Latest NICTA Publications

Discriminant Absorption Feature Learning for Material Classification

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Zhouyu Fu, Antonio Robles-Kelly

Abstract: In this paper, we develop a novel approach to object material identification in spectral imaging by combining the use of invariant spectral absorption features and statistical machine learning techniques. Our method hinges in the relevance of spectral absorption features for material identification and casts the problem into a pattern recognition setting by making use of an invariant representation of the most discriminant band-segments in the spectra. Thus, here, we view the identification problem as a classification task, which is effected based upon those invariant absorption bands in the spectra which are most discriminative between the materials under study. To robustly recover those bands that are most relevant to the identification process, we make use of discriminant learning. To illustrate the utility of our method for purposes of material identification, we perform experiments on both terrestrial and remotely sensed hyperspectral imaging data and compare our results to those yielded by an alternative.

IEEE Transactions on Geoscience and Remote Sensing-November 2011

Graph Embedding via Riemannian Submersion Learning

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Haifeng Zhao, Antonio Robles-Kelly, Jun Zhou, Jianfeng Lu, Jing-yu Yang

Abstract: In this paper, we tackle the problem of embedding a set of relational structures into a metric space for purposes of matching and categorisation. To this end, we view the problem from a Riemannian perspective and make use of the concepts of charts on the manifold to define the embedding as a mixture of class-specific submersions. Formulated in this manner, the mixture weights are recovered using a probability density estimation on the embedded graph node coordinates. Further, we recover these class-specific submersions making use of an iterative trust-region method so as to minimise the L2 norm between the hard limit of the graph-vertex posterior probabilities and their estimated values. The method presented here is quite general in nature and allows tasks such as matching, categorisation and retrieval. We show results on graph matching, shape categorisation and digit classification on synthetic data, the MPEG-7 database and the MNIST dataset.

Computer Vision and Image Understanding-July 2011

Optimal Adaptive Uplink Attenuation Algorithms for WCDMA Femtocell

Wed, 08 Feb 2012 00:00:00 +0000

Authors: He Wang, Ming Zhao, Mark Reed

Abstract: Femtocells are assuming an increasing important role to extend the cellular network coverage and provide high speed data service for indoor environments. To cope with the extreme interfering cases in the WCDMA uplink scenario, Home NodeB (HNB) receivers need to employ uplink attenuation at the RF frontend to suppress the interfering signals introduced by unattached co-channel users, e.g., a penetrating signal from macro-connected users in the proximity. The optimal level of uplink attenuation at the HNB receiver should be designed to balance the requirements between the noise rise limitation in HNB and maximum transmit power cap for the femto-connected users. In this paper, we introduce two optimal adaptive schemes that apply the minimum and necessary level of attenuation to the received signal and show the outage probability improvement by the method of simulation.

IEEIEEE GLOBECOM 2011-December 2011

Monocular Template-based reconstruction of inextensible surfaces

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Perriollat Mathieu, Hartley Richard, Adrien Bartoli

International Conference on Computer Vision-March 2011

Compressive evaluation in human motion tracking

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Yifan Lu, Lei Wang, Hongdong Li, Richard Hartley, Dan Xu

Asian Conference on Computer Vision-January 2011

Gradual sampling and mutual information maximation for markerless motion capture

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Yifan Lu, Lei Wang, Richard Hartley, Hongdong Li, Dan Xu

Asian Conference on Computer Vision-January 2011

Pyramid center-symmetric local binary/trinary patterns for effective pedestrian detection

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Yongbin Zheng, Chunhua Shen, Richard Hartley, Xinsheng Huang

Asian Conference on Computer Vision-January 2011

Monocular Template-based reconstruction of smooth and inextensible surfaces

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Florent Brunet, Richard Hartley, Adrien Bartoli

Asian Conference on Computer Vision-January 2011

Ultra-Broadband Indoor Full-Duplex WDM Optical Wireless Communication with Multi-Mode Fiber.

Wed, 08 Feb 2012 00:00:00 +0000

Authors: K. Wang, A. Nirmalathas, C. Lim, S. Skafidas

2011 International Quantum Electronics/Lasers and Electro-Optics Conference Pacific Rim-September 2011

4 x 12.5 Gb/s WDM Optical Wireless Communication System for Indoor Applications.

Wed, 08 Feb 2012 00:00:00 +0000

Authors: K. Wang, A. Nirmalathas, C. Lim, E. Skafidas

JOURNAL OF LIGHTWAVE TECHNOLOGY-July 2011

On the Use of the Chi-squared Distance for the Structured Learning of Graph Embeddings

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Haifeng Zhao, Antonio Robles-Kelly, Jun Zhou

Abstract: In this paper, we describe the use of concepts from the areas of structural and statistical pattern recognition for the purposes of recovering a mapping which can be viewed as an operator on the graph attribute-set. This mapping can be used to embed graphs into spaces where tasks such as classification and retrieval can be effected. To do this, we depart from concepts in graph theory so as to introduce mappings as operators over graph spaces. This treatment leads tothe recovery of a mapping based upon the graph attributes which is related to the edge-space of the graphs under study. As a result, the recovered mappingis a linear operator over the attribute set which is associated with the graph topology. To recover this mapping, we employ an optimisation approach whose cost function is based upon the Chi-squared distance and is related to the target function used in discrete Markov Random Field approaches. Thus, the method presented here provides a link between concepts in graph theory, statistical inference and linear operators. We illustrate the utility of the recovered embedding for purposes of shape categorisation and retrieval. We also compare our results to those yielded by alternatives.

Digital Image Computing Techniques and Applications (DICTA)-December 2011

On the Recovery of Shape and Reflectance from a Single Multispectral Image

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Sejuti Rahman, Antonio Robles-Kelly

Abstract: This paper presents a method for recovering the reflectanceand shape of a surface from a single hyperspectralimage. To do this, we depart from a reflectance modelbased on a physical interpretation of the reflection process,where the geometry of the scene and the properties of theobject surface describe the image formation process. Withthe model in hand, we propose a solution to the recoveryof reflection parameters via an optimisation approach thataims at recoverying both, the surface normals and the reflectance.This, in turn, allows for the application of themethod by Frankot and Chellappa [8] so as to recover thesurface depth at each pixel. We demonstrate the utility ofour method for the recovery of both, reflectance and surfacedepth on both synthetic and real world images.

Digital Image Computing Techniques and Applications (DICTA)-December 2011

Cost-Effective Introduction and Energy-Efficient Operation of Long-Reach WDM/TDM PON Systems.

Wed, 08 Feb 2012 00:00:00 +0000

Authors: H. Feng, JC. Chae, AV. Tran, A. Nirmalathas

JOURNAL OF LIGHTWAVE TECHNOLOGY-November 2011

Specific Absorption Rate Distribution on a Human Head Model from Inductive Power Coils.

Wed, 08 Feb 2012 00:00:00 +0000

Authors: DC. Ng, X. Wang, GK Felic, S. Bai, C. Boyd, M. Halpern, E. Skafidas

10th Int. Symposium on Electromagnetic Compatibility EMC Europe 2011-September 2011

Estimation of Brain Myelin Water Fractions from T2 Distributions.

Wed, 08 Feb 2012 00:00:00 +0000

Authors: K. Layton, M. Morelande, PM. Farrell, B. Moran, LA. Johnston

The Australian and New Zealand Society for Magnetic Resonance Annual Meeting, 2011,ANZMAG 2011-October 2011

BetaSearch: a new method for querying beta-residue motifs.

Wed, 08 Feb 2012 00:00:00 +0000

Authors: HK. Ho, G. Gange, MJ. Kuiper, R. Kotagiri

BMC Bioinformatics-January 2012

An Ultra Low Power and Small size PLL for Wearable and Implantable Medical Sensors.

Wed, 08 Feb 2012 00:00:00 +0000

Authors: B. Ghafari, L. Koushaeian, F. Goodarzi

IEEE Consumer Comm. and Networking Conference. (CCNC)-January 2012

Provisioning In-House Mobility for FTTH Customers by Incorporating Modifications in Optical Network Unit (ONU).

Wed, 08 Feb 2012 00:00:00 +0000

Authors: M. Bakaul, AHMR. Islam, A. Nirmalathas, E. Skafidas, L. Mehedy

24th LEOS Annual Meeting-October 2011

On the optimality of sequential forward feature selection using class separability

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Lei Wang, Chunhua Shen, Richard Hartley

Digital Image Computing: Techniques and Applications-December 2011

Superpixels, Occlusion and Stereo

Wed, 08 Feb 2012 00:00:00 +0000

Authors: Yuhang Zhang, Richard Hartley, John Mashford, Stewart Burn

Digital Image Computing: Techniques and Applications-December 2011