A novel method for analysing lighting variance

Robust descriptor matching across varying lighting conditions is important for vision-based robotics. We present a novel strategy for quantifying the lighting variance of descriptors. The strategy works by utilising recovered low dimensional mappings from Isomap and our measure of the lighting variance of each of these mappings. The resultant metric allows different descriptors to be compared given a dataset and a set of keypoints. We demonstrate that the SIFT descriptor typically has lower lighting variance than other descriptors, although the result depends on semantic class and lighting conditions.

Advantages of exploiting projection structure for segmenting dense 3D point clouds

Timely and comprehensive scene segmentation is often a critical step for many high level mobile robotic tasks. This paper examines a projected area based neighbourhood lookup approach with the motivation towards faster unsupervised segmentation of dense 3D point clouds. The proposed algorithm exploits the projection geometry of a depth camera to find nearest neighbours which is time independent of the input data size. Points near depth discontinuations are also detected to reinforce object boundaries in the clustering process. The search method presented is evaluated using both indoor and outdoor dense depth images and demonstrates significant improvements in speed and precision compared to the commonly used Fast library for approximate nearest neighbour (FLANN) [Muja and Lowe, 2009].

Low cost localisation for agricultural robotics

This paper presents a pose estimation approach that is resilient to typical sensor failure and suitable for low cost agricultural robots. Guiding large agricultural machinery with highly accurate GPS/INS systems has become standard practice, however these systems are inappropriate for smaller, lower-cost robots. Our positioning system estimates pose by fusing data from a low-cost global positioning sensor, low-cost inertial sensors and a new technique for vision-based row tracking. The results first demonstrate that our positioning system will accurately guide a robot to perform a coverage task across a 6 hectare field. The results then demonstrate that our vision-based row tracking algorithm improves the performance of the positioning system despite long periods of precision correction signal dropout and intermittent dropouts of the entire GPS sensor.

μAV - Design and implementation of an open source micro quadrotor

This paper presents the design of μAV, a palm size open source micro quadrotor constructed on a single Printed Circuit Board. The aim of the micro quadrotor is to provide a lightweight (approximately 86g) and cheap robotic research platform that can be used for a range of robotic applications. One possible application could be a cheap test bed for robotic swarm research. The goal of this paper is to give an overview of the design and capabilities of the micro quadrotor. The micro quadrotor is complete with a 9 Degree of Freedom Inertial Measurement Unit, a Gumstix Overo® Computer-On-Module which can run the widely used Robot Operating System (ROS) for use with other research algorithms.

Training-free probability models for whole-image based place recognition

Whole-image descriptors such as GIST have been used successfully for persistent place recognition when combined with temporal filtering or sequential filtering techniques. However, whole-image descriptor localization systems often apply a heuristic rather than a probabilistic approach to place recognition, requiring substantial environmental-specific tuning prior to deployment. In this paper we present a novel online solution that uses statistical approaches to calculate place recognition likelihoods for whole-image descriptors, without requiring either environmental tuning or pre-training. Using a real world benchmark dataset, we show that this method creates distributions appropriate to a specific environment in an online manner. Our method performs comparably to FAB-MAP in raw place recognition performance, and integrates into a state of the art probabilistic mapping system to provide superior performance to whole-image methods that are not based on true probability distributions. The method provides a principled means for combining the powerful change-invariant properties of whole-image descriptors with probabilistic back-end mapping systems without the need for prior training or system tuning.

Improved line tracker using IMU and vision for visual servoing

This paper introduces an improved line tracker using IMU and vision data for visual servoing tasks. We utilize an Image Jacobian which describes motion of a line feature to corresponding camera movements. These camera motions are estimated using an IMU. We demonstrate impacts of the proposed method in challenging environments: maximum angular rate ~160 0/s, acceleration ~6m /s2 and in cluttered outdoor scenes. Simulation and quantitative tracking performance comparison with the Visual Servoing Platform (ViSP) are also presented.

Towards bio-inspired place recognition over multiple spatial scales

This paper presents a new multi-scale place recognition system inspired by the recent discovery of overlapping, multi-scale spatial maps stored in the rodent brain. By training a set of Support Vector Machines to recognize places at varying levels of spatial specificity, we are able to validate spatially specific place recognition hypotheses against broader place recognition hypotheses without sacrificing localization accuracy. We evaluate the system in a range of experiments using cameras mounted on a motorbike and a human in two different environments. At 100% precision, the multiscale approach results in a 56% average improvement in recall rate across both datasets. We analyse the results and then discuss future work that may lead to improvements in both robotic mapping and our understanding of sensory processing and encoding in the mammalian brain.

Towards condition-invariant, top-down visual place recognition

In this paper we present a novel place recognition algorithm inspired by recent discoveries in human visual neuroscience. The algorithm combines intolerant but fast low resolution whole image matching with highly tolerant, sub-image patch matching processes. The approach does not require prior training and works on single images (although we use a cohort normalization score to exploit temporal frame information), alleviating the need for either a velocity signal or image sequence, differentiating it from current state of the art methods. We demonstrate the algorithm on the challenging Alderley sunny day – rainy night dataset, which has only been previously solved by integrating over 320 frame long image sequences. The system is able to achieve 21.24% recall at 100% precision, matching drastically different day and night-time images of places while successfully rejecting match hypotheses between highly aliased images of different places. The results provide a new benchmark for single image, condition-invariant place recognition.

An analysis of regression models for predicting the speed of a wave glider autonomous surface vehicle

An important aspect of robotic path planning for is ensuring that the vehicle is in the best location to collect the data necessary for the problem at hand. Given that features of interest are dynamic and move with oceanic currents, vehicle speed is an important factor in any planning exercises to ensure vehicles are at the right place at the right time. Here, we examine different Gaussian process models to find a suitable predictive kinematic model that enable the speed of an underactuated, autonomous surface vehicle to be accurately predicted given a set of input environmental parameters.

The (im)possibility of objectivity: Narrating the past in young adult historiographic metafiction

This article uses two Australian historiographic metafictive texts, Into White Silence (Eaton, 2008) and The Lace Maker’s Daughter (Crew, 2005), to demonstrate how particular narrative strategies destabilize the relationship between history and fiction and the past and the present.

Contribución de la fuerza y el dolor en la función del paciente con artrosis trapecio metacarpiana. Estudio transversal

Objetivo Establecer una posible relación entre la fuerza (Jamar), el dolor (EVA) y la capacidad funcional referida por el paciente (DASH) determinando en qué grado influyen unas en otras. Estudio observacional transversal analítico. Participantes Muestra de 72 pacientes que presentaban una artrosis trapecio metacarpiana grado 2-3 de Eaton. Los pacientes fueron reclutados cuando acudían a la Unidad de Cirugía de mano. Método Se realizaron mediciones de fuerza de agarre, pinza, valoración del dolor y funcionalidad, y se establecieron las correlaciones entre cada una de ellas. Resultados El modelo más significativo para la función (R2 =0.83) incluye la variable dolor y la fuerza. Pero es la fuerza punta contra punta la que presenta una mayor correlación con el cuestionario DASH (B-estandarizado: –57). Respecto al dolor, influye en todas las mediciones de fuerza realizadas con el dinamómetro, siendo también la fuerza de la pinza punta contra punta la que presenta una mayor correlación. Conclusiones Los hallazgos corroboran que existe una correlación significativa entre la función referida por el paciente y variables que podemos medir en consulta, como la fuerza del puño y la pinza. Pero también esta correlación es significativa entre las variables función y dolor entre sí, pero es la pinza punta contra punta la que presenta una mayor asociación con el cuestionario DASH. Abstract in English Objective To assess the relationship between muscle strength (Jama), and pain (VAS) levels with hand function (DASH) in patients with trapeziometarcapal osteoarthritis. Cross-sectional study. Participants Sample of 72 patients with osteoarthritis stage 2-3 (Eaton) and trapeziometacarpal osteoarthritis. Patients were recruited when they came to the Hand Surgery Unit. Method Grip strength, pinch, pain and hand function were measured, and correlation and regression coefficients between them were obtained. Results For function, the most significant model (R2 = 0.83) included pain and strength. But it is tip to tip pinch force which has a stronger relationship with DASH (Standardized B: –57) questionnaire. Pain also influenced strength measured with the dynamometer but it was tip to tip pinch force that was the most affected. Conclusions Findings confirm that there is a significant correlation between function referred by the patient and variables that can be measured in the clinic such as grip strength and pinch. The correlation between pain intensity and function was also significant, but tip to tip pinch strength had the greatest impact on the function.

Review of bats and SARS

Bats have been identified as a natural reservoir for an increasing number of emerging zoonotic viruses, including henipaviruses and variants of rabies viruses. Recently, we and another group independently identified several horse-shoe bat species (genus Rhinolophus) as the reservoir host for a large number of viruses that have a close genetic relationship with the coronavirus associated with severe acute respiratory syndrome (SARS). Our current research focused on the identification of the reservoir species for the progenitor virus of the SARS coronaviruses responsible for outbreaks during 2002-2003 and 2003-2004. In addition to SARS-like coronaviruses, many other novel bat coronaviruses, which belong to groups 1 and 2 of the 3 existing coronavirus groups, have been detected by PCR. The discovery of bat SARS-like coronaviruses and the great genetic diversity of coronaviruses in bats have shed new light on the origin and transmission of SARS coronaviruses.

Spontaneous and reversible self-assembly of a polypeptide fragment of insulin-like growth factor binding protein-2 into fluorescent nanotubular structures

In this communication, we report the spontaneous and reversible in vitro self-assembly of a polypeptide fragment derived from the C-terminal domain of Insulin-like Growth Factor Binding Protein (IGFBP-2) into soluble nanotubular structures several micrometres long via a mechanism involving inter-molecular disulfide bonds and exhibiting enhanced fluorescence.

Averting biodiversity collapse in tropical forest protected areas

The rapid disruption of tropical forests probably imperils global biodiversity more than any other contemporary phenomenon(1-3). With deforestation advancing quickly, protected areas are increasingly becoming final refuges for threatened species and natural ecosystem processes. However, many protected areas in the tropics are themselves vulnerable to human encroachment and other environmental stresses(4-9). As pressures mount, it is vital to know whether existing reserves can sustain their biodiversity. A critical constraint in addressing this question has been that data describing a broad array of biodiversity groups have been unavailable for a sufficiently large and representative sample of reserves. Here we present a uniquely comprehensive data set on changes over the past 20 to 30 years in 31 functional groups of species and 21 potential drivers of environmental change, for 60 protected areas stratified across the world's major tropical regions. Our analysis reveals great variation in reserve `health': about half of all reserves have been effective or performed passably, but the rest are experiencing an erosion of biodiversity that is often alarmingly widespread taxonomically and functionally. Habitat disruption, hunting and forest-product exploitation were the strongest predictors of declining reserve health. Crucially, environmental changes immediately outside reserves seemed nearly as important as those inside in determining their ecological fate, with changes inside reserves strongly mirroring those occurring around them. These findings suggest that tropical protected areas are often intimately linked ecologically to their surrounding habitats, and that a failure to stem broad-scale loss and degradation of such habitats could sharply increase the likelihood of serious biodiversity declines.

Product Dimension of Forests and Bounded Treewidth Graphs

The product dimension of a graph G is defined as the minimum natural number l such that G is an induced subgraph of a direct product of l complete graphs. In this paper we study the product dimension of forests, bounded treewidth graphs and k-degenerate graphs. We show that every forest on n vertices has product dimension at most 1.441 log n + 3. This improves the best known upper bound of 3 log n for the same due to Poljak and Pultr. The technique used in arriving at the above bound is extended and combined with a well-known result on the existence of orthogonal Latin squares to show that every graph on n vertices with treewidth at most t has product dimension at most (t + 2) (log n + 1). We also show that every k-degenerate graph on n vertices has product dimension at most inverted right perpendicular5.545 k log ninverted left perpendicular + 1. This improves the upper bound of 32 k log n for the same by Eaton and Rodl.