Continuous speech recognition based on ICA and geometrical learning

We investigate the use of independent component analysis (ICA) for speech feature extraction in digits speech recognition systems. We observe that this may be true for recognition tasks based on Geometrical Learning with little training data. In contrast to image processing, phase information is not essential for digits speech recognition. We therefore propose a new scheme that shows how the phase sensitivity can be removed by using an analytical description of the ICA-adapted basis functions. Furthermore, since the basis functions are not shift invariant, we extend the method to include a frequency-based ICA stage that removes redundant time shift information. The digits speech recognition results show promising accuracy. Experiments show that the method based on ICA and Geometrical Learning outperforms HMM in a different number of training samples.

Continuous speech research based on HyperSausage Neuron

In this paper, we presents HyperSausage Neuron based on the High-Dimension Space(HDS), and proposes a new algorithm for speaker independent continuous digit speech recognition. At last, compared to HMM-based method, the recognition rate of HyperSausage Neuron method is higher than that of in HMM-based method.

crossmodal error correction of continuous handwriting recognition by speech

In recognition-based user interface, users’ satisfaction is determined not only by recognition accuracy but also by effort to correct recognition errors. In this paper, we introduce a crossmodal error correction technique, which allows users to correct errors of Chinese handwriting recognition by speech. The focus of the paper is a multimodal fusion algorithm supporting the crossmodal error correction. By fusing handwriting and speech recognition, the algorithm can correct errors in both character extraction and recognition of handwriting. The experimental result indicates that the algorithm is effective and efficient. Moreover, the evaluation also shows the correction technique can help users to correct errors in handwriting recognition more efficiently than the other two error correction techniques.

REDUCTION OF CRYSTALLINE DISORDER IN MOLECULAR-BEAM EPITAXY GAAS ON SI BY MEV ION-IMPLANTATION AND SUBSEQUENT ANNEALING

Molecular beam epitaxy GaAs films on Si, with thicknesses ranging from 0.9-2.0-mu-m, were implanted with Si ions at 1.2-2.6 MeV to doses in the range 10(15)-10(16) cm-2. Subsequent rapid infrared thermal annealing was carried out at 850-degrees-C for 15 s in a flowing N2 atmosphere. Crystalline quality was analyzed by using Rutherfold backscattering/channeling technique and Raman scattering spectrometry. The experimental results show that the recrystallization process greatly depends on the dose and energy of implanted ions. Complete recrystallization with better crystalline quality can be obtained under proper implantation and subsequent annealing. In the improved layer the defect density was much lower than in the as-grown layer, especially near the interface.

INFLUENCES OF ALLOY DISORDER AND INTERFACE ROUGHNESS ON OPTICAL-SPECTRA OF INGAAS/GAAS STRAINED-LAYER QUANTUM-WELLS

We present studies of alloy composition and layer thickness dependences of excitonic linewidths in InGaAs/GaAs strained-layer quantum wells grown by MBE, using both photoluminescence and optical absorption. It is observed that linewidths of exciton spectra increase with indium content and well size. Using the virtual crystal approximation, the experimental data are analyzed. The results obtained show that the alloy disorder is the dominant mechanism for line broadening at low temperature. In addition, it is found that the absorption spectra related to light hole transitions have varied from a peak to a step-like structure as temperature increases. This behavior can be understood by the indirect space transitions of light holes.

DISORDER POINT FOR 2D ELECTRONS IN A HALF-FILLED LANDAU-LEVEL

Short-range correlations of two-dimensional electrons in a strong magnetic field are shown to be triangular in nature well below half-filling, but honeycomb well above half-filling. The half-filling point is thus proposed, and qualitatively confirmed by three-body correlation calculations, to be a new type of disorder point where short-range correlations change character. A wavefunction study also suggests that nodes become unbound at half-filling. Evidence for incompressibility but deformability of the half-filling state earlier suggested by Fano, Ortolani and Tosatti, is also presented and found to be in agreement with recent experiments.