QUT speaker identity verification system for EVALITA 2009

This document outlines the system submitted by the Speech and Audio Research Laboratory at the Queensland University of Technology (QUT) for the Speaker Identity Verication: Application task of EVALITA 2009. This submission consisted of a score-level fusion of three component systems, a joint-factor GMM system and two SVM systems using GLDS and GMM supervector kernels. Development and evaluation results are presented, demonstrating the effectiveness of this fused system approach.

Gate 38

Airports are a place of transition, empty halls of fleeting comings, goings and waitings. 'Gate 38' follows the experience of four groups of young people trapped at this point of departure. As contact with the outside world is cut off, the focus is placed squarely on what they’re doing, and where they’re going. A non-traditional musical set at the end of the world. Commissioned by MacGregor State High School's Centre of Artistic Development, script development included workshops with the CAD class of 2007. No musical score required.

Telephone reliability of the Frenchay Activity Index and EQ-5D amongst older adults

Background Older adults may find it problematic to attend hospital appointments due to the difficulty associated with travelling to, within and from a hospital facility for the purpose of a face-to-face assessment. This study aims to investigate equivalence between telephone and face-to-face administration for the Frenchay Activities Index (FAI) and the Euroqol-5D (EQ-5D) generic health-related quality of life instrument amongst an older adult population. Methods Patients aged >65 (n = 53) who had been discharged to the community following an acute hospital admission underwent telephone administration of the FAI and EQ-5D instruments seven days prior to attending a hospital outpatient appointment where they completed a face-to-face administration of these instruments. Results Overall, 40 subjects' datasets were complete for both assessments and included in analysis. The FAI items had high levels of agreement between the two modes of administration (item kappa's ranged 0.73 to 1.00) as did the EQ-5D (item kappa's ranged 0.67–0.83). For the FAI, EQ-5D VAS and EQ-5D utility score, intraclass correlation coefficients were 0.94, 0.58 and 0.82 respectively with paired t-tests indicating no significant systematic difference (p = 0.100, p = 0.690 and p = 0.290 respectively). Conclusion Telephone administration of the FAI and EQ-5D instruments provides comparable results to face-to-face administration amongst older adults deemed to have cognitive functioning intact at a basic level, indicating that this is a suitable alternate approach for collection of this information.

Validating the functional capacity index : a comparison of predicted versus observed total body scores

Background: The Functional Capacity Index (FCI) was designed to predict physical function 12 months after injury. We report a validation study of the FCI. Methods: This was a consecutive case series registered in the Queensland Trauma Registry who consented to the prospective 12-month telephone-administered follow-up study. FCI scores measured at 12 months were compared with those originally predicted. Results: Complete Abbreviated Injury Scale score information was available for 617 individuals, of whom 587 (95%) could be assigned at least one FCI score (range, 1-17). Agreement between the largest predicted FCI and observed FCI score was poor ([kappa] = 0.05; 95% confidence interval, 0.00-0.10) and explained only 1% of the variability in observed FCI. Using an encompassing model that included all FCI assignments, agreement remained poor ([kappa] = 0.05; 95% confidence interval, -0.02-0.12), and the model explained only 9% of the variability in observed FCI. Conclusion: The predicted functional capacity poorly agrees with actual functional outcomes. Further research should consider including other (noninjury) explanatory factors in predicting FCI at 12 months.

Portfolio assessment

At a time when global uncertainty is paramount and when a new form or re-form of curriculum is emerging – with content displaced by skills and knowledge acquisition by learning - assessment, too, begins to take on a new from or re-form. The focus for assessment has shifted to that which engages and promotes learning as s process rather than an assessment that focuses solely on measuring and reporting learning as product or score. The use of the portfolio for assessment offers the potential for the process and progress – integral to learning - to be included.

Osteoarthritis Index delivered by mobile phone (m-WOMAC) is valid, reliable, and responsive

Objectives: To evaluate the validity, reliability and responsiveness of EDC using the WOMAC® NRS 3.1 Index on Motorola V3 mobile phones. ---------- Methods: Patients with osteoarthritis (OA) undergoing primary unilateral hip or knee joint replacement surgery were assessed pre-operatively and 3-4 months post-operatively. Patients completed the WOMAC® Index in paper (p-WOMAC®) and electronic (m-WOMAC®) format in random order. ---------- Results: 24 men and 38 women with hip and knee OA participated and successfully completed the m-WOMAC® questionnaire. Pearson correlations between the summated total index scores for the p-WOMAC® and m-WOMAC® pre- and post-surgery were 0.98 and 0.99 (p<0.0001). There was no clinically important or statistically significant between-method difference in the adjusted total summated scores, pre- and post-surgery (adjusted mean difference = 4.44, p = 0.474 and 1.73, p = 0.781). Internal consistency estimates of m-WOMAC® reliability were 0.87 – 0.98. The m-WOMAC® detected clinically important, statistically significant (p<0.0001) improvements in pain, stiffness, function and total index score. ---------- Conclusions: Sixty-two patients with hip and knee OA successfully completed EDC by Motorola V3 mobile phone using the m-WOMAC® NRS3.1 Index; completion times averaging only 1-1.5 minutes longer than the p-WOMAC® Index. Data were successfully and securely transmitted from patients in Australia to a server in the USA. There was close agreement and no significant differences between m-WOMAC® and p-WOMAC® scores. This study confirms the validity, reliability and responsiveness of the Exco InTouch engineered, Java-based m-WOMAC® Index application. EDC with the m-WOMAC® Index provides unique opportunities for using quantitative measurement in clinical research and practice.

The Validity of the Brief Michigan Alcohol Screening Test (bMAST) as a Problem Drinking Severity Measure

Objective: The Brief Michigan Alcoholism Screening Test (bMAST) is a 10-item test derived from the 25-item Michigan Alcoholism Screening Test (MAST). It is widely used in the assessment of alcohol dependence. In the absence of previous validation studies, the principal aim of this study was to assess the validity and reliability of the bMAST as a measure of the severity of problem drinking. Method: There were 6,594 patients (4,854 men, 1,740 women) who had been referred for alcohol-use disorders to a hospital alcohol and drug service who voluntarily participated in this study. Results: An exploratory factor analysis defined a two-factor solution, consisting of Perception of Current Drinking and Drinking Consequences factors. Structural equation modeling confirmed that the fit of a nine-item, two-factor model was superior to the original one-factor model. Concurrent validity was assessed through simultaneous administration of the Alcohol Use Disorders Identification Test (AUDIT) and associations with alcohol consumption and clinically assessed features of alcohol dependence. The two-factor bMAST model showed moderate correlations with the AUDIT. The two-factor bMAST and AUDIT were similarly associated with quantity of alcohol consumption and clinically assessed dependence severity features. No differences were observed between the existing weighted scoring system and the proposed simple scoring system. Conclusions: In this study, both the existing bMAST total score and the two-factor model identified were as effective as the AUDIT in assessing problem drinking severity. There are additional advantages of employing the two-factor bMAST in the assessment and treatment planning of patients seeking treatment for alcohol-use disorders. (J. Stud. Alcohol Drugs 68: 771-779,2007)

A novel technique to demonstrate disturbed appetite profiles in haemodialysis patients

Background Malnutrition is common among dialysis patients and is associated with an adverse outcome. One cause of this is a persistent reduction in nutrient intake, suggesting an abnormality of appetite regulation. Methods We used a novel technique to describe the appetite profile in 46 haemodialysis (HD) patients and 40 healthy controls. The Electronic Appetite Rating System (EARS) employs a palmtop computer to collect hourly ratings of motivation to eat and mood. We collected data on hunger, desire to eat, fullness, and tiredness. HD subjects were monitored on the dialysis day and the interdialytic day. Controls were monitored for 1 or 2 days. Results Temporal profiles of motivation to eat for the controls were similar on both days. Temporal profiles of motivation to eat for the HD group were lower on the dialysis day. Mean HD scores were not significantly different from controls. Dietary records indicated that dialysis patients consumed less food than controls. Conclusions Our data indicate that the EARS can be used to monitor subjective appetite states continuously in a group of HD patients. A HD session reduces hunger and desire to eat. Patients feel more tired after dialysis. This does not correlate with their hunger score, but does correlate with their fullness rating. Nutrient intake is reduced, suggesting a resetting of appetite control for the HD group. The EARS may be useful for intervention studies.

Application of data mining techniques in the prediction of coronary artery disease : use of anaesthesia time-series and patient risk factor data

The high morbidity and mortality associated with atherosclerotic coronary vascular disease (CVD) and its complications are being lessened by the increased knowledge of risk factors, effective preventative measures and proven therapeutic interventions. However, significant CVD morbidity remains and sudden cardiac death continues to be a presenting feature for some subsequently diagnosed with CVD. Coronary vascular disease is also the leading cause of anaesthesia related complications. Stress electrocardiography/exercise testing is predictive of 10 year risk of CVD events and the cardiovascular variables used to score this test are monitored peri-operatively. Similar physiological time-series datasets are being subjected to data mining methods for the prediction of medical diagnoses and outcomes. This study aims to find predictors of CVD using anaesthesia time-series data and patient risk factor data. Several pre-processing and predictive data mining methods are applied to this data. Physiological time-series data related to anaesthetic procedures are subjected to pre-processing methods for removal of outliers, calculation of moving averages as well as data summarisation and data abstraction methods. Feature selection methods of both wrapper and filter types are applied to derived physiological time-series variable sets alone and to the same variables combined with risk factor variables. The ability of these methods to identify subsets of highly correlated but non-redundant variables is assessed. The major dataset is derived from the entire anaesthesia population and subsets of this population are considered to be at increased anaesthesia risk based on their need for more intensive monitoring (invasive haemodynamic monitoring and additional ECG leads). Because of the unbalanced class distribution in the data, majority class under-sampling and Kappa statistic together with misclassification rate and area under the ROC curve (AUC) are used for evaluation of models generated using different prediction algorithms. The performance based on models derived from feature reduced datasets reveal the filter method, Cfs subset evaluation, to be most consistently effective although Consistency derived subsets tended to slightly increased accuracy but markedly increased complexity. The use of misclassification rate (MR) for model performance evaluation is influenced by class distribution. This could be eliminated by consideration of the AUC or Kappa statistic as well by evaluation of subsets with under-sampled majority class. The noise and outlier removal pre-processing methods produced models with MR ranging from 10.69 to 12.62 with the lowest value being for data from which both outliers and noise were removed (MR 10.69). For the raw time-series dataset, MR is 12.34. Feature selection results in reduction in MR to 9.8 to 10.16 with time segmented summary data (dataset F) MR being 9.8 and raw time-series summary data (dataset A) being 9.92. However, for all time-series only based datasets, the complexity is high. For most pre-processing methods, Cfs could identify a subset of correlated and non-redundant variables from the time-series alone datasets but models derived from these subsets are of one leaf only. MR values are consistent with class distribution in the subset folds evaluated in the n-cross validation method. For models based on Cfs selected time-series derived and risk factor (RF) variables, the MR ranges from 8.83 to 10.36 with dataset RF_A (raw time-series data and RF) being 8.85 and dataset RF_F (time segmented time-series variables and RF) being 9.09. The models based on counts of outliers and counts of data points outside normal range (Dataset RF_E) and derived variables based on time series transformed using Symbolic Aggregate Approximation (SAX) with associated time-series pattern cluster membership (Dataset RF_ G) perform the least well with MR of 10.25 and 10.36 respectively. For coronary vascular disease prediction, nearest neighbour (NNge) and the support vector machine based method, SMO, have the highest MR of 10.1 and 10.28 while logistic regression (LR) and the decision tree (DT) method, J48, have MR of 8.85 and 9.0 respectively. DT rules are most comprehensible and clinically relevant. The predictive accuracy increase achieved by addition of risk factor variables to time-series variable based models is significant. The addition of time-series derived variables to models based on risk factor variables alone is associated with a trend to improved performance. Data mining of feature reduced, anaesthesia time-series variables together with risk factor variables can produce compact and moderately accurate models able to predict coronary vascular disease. Decision tree analysis of time-series data combined with risk factor variables yields rules which are more accurate than models based on time-series data alone. The limited additional value provided by electrocardiographic variables when compared to use of risk factors alone is similar to recent suggestions that exercise electrocardiography (exECG) under standardised conditions has limited additional diagnostic value over risk factor analysis and symptom pattern. The effect of the pre-processing used in this study had limited effect when time-series variables and risk factor variables are used as model input. In the absence of risk factor input, the use of time-series variables after outlier removal and time series variables based on physiological variable values’ being outside the accepted normal range is associated with some improvement in model performance.

An investigation of foot and ankle problems experienced by nurses

Relatively little information has been reported about foot and ankle problems experienced by nurses, despite anecdotal evidence which suggests they are common ailments. The purpose of this study was to improve knowledge about the prevalence of foot and ankle musculoskeletal disorders (MSDs) and to explore relationships between these MSDs and proposed risk factors. A review of the literature relating to work-related MSDs, MSDs in nursing, foot and lower-limb MSDs, screening for work-related MSDs, foot discomfort, footwear and the prevalence of foot problems in the community was undertaken. Based on the review, theoretical risk factors were proposed that pertained to the individual characteristics of the nurses, their work activity or their work environment. Three studies were then undertaken. A cross-sectional survey of 304 nurses, working in a large tertiary paediatric hospital, established the prevalence of foot and ankle MSDs. The survey collected information about self-reported risk factors of interest. The second study involved the clinical examination of a subgroup of 40 nurses, to examine changes in body discomfort, foot discomfort and postural sway over the course of a single work shift. Objective measurements of additional risk factors, such as individual foot posture (arch index) and the hardness of shoe midsoles, were performed. A final study was used to confirm the test-retest reliability of important aspects of the survey and key clinical measurements. Foot and ankle problems were the most common MSDs experienced by nurses in the preceding seven days (42.7% of nurses). They were the second most common MSDs to cause disability in the last 12 months (17.4% of nurses), and the third most common MSDs experienced by nurses in the last 12 months (54% of nurses). Substantial foot discomfort (Visual Analogue Scale (VAS) score of 50mm or more) was experienced by 48.5% of nurses at sometime in the last 12 months. Individual risk factors, such as obesity and the number of self-reported foot conditions (e.g., callouses, curled toes, flat feet) were strongly associated with the likelihood of experiencing foot problems in the last seven days or during the last 12 months. These risk factors showed consistent associations with disabling foot conditions and substantial foot discomfort. Some of these associations were dependent upon work-related risk factors, such as the location within the hospital and the average hours worked per week. Working in the intensive care unit was associated with higher odds of experiencing foot problems within the last seven days, foot problems in the last 12 months and foot problems that impaired activity in the last 12 months. Changes in foot discomfort experienced within a day, showed large individual variability. Fifteen of the forty nurses experienced moderate/substantial foot discomfort at the end of their shift (VAS 25+mm). Analysis of the association between risk factors and moderate/substantial foot discomfort revealed that foot discomfort was less likely for nurses who were older, had greater BMI or had lower foot arches, as indicated by higher arch index scores. The nurses’ postural sway decreased over the course of the work shift, suggesting improved body balance by the end of the day. These findings were unexpected. Further clinical studies examining individual nurses on several work shifts are needed to confirm these results, particularly due to the small sample size and the single measurement occasion. There are more than 280,000 nurses registered to practice in Australia. The nursing workforce is ageing and the prevalence of foot problems will increase. If the prevalence estimates from this study are extrapolated to the profession generally, more than 70,000 hospital nurses have experienced substantial foot discomfort and 25-30,000 hospital nurses have been limited in their activity due to foot problems during the last 12 months. Nurses with underlying foot conditions were more likely to report having foot problems at work. Strategies to prevent or manage foot conditions exist and they should be disseminated to nurses. Obesity is a significant risk factor for foot and ankle MSDs and these nurses may need particular assistance to manage foot problems. The risk of foot problems for particular groups of nurses, e.g. obese nurses, may vary depending upon the location within the hospital. Further research is needed to confirm the findings of this study. Similar studies should be conducted in other occupational groups that require workers to stand for prolonged periods.

Automatic speaker recognition under adverse conditions

Speaker verification is the process of verifying the identity of a person by analysing their speech. There are several important applications for automatic speaker verification (ASV) technology including suspect identification, tracking terrorists and detecting a person’s presence at a remote location in the surveillance domain, as well as person authentication for phone banking and credit card transactions in the private sector. Telephones and telephony networks provide a natural medium for these applications. The aim of this work is to improve the usefulness of ASV technology for practical applications in the presence of adverse conditions. In a telephony environment, background noise, handset mismatch, channel distortions, room acoustics and restrictions on the available testing and training data are common sources of errors for ASV systems. Two research themes were pursued to overcome these adverse conditions: Modelling mismatch and modelling uncertainty. To directly address the performance degradation incurred through mismatched conditions it was proposed to directly model this mismatch. Feature mapping was evaluated for combating handset mismatch and was extended through the use of a blind clustering algorithm to remove the need for accurate handset labels for the training data. Mismatch modelling was then generalised by explicitly modelling the session conditions as a constrained offset of the speaker model means. This session variability modelling approach enabled the modelling of arbitrary sources of mismatch, including handset type, and halved the error rates in many cases. Methods to model the uncertainty in speaker model estimates and verification scores were developed to address the difficulties of limited training and testing data. The Bayes factor was introduced to account for the uncertainty of the speaker model estimates in testing by applying Bayesian theory to the verification criterion, with improved performance in matched conditions. Modelling the uncertainty in the verification score itself met with significant success. Estimating a confidence interval for the "true" verification score enabled an order of magnitude reduction in the average quantity of speech required to make a confident verification decision based on a threshold. The confidence measures developed in this work may also have significant applications for forensic speaker verification tasks.

Automatic spoken language identification utilizing acoustic and phonetic speech information

Automatic spoken Language Identi¯cation (LID) is the process of identifying the language spoken within an utterance. The challenge that this task presents is that no prior information is available indicating the content of the utterance or the identity of the speaker. The trend of globalization and the pervasive popularity of the Internet will amplify the need for the capabilities spoken language identi¯ca- tion systems provide. A prominent application arises in call centers dealing with speakers speaking di®erent languages. Another important application is to index or search huge speech data archives and corpora that contain multiple languages. The aim of this research is to develop techniques targeted at producing a fast and more accurate automatic spoken LID system compared to the previous National Institute of Standards and Technology (NIST) Language Recognition Evaluation. Acoustic and phonetic speech information are targeted as the most suitable fea- tures for representing the characteristics of a language. To model the acoustic speech features a Gaussian Mixture Model based approach is employed. Pho- netic speech information is extracted using existing speech recognition technol- ogy. Various techniques to improve LID accuracy are also studied. One approach examined is the employment of Vocal Tract Length Normalization to reduce the speech variation caused by di®erent speakers. A linear data fusion technique is adopted to combine the various aspects of information extracted from speech. As a result of this research, a LID system was implemented and presented for evaluation in the 2003 Language Recognition Evaluation conducted by the NIST.

Vision-based classification of solder joint defects

Inspection of solder joints has been a critical process in the electronic manufacturing industry to reduce manufacturing cost, improve yield, and ensure product quality and reliability. The solder joint inspection problem is more challenging than many other visual inspections because of the variability in the appearance of solder joints. Although many research works and various techniques have been developed to classify defect in solder joints, these methods have complex systems of illumination for image acquisition and complicated classification algorithms. An important stage of the analysis is to select the right method for the classification. Better inspection technologies are needed to fill the gap between available inspection capabilities and industry systems. This dissertation aims to provide a solution that can overcome some of the limitations of current inspection techniques. This research proposes two inspection steps for automatic solder joint classification system. The “front-end” inspection system includes illumination normalisation, localization and segmentation. The illumination normalisation approach can effectively and efficiently eliminate the effect of uneven illumination while keeping the properties of the processed image. The “back-end” inspection involves the classification of solder joints by using Log Gabor filter and classifier fusion. Five different levels of solder quality with respect to the amount of solder paste have been defined. Log Gabor filter has been demonstrated to achieve high recognition rates and is resistant to misalignment. Further testing demonstrates the advantage of Log Gabor filter over both Discrete Wavelet Transform and Discrete Cosine Transform. Classifier score fusion is analysed for improving recognition rate. Experimental results demonstrate that the proposed system improves performance and robustness in terms of classification rates. This proposed system does not need any special illumination system, and the images are acquired by an ordinary digital camera. In fact, the choice of suitable features allows one to overcome the problem given by the use of non complex illumination systems. The new system proposed in this research can be incorporated in the development of an automated non-contact, non-destructive and low cost solder joint quality inspection system.

Pretreatment malnutrition and quality of life - association with prolonged length of hospital stay among patients with gynecological cancer: A cohort study

Background Length of hospital stay (LOS) is a surrogate marker for patients' well-being during hospital treatment and is associated with health care costs. Identifying pretreatment factors associated with LOS in surgical patients may enable early intervention in order to reduce postoperative LOS. Methods This cohort study enrolled 157 patients with suspected or proven gynecological cancer at a tertiary cancer centre (2004-2006). Before commencing treatment, the scored Patient Generated - Subjective Global Assessment (PG-SGA) measuring nutritional status and the Functional Assessment of Cancer Therapy-General (FACT-G) scale measuring quality of life (QOL) were completed. Clinical and demographic patient characteristics were prospectively obtained. Patients were grouped into those with prolonged LOS if their hospital stay was greater than the median LOS and those with average or below average LOS. Results Patients' mean age was 58 years (SD 14 years). Preoperatively, 81 (52%) patients presented with suspected benign disease/pelvic mass, 23 (15%) with suspected advanced ovarian cancer, 36 (23%) patients with suspected endometrial and 17 (11%) with cervical cancer, respectively. In univariate models prolonged LOS was associated with low serum albumin or hemoglobin, malnutrition (PG-SGA score and PG-SGA group B or C), low pretreatment FACT-G score, and suspected diagnosis of cancer. In multivariable models, PG-SGA group B or C, FACT-G score and suspected diagnosis of advanced ovarian cancer independently predicted LOS. Conclusions Malnutrition, low quality of life scores and being diagnosed with advanced ovarian cancer are the major determinants of prolonged LOS amongst gynecological cancer patients. Interventions addressing malnutrition and poor QOL may decrease LOS in gynecological cancer patients.