Vocal Tract Length Normalization Features for Audio Search

This paper presents speaker normalization approaches for audio search task. Conventional state-of-the-art feature set, viz., Mel Frequency Cepstral Coefficients (MFCC) is known to contain speaker-specific and linguistic information implicitly. This might create problem for speaker-independent audio search task. In this paper, universal warping-based approach is used for vocal tract length normalization in audio search. In particular, features such as scale transform and warped linear prediction are used to compensate speaker variability in audio matching. The advantage of these features over conventional feature set is that they apply universal frequency warping for both the templates to be matched during audio search. The performance of Scale Transform Cepstral Coefficients (STCC) and Warped Linear Prediction Cepstral Coefficients (WLPCC) are about 3% higher than the state-of-the-art MFCC feature sets on TIMIT database.

PolyMage: Automatic Optimization for Image Processing Pipelines

This paper presents the design and implementation of PolyMage, a domain-specific language and compiler for image processing pipelines. An image processing pipeline can be viewed as a graph of interconnected stages which process images successively. Each stage typically performs one of point-wise, stencil, reduction or data-dependent operations on image pixels. Individual stages in a pipeline typically exhibit abundant data parallelism that can be exploited with relative ease. However, the stages also require high memory bandwidth preventing effective utilization of parallelism available on modern architectures. For applications that demand high performance, the traditional options are to use optimized libraries like OpenCV or to optimize manually. While using libraries precludes optimization across library routines, manual optimization accounting for both parallelism and locality is very tedious. The focus of our system, PolyMage, is on automatically generating high-performance implementations of image processing pipelines expressed in a high-level declarative language. Our optimization approach primarily relies on the transformation and code generation capabilities of the polyhedral compiler framework. To the best of our knowledge, this is the first model-driven compiler for image processing pipelines that performs complex fusion, tiling, and storage optimization automatically. Experimental results on a modern multicore system show that the performance achieved by our automatic approach is up to 1.81x better than that achieved through manual tuning in Halide, a state-of-the-art language and compiler for image processing pipelines. For a camera raw image processing pipeline, our performance is comparable to that of a hand-tuned implementation.

Comparison of Matching Pursuit Algorithm with Other Signal Processing Techniques for Computation of the Time-Frequency Power Spectrum of Brain Signals

Signals recorded from the brain often show rhythmic patterns at different frequencies, which are tightly coupled to the external stimuli as well as the internal state of the subject. In addition, these signals have very transient structures related to spiking or sudden onset of a stimulus, which have durations not exceeding tens of milliseconds. Further, brain signals are highly nonstationary because both behavioral state and external stimuli can change on a short time scale. It is therefore essential to study brain signals using techniques that can represent both rhythmic and transient components of the signal, something not always possible using standard signal processing techniques such as short time fourier transform, multitaper method, wavelet transform, or Hilbert transform. In this review, we describe a multiscale decomposition technique based on an over-complete dictionary called matching pursuit (MP), and show that it is able to capture both a sharp stimulus-onset transient and a sustained gamma rhythm in local field potential recorded from the primary visual cortex. We compare the performance of MP with other techniques and discuss its advantages and limitations. Data and codes for generating all time-frequency power spectra are provided.

A Numerical Model of Rapid Solidification Processing of Ni-Al Alloy in Planar Flow Casting

A numerical model has been developed for simulating the rapid solidification processing (RSP) of Ni-Al alloy in order to predict the resultant phase composition semi-quantitatively during RSP. The present model couples the initial nucleation temperature evaluating method based on the time dependent nucleation theory, and solidified volume fraction calculation model based on the kinetics model of dendrite growth in undercooled melt. This model has been applied to predict the cooling curve and the volume fraction of solidified phases of Ni-Al alloy in planar flow casting. The numerical results agree with the experimental results semi-quantitatively.

Tailoring of dendritic microstructure in solidification processing by crucible vibration

We analyzed the effects of both natural convection and forced flows on solid–liquid interface morphology during upward Bridgman solidification of metallic alloys. Experiments were carried out on Al–3.5wt% Ni alloy, for a cylindrical sample. The influence of natural convection induced by radial thermal gradient on solidified microstructure was first analyzed as a function of the pulling rate. Then, the influence of axial vibration on solidification microstructure was experimentally investigated by varying vibration parameters (frequency and amplitude). Experimental results demonstrated that vibrations could be used to either attenuate fluid flow in the melt and obtain a uniform dendritic pattern or to promote a fragmented dendritic microstructure. However, no marked effect was observed for cellular growth. This pointed out the critical role of the mushy zone in the interaction between fluid flow and solidification microstructure.

Influence of the Processing Conditions on the Characteristics of the Clad Layers Produced with Laminar Plasma Technology

Laminar plasma technology was used to produce ceramic hardened layers of Al2O3-40% mass Ni composite powders on stainless steel substrates. In order to investigate the influences of processing conditions on the morphologies of the surface modified layers, two different powder-feeding methods were tested, one with carrier gas called the powder injection method, and the other without carrier gas called powder transfers method. The microscopic investigations demonstrate that the cross-section of the clad layers consists of two distinct microstructural regions, in which the Al2O3 phases exhibit different growth mechanisms. When the powder transfers method is adopted, the number density and volume fraction of the Al2O3 particles increase considerably and their distributions exhibit zonal periodical characteristics. When the powder-feeding rate increases, the microstructure of the Al2O3 phases changes from a small globular to a long needle shape. Finite element simulations show that the transient thermo-physical features of the pool substances, such as solidification rate and cooling rate, influence strongly the mechanisms of the nucleation and the directional growth of the Al2O3 phases in the thermal processing.

Determination of Proper Processing Conditions of Material Surface Heat Treatment Through Finite Element Method

A two-dimensional model has been developed based on the experimental results of stainless steel remelting with the laminar plasma technology to investigate the transient thermo-physical characteristics of the melt pool liquids. The influence of the temperature field, temperature gradient, solidification rate and cooling rate on the processing conditions has been investigated numerically. Not only have the appropriate processing conditions been determined according to the calculations, but also they have been predicted with a criterion established based on the concept of equivalent temperature area density (ETAD) that is actually a function of the processing parameters and material properties. The comparison between the resulting conditions shows that the ETAD method can better predict the optimum condition.

Three-Dimensional Modeling of Heat Transfer and Fluid Flow in Laminar-Plasma Material Re-Melting Processing

Modeling study is performed concerning the heat transfer and fluid flow for a laminar argon plasma jet impinging normally upon a flat workpiece exposed to the ambient air. The diffusion of the air into the plasma jet is handled by using the combined-diffusion-coefficient approach. The heat flux density and jet shear stress distributions at the workpiece surface obtained from the plasma jet modeling are then used to study the re-melting process of a carbon steel workpiece. Besides the heat conduction within the workpiece, the effects of the plasma-jet inlet parameters (temperature and velocity), workpiece moving speed, Marangoni convection, natural convection etc. on the re-melting process are considered. The modeling results demonstrate that the shapes and sizes of the molten pool in the workpiece are influenced appreciably by the plasma-jet inlet parameters, workpiece moving speed and Marangoni convection. The jet shear stress manifests its effect at higher plasma-jet inlet velocities, while the natural convection effect can be ignored. The modeling results of the molten pool sizes agree reasonably with available experimental data.

Numerical Simulation of Transient Thermal Features in the Remelting Processing of Stainless Steel

应用有限元方法对层流等离子体射流不锈钢表面重熔工艺中的瞬态热物理现象进行了数值模拟研究.针对不同加热距离,确定了材料熔化和凝固过程中的瞬态温度场、温度梯度和凝固率的时间和空间分布特征.通过引入等效温度面积密度概念,研究了不锈钢重熔热处理的适合条件.结果表明,9～13mm的范围是较为适宜的加热距离,该结果与试验观察基本符合.

Preliminary Investigations on Low-Pressure Laminar Plasma Spray Processing

The usual plasma spraying methods often involve entrainment of the surrounding air into the turbulent plasma core and result in coated materials having relatively high porosity and low adhesive strength. Therefore, exploration of new plasma spraying methods for fabricating high quality coatings to meet the requirement of special applications will be quite important. In this study, an alternative plasma spraying method, i.e. the low-pressure laminar plasma spraying process, is investigated and used in an attempt for spraying thermal barrier coatings (TBCs). Investigations on the characteristics of the laminar plasma jets, feeding methods for the ceramic powder and the formation process of the individual quenched splats have been carried out. The properties of the TBCs sprayed by laminar plasma jet process, such as the adhesive strength at the interface of the ceramic coating/bond coat, the surface roughness and microstructure, are examined by tensile tests and scanning electron microscope (SEM) observations.