Improved recognition of aged Kannada documents by effective segmentation of merged characters

In optical character recognition of very old books, the recognition accuracy drops mainly due to the merging or breaking of characters. In this paper, we propose the first algorithm to segment merged Kannada characters by using a hypothesis to select the positions to be cut. This method searches for the best possible positions to segment, by taking into account the support vector machine classifier's recognition score and the validity of the aspect ratio (width to height ratio) of the segments between every pair of cut positions. The hypothesis to select the cut position is based on the fact that a concave surface exists above and below the touching portion. These concave surfaces are noted down by tracing the valleys in the top contour of the image and similarly doing it for the image rotated upside-down. The cut positions are then derived as closely matching valleys of the original and the rotated images. Our proposed segmentation algorithm works well for different font styles, shapes and sizes better than the existing vertical projection profile based segmentation. The proposed algorithm has been tested on 1125 different word images, each containing multiple merged characters, from an old Kannada book and 89.6% correct segmentation is achieved and the character recognition accuracy of merged words is 91.2%. A few points of merge are still missed due to the absence of a matched valley due to the specific shapes of the particular characters meeting at the merges.

Behavior of Brick-Mortar Interfaces in FRP-Strengthened Masonry Assemblages under Normal Loading and Shear Loading

Determination of shear strength of brick-mortar bed joint is critical to overcome the sliding-shear or joint-shear failure in masonry. In the recent past, researchers have attempted to enhance the shear strength and deformation capacity of brick-mortar bed joints by gluing fiber-reinforced polymer (FRP) composite across the bed joint. FRP composites offer several advantages like high strength-to-weight ratio, and ease of application in terms of labor, time, and reduced curing period. Furthermore, FRP composites are desirable for strengthening old masonry buildings having heritage value because of its minimal interference with the existing architecture. A majority of earlier studies on shear strengthening of masonry available in the literature adopted masonry having the ratio of modulus of elasticity of masonry unit (Emu) to modulus of elasticity of mortar (Em) greater than one. Information related to shear behavior of FRP glued masonry composed of masonry units having Young's modulus lower than mortar is limited. Hence the present study is focused on characterizing the interfacial behavior of brick-mortar bed joint of masonry assemblages composed of solid burnt clay bricks and cement-sand mortar (E-mu/E-m ratio less than one), strengthened with FRP composites. Masonry triplets and prisms with bed joint inclined to loading axis (0 degrees, 30 degrees, 45 degrees, 60 degrees and 90 degrees) are employed in this study. Glass and carbon FRP composites composed of bidirectional FRP fabric with equal density in both directions are used for strengthening masonry. Masonry triplets are glued with glass and carbon FRP composites in two configurations: (1) both faces of the triplet specimens are fully glued with GFRP composites; and (2) both faces of the triplet specimens are glued with GFRP and CFRP composites in strip form. The performance of masonry assemblages strengthened with FRP composites is assessed in terms of gain in shear strength, shear displacement, and postpeak behavior for various configurations and types of FRP composites considered. A semianalytical model is proposed for the prediction of shear strength of masonry bed joints glued with FRP composites. A composite failure envelope consisting of a Coulomb friction model and a compression cap is obtained for unreinforced masonry and GFRP-strengthened masonry based on the test results of masonry triplets and masonry prisms with bed joints having various inclinations to the loading (C) 2015 American Society of Civil Engineers.

SPEECH ENHANCEMENT USING DISCRIMINATIVE RANDOM FIELDS

Speech enhancement in stationary noise is addressed using the ideal channel selection framework. In order to estimate the binary mask, we propose to classify each time-frequency (T-F) bin of the noisy signal as speech or noise using Discriminative Random Fields (DRF). The DRF function contains two terms - an enhancement function and a smoothing term. On each T-F bin, we propose to use an enhancement function based on likelihood ratio test for speech presence, while Ising model is used as smoothing function for spectro-temporal continuity in the estimated binary mask. The effect of the smoothing function over successive iterations is found to reduce musical noise as opposed to using only enhancement function. The binary mask is inferred from the noisy signal using Iterated Conditional Modes (ICM) algorithm. Sentences from NOIZEUS corpus are evaluated from 0 dB to 15 dB Signal to Noise Ratio (SNR) in 4 kinds of additive noise settings: additive white Gaussian noise, car noise, street noise and pink noise. The reconstructed speech using the proposed technique is evaluated in terms of average segmental SNR, Perceptual Evaluation of Speech Quality (PESQ) and Mean opinion Score (MOS).

l(1)-K-SVD: A robust dictionary learning algorithm with simultaneous update

We develop a new dictionary learning algorithm called the l(1)-K-svp, by minimizing the l(1) distortion on the data term. The proposed formulation corresponds to maximum a posteriori estimation assuming a Laplacian prior on the coefficient matrix and additive noise, and is, in general, robust to non-Gaussian noise. The l(1) distortion is minimized by employing the iteratively reweighted least-squares algorithm. The dictionary atoms and the corresponding sparse coefficients are simultaneously estimated in the dictionary update step. Experimental results show that l(1)-K-SVD results in noise-robustness, faster convergence, and higher atom recovery rate than the method of optimal directions, K-SVD, and the robust dictionary learning algorithm (RDL), in Gaussian as well as non-Gaussian noise. For a fixed value of sparsity, number of dictionary atoms, and data dimension, l(1)-K-SVD outperforms K-SVD and RDL on small training sets. We also consider the generalized l(p), 0 < p < 1, data metric to tackle heavy-tailed/impulsive noise. In an image denoising application, l(1)-K-SVD was found to result in higher peak signal-to-noise ratio (PSNR) over K-SVD for Laplacian noise. The structural similarity index increases by 0.1 for low input PSNR, which is significant and demonstrates the efficacy of the proposed method. (C) 2015 Elsevier B.V. All rights reserved.

Gasification of wood particles in a co-current packed bed: Experiments and model analysis

This study focuses on addressing the propagation front movement in a co-current downdraft gasification system. A detailed single particle modeling analysis extended to the packed bed reactor is used to compare with the experimental measurement as well those available in the literature. This model for biomass gasification systems considered pyrolysis process, gas phase volatile combustion, and heterogeneous char reactions along with gas phase reactions in the packed bed. The pyrolysis kinetics has a critical influence on the gasification process. The propagation front has been shown to increase with air mass flux, attains a peak and then decreases with further increase in air mass flux and finally approaches negative propagation rate. This indicates that front is receding, or no upward movement() bra her it is moving downward towards the char bed. The propagation rate correlates with mass flux as (m) over dot `'(0.883) during the increasing regimes of the front movement The study clearly identifies that bed movement is an important parameter for consideration in a co-current configuration towards establishing the effective bed movement. The study also highlights the importance of surface area to volume ratio of the particles in the packed bed and its influence on the volatile generation. Finally, the gas composition for air gasification under various air mass fluxes is compared with the experimental results. (C) 2016 Elsevier B.V. All rights reserved.

Numerical study of driven flows of shear thinning viscoelastic fluids in rectangular cavities

In this work, we present a numerical study of flow of shear thinning viscoelastic fluids in rectangular lid driven cavities for a wide range of aspect ratios (depth to width ratio) varying from 1/16 to 4. In particular, the effect of elasticity, inertia, model parameters and polymer concentration on flow features in rectangular driven cavity has been studied for two shear thinning viscoelastic fluids, namely, Giesekus and linear PTT. We perform numerical simulations using the symmetric square root representation of the conformation tensor to stabilize the numerical scheme against the high Weissenberg number problem. The variation in flow structures associated with merging and splitting of elongated vortices in shallow cavities and coalescence of corner eddies to yield a second primary vortex in deep cavities with respect to the variation in flow parameters is discussed. We discuss the effect of the dominant eigenvalues and the corresponding eigenvectors on the location of the primary eddy in the cavity. We also demonstrate, by performing numerical simulations for shallow and deep cavities, that where the Deborah number (based on convective time scale) characterizes the elastic behaviour of the fluid in deep cavities, Weissenberg number (based on shear rate) should be used for shallow cavities. (C) 2016 Elsevier B.V. All rights reserved.