Experimental comparison of odometry approaches

Odometry is an important input to robot navigation systems, and we are interested in the performance of vision-only techniques. In this paper we experimentally evaluate and compare the performance of wheel odometry, monocular feature-based visual odometry, monocular patch-based visual odometry, and a technique that fuses wheel odometry and visual odometry, on a mobile robot operating in a typical indoor environment.

Adaptive evolution of color vision as seen through the eyes of butterflies

Butterflies and primates are interesting for comparative color vision studies, because both have evolved middle- (M) and long-wavelength- (L) sensitive photopigments with overlapping absorbance spectrum maxima (lambda(max) values). Although positive selection is important for the maintenance of spectral variation within the primate pigments, it remains an open question whether it contributes similarly to the diversification of butterfly pigments. To examine this issue, we performed epimicrospectrophotometry on the eyes of five Limenitis butterfly species and found a 31-nm range of variation in the lambda(max) values of the L-sensitive photopigments (514-545 nm). We cloned partial Limenitis L opsin gene sequences and found a significant excess of replacement substitutions relative to polymorphisms among species. Mapping of these L photopigment lambda(max) values onto a phylogeny revealed two instances within Lepidoptera of convergently evolved L photopigment lineages whose lambda(max) values were blue-shifted. A codon-based maximum-likelihood analysis indicated that, associated with the two blue spectral shifts, four amino acid sites (Ile17Met, Ala64Ser, Asn70Ser, and Ser137Ala) have evolved substitutions in parallel and exhibit significant d(N)/d(S) >1. Homology modeling of the full-length Limenitis arthemis astyanax L opsin placed all four substitutions within the chromophore-binding pocket. Strikingly, the Ser137Ala substitution is in the same position as a site that in primates is responsible for a 5- to 7-nm blue spectral shift. Our data show that some of the same amino acid sites are under positive selection in the photopigments of both butterflies and primates, spanning an evolutionary distance >500 million years.

GABAergic modification of myopic and hyperopic ocular tissue effects on scleral fibroblast growth

Purpose: Myopia is a common eye disorder affecting up to 90% of children in South East Asia and 30% of the population worldwide. Myopia of high severity is a leading cause of blindness around the world (4th to 5th most common). Changes and remodelling of the sclera i.e. increase cellular proliferation & increase protein synthesis within scleral cells (↑ scleral DNA) and thinning and lose of extracellular matrix of sclera (↓ scleral GAG synthesis) have been linked to myopic eye growth in animal models. Signals acting on the sclera are thought to originate in the retina, and are modulated by the retinal pigment epithelium (RPE) with limited evidence suggesting that the RPE can modify scleral cell growth in culture. However, the mechanism of retinal signal transmission and the role of posterior eye cup tissue, including the RPE, in mediating changes in scleral fibroblast growth during myopia development are unclear. Retinal transmitter systems are critically involved in pathways regulating eye growth, which ultimately lead to alterations in the sclera if eye size is to change. A dopaminergic agonist and muscarinic antagonists decrease the proliferation of scleral chondrocytes when co-cultured with chick’s retinal pigment epithelium (RPE). GABA receptors have recently been localised to chick sclera. We therefore hypothesised that posterior eye cup tissue from myopic eyes would stimulate and from hyperopic eyes would inhibit growth of scleral fibroblasts in vitro and that GABAergic agents could directly interact with scleral cells or indirectly modify the effects of myopic and hyperopic posterior eye cup tissue on scleral fibroblast growth. Method: Fibroblastic cells obtained from 8-day-old chick sclera were used to establish cell banks. Two major experiments were performed. Experiment 1: To determine if posterior eye cup tissues from myopic eye stimulates and hyperopic eye inhibits scleral cell proliferation, when co-cultured with scleral cells in vitro. This study comprised two linked experiments, i) monocular visual treatments of FDM (form-deprivation myopia), LIM (lens-induced myopia) and LIH (lens-induced hyperopia) with assessment of the effect of full punch eye cup tissue on DNA and GAG synthesis by cultured chick scleral fibroblasts, and ii) binocular visual treatments comprising LIM and LIH with assessment of the effect of individual layers of eye cup tissues (neural retina, RPE and choroid) on cultured chick scleral fibroblasts. Visual treatment was applied for 3 days. Experiment 2: To determine the direct interaction of GABA agents on scleral cell growth and to establish whether GABA agents modify the stimulatory/inhibitory effect of myopic and hyperopic posterior eye cup tissues on cultured scleral cell growth in vitro. Two linked experiments were performed. i) GABA agonists (muscimol and baclofen) and GABA antagonists (bicuculine (-), CGP46381 and TPMPA) were added to scleral cell culture medium to determine their direct effect on scleral cells. ii) GABAergic agents (agonists and antagonists) were administered to scleral fibroblasts co-cultured with posterior eye cup tissue (retina, RPE, retina/RPE, RPE/choroid). Ocular tissues were obtained from chick eyes wearing +15D (LIH) or -15D lenses (LIM) for 3 days. In both experiments, tissues were added to hanging cell culture insert (pore size 1.0ìm) placed over each well of 24 well plates while scleral cells were cultured in DMEM/F12, Glutamax (Gibco) plus 10% FBS and penicillin/streptomycin (50U/ml)) and fungizone (1.25ug/ml) (Gibco), at seeding density of 30,000 cells/well at the bottom of the well and allowed to grow for 3 days. Scleral cells proliferation rate throughout the study was evaluated by determining GAG and DNA content of scleral cells using Dimethylmethylene blue (DMMB) dye and Quant-iTTm Pico Green® dsDNA reagent respectively. Results and analysis: Based on DNA and GAG content, there was no significant difference in tissue effect of LIM and LIH eyes on scleral fibroblast growth (DNA: 8.4 ± 1.1μg versus 9.3 ± 2.3 μg, p=0.23; GAG: 10.13 ± 1.4 μg versus 12.67 ± 1.2 μg, F2,23=6.16, p=0.0005) when tissues were obtained from monocularly treated chick eyes (FDM or +15D lens or -15D lens over right eyes with left eyes untreated) and co-cultured as full punch. When chick eyes were treated binocularly with -15D lens (LIM) right eye and +15D lens (LIH) left eyes and tissue layers were separated, the retina from LIM eyes did not stimulate scleral cell proliferation compared to LIH eyes (DNA: 27.2 ± 6.7 μg versus 23.2 ± 1.5 μg, p=0.23; GAG: 28.1 ±3.7 μg versus 28.7 ± 4.2 μg, p=0.21). Similarly, the LIH and LIM choroid did not produce a differential effect based on DNA (LIM 46.9 ± 6.4 μg versus LIH 53.5 ± 4.7 μg, p=0.18), however the choroid from LIH eyes induced higher scleral GAG content than from LIM eyes (32.5 ± 6.7 μg versus 18.9 ± 1.2 μg, p=0.023). In contrast, the RPE from LIM eyes caused a significant increase in fibroblast proliferation whereas the RPE from LIH eyes was relatively inhibitory (72.4 ± 6.3 μg versus 27.9 ± 2.3 μg, F1, 6=69.99, p=0.0005). GAG data were opposite to DNA data e.g. the RPE from LIH eyes increased (33.7 ± 7.9 μg) while the RPE from LIM eyes decreased (28.2 ± 3.0 μg) scleral cell growth (F1, 6=13.99, p=0.010). Based on DNA content, GABA agents had a small direct effect on scleral cell growth; GABA agonists increased (21.4 ± 1.0% and 18.3 ± 1.0% with muscimol and baclofen, p=0.0021), whereas GABA antagonists decreased fibroblast proliferation (-23.7 ± 0.9% with bicuculine & CGP46381 and -28.1 ± 0.5% with TPMPA, p=0.0004). GABA agents also modified the effect of LIM and LIH tissues (p=0.0005).The increase in proliferation rate of scleral fibroblasts co-cultured with tissues (RPE, retina, RPE/retina and RPE/choroid) from LIM treated eyes was enhanced by GABA agonists (muscimol: 27.4 ± 1.2%, 35.8 ± 1.6%, 8.4 ± 0.3% and 11.9 ± 0.6%; baclofen: 27.0 ± 1.0%, 15.8 ± 1.5%, 16.8 ± 1.2% and 15.4 ± 0.4%, p=0.014) whereas GABA antagonists further reduced scleral fibroblasts growth (bicuculine: -52.5 ± 2.5%, -36.9 ± 1.4%, -37.5 ± 0.6% and -53.7 ± 0.9%; TPMPA: 57.3 ± 1.3%, -15.7 ± 1.2%, -33.5 ± 0.4% and -45.9 ± 1.5%; CGP46381: -51.9 ± 1.6%, -28.5 ± 1.5%, -25.4 ± 2.0% and -45.5 ± 1.9% respectively, p=0.0034). GAG data were opposite to DNA data throughout the experiment e.g. GABA agonists further inhibited while antagonists relatively enhanced scleral fibroblasts growth for both LIM and LIH tissue co-culture. The effect of GABA agents was relatively lower (p=0.0004) for tissue from LIH versus LIM eyes but was in a similar direction. There was a significant drug effect on all four tissue types e.g. RPE, retina, RPE/retina and RPE/choroid for both LIM and LIH tissue co-culture (F20,92=3.928, p=0.0005). However, the effect of GABA agents was greatest in co-culture with RPE tissue (F18,36=4.865, p=0.0005). Summary and Conclusion: 1) Retinal defocus signals are transferred to RPE and choroid which then exert their modifying effect on scleral GAG and DNA synthesis either through growth stimulating factors or directly interacting with scleral cells in process of scleral remodeling during LIM and LIH visual conditions. 2) GABAergic agents affect the proliferation of scleral fibroblasts both directly and when co-cultured with ocular tissues in vitro.

Quantitative evaluation of matching methods and validity measures for stereo vision

The authors present a qualitative and quantitative comparison of various similarity measures that form the kernel of common area-based stereo-matching systems. The authors compare classical difference and correlation measures as well as nonparametric measures based on the rank and census transforms for a number of outdoor images. For robotic applications, important considerations include robustness to image defects such as intensity variation and noise, the number of false matches, and computational complexity. In the absence of ground truth data, the authors compare the matching techniques based on the percentage of matches that pass the left-right consistency test. The authors also evaluate the discriminatory power of several match validity measures that are reported in the literature for eliminating false matches and for estimating match confidence. For guidance applications, it is essential to have and estimate of confidence in the three-dimensional points generated by stereo vision. Finally, a new validity measure, the rank constraint, is introduced that is capable of resolving ambiguous matches for rank transform-based matching.

Review of guidelines for children’s vision screenings

The aim of children's vision screenings is to detect visual problems that are common in this age category through valid and reliable tests. Nevertheless, the cost effectiveness of paediatric vision screenings, the nature of the tests included in the screening batteries and the ideal screening age has been the cause of much debate in Australia and worldwide. Therefore, the purpose of this review is to report on the current practice of children's vision screenings in Australia and other countries, as well as to evaluate the evidence for and against the provision of such screenings. This was undertaken through a detailed investigation of peer-reviewed publications on this topic. The current review demonstrates that there is no agreed vision screening protocol for children in Australia. This appears to be a result of the lack of strong evidence supporting the benefit of such screenings. While amblyopia, strabismus and, to a lesser extent refractive error, are targeted by many screening programs during pre-school and at school entry, there is less agreement regarding the value of screening for other visual conditions, such as binocular vision disorders, ocular health problems and refractive errors that are less likely to reduce distance visual acuity. In addition, in Australia, little agreement exists in the frequency and coverage of screening programs between states and territories and the screening programs that are offered are ad hoc and poorly documented. Australian children stand to benefit from improved cohesion and communication between jurisdictions and health professionals to enable an equitable provision of validated vision screening services that have the best chance of early detection and intervention for a range of paediatric visual problems.

From vision to reality : views of primary school principals on inclusive education in New South Wales, Australia

This paper discusses the findings of a research study that used semi-structured interviews to explore the views of primary school principals on inclusive education in New South Wales, Australia. Content analysis of the transcript data indicates that principals’ attitudes towards inclusive education and their success in engineering inclusive practices within their school are significantly affected by their own conception of what “inclusion” means, as well as the characteristics of the school community, and the attitudes and capacity of staff. In what follows, we present two parallel conversations that arose from the interview data to illustrate the main conceptual divisions existing between our participants’ conceptions of inclusion. First, we discuss the act of “being inclusive” which was perceived mainly as an issue of culture and pedagogy. Second, we consider the mechanics of “including,” which reflected a more instrumentalist position based on perceptions of individual student deficit, the level of support they may require and the amount of funding they can attract.

100Hz onboard vision for quadrotor state estimation

This paper introduces a high-speed, 100Hz, visionbased state estimator that is suitable for quadrotor control in close quarters manoeuvring applications. We describe the hardware and algorithms for estimating the state of the quadrotor. Experimental results for position, velocity and yaw angle estimators are presented and compared with motion capture data. Quantitative performance comparison with state-of-the-art achievements are also presented.

Translating vision research into policy and practice to improve visibility, and hence safety, of road workers at night

This case study report describes the stages involved in the translation of research on night-time visibility into standards for the safety clothing worn by roadworkers. Vision research demonstrates that when lights are placed on the moveable joints of the body and the person moves in a dark setting, the phenomenon known as “biological motion or biomotion” occurs, enabling rapid and accurate recognition of the human form although only the lights can be seen. QUT was successful in gaining funding from the Australian Research Council for a Linkage grant due to the support of the predecessors of the Queensland Department of Transport and Main Roads (TMR) to research the biomotion effect in on-road settings using materials that feature in roadworker clothing. Although positive results were gained, the process of translating the research results into policy, practices and standards relied strongly on the supportive efforts of TMR staff engaged in the review and promulgation of national standards. The ultimate result was the incorporation of biomotion marking into AS/NZS 4602.1 2011. The experiences gained in this case provide insights into the processes involved in translating research into practice.

Vision based control for micro aerial vehicles : Application to sense and avoid

This work presents a collision avoidance approach based on omnidirectional cameras that does not require the estimation of range between two platforms to resolve a collision encounter. Our method achieves minimum separation between the two vehicles involved by maximising the view-angle given by the omnidirectional sensor. Only visual information is used to achieve avoidance under a bearing- only visual servoing approach. We provide theoretical problem formulation, as well as results from real flights using small quadrotors

Diurnal variation of ocular biometrics under natural and defocused conditions

It is well known that a broad range of ocular anatomical and physiological parameters undergo significant diurnal variation. However, the natural diurnal variations that occur in the length of the human eye (axial length) and their underlying causes have been less well studied. Improvements in optical methods for the measurement of ocular biometrics now allow more precise and comprehensive measurements of axial length to be performed than has previously been possible. Research from animal models also suggests a link between diurnal axial length variations and longer term myopic eye growth, and that retinal image defocus can disrupt these diurnal rhythms in axial length. This research programme has examined the diurnal variations in axial length in young normal eyes, the contributing components and the influence of optical stimuli on these changes. In the first experiment, the normal pattern and consistency of the diurnal variations in axial length were examined at 10 different times (5 measurements each day, at ~ 3-hour intervals from ~ 9 am to ~ 9 pm) over 2 consecutive days on 30 young adult subjects (15 myopes, 15 emmetropes). Additionally, variations in a range of other ocular biometric measurements such as choroidal thickness, intraocular pressure, and other ocular biometrics were also explored as potential factors that may be associated with the observed variations in axial length. To investigate the potential influence of refractive error on diurnal axial length variations, the differences in the magnitude and pattern of diurnal variations in axial length between the myopic and emmetropic subjects were examined. Axial length underwent significant diurnal variation that was consistently observed over the 2 consecutive days of measurements, with the longest axial length typically occurring during the day, and the shortest at night. Significant diurnal variations were also observed in choroidal thickness, IOP and other ocular biometrics (such as central corneal thickness, anterior chamber depth and vitreous chamber depth) of the eye. Diurnal variations in vitreous chamber depth, IOP (positive associations) and choroidal thickness (negative association) were all significantly correlated with the diurnal changes in axial length. Choroidal thickness was found to fluctuate approximately in antiphase to the axial length changes, with the average timing of the longest axial length coinciding with the thinnest choroid and vice versa. There were no significant differences in the ocular diurnal variations associated with refractive error. Given that the diurnal changes in axial length could be associated with the changes in the eye’s optical quality, whether the optical quality of the eye also undergoes diurnal variation in the same cohort of young adult myopes and emmetropes over 2 consecutive days was also examined. Significant diurnal variations were observed only in the best sphere refraction (power vector M) and in the spherical aberration of the eye over two consecutive days of testing. The changes in the eyes lower and higher order ocular optics were not significantly associated with the diurnal variations in axial length and the other measured ocular biometric parameters. No significant differences were observed in the magnitude and timing of diurnal variations in lower-order and higher-order optics associated with refractive error. Since the small natural fluctuations in the eye’s optical quality did not appear to be sufficient to influence the natural diurnal fluctuations in ocular biometric parameters, in the next experiment, the influence of monocular myopic defocus (+1.50 DS) upon the normal diurnal variations in axial length and choroidal thickness of young adult emmetropic human subjects (n=13) imposed over a 12 hour period was examined. A series of axial length and choroidal thickness measurements (collected at ~3 hourly intervals, with the first measurement at ~9 am and the final measurement at ~9 pm) were obtained over three consecutive days. The natural diurnal rhythms (Day 1, no defocus), diurnal rhythms with monocular myopic defocus (Day 2, +1.50 DS spectacle lens over the right eye), and the recovery from any defocus induced changes (Day 3, no defocus) were examined. Significant diurnal variations over the course of the day were observed in both axial length and choroidal thickness on each of the three measurement days. The introduction of monocular myopic defocus led to significant reductions in the mean amplitude of diurnal change, and phase shifts in the peak timing of the diurnal rhythms in axial length and choroidal thickness. These defocus induced changes were found to be transient in nature and returned to normal the day following removal of the defocus. To further investigate the influence of optical stimuli on human diurnal rhythms, in the final experiment, the influence of monocular hyperopic defocus on the normal diurnal rhythms in axial length and choroidal thickness was examined in young adult emmetropic subjects (n=15). Similar to the previous experiment, the natural diurnal rhythms (Day 1, no defocus), diurnal rhythms with monocular hyperopic defocus (Day 2, -2.00 DS spectacle lens over the right eye), and the recovery from any defocus induced changes (Day 3, no defocus) were examined over three consecutive days. Both axial length and choroidal thickness underwent significant diurnal variations on each of the three days. The introduction of monocular hyperopic defocus resulted in a significant increase in the amplitude of diurnal change, but no change in the peak timing of diurnal rhythms in both parameters. The ocular changes associated with hyperopic defocus returned to normal, the day following removal of the defocus. This research has shown that axial length undergoes significant diurnal variation in young adult human eyes, and has shown that the natural diurnal variations in choroidal thickness and IOP are significantly associated, and may underlie these diurnal fluctuations in axial length. This work also demonstrated for the first time that exposing young human eyes to monocular myopic and hyperopic defocus leads to a significant disruption in the normal diurnal rhythms of axial length and choroidal thickness. These changes in axial length with defocus may reflect underlying mechanisms in the human eye that are involved in the regulation of longer term eye growth.

Automatic object segmentation of unstructured scenes using colour and depth maps

This study presents a segmentation pipeline that fuses colour and depth information to automatically separate objects of interest in video sequences captured from a quadcopter. Many approaches assume that cameras are static with known position, a condition which cannot be preserved in most outdoor robotic applications. In this study, the authors compute depth information and camera positions from a monocular video sequence using structure from motion and use this information as an additional cue to colour for accurate segmentation. The authors model the problem similarly to standard segmentation routines as a Markov random field and perform the segmentation using graph cuts optimisation. Manual intervention is minimised and is only required to determine pixel seeds in the first frame which are then automatically reprojected into the remaining frames of the sequence. The authors also describe an automated method to adjust the relative weights for colour and depth according to their discriminative properties in each frame. Experimental results are presented for two video sequences captured using a quadcopter. The quality of the segmentation is compared to a ground truth and other state-of-the-art methods with consistently accurate results.

Vision-only autonomous navigation using topometric maps

This paper presents a mapping and navigation system for a mobile robot, which uses vision as its sole sensor modality. The system enables the robot to navigate autonomously, plan paths and avoid obstacles using a vision based topometric map of its environment. The map consists of a globally-consistent pose-graph with a local 3D point cloud attached to each of its nodes. These point clouds are used for direction independent loop closure and to dynamically generate 2D metric maps for locally optimal path planning. Using this locally semi-continuous metric space, the robot performs shortest path planning instead of following the nodes of the graph --- as is done with most other vision-only navigation approaches. The system exploits the local accuracy of visual odometry in creating local metric maps, and uses pose graph SLAM, visual appearance-based place recognition and point clouds registration to create the topometric map. The ability of the framework to sustain vision-only navigation is validated experimentally, and the system is provided as open-source software.

Vision-based place recognition : how low can you go?

In this paper we use the algorithm SeqSLAM to address the question, how little and what quality of visual information is needed to localize along a familiar route? We conduct a comprehensive investigation of place recognition performance on seven datasets while varying image resolution (primarily 1 to 512 pixel images), pixel bit depth, field of view, motion blur, image compression and matching sequence length. Results confirm that place recognition using single images or short image sequences is poor, but improves to match or exceed current benchmarks as the matching sequence length increases. We then present place recognition results from two experiments where low-quality imagery is directly caused by sensor limitations; in one, place recognition is achieved along an unlit mountain road by using noisy, long-exposure blurred images, and in the other, two single pixel light sensors are used to localize in an indoor environment. We also show failure modes caused by pose variance and sequence aliasing, and discuss ways in which they may be overcome. By showing how place recognition along a route is feasible even with severely degraded image sequences, we hope to provoke a re-examination of how we develop and test future localization and mapping systems.