988 resultados para Dynamic texture recognition


Relevância:

30.00% 30.00%

Publicador:

Resumo:

We are addressing the novel problem of jointly evaluating multiple speech patterns for automatic speech recognition and training. We propose solutions based on both the non-parametric dynamic time warping (DTW) algorithm, and the parametric hidden Markov model (HMM). We show that a hybrid approach is quite effective for the application of noisy speech recognition. We extend the concept to HMM training wherein some patterns may be noisy or distorted. Utilizing the concept of ``virtual pattern'' developed for joint evaluation, we propose selective iterative training of HMMs. Evaluating these algorithms for burst/transient noisy speech and isolated word recognition, significant improvement in recognition accuracy is obtained using the new algorithms over those which do not utilize the joint evaluation strategy.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

We are addressing a new problem of improving automatic speech recognition performance, given multiple utterances of patterns from the same class. We have formulated the problem of jointly decoding K multiple patterns given a single Hidden Markov Model. It is shown that such a solution is possible by aligning the K patterns using the proposed Multi Pattern Dynamic Time Warping algorithm followed by the Constrained Multi Pattern Viterbi Algorithm The new formulation is tested in the context of speaker independent isolated word recognition for both clean and noisy patterns. When 10 percent of speech is affected by a burst noise at -5 dB Signal to Noise Ratio (local), it is shown that joint decoding using only two noisy patterns reduces the noisy speech recognition error rate to about 51 percent, when compared to the single pattern decoding using the Viterbi Algorithm. In contrast a simple maximization of individual pattern likelihoods, provides only about 7 percent reduction in error rate.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

We describe a novel method for human activity segmentation and interpretation in surveillance applications based on Gabor filter-bank features. A complex human activity is modeled as a sequence of elementary human actions like walking, running, jogging, boxing, hand-waving etc. Since human silhouette can be modeled by a set of rectangles, the elementary human actions can be modeled as a sequence of a set of rectangles with different orientations and scales. The activity segmentation is based on Gabor filter-bank features and normalized spectral clustering. The feature trajectories of an action category are learnt from training example videos using dynamic time warping. The combined segmentation and the recognition processes are very efficient as both the algorithms share the same framework and Gabor features computed for the former can be used for the later. We have also proposed a simple shadow detection technique to extract good silhouette which is necessary for good accuracy of an action recognition technique.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

The evolution of crystallographic texture in polycrystalline copper and nickel has been studied. The deformation texture evolution in these two materials over seven orders of magnitude of strain rate from 3 x 10(-4) to similar to 2.0 x 10(+3) s(-1) show little dependence on the stacking fault energy (SFE) and the amount of deformation. Higher strain rate deformation in nickel leads to weakerh < 101 > texture because of extensive microband formation and grain fragmentation. This behavior, in turn, causes less plastic spin and hence retards texture evolution. Copper maintains the stable end < 101 > component over large strain rates (from 3 x 10(-4) to 10(+2) s(-1)) because of its higher strain-hardening rate that resists formation of deformation heterogeneities. At higher strain rates of the order of 2 x 10(+3) s(-1), the adiabatic temperature rise assists in continuous dynamic recrystallization that leads to an increase in the volume fraction of the < 101 > component. Thus, strain-hardening behavior plays a significant role in the texture evolution of face-centered cubic materials. In addition, factors governing the onset of restoration mechanisms like purity and melting point govern texture evolution at high strain rates. SFE may play a secondary role by governing the propensity of cross slip that in turn helps in the activation of restoration processes.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Texture development in commercially pure titanium during equal channel angular extrusion (ECAE) through Routes A, Be and C has been studied up to three passes at 400 C. Textures were measured using X-ray diffraction, while the microstructural analyses were performed using electron back-scattered diffraction as well as transmission electron microscopy. Occurrences of dynamic restoration processes (recovery and recrystallization) were clearly noticed at all levels of deformations. Finally, the textures were simulated using a viscoplastic polycrystal self-consistent (VPSC) model. Simulations were performed incorporating basal, prismatic and pyramidal slip systems as well as tensile and compressive twinning. The simulated textures corroborate well with experimental textures in spite of the occurrence of dynamic restoration processes. (C) 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Cylindrical specimens of textured commercial pure alpha-titanium plate, cut with the cylinder axis along the rolling direction for one set of experiments and in the long transverse direction for the other set, were compressed at strain rates in the range of 0.001 to 100 s-1 and temperatures in the range of 25-degrees-C to 400-degrees-C. At strain rates greater-than-or-equal-to 1 s-1, both sets of specimens exhibited adiabatic shear bands, but the intensity of shear bands was found to be higher in the rolling direction specimens than in the long transverse direction specimens. At strain rates -0.1 s-1, the material deformed in a microstructurally inhomogeneous fashion. For the rolling direction specimens, cracking was observed at 100-degrees-C and at strain rates -0.1 s-1. This is attributed to dynamic strain aging. Such cracking was not observed in the long transverse specimens. The differences in the intensity of adiabatic shear bands and that of dynamic strain aging between the two sets of test specimens are attributed to the strong crystallographic texture present in these plates.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Evolution of crystallographic texture during high strain rate deformation in FCC materials with different stacking fault energy (Ni, Cu, and Cu-10Zn alloy) has been studied. The texture evolved in FCC materials at these strain rates show little dependence on the Stacking Fault Energy and the amount of deformation. Copper shows an anomalous behavior that is attributed to the ease of cross slip and continuous Dynamic Recrystallization that are operative under the experimental conditions.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

The evolution of microstructure and texture during extrusion of pure magnesium and its single phase alloy AM30 has been studied experimentally as well as by crystal plasticity simulation. Microstructure and micro-texture were characterized by electron back scattered diffraction (EBSD), bulk-texture was measured using X-ray diffraction and deformation texture simulations were carried out using visco-plastic self consistent (VPSC) model. In spite of clear indications of the occurrence of dynamic recrystallization (DRX), simulations were able to reproduce the experimental textures successfully. This was attributed to the fact that the textures were c-type fibers with their axis of rotation parallel to the c-axis and DRX leads to simply rotate the texture around the c-axis. (C) 2011 Elsevier B.V. All rights reserved.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Commercially Pure Magnesium initially hot rolled and having a basal texture was deformed by Equal Channel Angular Extrusion (ECAE). ECAE was carried out upto 8 passes in a 90° die following routes A and Bc through a processing sequence involving two temperatures, namely 523 and 473 K. Texture and microstructure formed were studied using electron back scatter diffraction (EBSD) technique. In addition to significant reduction in grain size, strong <0002> fiber texture inclined at an angle ~ 45o from the extrusion axis formed in the material. Texture was also analyzed by orientation distribution function (ODF) and compared vis-à-vis shear texture. A significant amount of dynamic recrystallization occurred during ECAE, which apparently did not influence texture.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

In this paper, we present an unrestricted Kannada online handwritten character recognizer which is viable for real time applications. It handles Kannada and Indo-Arabic numerals, punctuation marks and special symbols like $, &, # etc, apart from all the aksharas of the Kannada script. The dataset used has handwriting of 69 people from four different locations, making the recognition writer independent. It was found that for the DTW classifier, using smoothed first derivatives as features, enhanced the performance to 89% as compared to preprocessed co-ordinates which gave 85%, but was too inefficient in terms of time. To overcome this, we used Statistical Dynamic Time Warping (SDTW) and achieved 46 times faster classification with comparable accuracy i.e. 88%, making it fast enough for practical applications. The accuracies reported are raw symbol recognition results from the classifier. Thus, there is good scope of improvement in actual applications. Where domain constraints such as fixed vocabulary, language models and post processing can be employed. A working demo is also available on tablet PC for recognition of Kannada words.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

In this paper, we compare the experimental results for Tamil online handwritten character recognition using HMM and Statistical Dynamic Time Warping (SDTW) as classifiers. HMM was used for a 156-class problem. Different feature sets and values for the HMM states & mixtures were tried and the best combination was found to be 16 states & 14 mixtures, giving an accuracy of 85%. The features used in this combination were retained and a SDTW model with 20 states and single Gaussian was used as classifier. Also, the symbol set was increased to include numerals, punctuation marks and special symbols like $, & and #, taking the number of classes to 188. It was found that, with a small addition to the feature set, this simple SDTW classifier performed on par with the more complicated HMM model, giving an accuracy of 84%. Mixture density estimation computations was reduced by 11 times. The recognition is writer independent, as the dataset used is quite large, with a variety of handwriting styles.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Commercial purity (99.8%) magnesium single crystals were subjected to plane strain compression (PSC) along the c-axis at 200 and 370 degrees C and a constant strain rate of 10(-3) s(-1). Extension was confined to the < 1 1 (2) over bar 0 > direction and the specimens were strained up to a logarithmic true strain of -1. The initial rapid increase in flow stress was followed by significant work softening at different stresses and comparable strains of about -0.05 related to macroscopic twinning events. The microstructure of the specimen after PSC at 200 degrees C was characterized by a high density of {1 0 (1) over bar 1} and {1 0 (1) over bar 3} compression twins, some of which were recrystallized. After PSC at 370 degrees C, completely recrystallized twin bands were the major feature of the observed microstructure. All new grains in these bands retained the same c-axis orientation of their compression twin hosts. The basal plane in these grains was randomly rotated around the c-axis, forming a fiber texture component. The obtained results are discussed with respect to the mechanism of recrystallization, the specific character of the boundaries between new grains and the initial matrix, and the importance of the dynamically recrystallized bands for strain accommodation in these deformed magnesium single crystals. (C) 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Boron addition to conventional titanium alloys below the eutectic limit refines the cast microstructure and improves mechanical properties. The present work explores the influence of hypoeutectic boron addition on the microstructure and texture evolution in Ti-6Al-4V alloy under beta extrusion. The beta extruded microstructure of Ti-6Al-4V is characterized by shear bands parallel to the extrusion direction. In contrast, the extruded Ti-6Al-4V-0.1B alloy shows a regular beta worked microstructure consisting of fine prior beta grains and acicular alpha-lamellae with no signs of the microstructural instability. Crystallographic texture after extrusion was almost identical for the two alloys indicating the similarity in their transformation behavior, which is attributed to complete dynamic recrystallization during beta processing. Microstructural features as well as crystallographic texture indicate dominant grain boundary related deformation processes for the boron modified alloy that leads to homogeneous deformation without instability formation. The absence of shear bands has significant technological importance as far as the secondary processing of boron added alloys in (alpha + beta)-phase field are concerned. (C) 2012 Elsevier B.V. All rights reserved.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

In the present study, asymmetric rolling was carried out for incorporating a shear component during the rolling at different temperatures, and was compared with conventional (symmetric) rolling. The microstructures were investigated using electron back-scatter diffraction (EBSD). The strain incorporated was compared with the help of grain orientation spread (GOS). GOS was eventually used as a criterion to partition the microstructure for separating the deformed and the dynamically recrystallized (DRX) grains. The texture of the partitioned DRX grains was shifted by similar to 30 degrees along the c-axis from the deformed grains. The mechanism of dynamic recrystallization (DRX) has been identified as continuous dynamic recovery and recrystallization (CDRR). The partitioned deformed grains for the higher temperature rolled specimens exhibited a texture similar to the room temperature rolled specimen. The asymmetric rolling introduces a shear component which shifts the texture fibre by similar to 5-10 degrees from the conventional rolling texture. This led to an increase in ductility with little compromise on strength. (c) 2012 Elsevier B.V. All rights reserved.

Relevância:

30.00% 30.00%

Publicador:

Resumo:

Multi-layered materials have been made from Cu-Fe with approximately equal volume fractions using the Accumulated Roll Bonding (ARB) technique with less than 1 μm thickness of the individual layers. The so-obtained multi-layers have been subjected to deformation by cold rolling to 25, 50, 75, 87 and 93% reduction in thickness. A detailed characterization has been carried out using X-ray diffraction (line profile analysis and texture measurement) and electron (scanning and transmission) microscopy. It has been found that Fe layers are disintegrated whereas Cu retains its continuity within a layer. Microstructural Characterization from X-Ray Line profile Analysis (XRDLPA) through Variance Method reveals that large amount of strain is initially carried by Cu layers during deformation. In the Cu-Fe layer, the texture is comparatively weaker in Cu layer and strong in Fe layers. Brass Component increases up to 75% reduction and then decreases, while the ratio of Cu/S and Bs/S remains almost constant through out the deformation. After 50% reduction, dynamic recovery is predominant as indicated by the increase in the amount of low angle grain boundaries and decrease in dislocation density. The presence of R component indicates continuous dynamic recovery and recrystallization (CDRR) at the advanced stage of deformation.