Single-channel, speaker-independent speech separation methods have recently seen great progress. However, the accuracy, latency, and computational cost of such methods remain insufficient. The ...majority of the previous methods have formulated the separation problem through the time-frequency representation of the mixed signal, which has several drawbacks, including the decoupling of the phase and magnitude of the signal, the suboptimality of time-frequency representation for speech separation, and the long latency in calculating the spectrograms. To address these shortcomings, we propose a fully convolutional time-domain audio separation network (Conv-TasNet), a deep learning framework for end-to-end time-domain speech separation. Conv-TasNet uses a linear encoder to generate a representation of the speech waveform optimized for separating individual speakers. Speaker separation is achieved by applying a set of weighting functions (masks) to the encoder output. The modified encoder representations are then inverted back to the waveforms using a linear decoder. The masks are found using a temporal convolutional network consisting of stacked one-dimensional dilated convolutional blocks, which allows the network to model the long-term dependencies of the speech signal while maintaining a small model size. The proposed Conv-TasNet system significantly outperforms previous time-frequency masking methods in separating two- and three-speaker mixtures. Additionally, Conv-TasNet surpasses several ideal time-frequency magnitude masks in two-speaker speech separation as evaluated by both objective distortion measures and subjective quality assessment by human listeners. Finally, Conv-TasNet has a significantly smaller model size and a shorter minimum latency, making it a suitable solution for both offline and real-time speech separation applications. This study, therefore, represents a major step toward the realization of speech separation systems for real-world speech processing technologies.
Protecting the health and safety of health care workers is vital to the health of each of us. Preparing for and responding to a future influenza pandemic or to a sustained outbreak of an airborne ...transmissible disease requires a high-level commitment to respiratory protection for health care workers across the wide range of settings in which they work and the jobs that they perform. Keeping health care workers healthy is an ethical commitment both in terms of addressing the occupational risks faced by health care workers and of providing for the continuity of patient care and services needed to maintain the health of individuals and communities.
During a public health emergency, challenges will arise concerning the availability of respiratory protective devices (i.e., respirators). Reusable respirators (specifically, reusable half-facepiece elastomeric respirators) are the standard respiratory protection device used in many industries, and they provide an option for use in health care that has to date not been fully explored. The durability and reusability of elastomeric respirators make them desirable for stockpiling for emergencies, where the need for large volumes of respirators can be anticipated. However, they are used infrequently in health care.
Reusable Elastomeric Respirators in Health Care explores the potential for the use of elastomeric respirators in the U.S. health care system with a focus on the economic, policy, and implementation challenges and opportunities. This report examines the practicability of elastomeric use in health care on a routine basis and during an influenza pandemic or other large aerosol-transmissible outbreak, when demand for respiratory protective devices by U.S. health care personnel may be larger than domestic supplies. The report also addresses the issues regarding emergency stockpile management of elastomeric respiratory protective devices.
Purpose
The efficacy of noninvasive continuous positive airway pressure (CPAP) to improve outcomes in severe hypoxemic acute respiratory failure (hARF) due to pneumonia has not been clearly ...established. The aim of this study was to compare CPAP vs. oxygen therapy to reduce the risk of meeting criteria for endotracheal intubation (ETI).
Methods
In a multicenter randomized controlled trial conducted in four Italian centers patients with severe hARF due to pneumonia were randomized to receive helmet CPAP (CPAP group) or oxygen delivered with a Venturi mask (control group). The primary endpoint was the percentage of patients meeting criteria for ETI, including either one or more major criteria (respiratory arrest, respiratory pauses with unconsciousness, severe hemodynamic instability, intolerance) or at least two minor criteria (reduction of at least 30 % of basal PaO
2
/FiO
2
ratio, increase of 20 % of PaCO
2
, worsening of alertness, respiratory distress, SpO
2
less than 90 %, exhaustion).
Results
Between February 2010 and 2013, 40 patients were randomized to CPAP and 41 to Venturi mask. The proportion of patients meeting ETI criteria in the CPAP group was significantly lower compared to those in the control group (6/40 = 15 % vs. 26/41 = 63 %, respectively,
p
< 0.001; relative risk 0.24, 95 % CI 0.11–0.51; number needed to treat, 2) two patients were intubated in the CPAP group and one in the control group. The CPAP group showed a faster and greater improvement in oxygenation in comparison to controls (
p
< 0.001). In either study group, no relevant adverse events were detected.
Conclusions
Helmet CPAP reduces the risk of meeting ETI criteria compared to oxygen therapy in patients with severe hARF due to pneumonia.
Mie resonance wavelengths of a dielectric structure are strongly dependent on the inherent material property and structural geometry. In particular, a high-index nanostructure enables light ...confinement within itself over the range of visible wavelengths, which allows the Mie resonator to be applied to a pixel in color printing of subwavelength-scale resolution. However, if the Mie resonator is packed into a smaller area in order to achieve better resolution, the interaction between adjacent resonators occurs depending on these spatial distances, leading to unexpected color changes. Here, we demonstrate metal-masked Mie-resonant color printing for suppressing undesirable color changes. We observed that the interaction between monocrystalline Si resonators can be suppressed by the addition of a Cr mask. The pixels with this functionality can produce individual colors even if operating as a single element or in other periodic arrays, resulting in the realization of higher resolution encoding. The coincidence of resonance peak positions derived from excited electric/magnetic dipoles enables the demonstration of brilliant full-color printing with higher color purity. Furthermore, a vivid printing image with a resolution of more than 100 000 dpi was achieved using the designed subwavelength pixels. This study can contribute not only to the improvement of the resolution of color printing but also to the suppression of unwanted interactions of Mie resonance in optical devices.
PDF
