Repulsive fermions in optical lattices: Phase separation versus coexistence of antiferromagnetism and d-wave superfluidity

We investigate a system of fermions on a two-dimensional optical square lattice in the strongly repulsive coupling regime. In this case, the interactions can be controlled by laser intensity as well as by Feshbach resonance. We compare the energetics of states with resonating valence bond d-wave superfluidity, antiferromagnetic long-range order, and a homogeneous state with coexistence of superfluidity and antiferromagnetism. Using a variational formalism, we show that the energy density of a hole e(hole)(x) has a minimum at doping x = x(c) that signals phase separation between the antiferromagnetic and d-wave paired superfluid phases. The energy of the phase-separated ground state is, however, found to be very close to that of a homogeneous state with coexisting antiferromagnetic and superfluid orders. We explore the dependence of the energy on the interaction strength and on the three-site hopping terms and compare with the nearest-neighbor hopping t-J model.

Prototype learning methods for online handwriting recognition

In this paper, we study different methods for prototype selection for recognizing handwritten characters of Tamil script. In the first method, cumulative pairwise- distances of the training samples of a given class are used to select prototypes. In the second method, cumulative distance to allographs of different orientation is used as a criterion to decide if the sample is representative of the group. The latter method is presumed to offset the possible orientation effect. This method still uses fixed number of prototypes for each of the classes. Finally, a prototype set growing algorithm is proposed, with a view to better model the differences in complexity of different character classes. The proposed algorithms are tested and compared for both writer independent and writer adaptation scenarios.

Language identification using parallel sub-word recognition

Parallel sub-word recognition (PSWR) is a new model that has been proposed for language identification (LID) which does not need elaborate phonetic labeling of the speech data in a foreign language. The new approach performs a front-end tokenization in terms of sub-word units which are designed by automatic segmentation, segment clustering and segment HMM modeling. We develop PSWR based LID in a framework similar to the parallel phone recognition (PPR) approach in the literature. This includes a front-end tokenizer and a back-end language model, for each language to be identified. Considering various combinations of the statistical evaluation scores, it is found that PSWR can perform as well as PPR, even with broad acoustic sub-word tokenization, thus making it an efficient alternative to the PPR system.

Experiments on three-dimensional wire-frame object recognition

Some experimental results on the recognition of three-dimensional wire-frame objects are presented. In order to overcome the limitations of a recent model, which employs radial basis functions-based neural networks, we have proposed a hybrid learning system for object recognition, featuring: an optimization strategy (simulated annealing) in order to avoid local minima of an energy functional; and an appropriate choice of centers of the units. Further, in an attempt to achieve improved generalization ability, and to reduce the time for training, we invoke the principle of self-organization which utilises an unsupervised learning algorithm.

A novel interrogation system for fiber Bragg grating sensor array

Interrogation techniques for fiber Bragg grating sensor arrays need particular attention in the case of structural health monitoring applications involving dynamic strain measurement. Typically the performance of the sensing system is dependent on both the sensor type and the interrogation method employed. A novel interrogation system is proposed here that consists of different interrogation units for each sensor in the array, each unit comprising of a circulator, chirped grating and a Mach-Zehnder interferometer. We present an analysis that consists of tracking the spectral changes as the light passes through various elements in the interrogation system. This is expected to help in the optimization of sensor and interrogation elements leading to improved performance of the health monitoring system.

High resolution multiple excitation spot optical microscopy

We propose fundamental improvements in three-dimensional (3D) resolution of multiple excitation spot optical microscopy. The excitation point spread function (PSF) is generated by two interfering counter-propagating depth-of-focus beams along the optical axis. Detection PSF is obtained by coherently interfering the emitted fluorescent light (collected by both the objectives) at the detector. System PSF shows upto 14-fold reduction in focal volume as compared to confocal, and almost 2-fold improvement in lateral resolution. Proposed PSF has the ability to simultaneously excite multiple 3D-spots of sub-femtoliter volume. Potential applications are in fluorescence microscopy and nanobioimaging. Copyright 2011 Author(s). This article is distributed under a Creative Commons Attribution 3.0 Unported License. [doi:10.1063/1.3598413]

Human action recognition using a fast learning fully complex-valued classifier

In this paper, we use optical flow based complex-valued features extracted from video sequences to recognize human actions. The optical flow features between two image planes can be appropriately represented in the Complex plane. Therefore, we argue that motion information that is used to model the human actions should be represented as complex-valued features and propose a fast learning fully complex-valued neural classifier to solve the action recognition task. The classifier, termed as, ``fast learning fully complex-valued neural (FLFCN) classifier'' is a single hidden layer fully complex-valued neural network. The neurons in the hidden layer employ the fully complex-valued activation function of the type of a hyperbolic secant function. The parameters of the hidden layer are chosen randomly and the output weights are estimated as the minimum norm least square solution to a set of linear equations. The results indicate the superior performance of FLFCN classifier in recognizing the actions compared to real-valued support vector machines and other existing results in the literature. Complex valued representation of 2D motion and orthogonal decision boundaries boost the classification performance of FLFCN classifier. (c) 2012 Elsevier B.V. All rights reserved.

Practical fully three-dimensional reconstruction algorithms for diffuse optical tomography

We have developed an efficient fully three-dimensional (3D) reconstruction algorithm for diffuse optical tomography (DOT). The 3D DOT, a severely ill-posed problem, is tackled through a pseudodynamic (PD) approach wherein an ordinary differential equation representing the evolution of the solution on pseudotime is integrated that bypasses an explicit inversion of the associated, ill-conditioned system matrix. One of the most computationally expensive parts of the iterative DOT algorithm, the reevaluation of the Jacobian in each of the iterations, is avoided by using the adjoint-Broyden update formula to provide low rank updates to the Jacobian. In addition, wherever feasible, we have also made the algorithm efficient by integrating along the quadratic path provided by the perturbation equation containing the Hessian. These algorithms are then proven by reconstruction, using simulated and experimental data and verifying the PD results with those from the popular Gauss-Newton scheme. The major findings of this work are as follows: (i) the PD reconstructions are comparatively artifact free, providing superior absorption coefficient maps in terms of quantitative accuracy and contrast recovery; (ii) the scaling of computation time with the dimension of the measurement set is much less steep with the Jacobian update formula in place than without it; and (iii) an increase in the data dimension, even though it renders the reconstruction problem less ill conditioned and thus provides relatively artifact-free reconstructions, does not necessarily provide better contrast property recovery. For the latter, one should also take care to uniformly distribute the measurement points, avoiding regions close to the source so that the relative strength of the derivatives for measurements away from the source does not become insignificant. (c) 2012 Optical Society of America

Linearization and Reconstruction of Non-linear Diffuse Optical Tomographic Image

Diffuse optical tomography (DOT) is one of the ways to probe highly scattering media such as tissue using low-energy near infra-red light (NIR) to reconstruct a map of the optical property distribution. The interaction of the photons in biological tissue is a non-linear process and the phton transport through the tissue is modelled using diffusion theory. The inversion problem is often solved through iterative methods based on nonlinear optimization for the minimization of a data-model misfit function. The solution of the non-linear problem can be improved by modeling and optimizing the cost functional. The cost functional is f(x) = x(T)Ax - b(T)x + c and after minimization, the cost functional reduces to Ax = b. The spatial distribution of optical parameter can be obtained by solving the above equation iteratively for x. As the problem is non-linear, ill-posed and ill-conditioned, there will be an error or correction term for x at each iteration. A linearization strategy is proposed for the solution of the nonlinear ill-posed inverse problem by linear combination of system matrix and error in solution. By propagating the error (e) information (obtained from previous iteration) to the minimization function f(x), we can rewrite the minimization function as f(x; e) = (x + e)(T) A(x + e) - b(T)(x + e) + c. The revised cost functional is f(x; e) = f(x) + e(T)Ae. The self guided spatial weighted prior (e(T)Ae) error (e, error in estimating x) information along the principal nodes facilitates a well resolved dominant solution over the region of interest. The local minimization reduces the spreading of inclusion and removes the side lobes, thereby improving the contrast, localization and resolution of reconstructed image which has not been possible with conventional linear and regularization algorithm.

Nanocrystallization of TiO2 Anatase Phase in Hydrophobic BaO-TiO2-B2O3 Glass System

Transparent colorless glasses in the ternary BaOTiO2B2O3 system were fabricated via conventional melt-quenching technique. The glasses with certain molar concentrations of BaO and TiO2 on heat treatment at appropriate temperatures yielded nanocrystalline phase of TiO2 associated with the crystallite size in the 515 nm range. Nanocrystallized glasses exhibited high refractive index (n = 2.15) measured at lambda = 543 nm. These glasses were found to be hydrophobic in nature associated with the contact angle of 90 degrees. These high-index glass nanocrystal composites would be of potential interest for optical device applications.

Polymer microfabrication by scanning based microstereolithography: Optical design and material functionality

Several research groups have attempted to optimize photopolymerization parameters to increase the throughput of scanning based microstereolithography (MSL) systems through modified beam scanning techniques. Efforts in reducing the curing line width have been implemented through high numerical aperture (NA) optical setups. However, the intensity contour symmetry and the depth of field of focus have led to grossly non-vertical and non-uniform curing profiles. This work tries to review the photopolymerization process in a scanning based MSL system from the aspect of material functionality and optical design. The focus has been to exploit the rich potential of photoreactor scanning system in achieving desired fabrication modalities (minimum curing width, uniform depth profile, and vertical curing profile) even with a reduced NA optical setup and a single movable stage. The present study tries to manipulate to its advantage the effect of optimized lower c] (photoinitiator (PI) concentration) in reducing the minimum curing width to similar to 10-20 mu m even with a higher spot size (similar to 21.36 mu m) through a judiciously chosen ``monomer-PI'' system. Optimization on grounds of increasing E-max (maximum laser exposure energy at surface) by optimizing the scan rate provides enough time for the monomer or resin to get cured across the entire resist thickness (surface to substrate similar to 10-100 mu m), leading to uniform depth profiles along the entire scan lengths. (C) 2012 American Institute of Physics. http://dx.doi.org/10.1063/1.4750975]

Fast Likelihood Computation in Speech Recognition using Matrices

Acoustic modeling using mixtures of multivariate Gaussians is the prevalent approach for many speech processing problems. Computing likelihoods against a large set of Gaussians is required as a part of many speech processing systems and it is the computationally dominant phase for Large Vocabulary Continuous Speech Recognition (LVCSR) systems. We express the likelihood computation as a multiplication of matrices representing augmented feature vectors and Gaussian parameters. The computational gain of this approach over traditional methods is by exploiting the structure of these matrices and efficient implementation of their multiplication. In particular, we explore direct low-rank approximation of the Gaussian parameter matrix and indirect derivation of low-rank factors of the Gaussian parameter matrix by optimum approximation of the likelihood matrix. We show that both the methods lead to similar speedups but the latter leads to far lesser impact on the recognition accuracy. Experiments on 1,138 work vocabulary RM1 task and 6,224 word vocabulary TIMIT task using Sphinx 3.7 system show that, for a typical case the matrix multiplication based approach leads to overall speedup of 46 % on RM1 task and 115 % for TIMIT task. Our low-rank approximation methods provide a way for trading off recognition accuracy for a further increase in computational performance extending overall speedups up to 61 % for RM1 and 119 % for TIMIT for an increase of word error rate (WER) from 3.2 to 3.5 % for RM1 and for no increase in WER for TIMIT. We also express pairwise Euclidean distance computation phase in Dynamic Time Warping (DTW) in terms of matrix multiplication leading to saving of approximately of computational operations. In our experiments using efficient implementation of matrix multiplication, this leads to a speedup of 5.6 in computing the pairwise Euclidean distances and overall speedup up to 3.25 for DTW.

Optical properties change with the addition and diffusion of Bi to As2S3 in the Bi/As2S3 bilayer thin film

The role of Bi layer (thickness similar to 7 nm) on As2S3 film was extensively studied for different optical applications in which Bi (top layer) as active and diffusing layer and As2S3 as barrier (matrix) layer. Bilayer thin films of Bi/As2S3 were prepared from Bi and As2S3 by thermal evaporation technique under high vacuum. The decrease of optical band gap with the addition of Bi to As2S3 has been explained on the basis of density of states and the increase in disorder in the system. It was found that the efficient changes of optical parameters (transmission, optical band gap, refraction) could be realized due to the photo induced diffusion activated by the focused 532 nm laser irradiation and formation of different bonds. The diffusion of Bi into As2S3 matrix increases the optical band gap producing photo bleaching effect. The changes were characterised by different experimental techniques. (C) 2012 Elsevier B.V. All rights reserved.

Flux density calibration in diffuse optical tomographic systems

The solution of the forward equation that models the transport of light through a highly scattering tissue material in diffuse optical tomography (DOT) using the finite element method gives flux density (Phi) at the nodal points of the mesh. The experimentally measured flux (U-measured) on the boundary over a finite surface area in a DOT system has to be corrected to account for the system transfer functions (R) of various building blocks of the measurement system. We present two methods to compensate for the perturbations caused by R and estimate true flux density (Phi) from U-measured(cal). In the first approach, the measurement data with a homogeneous phantom (U-measured(homo)) is used to calibrate the measurement system. The second scheme estimates the homogeneous phantom measurement using only the measurement from a heterogeneous phantom, thereby eliminating the necessity of a homogeneous phantom. This is done by statistically averaging the data (U-measured(hetero)) and redistributing it to the corresponding detector positions. The experiments carried out on tissue mimicking phantom with single and multiple inhomogeneities, human hand, and a pork tissue phantom demonstrate the robustness of the approach. (C) 2013 Society of Photo-Optical Instrumentation Engineers (SPIE) DOI: 10.1117/1.JBO.18.2.026023]

Fully complex-valued ELM classifiers for human action recognition

In this paper, we present a fast learning neural network classifier for human action recognition. The proposed classifier is a fully complex-valued neural network with a single hidden layer. The neurons in the hidden layer employ the fully complex-valued hyperbolic secant as an activation function. The parameters of the hidden layer are chosen randomly and the output weights are estimated analytically as a minimum norm least square solution to a set of linear equations. The fast leaning fully complex-valued neural classifier is used for recognizing human actions accurately. Optical flow-based features extracted from the video sequences are utilized to recognize 10 different human actions. The feature vectors are computationally simple first order statistics of the optical flow vectors, obtained from coarse to fine rectangular patches centered around the object. The results indicate the superior performance of the complex-valued neural classifier for action recognition. The superior performance of the complex neural network for action recognition stems from the fact that motion, by nature, consists of two components, one along each of the axes.