A robust classification procedure based on mixture classifiers and nonparametric weighted feature extraction

There are many factors to consider in carrying out a hyperspectral data classification. Perhaps chief among them are class training sample size, dimensionality, and distribution separability. The intent of this study is to design a classification procedure that is robust and maximally effective, but which provides the analyst with significant assists, thus simplifying the analyst's task. The result is a quadratic mixture classifier based on Mixed-LOOC2 regularized discriminant analysis and nonparametric weighted feature extraction. This procedure has the advantage of providing improved classification accuracy compared to typical previous methods but requires minimal need to consider the factors mentioned above. Experimental results demonstrating these properties are presented.     

Characterization of residual information for SeaWinds quality control

Recent work has shown the important properties of the wind inversion residual or maximum-likelihood estimator (MLE) for quality Control (QC) of QuikSCAT Hierarchical Data Format (HDF) observations. Since March 2000, the QuikSCAT near-real-time (NRT) Binary Universal Format Representation (BUFR) product is available. As this product is used for numerical weather prediction (NWP) assimilation purposes, a QC procedure for the BUFR product is needed. We study the behavior of the MLE in order to determine whether the HDF QC procedure is appropriate for BUFR data. A comparison using real HDF and BUFR data reveals that the MLE distributions of HDF and BUFR differ and are actually poorly correlated. One important difference between BUFR and HDF is the amount of signal averaging prior to wind inversion. The averaging reduces the number of observations used in the wind retrieval for the BUFR product as compared to HDF. We show with a simple example that different MLE distributions are indeed expected due to this averaging. We also run a simulation in order to link theory and reality and better understand the behavior of the MLE. Despite the different MLE behavior in BUFR and HDF, the quality of the retrieved winds, as compared with the European Centre for Medium-Range Weather Forecasts winds, is very similar. We develop an MLE-based QC procedure for BUFR, similarly to the one in HDF, and we compare both. The skill of the QC in BUFR is again very similar to the one in HDF, showing that despite the different MLE behavior in both formats, the properties of the MLE as a QC indicator remain very similar.     

Capacity-achieving sequences for the erasure channel

This paper starts a systematic study of capacity-achieving (c.a.) sequences of low-density parity-check codes for the erasure channel. We introduce a class A of analytic functions and develop a procedure to obtain degree distributions for the codes. We show various properties of this class which help us to construct new distributions from old ones. We then study certain types of capacity-achieving sequences and introduce new measures for their optimality. For instance, it turns out that the right-regular sequence is c.a. in a much stronger sense than, e.g., the Tornado sequence. This also explains why numerical optimization techniques tend to favor graphs with only one degree of check nodes.     

Electromigration-induced damage in bamboo Al interconnects

Depletion and hillock formation were examined in-situ in a scanning electron microscope (SEM) during electromigration of bamboo Al interconnect segments. Hillocks formed directly at the anode ends of the segments by epitaxial addition of Al at the bottom Al/TiN interface. Depletion occurred nonuniformly from the cathode end and stopped once the distance between the leading void and the hillock reached the critical length for electromigration at the given current density. A modified equation for the drift velocity is proposed, which includes the effect of nonuniform depletion and predicts that interconnects with nonuniform depletion are more reliable than those with uniform depletion.     

Logarithmic temperature dependence of the conductivity of the two-dimensional metal

We report on the first observation and studies of a weak delocalizing logarithmic temperature dependence of the conductivity, which causes the conductivity of the 2D metal to increase as T decreases down to 16 mK. The prefactor of the logarithmic dependence is found to decrease gradually with density, to vanish at a critical density n _c , 2∼2×10¹² cm⁻², and then to have the opposite sign at n>n _c ,2. The second critical density sets the upper limit on the existence region of the 2D metal, whereas the conductivity at the critical point, G _c ,2∼120e ²/h, sets an upper (low-temperature) limit on its conductivity. Pis’ma Zh. éksp. Teor. Fiz. 68, No. 6, 497–501 (25 September 1998) Published in English in the original Russian journal. Edited by Steve Torstveit.     

Positive temperature dependence of the electron impact ionizationcoefficient in In0.53Ga0.47As/InP HBTs

Impact ionization is a major limiting factor to the maximum operating voltage of InGaAs-based, high-speed transistors. In this work, data on the positive temperature dependence of the electron impact ionization coefficient α_n in In_0.53Ga_0.47As at medium-low electric fields are reported for the first time. The increase of α_n with temperature is opposite to the behavior normally observed in most semiconductors. This anomalous behavior implies the onset of a positive feedback between power dissipation and avalanche generation which may adversely affect the power handling capability of In_0.53Ga _0.47As-based devices, and which should be taken into account in device thermal modeling. In the experimental procedure, based on the measurement of the multiplication factor M-1 in npn In_0.53Ga _0.47As/InP Heterojunction Bipolar Transistors (HBT), particular care has been taken in order to rule out possible spurious, temperature-dependent contributions to the measured multiplication current     

Quadratic interpolation for image resampling

Nearest-neighbor, linear, and various cubic interpolation functions are frequently used in image resampling. Quadratic functions have been disregarded largely because they have been thought to introduce phase distortions. This is shown not to be the case, and a family of quadratic functions is derived. The interpolating member of this family has visual quality close to that of the Catmull-Rom cubic, yet requires only 60% of the computation time.     

Error-resilience improvement for block-transform video coders

An algorithm is presented to enhance the resilience of block-based video-coding algorithms against channel errors. The error-resilience algorithm described suggests the use of fixed length coding to alleviate the effect of Huffman coding on the synchronisation of the decode under erroneous conditions. The proposed mechanism modifies the order of transmission of the fixed-length video parameters to increase the chance of their arrival. Synchronisation words are sent at fixed-length intervals within the bit stream to reduce the possibility of the decoder falling on similar bit patterns within the bit stream. To limit the effect of error accumulation, the motion prediction process is halted and the differential coding of motion vectors is not applied. FEC techniques are applied on some error-sensitive segments of the reordered video bitstream. The effectiveness of the proposed error-resilience algorithm is evidenced by both subjective and objective results     

Four-channel polarization-insensitive optically transparent wavelength converter

Multichannel wavelength converters may be important components in the cross-connects in future wavelength-division multiplexed (WDM) transport networks. We demonstrate a multichannel, polarization-insensitive, optically transparent wavelength converter, based on four-wave mixing in two semiconductor optical amplifiers in a polarization-diversity arrangement. Bit-error-rate (BER) measurements with four input 2.5-Gb/s WDM channels, spaced by 2 nm, show penalties for wavelength conversion less than 2.6 dB at 10/sup -9/ BER. Changes in the state of polarization of the input signals cause the output power to change by less than 1.2 dB, and the corresponding power penalties change by less than 0.9 dB.     

Wavelet packet modulation for orthogonally multiplexedcommunication

Utilizing multidimensional signaling techniques, a generalized multirate wavelet-based modulation format for orthogonally multiplexed communication systems is presented. Wavelet packet modulation (WPM) employs the basis functions from an arbitrary pruning of a dyadic tree structured filter bank as orthogonal pulse shapes for conventional quadrature amplitude modulation (QAM) symbols. This generalized framework affords an entire library of basis sets with increased flexibility in time-frequency (T-F) partitioning. The bandwidth efficiency and power spectral density figures of merit for the general signal are derived and shown to be that of standard QAM