Single-Channel Speech Separation Using Soft Mask Filtering

We present an approach for separating two speech signals when only one single recording of their linear mixture is available. For this purpose, we derive a filter, which we call the soft mask filter, using minimum mean square error (MMSE) estimation of the log spectral vectors of sources given the mixture's log spectral vectors. The soft mask filter's parameters are estimated using the mean and variance of the underlying sources which are modeled using the Gaussian composite source modeling (CSM) approach. It is also shown that the binary mask filter which has been empirically and extensively used in single-channel speech separation techniques is, in fact, a simplified form of the soft mask filter. The soft mask filtering technique is compared with the binary mask and Wiener filtering approaches when the input consists of male+male, female+female, and male+female mixtures. The experimental results in terms of signal-to-noise ratio (SNR) and segmental SNR show that soft mask filtering outperforms binary mask and Wiener filtering.     

Monaural Speech Separation Based on Gain Adapted Minimum Mean Square Error Estimation

We present a new model-based monaural speech separation technique for separating two speech signals from a single recording of their mixture. This work is an attempt to solve a fundamental limitation in current model-based monaural speech separation techniques in which it is assumed that the data used in the training and test phases of the separation model have the same energy level. To overcome this limitation, a gain adapted minimum mean square error estimator is derived which estimates sources under different signal-to-signal ratios. Specifically, the speakers’ gains are incorporated as unknown parameters into the separation model and then the estimator is derived in terms of the source distributions and the signal-to-signal ratio. Experimental results show that the proposed system improves the separation performance significantly when compared with a similar model without gain adaptation as well as a maximum likelihood estimator with gain estimation.     

Analysis of sliding and pressure distribution during arepositioning of persons in a simulator chair

This study was undertaken to investigate the effect of system tilt and back recline angles on sliding and pressure distribution of seated subjects. Ten able-bodied subjects adopted successively 12 postures on a multiadjustable simulator chair. The system tilt angle was varied from 0° to 45° posterior tilt, while the seat to back angle varied from 90° to 120°. A maximum of 40.2% of weight shift was found when combining a system tilt angle of 45° to a seat to back angle of 120°. Maximum value of 74 mm of sliding was observed for the acromion marker during repositioning. Significant weight shift at the level of the seat is obtained only when the system tilt angle exceeds 15° in a posterior direction. We can put forward here that a small tilt ⩽15° can be used to adjust back pressure distribution, whereas large posterior tilts are used for an effective weight shift at the seat level. The peak pressure gradient remains in general in the interval of ±30% from the neutral posture for the able-bodied subjects and is fairly constant at 15° of tilt. A significant amount of displacement along the back and seat reference plane were found for the shoulder and hip markers, but this displacement does not necessarily correspond to a pure translation motion of the pelvic segment     

Residential drug abuse treatment for probationers. Use of node-link mapping to enhance participation and progress

OBJECTIVES: the study was set up to explore regional disparities in France in the annual prevalence of cardiovascular risk factors (hypertension, high lipids, obesity, diabetes and tobacco consumption) and protective factors (alcohol intake) in members of the GAZEL Cohort and to determine which, if any, risk factors followed the same north-south gradient as the one known for cardiovascular mortality. METHODS: the population studied comprised 18,070 subjects aged from 35 to 50 years, employed by the French national gas and electricity company. Every year from 1989 to 1993 they received a self-administered questionnaire by mail. From the replies, differences in risk factor prevalence were evaluated by an odds ratio adjusted for sex and age, using a marginal model. RESULTS: a regional effect was found (p < 0.001) for arterial hypertension, with high odds ratios in northern France and low ratios in the south east, for high lipids, with high odds ratios in Champagne, the Loire region and Upper Normandy and low ratios in the south east and the Paris area, and for obesity, with high odds ratios in the north and north east and low ratios in the south, south east and west. CONCLUSION: these results show a gradient between the north and south east for hypertension, obesity, and to a lesser degree high lipids. This gradient corresponds to the known regional disparities for cardiovascular mortality.     

Cardiopulmonary thromboembolism detected by Tc-99m MH-1 antifibrin antibody

One hundred eighty three patients (73.8%) in the SFNC series received radiation therapy alone, for 60 patients and with additional chemotherapy for 129 patients. From this review, a dose-response relationship with an improved survival time with doses greater than 50 Gy could not be demonstrated (p = 0.97). The multivariate risk analysis identified the radiation therapy as one of the factors which has the most favorable impact on survival. Radiation therapy remains the treatment method of choice. These results are compared with data from the literature.     

Individual differences in dyadic cooperative learning.

The impact of individual differences on the performance of 2 roles—learner and learning facilitator—was assessed during dyadic cooperative learning. Eighty university students, 40 men and 40 women, participated in same-sex groups of 4. Each student cooperatively learned a text passage with 1 partner and then learned a 2nd passage with another partner. In a later session, the students recalled the information contained within both text passages and completed several personality measures. A social relations analysis (D. A. Kenny & L. LaVoie, 1984) was used to partition the variability in recall for the passages into various sources. Variability in recall depended strongly on individual differences in learning ability and (to a lesser extent) on individual differences in the ability to facilitate others' learning. Differences in the ability to learn text passages were independent of individual differences in the ability to facilitate others' learning. Effective learners were high in verbal ability, whereas effective learning facilitators were low in public self-consciousness and in self-monitoring. The influence of cognitive and rapport factors on the performance of the learner role and the learning facilitator role is discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effects of communication aids and strategies on cooperative teaching.

The purpose of this study was to extend the research on the use of communication aids (types of visual aids used by persons to communicate information to one another) and strategies for using communication aids during cooperative teaching. College students were the participants. One half studied a knowledge map (a multirelational semantic map that displays partial distributions of verbal information), and one half studied a text. Within each half, one-third used a communication aid and were given a strategy, one-third used a communication aid without a strategy, and one-third did not use a communication aid and were given no strategy. Among the results, Ss interacting with a knowledge map were found to perform better than Ss using a text, and Ss who were given a strategy performed better than those who were not given a strategy. The theoretical and pragmatic implications of the results are considered. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Downward Spiral: a pedagogical game depicting the dangers of substance abuse

We have developed a pedagogical game called Downward Spiral that illuminates the dangers associated with continued drug use (including alcohol). The game provides facts, quotes, and episodic consequences that can adversely affect each player's mental and physical health, social support, and financial/legal situation over the course of the game. The conceptual and empirical foundations of the game (both a college prevention version and substance abuse treatment version) are discussed. Pilot studies with college students and residents at a substance abuse treatment facility for probationers indicate that Downward Spiral is engaging, informative, and thought provoking. Plans for examining the impact of this game as a preventive technique (for the college setting) and as an approach to increasing motivation for substance abuse treatment are described.     

Development and evaluation of a learning strategy training program.

Developed an interactive learning strategy system consisting of primary and support strategies. A set of training procedures was also created to facilitate the communication of this system to college-age students. This system was assessed in the context of a 15-wk (2 hrs/wk) learning skills course with 108 Ss. Results indicate that Ss receiving this training exhibited (p?