Modeling subjectively perceived annoyance of H.264/AVC video as a function of perceived artifact strength

This paper is concerned with the subjective perception of video coding artifacts in H.264/AVC encoded and decoded video. Our objective is to model the perceived annoyance of such low bit rate video sequences as a function of perceived artifact strength. We introduce a new method for determining this function and apply it to the data from two psychophysical experiments. Both experiments produced numerical judgments of the subjectively perceived annoyance of artifact combinations and the perceived strength of the respective coding artifacts. Our method produces a discrete multi-dimensional representation of the relationships in the data from which the function relating annoyance to artifact strength is derived. The method is applied to the data of the first experiment and the resulting function is shown to describe the data from the second experiment as well.     

Pentacene-Based Low-Voltage Strain Sensors With PVP/$ hbox{Ta}_{2}hbox{O}_{5}$ Hybrid Gate Dielectrics

Field-controllable pentacene-semiconductor-based strain sensors were fabricated with hybrid gate dielectrics using polyvinyl phenol (PVP) and high-$k$ inorganic tantalum pentoxide $(hbox{Ta}_{2}hbox{O}_{5})$ onto polyethylene naphthalate films. The $hbox{Ta}_{2}hbox{O}_{5}$ gate-dielectric layer combined with a thin PVP layer to form very smooth and hydrophobic surfaces turns out to improve the molecular structures of pentacene films significantly. The PVP– $hbox{Ta}_{2}hbox{O}_{5}$ hybrid-gate-dielectric films exhibit a high dielectric constant of 19.27 and a leakage-current density of as low as 100 $hbox{nA/cm}^{2}$ . The sensors employing a thin-film-transistor-like Wheatstone bridge configuration able to operate at reduced voltage ($sim$4 V) show good device characteristics with a field-effect mobility of 1.89 $hbox{cm}^{2}/hbox{V} cdot hbox{s}$ and a threshold voltage of $-$0.5 V. The strain sensor characterized with bending at 45$^{circ}$ with respect to the bridge bias direction with different bending radii of 50-, 40-, 30-, 20-, and 8-mm displays output signals improved in linearity in a low range of operating voltages.      

Stable single-frequency fiber ring laser for 25-GHz ITU-T grids utilizing saturable absorber filter

A stable single-frequency fiber ring laser is proposed that operates in a single mode for more than an hour by incorporating unpumped erbium-doped fiber (EDF) as a saturable absorber filter and optimizing the length of EDF used as gain medium. This laser can be continuously tuned to 25-GHz spacing that precisely matches the ITU-T grids by temperature control of etalon filter. This laser had a signal-to-source spontaneous emission ratio higher than 70 dB, and lasing frequencies of 361 channels was matched to ITU-T grids with excellent flatness. Frequency offset from the ITU-T grid was less than 0.14 GHz. The linewidth and the relative intensity noise value was less than 1.3 kHz and 130 dB/Hz (above 250 kHz), respectively.     

EEG-based drowsiness estimation for safety driving using independent component analysis

Preventing accidents caused by drowsiness has become a major focus of active safety driving in recent years. It requires an optimal technique to continuously detect drivers' cognitive state related to abilities in perception, recognition, and vehicle control in (near-) real-time. The major challenges in developing such a system include: 1) the lack of significant index for detecting drowsiness and 2) complicated and pervasive noise interferences in a realistic and dynamic driving environment. In this paper, we develop a drowsiness-estimation system based on electroencephalogram (EEG) by combining independent component analysis (ICA), power-spectrum analysis, correlation evaluations, and linear regression model to estimate a driver's cognitive state when he/she drives a car in a virtual reality (VR)-based dynamic simulator. The driving error is defined as deviations between the center of the vehicle and the center of the cruising lane in the lane-keeping driving task. Experimental results demonstrate the feasibility of quantitatively estimating drowsiness level using ICA-based multistream EEG spectra. The proposed ICA-based method applied to power spectrum of ICA components can successfully (1) remove most of EEG artifacts, (2) suggest an optimal montage to place EEG electrodes, and estimate the driver's drowsiness fluctuation indexed by the driving performance measure. Finally, we present a benchmark study in which the accuracy of ICA-component-based alertness estimates compares favorably to scalp-EEG based.     

Numerical Study of Self-Complementary Antenna Characteristics on Substrate Lenses at Terahertz Frequency

This paper presents a numerical study of self-complementary antennas on substrate lenses made of high-permittivity dielectric material. Bowtie, logarithmically periodic, and logarithmic spiral antennas with the same outer and inner dimensions were selected for study, and their overall performances were compared in the terahertz band at frequencies up to 5.0?THz. The resonance and radiation characteristics of the three antennas were investigated in terms of input impedance, directivity, and radiation efficiency, using a full electromagnetic simulator. This study provides useful guidelines and partially solves the difficult problems of choosing the proper feed and optimizing the lens structure for a THz broadband integrated lens antenna.     

Doping Effects on Thermoelectric Properties in the Mg2Sn System

A weak point of Mg₂X thermoelectrics is the absence of a p-type composition, which motivates research into the Mg₂Sn system. Mg₂Sn thermoelectrics were fabricated by a vacuum melting method and a spark plasma sintering process. As a result, Mg₂Sn single phases were acquired in a wide range of Mg-to-Sn atomic ratios (67:33 to 71:29), showing slightly different thermoelectric characteristics. However, the thermoelectric properties of the undoped system were not sufficient for application in commercial production. To maximize the p-type characteristics, many atoms [Ni (VIIIA), Cu (IB), Ag (IB), Zn (IIB), and In (IIIB)] were doped into the Mg₂Sn phase. Among them, the power factor values increased only in the Ag-doped case. Ag-doping resulted in a power factor that was more than 10 times larger than the value in the undoped case. This result could be important for developing p-type polycrystalline thermoelectrics in the Mg₂X (X?=?Si, Sn) system. However, other atoms [Ni (VIIIA), Cu (IB), Zn (IIB), and In (IIIB)] were not determined to act as acceptor atoms. The maximum ZT value for the Ag-doped Mg₂Sn thermoelectric was more than 0.18, which is comparable to the value for the n-type Mg₂Si system.     

Selective Wavelength Enhanced Photochemical and Photothermal H2 Generation of Classical Oxide Supported Metal Catalyst

Conventional views of constructing simply broadband catalysts for photothermal-enhanced catalysis do not realize that without designating photochemical and photothermal conversion to their optimal working spectra can lead to a performance trade-off. Here, spectrally selective designed photoredox and photothermal heating functions of a classical oxide supported metal catalyst are demonstrated, which exhibits markedly improved hydrogen reactivity. While photothermal hydrogen producing catalysis is previously demonstrated, distinctive wavelength dominant redox and thermal phenomena are not studied due to the complex interdependent behavior they exhibit. The exceptionally high H₂ evolution rate of 30.2 mmol g⁻¹ h⁻¹ (≈74 times that of the control sample) is attributed to the nonoverlapped light absorption and undisrupted charge transfer rationales. This study presents a proof-by-existence that spectrally tailored solar utilization strategy is broadly impactful for the hybrid photothermal–photochemical catalysis. Moreover, the spatially decoupled structural configuration may open up discrete parametric control over photoredox and photoheating functionalities.     

Development of Metallic Hermetic Sealing for MEMS Packaging for Harsh Environment Applications

Hermetic sealing of microelectromechanical system sensors is indispensable to ensure their reliable operation and also to provide protection during fabrication. This work proposes two prospective candidates for hermetic sealing for rugged environment applications, i.e., Al-Ge and Pt-In. Al-Ge was chosen due to its compatibility with complementary metal–oxide–semiconductor technology. Pt-In possesses the highest remelting temperature among all the solder systems, which is desired for high-temperature applications in both the energy and aerospace industries. The various bonding parameters for Al-Ge eutectic bonding and Pt-In transient liquid-phase (TLP) bonding have been optimized, and their influence on the bond quality is reported. Optimization of bonding parameters has been carried out with the objective of ensuring void-free bonds. A new configuration for stacking Al-Ge thin films has been demonstrated to tackle the issue of loss of Ge prior to bonding, since native Ge oxides are soluble in deionized water. The impact of solid-state aging prior to Al-Ge eutectic bonding has been investigated. The method of tailoring the phases in the Pt-In joint is also discussed. The prospects and constraints of eutectic and TLP bonding from the hermeticity perspective are discussed in detail. Furthermore, changes in the microstructure under aging at 300°C up to 500 h and the resulting influence on the mechanical properties are presented. The overall finding of this work is that Al-Ge can achieve better mechanical and hermetic performance for high-temperature applications.     

Fast mode decision algorithm for H.264 inter-prediction

A fast mode decision algorithm for H.264/AVC inter-prediction to reduce computational complexity of the H.264 encoder is presented. Experimental results show that the algorithm can save the entire encoding time by 77% on average while introducing only negligible loss in PSNR value and small increment of bit rate.     

Panorama field rendering with scene depth estimation

A novel approach to image-based modelling and rendering is proposed. A set of randomly placed panoramas and omnidirectional depth information of the surrounding scene are used to generate an output image at a vantage viewpoint. Implementation results show that the proposed approach achieves smooth walkthrough at a realtime