This paper presents new results by using our recently proposed on-line Bayesian learning approach for affine transformation parameter estimation in speaker adaptation. The on-line Bayesian learning technique allows updating parameter estimates after each utterance and i t can accommodate flexible forms of transformation functions as well as prior probability density function. We show through ex...

In these notes we study Lorentz transformations and focus on the group of proper, orthochronous Lorentz transformations, donated by L+. (These Lorentz transformations have determinant one and preserve the direction of time.) The 2×2 matrices with determinant one, denoted SL(2, C), play a key role, as there is a map from SL(2,C) onto L+, that is 2-to-1 and a homomorphism. In this correspondence,...

Abstract A new transformation, a generalization of the Givens rotation, is introduced here. Its properties are studied. As application of the new transformation, we discuss how to employ it to tridiagonalize matrices similarly, without requiring an intermediate transformation to Hessenberg form. This approach to tridiagonalization is cheaper than usual strategies, specially for sparse case in t...

Point pattern matching (PPM) is an important problem in pattern recognition, digital video processing and computer vision. In this paper, novel and fast procedure based on Genetic Algorithm, for afine PPM is described. Most matching techniques solved the PPM problem by determining the correspondence between points localized spatially within two sets, then to get the proper transformation parame...

To narrow down the gap between low-level visual features and high-level semantic concepts in content-based image retrieval(CBIR) system, a new dimensionality reduction algorithm called tensor biased discriminant Euclidean embedding (TBDEE) is proposed in this paper. The key idea of this algorithm is as follows: First, the image data are represented with high order tensor structure so that the c...

A novel approach for analyzing multiple protein structures is presented. A family of related protein structures may be characterized by an affine model, obtained by applying transformation matrices that permit both rotation and shear. The affine model and transformation matrices can be computed efficiently using a single eigen-decomposition. A novel method for finding correspondences is also in...

Point pattern matching (PPM) is an important problem in pattern recognition, digital video processing and computer vision. In this paper novel robust and fast procedure based on Genetic Algorithm, for n-dimensional PPM is described. Most matching techniques solved the PPM problem by determining the correspondence between points localized spatially within two sets, then to get the proper transfo...

We present an approach to spatial inference which is based on the procedural semantics of spatial relations. In contrast to qualitative reasoning, we do not use discrete symbolic models. Instead, relations between pairs of objects are represented by parameterized homogeneous transformation matrices with numerical constraints. A textual description of a spatial scene is transformed into a graph ...

In this paper we present the following definitions P-asymptotically equivalent probability of multiple L and P-asymptotically probability regular. In addition to these definitions we asked and provide answers for the following questions. (1) If x Probability ≈ y then what type of four dimensional matrices transformation will satisfy the following μ(Ax) Probability ≈ μ(Ay)? (2) If [x] and [y] ar...

In this paper, we present ideas to combine VTLN and SAT to improve the performance of automatic speech recognition. We show that VTLN matrices can be used as SAT transformation matrices in recognition, though the training still follows conventional SAT. This will be useful when there is very little adaptation data and the SAT transformation matrix can not be estimated to perform the required ad...