Knowledge-Based Transformation Ordering - Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE Inte
نویسندگان
چکیده
We propose a two-step approach for transformation ordering which combines the use of optimization-intensive CAD techniques with knowledge-based user-driven search strategy. The first step is development of basic building blocks which target small sets of transformations which are well suited for optimization intensive CAD treatment. Next, transformation orderings are developed using knowledge about mathematical laws, an application domain, and the relationship among transformations. Transformation ordering scripts combine several building blocks: to form effective approaches for optimization of several design metrics in many common computational structures. As the highlight of the approach, we developed a method which efficiently simultaneously optimizes throughput, latency, power, and area of linear computations.
منابع مشابه
ESTIMATION OF RIGID BODY MOTION AND SCENE STRUCTURE FROM IMAGE SEQUENCES USING A NOVEL EPIPOLAR TRAN - Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE Inte
We present a new algorithm for estimation of rigid body motion parameters and scene structure from monocular image sequences. A novel Epipolar Transform is utilized to preserve the relevant information from mean squared displaced frame difference (DFD) surfaces and thus overcomes the inherent limitations of feature correspondence methods. Our algorithm conducts a coarse-to-fine search in 5-dime...
متن کاملOPTIMAL PHASE KERNELS FOR TIME-FREQUENCY ANALYSIS - Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE Inte
We consider the design of kernels for time-frequency distributions through the phase, rather than amplitude, response. While phase kernels do not attenuate troublesome crosscomponents, they can translate them in the time-frequency plane. In contrast to previous work on phase kernels that concentrated on placing the cross-components on top of the auto-components, we set up a “don’t care” region ...
متن کاملDesign of Efficient FIR Filters with Cyclotomic Polynomial Prefilters Using Mixed Integer Linear Pro - Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE Inte
s-The Cyclotomic Polynomial (CP) prefilter design problem is formulated as an optimization problem with linear objective functions by applying the logarithm to the transfer function of the CP prefilter. Then this problem is solved by mixed integer linear programming (MILP). Design examples demonstrate that this method leads to more efficient cascaded FIR prefilter-equalizers than existing methods.
متن کاملKNOWLEDGE-BASED PARAMETERS FOR HMM SPEECH RECOGNITION - Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE Inte
This paper presents acoustic parameters ( AP’s) that were motivated by phonetic feature theory and employed as a signal representation of speech in a Hidden Markov Model (HMM) recognition framework. Presently, the phonetic features considered are the manner features: sonorant, syllabic, nonsyllabic, noncontinuant and fricated. The objective of the parameters is to directly target the linguistic...
متن کاملSPECTRAL ESTIMATION BASED ON STRUCTURED LOW RANK MATRIX PENCIL - Acoustics, Speech, and Signal Processing, 1996. ICASSP-96. Conference Proceedings., 1996 IEEE Inte
This paper proposes a new parameter estimation algorithm for damped sinusoidal signals. Parameter estimation for damped sinusoidal signals with additive white noise is a problem of significant interest in many signal processing applications, like analysis of NMR data and system identification. The new algorithm estimates the signal parameters using a matrix pencil constructed from the measured ...
متن کامل