A deep bidirectional LSTM approach for video-realistic talking head

This paper proposes a deep bidirectional long short-term memory approach in modeling the long contextual, nonlinear mapping between audio and visual streams for video-realistic talking head. In training stage, an audio-visual stereo database is firstly recorded as a subject talking to a camera. The audio streams are converted into acoustic feature, i.e. Mel-Frequency Cepstrum Coefficients (MFCCs), and their textual labels are also extracted. The visual streams, in particular, the lower face region, are compactly represented by active appearance model (AAM) parameters by which the shape and texture variations can be jointly modeled. Given pairs of the audio and visual parameter sequence, a DBLSTM model is trained to learn the sequence mapping from audio to visual space. For any unseen speech audio, whether it is original recorded or synthesized by text-to-speech (TTS), the trained DBLSTM model can predict a convincing AAM parameter trajectory for the lower face animation. To further improve the realism of the proposed talking head, the trajectory tiling method is adopted to use the DBLSTM predicted AAM trajectory as a guide to select a smooth real sample image sequence from the recorded database. We then stitch the selected lower face image sequence back to a background face video of the same subject, resulting in a video-realistic talking head. Experimental results show that the proposed DBLSTM approach outperforms the existing HMM-based approach in both objective and subjective evaluations.     

Characterization of the Advanced Satellite for Cosmology and Astrophysics x-ray telescope: preflight calibration and ray tracing

The x-ray properties of multinested thin-foil mirror x-ray telescopes (XRT's) on board ASCA, the Advanced Satellite for Cosmology and Astrophysics, were fully evaluated with an x-ray pencil beam.Scanning over the telescope aperture of 35 cm in diameter with an x-ray pencil beam, we found the effective area of a set of XRT's to be 325, 200, and 113 cm(2) at energies of 1.5, 4.5, and 8.0 keV, respectively. We derive the point-spread functions (PSF's) of the XRT's by measuring the image profile at the focal plane with an x-ray CCD. The PSF is found to exhibit a sharp core concentrated within 30 arcsec and a broad wing extended to 3 arcmin in half-power diameter. We also evaluate the contribution of stray light, which is caused by the single reflection of x rays by primary or secondary mirrors and by the backside reflection of the mirrors. To obtain the characteristics of the XRT in the energy region of 0.5-10.0 keV, incorporated with the measurements at discrete energies, we develop a ray-tracing method with the telescope design parameter, the PSF, and optical constants. In particular, we obtain the optical constants around the gold-atom M shell (Au-M) absorption-edge energies by measuring the reflectivity of our mirror sample, with monochromatized x-rays in the energy range of 2.0-3.5 keV from synchrotron radiation. Taking into account the PSF's and optical constants, we find that our ray-tracing program can reproduce all these XRT performances.     

Designing Metallic and Insulating Nanocrystal Heterostructures to Fabricate Highly Sensitive and Solution Processed Strain Gauges for Wearable Sensors

All‐solution processed, high‐performance wearable strain sensors are demonstrated using heterostructure nanocrystal (NC) solids. By incorporating insulating artificial atoms of CdSe quantum dot NCs into metallic artificial atoms of Au NC thin film matrix, metal–insulator heterostructures are designed. This hybrid structure results in a shift close to the percolation threshold, modifying the charge transport mechanism and enhancing sensitivity in accordance with the site percolation theory. The number of electrical pathways is also manipulated by creating nanocracks to further increase its sensitivity, inspired from the bond percolation theory. The combination of the two strategies achieves gauge factor up to 5045, the highest sensitivity recorded among NC‐based strain gauges. These strain sensors show high reliability, durability, frequency stability, and negligible hysteresis. The fundamental charge transport behavior of these NC solids is investigated and the combined site and bond percolation theory is developed to illuminate the origin of their enhanced sensitivity. Finally, all NC‐based and solution‐processed strain gauge sensor arrays are fabricated, which effectively measure the motion of each finger joint, the pulse of heart rate, and the movement of vocal cords of human. This work provides a pathway for designing low‐cost and high‐performance electronic skin or wearable devices.     

What is the best method of surgical training?: A report of America's leading senior surgeons

OBJECTIVES: To characterize the career choices and developments made by leading senior surgeons in this country and to examine hypothetically whether application of a short tracking program would have hindered their career decisions. DESIGN: A survey pertaining to each surgeon's career, decisions, and opinions concerning surgical training. SETTING AND PARTICIPANTS: Senior surgeons of regional and national surgical societies. MAIN OUTCOME MEASURE: Survey responses. RESULTS: A total of 352 surveys (41.4%) were received. Respondents answered that the most common reasons for choosing a specialty were role models or mentors (56%), research (51%), and available patient population (23%). The 2 most common stages in a career at which the respondents became interested in a specialty, or an area of expertise, were at the junior residency level (when the specialty was chosen) and at the assistant professor level (when a more specific topic within the specialty was chosen). The most common stage at which the group believed they acquired their expertise was also at the assistant professor level. Seventy-one percent of respondents believed broad training was superior to a short tracking system, although none had participated in shortened surgical training. CONCLUSIONS: Most leading senior surgeons in this country still believe that broad surgical training is superior and should be maintained. Because career specialties in this surveyed group were generally chosen in early residency, a hypothetical application of the short tracking system would have still allowed for these important decisions to be made. Also, it seems likely that specialty and career development would not have been hindered because "expertization" mostly occurred after training was completed. Regardless of training method, a role model or mentor seems most important in career choices and developments.     

Properties of whole saliva and dental plaque in relation to 40-month consumption of chewing gums containing xylitol, sorbitol of sucrose

The concern with nursing research in the future work of nurse graduates. Expectations from the viewpoint of the nurse researcher. Keeping in mind the concerns which occupy nurse researchers now and in the future, expectations are raised of the graduates of degree programmes in nursing science which are now also being established in the Federal Republic of Germany. This takes into account the structurally determined areas of self determined action in new fields of professional practice. Nurse graduates will be seen as "change agents" in their future areas of work and will have three essential aspects of involvement with nursing research: 1. they themselves will conduct research 2. they will receive, put into practice and transmit results of nursing research 3. they will initiate and commission nursing research projects. According to the broad topics of nursing research, requirements for nurses to act as professionals in their new working fields are formulated referring to the given structural conditions in Germany.     

Analysis of pressure-driven electrokinetic flows in hydrophobic microchannels with slip-dependent zeta potential

The present study is an analysis of pressure-driven electrokinetic flows in hydrophobic microchannels with emphasis on the slip effects under coupling of interfacial electric and fluid slippage phenomena. Commonly used linear model with slip-independent zeta potential and the nonlinear model at limiting (high-K) condition with slip-dependent zeta potential are solved analytically. Then, numerical solutions of the electrokinetic flow model with zeta potential varying with slip length are analyzed. Different from the general notion of “the more hydrophobic the channel wall, the higher the flowrate,” the results with slip-independent and slip-dependent zeta potentials both disclose that flowrate becomes insensitive to the wall hydrophobicity or fluid slippage at sufficiently large slip lengths. Boundary slip not only assists fluid motion but also enhances counter-ions transport in EDL and, thus, results in strong streaming potential as well as electrokinetic retardation. With slip-dependent zeta potential considered, flowrate varies non-monotonically with increasing slip length due to competition of the favorable and adverse effects with more complicated interactions. The influence of the slip on the electrokinetic flow is eventually nullified at large slip lengths for balance of the counter effects, and the flowrate becomes insensitive to further hydrophobicity of the microchannel. The occurrence of maximum, minimum, and insensitivity on the flowrate-slip curves can be premature at a higher zeta potential and/or larger electrokinetic separation distance.     

A Constrained Line Search Optimization Method for Discriminative Training of HMMs

In this paper, we propose a novel optimization algorithm called constrained line search (CLS) for discriminative training (DT) of Gaussian mixture continuous density hidden Markov model (CDHMM) in speech recognition. The CLS method is formulated under a general framework for optimizing any discriminative objective functions including maximum mutual information (MMI), minimum classification error (MCE), minimum phone error (MPE)/minimum word error (MWE), etc. In this method, discriminative training of HMM is first cast as a constrained optimization problem, where Kullback-Leibler divergence (KLD) between models is explicitly imposed as a constraint during optimization. Based upon the idea of line search, we show that a simple formula of HMM parameters can be found by constraining the KLD between HMM of two successive iterations in an quadratic form. The proposed CLS method can be applied to optimize all model parameters in Gaussian mixture CDHMMs, including means, covariances, and mixture weights. We have investigated the proposed CLS approach on several benchmark speech recognition databases, including TIDIGITS, Resource Management (RM), and Switchboard. Experimental results show that the new CLS optimization method consistently outperforms the conventional EBW method in both recognition performance and convergence behavior.     

Estimation of break location and size for loss of coolant accidents using neural networks

In this work, a probabilistic neural network (PNN) that has been applied well to the classification problems is used in order to identify the break locations of loss of coolant accidents (LOCA) such as hot-leg, cold-leg and steam generator tubes. Also, a fuzzy neural network (FNN) is designed to estimate the break size. The inputs to PNN and FNN are time-integrated values obtained by integrating measurement signals during a short time interval after reactor scram. An automatic structure constructor for the fuzzy neural network automatically selects the input variables from the time-integrated values of many measured signals, and optimizes the number of rules and its related parameters. It is verified that the proposed algorithm identifies very well the break locations of LOCAs and also, estimate their break size accurately.     

Numerical method for the analysis of queuing models with priority jumps1

An algorithmic approach to study queuing models with common finite and infinite buffer and jump priorities is developed. It is assumed that upon arrival of a low-priority call, one call of such kind can be transferred with some probability to the end of the queue of high-priority calls. The transition probability depends on the state of the queue of heterogeneous calls. The algorithms are proposed to calculate the quality of service metrics of such queuing models.     

Comparison of alteplase (tissue plasminogen activator) high‐dose vs. low‐dose protocol in restoring hemodialysis catheter function: The ALTE‐DOSE study

Hemodialysis catheter (HDC) dysfunction due to thrombosis is common, and dysfunction incidence can reach up to 50% within 1 year of use. Although administration of intraluminal alteplase (tissue plasminogen activator [tPA]) is the standard of practice to pharmacologically restore HDC function, there are no evidence‐based guidelines concerning the optimal tPA dose. The purpose of this study was to compare the efficacy of 1.0‐mg vs. 2.0‐mg tPA dwell protocols in restoring the HDC function in thrombotic dysfunctional catheters. A retrospective, single‐center study was conducted on two independent cohorts of patients; the first (n = 129) received 2.0 mg tPA/catheter lumen, while the second (n = 108) received 1.0 mg tPA/catheter lumen. Kaplan–Meier and Cox regression analyses were performed to compare the catheter survival time between patients who received 1.0 mg tPA and those who received 2.0 mg tPA. Catheter removal occurred in 25 (19.4%) of those catheters treated with 1.0 mg tPA compared with 11 (10.2%) of catheters treated with 2.0 mg tPA (P = 0.05). The hazard ratio (HR) for catheter removal was 2.75 (95% confidence interval [^95%CI] = 1.25–6.04) for the 1.0‐mg tPA cohort compared with the 2.0‐mg tPA cohort. Correction added on 3 December 2012, after first online publication: The tPA cohort values were changed. Female gender (HR = 2.51; ^95%CI = 1.20–5.27) and age (HR = 0.96; ^95%CI = 0.94–0.98) were also associated with catheter survival. Our findings suggest that treatment of dysfunctional HDC with 2.0‐mg tPA dwells is superior to 1.0‐mg tPA dwells.