vertical misalignment is equally shared by the sensory (binocular fusion) and the motor (vertical vergence) components of binocular vision. The sensitivity of each participant varied, both in terms of perceived discomfort and misalignment tolerance, suggesting that a per-user calibration might be useful for a comfortable visual experience.
New Approaches to 3D Vision by Linton P, Morgan MJ, Read JCA, Vishwanath D, Creem-Regehr SH, Domini F, LintonMorganReadVishwanathCreemRegehrDomini2022.pdf (3.2 MiB) - New approaches to 3D vision are enabling new advances in artificial intelligence and autonomous vehicles, a better understanding of how animals navigate the 3D world, and new insights into human perception in virtual and augmented reality. Whilst traditional approaches to 3D vision in computer vision (SLAM: simultaneous localization and mapping), animal navigation (cognitive maps), and human vision (optimal cue integration) start from the assumption that the aim of 3D vision is to provide an accurate 3D model of the world, the new approaches to 3D vision explored in this issue challenge this assumption. Instead, they investigate the possibility that computer vision, animal navigation, and human vision can rely on partial or distorted models or no model at all. This issue also highlights the implications for artificial intelligence, autonomous vehicles, human perception in virtual and augmented reality, and the treatment of visual disorders, all of which are explored by individual articles.
This article is part of a discussion meeting issue ‘New approaches to 3D vision’.
Stereopsis without correspondence by Read JCA, Read2022.pdf (1.9 MiB) - Stereopsis has traditionally been considered a complex visual ability, restricted
to large-brained animals. The discovery in the 1980s that insects, too, have
stereopsis, therefore, challenged theories of stereopsis. How can such simple
brains see in three dimensions? A likely answer is that insect stereopsis has
evolved to produce simple behaviour, such as orienting towards the closer
of two objects or triggering a strike when prey comes within range. Scientific
thinking about stereopsis has been unduly anthropomorphic, for example
assuming that stereopsis must require binocular fusion or a solution of the
stereo correspondence problem. In fact, useful behaviour can be produced
with very basic stereoscopic algorithms which make no attempt to achieve
fusion or correspondence, or to produce even a coarse map of depth across
the visual field. This may explain why some aspects of insect stereopsis
seem poorly designed from an engineering point of view: for example,
paying no attention to whether interocular contrast or velocities match. Such
algorithms demonstrably work well enough in practice for their species, and
may prove useful in particular autonomous applications.
This article is part of a discussion meeting issue ‘New approaches to 3D
vision’.
A computational model of stereoscopic prey capture in praying mantises by O’Keeffe J, Yap SH, Llamas-Cornejo I, Nityananda V, Read JCA, OKeeffeYapLlamasCornejoNityanandaRead2022.pdf (4.2 MiB) - We present a simple model which can account for the stereoscopic sensitivity of praying mantis predatory strikes. The model consists of a single “disparity sensor”: a binocular neuron sensitive to stereoscopic disparity and thus to distance from the animal. The model is based closely on the known behavioural and neurophysiological properties of mantis stereopsis. The monocular inputs to the neuron reflect temporal change and are insensitive to contrast sign, making the sensor insensitive to interocular correlation. The monocular receptive fields have a excitatory centre and inhibitory surround, making them tuned to size. The disparity sensor combines inputs from the two eyes linearly, applies a threshold and then an exponent output nonlinearity. The activity of the sensor represents the model mantis’s instantaneous probability of striking. We integrate this over the stimulus duration to obtain the expected number of strikes in response to moving targets with different stereoscopic disparity, size and vertical disparity. We optimised the parameters of the model so as to bring its predictions into agreement with our empirical data on mean strike rate as a function of stimulus size and disparity. The model proves capable of reproducing the relatively broad tuning to size and narrow tuning to stereoscopic disparity seen in mantis striking behaviour. Although the model has only a single centre-surround receptive field in each eye, it displays qualitatively the same interaction between size and disparity as we observed in real mantids: the preferred size increases as simulated prey distance increases beyond the preferred distance. We show that this occurs because of a stereoscopic “false match” between the leading edge of the stimulus in one eye and its trailing edge in the other; further work will be required to find whether such false matches occur in real mantises. Importantly, the model also displays realistic responses to stimuli with vertical disparity and to pairs of identical stimuli offering a “ghost match”, despite not being fitted to these data. This is the first image-computable model of insect stereopsis, and reproduces key features of both neurophysiology and striking behaviour.
Binocular Vision and Stereopsis Across the Animal Kingdom by Read JCA, Read2021.pdf (3.4 MiB) - Most animals have at least some binocular overlap, i.e., a region of space that is viewed by both eyes. This reduces the overall visual field and raises the problem of combining two views of the world, seen from different vantage points, into a coherent whole. However, binocular vision also offers many potential advantages, including increased ability to see around obstacles and increased contrast sensitivity. One particularly interesting use for binocular vision is comparing information from both eyes to derive information about depth. There are many different ways in which this might be done, but in this review, I refer to them all under the general heading of stereopsis. This review examines the different possible uses of binocular vision and stereopsis and compares what is currently known about the neural basis of stereopsis in different taxa. Studying different animals helps us break free of preconceptions stemming from the way that stereopsis operates in human vision and provides new insights into the different possible forms of stereopsis.
ASTEROID stereotest v1.0: lower stereo thresholds using smaller, denser and faster dots by Read JCA, Wong ZY, Yek X, Wong YX, Bachtoula O, Llamas-Cornejo I, Serrano-Pedraza, ReadWongYekWongBachtoulaLlamasCornejoSerranoPedraza2020_compressed.pdf (0.6 MiB) - Purpose: In 2019, we described ASTEROID, a new stereotest run on a 3D tablet
computer which involves a four-alternative disparity detection task on a dynamic
random-dot stereogram. Stereo thresholds measured with ASTEROID were well
correlated with, but systematically higher than (by a factor of around 1.5), thresholds
measured with previous laboratory stereotests or the Randot Preschool clinical
stereotest. We speculated that this might be due to the relatively large, sparse
dots used in ASTEROID v0.9. Here, we introduce and test the stereo thresholds
and test-repeatability of the new ASTEROID v1.0, which uses precomputed
images to allow stereograms made up of much smaller, denser dots.
Methods: Stereo thresholds and test/retest repeatability were tested and compared
between the old and new versions of ASTEROID (n = 75) and the Randot Circles
(n = 31) stereotest, in healthy young adults.
Results: Thresholds on ASTEROID v1.0 are lower (better) than on ASTEROID
v0.9 by a factor of 1.4, and do not differ significantly from thresholds on the Randot
Circles. Thresholds were roughly log-normally distributed with a mean of
1.54 log10 arcsec (35 arcsec) on ASTEROID v1.0 compared to 1.70 log10 arcsec
(50 arcsec) on ASTEROID v0.9. The standard deviation between observers was
the same for both versions, 0.32 log10 arcsec, corresponding to a factor of 2 above
and below the mean. There was no difference between the versions in their test/
retest repeatability, with 95% coefficient of repeatability = 0.46 log10 arcsec (a
factor of 2.9 or 1.5 octaves) and a Pearson correlation of 0.8 (comparable to other
clinical stereotests).
Conclusion: The poorer stereo thresholds previously reported with ASTEROID
v0.9 appear to have been due to the relatively large, coarse dots and low density
used, rather than to some other aspect of the technology. Employing the small
dots and high density used in ASTEROID v1.0, thresholds and test/retest repeatability
are similar to other clinical stereotests.
Stereotest Comparison: Efficacy, Reliability, and Variability of a New Glasses-Free Stereotest by McCaslin AG, Vancleef K, Hubert L, Read JCA, Port NL, McCaslinVancleefHubertReadPort2020_compressed.pdf (0.6 MiB) - Purpose: To test the validity of the ASTEROID stereotest as a clinical test of depth perception by comparing it to clinical and research standard tests.
Methods: Thirty-nine subjects completed four stereotests twice: the ASTEROID test on an autostereo 3D tablet, a research standard on a VPixx PROPixx 3D projector, Randot Circles, and Randot Preschool. Within 14 days, subjects completed each test for a third time.
Results: ASTEROID stereo thresholds correlated well with research standard thresholds (r = 0.87, P < 0.001), although ASTEROID underestimated standard threshold (mean difference = 11 arcsec). ASTEROID results correlated less strongly with Randot Circles (r = 0.54, P < 0.001) and Randot Preschool (r = 0.64, P < 0.001), due to the greater measurement range of ASTEROID (1–1000 arcsec) compared to Randot Circles or Randot Preschool. Stereo threshold variability was low for all three clinical stereotests (Bland–Altman 95% limits of agreement between test and retest: ASTEROID, ±0.37; Randot Circles, ±0.24; Randot Preschool, ±0.23). ASTEROID captured the largest range of stereo in a normal population with test–retest reliability comparable to research standards (immediate r = 0.86 for ASTEROID vs. 0.90 for PROPixx; follow-up r = 0.68 for ASTEROID vs. 0.88 for PROPixx).
Conclusions: Compared to clinical and research standards for assessing depth perception, ASTEROID is highly accurate, has good test–retest reliability, and measures a wider range of stereo threshold.
Translational Relevance: The ASTEROID stereotest is a better clinical tool for determining baseline stereopsis and tracking changes during treatment for amblyopia and strabismus compared to current clinical tests.
The impact of active research involvement of young children in the design of a new stereotest by Casanova T, Black C, Rafiq S, Hugill-Jones J, Read JCA, Vancleef K, CasanovaBlackRafiqHugillJonesReadVancleef2020.pdf (1.4 MiB) - Background: Although considered important, the direct involvement of young children in research design is scarce and to our knowledge its impact has never been measured. We aim to demonstrate impact of young children’s involvement in improving the understanding of a new 3D eye test or stereotest.
Methods: After a pre-measure of understanding was taken, we explored issues with the test instructions in patient and public involvement (PPI) sessions where children acted as advisers in the test design. Feedback was collected via observations, rating scales and verbal comments. An interdisciplinary panel reviewed the feedback, discussed potential changes to the test design, and decided on the implementation. Subsequently, a post-measure of understanding (Study 1–2) and engagement (Study 3) was collected in a pre-post study design. Six hundred fifty children (2–11.8 years old) took part in the pre-measure, 111 children (1–12 years old) in the subsequent PPI sessions, and 52 children (4–6 years old) in the first post-measure. One hundred twenty-two children (1–12 years old) and unrelated adults took then part in a second series of PPI sessions, and 53 people (2–39 years old) in the final post-measure. Adults were involved to obtain verbal descriptions of the target that could be used to explain the task to children.
Results: Following feedback in Study 1, we added a frame cue and included a shuffle animation. This increased the percentage of correct practice trials from 76 to 97% (t (231) = 14.29, p < .001), but more encouragements like ‘Keep going!’ were needed (t (64) = 8.25, p < .001). After adding a cardboard demo in Study 2, the percentage of correct trials remained stable but the number of additional instructions given decreased (t (103) = 3.72, p < .001) as did the number of encouragements (t (103) = 8.32, p < .001). Therefore, changes in test design following children’s feedback significantly improved task understanding.
Conclusions: Our study demonstrates measurable impact of involvement of very young children in research design through accessible activities. The changes implemented following their feedback significantly improved the understanding of our test. Our approach can inform researchers on how to involve young children in research design and can contribute to developing guidelines for involvement of young children in research.
Efficient estimation of stereo thresholds: What slope should be assumed for the psychometric function? by Serrano-Pedraza I, Vancleef K, Herbert W, Goodship N, Woodhouse M, Read JCA, StrangGilmartinGrayWinfieldWinn2000.pdf (0.3 MiB) - Bayesian staircases are widely used in psychophysics to estimate detection thresholds. Simulations have revealed the importance of the parameters selected for the assumed subject's psychometric function in enabling thresholds to be estimated with small bias and high precision. One important parameter is the slope of the psychometric function, or equivalently its spread. This is often held fixed, rather than estimated for individual subjects, because much larger numbers of trials are required to estimate the spread as well as the threshold. However, if this fixed value is wrong, the threshold estimate can be biased. Here we determine the optimal slope to minimize bias and maximize precision when measuring stereoacuity with Bayesian staircases. We performed 2- and 4AFC disparity detection stereo experiments in order to measure the spread of the disparity psychometric function in human observers assuming a Logistic function. We found a wide range, between 0.03 and 3.5 log10 arcsec, with little change with age. We then ran simulations to examine the optimal spread using the empirical data. From our simulations and for three different experiments, we recommend selecting assumed spread values between the percentiles 60-80% of the population distribution of spreads (these percentiles can be extended to other type of thresholds). For stereo thresholds, we recommend a spread around the value σ = 1.7 log10 arcsec for 2AFC (slope β = 4.3 /log10 arcsec), and around σ = 1.5 log10 arcsec for 4AFC (β = 4.9 /log10 arcsec). Finally, we compared a Bayesian procedure (ZEST using the optimal σ) with five Bayesian procedures that are versions of ZEST-2D, Psi, and Psi-marginal. In general, for the conditions tested, ZEST optimal σ showed the lowest threshold bias and highest precision.
