Queensland Research Laboratory

St Lucia, Australia

Queensland Research Laboratory

St Lucia, Australia
Time filter
Source Type

Zhang B.,Beihang University | Gao Y.,Griffith University | Gao Y.,Queensland Research Laboratory | Zhao S.,Griffith University | And 2 more authors.
IEEE Transactions on Circuits and Systems for Video Technology | Year: 2011

This paper proposes a novel kernel similarity modeling of texture pattern flow (KSM-TPF) for background modeling and motion detection in complex and dynamic environments. The texture pattern flow encodes the binary pattern changes in both spatial and temporal neighborhoods. The integral histogram of texture pattern flow is employed to extract the discriminative features from the input videos. Different from existing uniform threshold based motion detection approaches which are only effective for simple background, the kernel similarity modeling is proposed to produce an adaptive threshold for complex background. The adaptive threshold is computed from the mean and variance of an extended Gaussian mixture model. The proposed KSM-TPF approach incorporates machine learning method with feature extraction method in a homogenous way. Experimental results on the publicly available video sequences demonstrate that the proposed approach provides an effective and efficient way for background modeling and motion detection. © 2011 IEEE.

Thompson M.B.,University of Queensland | Thompson M.B.,Queensland Research Laboratory | Tangen J.M.,University of Queensland | Mccarthy D.J.,Forensic Services Branch
Journal of Forensic Sciences | Year: 2013

Although fingerprint experts have presented evidence in criminal courts for more than a century, there have been few scientific investigations of the human capacity to discriminate these patterns. A recent latent print matching experiment shows that qualified, court-practicing fingerprint experts are exceedingly accurate (and more conservative) compared with novices, but they do make errors. Here, a rationale for the design of this experiment is provided. We argue that fidelity, generalizability, and control must be balanced to answer important research questions; that the proficiency and competence of fingerprint examiners are best determined when experiments include highly similar print pairs, in a signal detection paradigm, where the ground truth is known; and that inferring from this experiment the statement "The error rate of fingerprint identification is 0.68%" would be unjustified. In closing, the ramifications of these findings for the future psychological study of forensic expertise and the implications for expert testimony and public policy are considered. © 2013 American Academy of Forensic Sciences.

Horvath I.,Queensland University of Technology | Horvath I.,Queensland Research Laboratory | Lovell B.C.,Queensland University of Technology | Lovell B.C.,Queensland Research Laboratory
Journal of Geophysical Research: Space Physics | Year: 2010

Field-aligned passes track true profiles. Such Defense Meteorological Satellite Program passes permitted investigating storm-enhanced plasma density (SED) feature development during the Bastille Day Superstorm in a comprehensive way. We tracked equatorial ionization anomaly (EIA) and SED features and their underlying forward fountain circulation and downward SED plume plasma flows, respectively. Northward subauroral polarization stream E fields detaching plasma and producing SED plumes were also detected. We assessed the effects of South Atlantic Magnetic Anomaly and summer-towinter interhemispheric plasma flows on the EIA and found a southward dipping gradient in drift/flow when no storm/evening-related fountain strengthening occurred. We investigated the relative importance of different plasma sources in SED development. An extremely large plasma enhancement seen over Florida at 2200 UT on 15 July 2000 was a SED feature that was tracked by many GPS total electron content (TEC) maps as a 200 TEC unit (TECU) enhancement. We tracked its equally large conjugate pair over Trelew (Argentine Patagonia) and unraveled their development. Their underlying SED plume supplied most of the plasma. Appearing between these two SED features, a small and highly asymmetrical EIA offered on each side a low baseline upon which the downward streaming SED plume plasma piled up. Contradicting a currently accepted explanation, there was no enhanced fountain action detected to contribute 150 TECU to the 200 TECU. Later (∼2400 UT), there was enhanced fountain action, but SED plume contribution still dominated. Proven by observational evidence, SED development is a complex process of SED plume plasma flows and equatorward wind effects that cannot be described by one single explanation. Copyright © 2010 by the American Geophysical Union.

Lam B.S.Y.,Griffith University | Gao Y.,Griffith University | Gao Y.,Queensland Research Laboratory | Liew A.W.-C.,Queensland Research Laboratory | Liew A.W.-C.,Griffith University
IEEE Transactions on Medical Imaging | Year: 2010

Detecting blood vessels in retinal images with the presence of bright and dark lesions is a challenging unsolved problem. In this paper, a novel multiconcavity modeling approach is proposed to handle both healthy and unhealthy retinas simultaneously. The differentiable concavity measure is proposed to handle bright lesions in a perceptive space. The line-shape concavity measure is proposed to remove dark lesions which have an intensity structure different from the line-shaped vessels in a retina. The locally normalized concavity measure is designed to deal with unevenly distributed noise due to the spherical intensity variation in a retinal image. These concavity measures are combined together according to their statistical distributions to detect vessels in general retinal images. Very encouraging experimental results demonstrate that the proposed method consistently yields the best performance over existing state-of-the-art methods on the abnormal retinas and its accuracy outperforms the human observer, which has not been achieved by any of the state-of-the-art benchmark methods. Most importantly, unlike existing methods, the proposed method shows very attractive performances not only on healthy retinas but also on a mixture of healthy and pathological retinas. © 2006 IEEE.

Thompson M.B.,University of Queensland | Thompson M.B.,Queensland Research Laboratory | Tangen J.M.,University of Queensland | McCarthy D.J.,Forensic Services Branch
Law and Human Behavior | Year: 2014

There has been very little research into the nature and development of fingerprint matching expertise. Here we present the results of an experiment testing the claimed matching expertise of fingerprint examiners. Expert (n = 37), intermediate trainee (n = 8), new trainee (n = 9), and novice (n = 37) participants performed a fingerprint discrimination task involving genuine crime scene latent fingerprints, their matches, and highly similar distractors, in a signal detection paradigm. Results show that qualified, court-practicing fingerprint experts were exceedingly accurate compared with novices. Experts showed a conservative response bias, tending to err on the side of caution by making more errors of the sort that could allow a guilty person to escape detection than errors of the sort that could falsely incriminate an innocent person. The superior performance of experts was not simply a function of their ability to match prints, per se, but a result of their ability to identify the highly similar, but nonmatching fingerprints as such. Comparing these results with previous experiments, experts were even more conservative in their decision making when dealing with these genuine crime scene prints than when dealing with simulated crime scene prints, and this conservatism made them relatively less accurate overall. Intermediate trainees-despite their lack of qualification and average 3.5 years experience-performed about as accurately as qualified experts who had an average 17.5 years experience. New trainees-despite their 5-week, full-time training course or their 6 months experience-were not any better than novices at discriminating matching and similar nonmatching prints, they were just more conservative. Further research is required to determine the precise nature of fingerprint matching expertise and the factors that influence performance. The findings of this representative, lab-based experiment may have implications for the way fingerprint examiners testify in court, but what the findings mean for reasoning about expert performance in the wild is an open, empirical, and epistemological question. © 2013 American Psychological Association.

Hajati F.,Amirkabir University of Technology | Hajati F.,Griffith University | Raie A.A.,Amirkabir University of Technology | Gao Y.,Griffith University | Gao Y.,Queensland Research Laboratory
Pattern Recognition | Year: 2012

In this paper, we propose a novel Patch Geodesic Distance (PGD) to transform the texture map of an object through its shape data for robust 2.5D object recognition. Local geodesic paths within patches and global geodesic paths for patches are combined in a coarse to fine hierarchical computation of PGD for each surface point to tackle the missing data problem in 2.5D images. Shape adjusted texture patches are encoded into local patterns for similarity measurement between two 2.5D images with different viewing angles and/or shape deformations. An extensive experimental investigation is conducted on 2.5 face images using the publicly available BU-3DFE and Bosphorus databases covering face recognition under expression and pose changes. The performance of the proposed method is compared with that of three benchmark approaches. The experimental results demonstrate that the proposed method provides a very encouraging new solution for 2.5D object recognition. © 2011 Published by Elsevier Ltd. All rights reserved.

Dahm N.,Queensland Research Laboratory | Dahm N.,Griffith University | Bunke H.,University of Bern | Caelli T.,Victoria Research Laboratory | And 2 more authors.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013

This paper presents techniques to address the complexity problem of subgraph isomorphism detection on large graphs. To overcome the inherently high computational complexity, the problem is simplified through the calculation and strengthening of topological node features. These features can be utilised, in principle, by any subgraph isomorphism algorithm. The design and capabilities of the proposed unified strengthening framework are discussed in detail. Additionally, the concept of an n-neighbourhood is introduced, which facilitates the development of novel features and provides an additional platform for feature strengthening. Through experiments performed with state-of-the-art subgraph isomorphism algorithms, the theoretical and practical advantages of using these techniques become evident. © 2013 Springer-Verlag.

Chen W.,Griffith University | Gao Y.,Griffith University | Gao Y.,Queensland Research Laboratory
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2010

Automatically recognizing human faces with partial occlusions is one of the most challenging problems in face analysis community. This paper presents a novel string-based face recognition approach to address the partial occlusion problem in face recognition. In this approach, a new face representation, Stringface, is constructed to integrate the relational organization of intermediate-level features (line segments) into a high-level global structure (string). The matching of two faces is done by matching two Stringfaces through a string-to-string matching scheme, which is able to efficiently find the most discriminative local parts (substrings) for recognition without making any assumption on the distributions of the deformed facial regions. The proposed approach is compared against the state-of-the-art algorithms using both the AR database and FRGC (Face Recognition Grand Challenge) ver2.0 database. Very encouraging experimental results demonstrate, for the first time, the feasibility and effectiveness of a high-level syntactic method in face recognition, showing a new strategy for face representation and recognition. © 2010 Springer-Verlag.

Folkman L.,Griffith University | Folkman L.,Queensland Research Laboratory | Stantic B.,Griffith University | Sattar A.,Griffith University | Sattar A.,Queensland Research Laboratory
BMC Bioinformatics | Year: 2013

Background: Even a single amino acid substitution in a protein sequence may result in significant changes in protein stability, structure, and therefore in protein function as well. In the post-genomic era, computational methods for predicting stability changes from only the sequence of a protein are of importance. While evolutionary relationships of protein mutations can be extracted from large protein databases holding millions of protein sequences, relevant evolutionary features for the prediction of stability changes have not been proposed. Also, the use of predicted structural features in situations when a protein structure is not available has not been explored. Results: We proposed a number of evolutionary and predicted structural features for the prediction of stability changes and analysed which of them capture the determinants of protein stability the best. We trained and evaluated our machine learning method on a non-redundant data set of experimentally measured stability changes. When only the direction of the stability change was predicted, we found that the best performance improvement can be achieved by the combination of the evolutionary features mutation likelihood and SIFT score in conjunction with the predicted structural feature secondary structure. The same two evolutionary features in the combination with the predicted structural feature accessible surface area achieved the lowest error when the prediction of actual values of stability changes was assessed. Compared to similar studies, our method achieved improvements in prediction performance. Conclusion: Although the strongest feature for the prediction of stability changes appears to be the vector of amino acid identities in the sequential neighbourhood of the mutation, the most relevant combination of evolutionary and predicted structural features further improves prediction performance. Even the predicted structural features, which did not perform well on their own, turn out to be beneficial when appropriately combined with evolutionary features. We conclude that a high prediction accuracy can be achieved knowing only the sequence of a protein when the right combination of both structural and evolutionary features is used. © 2013 Folkman et al.

Dahm N.,Griffith University | Gao Y.,Queensland Research Laboratory
Proceedings - International Conference on Pattern Recognition | Year: 2010

Many Face Recognition techniques focus on 2D-2D comparison or 3D-3D comparison, however few techniques explore the idea of cross-dimensional comparison. This paper presents a novel face recognition approach that implements cross-dimensional comparison to solve the issue of pose invariance. Our approach implements a Gabor representation during comparison to allow for variations in texture, illumination, expression and pose. Kernel scaling is used to reduce comparison time during the branching search, which determines the facial pose of input images. The conducted experiments prove the viability of this approach, with our larger kernel experiments returning 91.6% - 100% accuracy on a database comprised of both local data, and data from the USF HumanID 3D database. © 2010 IEEE.

Loading Queensland Research Laboratory collaborators
Loading Queensland Research Laboratory collaborators