Combat sports analytics: Boxing punch classification using overhead depth imagery

In competitive combat sporting environments like boxing, the statistics on a boxer's performance, including the amount and type of punches thrown, provide a valuable source of data and feedback which is routinely used for coaching and performance improvement purposes. This paper presents a robust framework for the automatic classification of a boxer's punches. Overhead depth imagery is employed to alleviate challenges associated with occlusions, and robust body-part tracking is developed for the noisy time-of-flight sensors. Punch recognition is addressed through both a multi-class SVM and Random Forest classifiers. A coarse-to-fine hierarchical SVM classifier is presented based on prior knowledge of boxing punches. This framework has been applied to shadow boxing image sequences taken at the Australian Institute of Sport with 8 elite boxers. Results demonstrate the effectiveness of the proposed approach, with the hierarchical SVM classifier yielding a 96% accuracy, signifying its suitability for analysing athletes punches in boxing bouts.

Hydrazine carboxylate precursors to fine particle titania, zirconia and zirconium titanate

Titanyl hydrazine carboxylate dihydrate, TiO(N2H3COO)2.2H2O, zirconyl hydrazine carboxylate dihydrate, ZrO(N2H3COO)2.2H2O and their solid solution, ZrTiO2(N2H3COO)4.4H2O have been prepared for the first time and investigated as precursors to fine particle TiO2, ZrO2 and ZrTiO4 respectively. Titania(anatase) formed has a very high surface area of 110 m2/g and zirconium titanate showed very low dielectric loss (4 x 10(-4)).

Contribution of the temporoammonic pathway to hippocampal processing

The temporoammonic (TA) pathway is the direct, monosynaptic projection from layer III of entorhinal cortex to the distal dendritic region of area CA1 of the hippo­ campus. Although this pathway has been implicated in various functions, such as memory encoding and retrieval, spatial navigation, generation of oscillatory activity, and control of hippocampal excitability, the details of its physiology are not well understood. In this thesis, I examine the contribution of the TA pathway to hippocampal processing. I find that, as has been previously reported, the TA pathway includes both excitatory, glutamatergic components and inhibitory, GABAergic components. Several new discoveries are reported in this thesis. I show that the TA pathway is subject to forms of short-term activity-dependent regulation, including paired-pulse and frequency­ dependent plasticity, similar to other hippocampal pathways such as the Schaffer collateral (SC) input from CA3 to CA1. The TA pathway provides a strongly excitatory input to stratum radiatum giant cells of CA1. The excitatory component of the TA pathway undergoes a long-lasting decrease in synaptic strength following low-frequency stimulation in a manner partially dependent on the activation of NMDA receptors. High­ frequency activation of the TA pathway recruits a feedforward inhibition that can prevent CA1 pyramidal cells from spiking in response to SC input; this spike-blocking effect shows that the TA pathway can act to regulate information flow through the hippocampal trisynaptic pathway.

CEDI: A Neural Model of Colour Vision, with Applications to Image Processing and Classification

Air Force Office of Scientific Research (F49620-01-1-0423); National Geospatial-Intelligence Agency (NMA 201-01-1-2016); National Science Foundation (SBE-035437, DEG-0221680); Office of Naval Research (N00014-01-1-0624)

Linking Visual Development and Learning to Information Processing: Preattentive and Attentive Brain Dynamics

National Science Foundation (SBE-0354378); Office of Naval Research (N00014-95-1-0657)

Concise introduction to image processing using C++

A Concise Intro to Image Processing using C++ presents state-of-the-art image processing methodology, including current industrial practices for image compression, image de-noising methods based on partial differential equations, and new image compression methods such as fractal image compression and wavelet compression. It includes elementary concepts of image processing and related fundamental tools with coding examples as well as exercises. With a particular emphasis on illustrating fractal and wavelet compression algorithms, the text covers image segmentation, object recognition, and morphology. An accompanying CD-ROM contains code for all algorithms.

Fluidised bed characterisation using Raman spectroscopy: Applications to pharmaceutical processing

This study investigates a model system for potential pharmaceutical materials in fluidised bed processes. In particular, this study proposes a novel use of Raman spectroscopy, which allows in situ measurement of the composition of the material within the fluidised bed in three spatial dimensions and as a function of time. This is achieved by recording Raman spectra from specific volumes of space. The work shows that Raman spectroscopy can be used to provide 3D maps of the concentration and chemical structure of the particles in a fluidised bed within a relatively short (120 s) time window. At the most basic level the technique measures particle density via the intensity of the Raman spectra, however this could be used. More importantly the data are also rich in spectroscopic information on the chemical structure of the fluidised particles which is useful either for monitoring a given granulation process or more generally for the analysis of the dynamics of the airflow if the data were incorporated into an appropriate model. The technique has the potential to give detailed in situ information on how the structure and composition of the granules/powders within the fluidised bed (dryer or granulator) vary with the position and evolve with time. (c) 2007 Elsevier Ltd. All rights reserved.

Multi-scale keypoint annotation: a biological approach

The primary visual cortex employs simple, complex and end-stopped cells to create a scale space of 1D singularities (lines and edges) and of 2D singularities (line and edge junctions and crossings called keypoints). In this paper we show first results of a biological model which attributes information of the local image structure to keypoints at all scales, ie junction type (L, T, +) and main line/edge orientations. Keypoint annotation in combination with coarse to fine scale processing facilitates various processes, such as image matching (stereo and optical flow), object segregation and object tracking.