Source author record

Arijit Dey

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

math.AG eess.AS Machine Learning math.CV Sound

Catalog footprint

Research overview

9works

5topics

4close collaborators

Actions

Is this your work?

Open graph Browse papers

preprint2020arXiv

Optimizing Speech Emotion Recognition using Manta-Ray Based Feature Selection

Emotion recognition from audio signals has been regarded as a challenging task in signal processing as it can be considered as a collection of static and dynamic classification tasks. Recognition of emotions from speech data has been heavily relied upon end-to-end feature extraction and classification using machine learning models, though the absence of feature selection and optimization have restrained the performance of these methods. Recent studies have shown that Mel Frequency Cepstral Coefficients (MFCC) have been emerged as one of the most relied feature extraction methods, though it circumscribes the accuracy of classification with a very small feature dimension. In this paper, we propose that the concatenation of features, extracted by using different existing feature extraction methods can not only boost the classification accuracy but also expands the possibility of efficient feature selection. We have used Linear Predictive Coding (LPC) apart from the MFCC feature extraction method, before feature merging. Besides, we have performed a novel application of Manta Ray optimization in speech emotion recognition tasks that resulted in a state-of-the-art result in this field. We have evaluated the performance of our model using SAVEE and Emo-DB, two publicly available datasets. Our proposed method outperformed all the existing methods in speech emotion analysis and resulted in a decent result in these two datasets with a classification accuracy of 97.06% and 97.68% respectively.

preprint2016arXiv

On the Rationality of Nagaraj-Seshadri Moduli Space

We show that each of the irreducible components of moduli of rank 2 torsion-free sheaves with odd Euler characteristic over a reducible nodal curve is rational.

preprint2016arXiv

Rationality of Moduli space over reducible curve

Let $M(2,\textbf{\underline{w}},χ)$ be the moduli space of rank $2$ torsion-free sheaves over a reducible nodal curve with each component having utmost two nodal singularities. We show that in each component of $M(2,\textbf{\underline{w}},χ)$ , the closure of rank $2$ vector bundles with fixed determinant is rational.

preprint2015arXiv

A classification of equivariant principal bundles over nonsingular toric varieties

We classify holomorphic as well as algebraic torus equivariant principal $G$ -bundles over a nonsingular toric variety $X$ , where $G$ is a complex linear algebraic group. It is shown that any such bundle over an affine, nonsingular toric variety admits a trivialization in equivariant sense. We also obtain some splitting results.

preprint2015arXiv

Equivariant principal bundles and logarithmic connections on toric varieties

Let $M$ be a smooth complex projective toric variety equipped with an action of a torus $T$ , such that the complement $D$ of the open $T$ --orbit in $M$ is a simple normal crossing divisor. Let $G$ be a complex reductive affine algebraic group. We prove that an algebraic principal $G$ --bundle $E_G\to M$ admits a $T$ --equivariant structure if and only if $E_G$ admits a logarithmic connection singular over $D$ . If $E_H\to M$ is a $T$ -equivariant algebraic principal $H$ --bundle, where $H$ is any complex affine algebraic group, then $E_H$ in fact has a canonical integrable logarithmic connection singular over $D$ .

preprint2013arXiv

Equivariant principal bundles on nonsingular toric varieties

We give a classification of the equivariant principal $G$ -bundles on a nonsingular toric variety when $G$ is a closed Abelian subgroup of $GL_k(\mathbb{C})$ . We prove that any such bundle splits, that is, admits a reduction of structure group to the intersection of $G$ with a torus. We give an explicit parametrization of the isomorphism classes of such bundles for a large family of $G$ when $X$ is complete.

preprint2012arXiv

SYZ duality for parabolic Higgs moduli spaces

We prove the SYZ (Strominger-Yau-Zaslow) duality for the moduli space of full flag parabolic Higgs bundles over a compact Riemann surface.The SYZ duality was proved for moduli spaces of Higgs vector bundles over a compact Riemann surface by Hausel and Thaddeus.

preprint2010arXiv

Brauer group of a moduli space of parabolic vector bundles over a curve

Let ${\mathcal P}{\mathcal M}^α_s$ be a moduli space of stable parabolic vector bundles of rank $n \geq 2$ and fixed determinant of degree $d$ over a compact connected Riemann surface $X$ of genus $g(X) \geq 2$ . If $g(X) = 2$ , then we assume that $n > 2$ . Let $m$ denote the greatest common divisor of $d$ , $n$ and the dimensions of all the successive quotients of the quasi-parabolic filtrations. We prove that the cohomological Brauer group ${\rm Br}({\mathcal P}{\mathcal M}^α_s)$ is isomorphic to the cyclic group ${\mathbb Z}/ m{\mathbb Z}$ . We also show that ${\rm Br}({\mathcal P}{\mathcal M}^α_s)$ is generated by the Brauer class of the projective bundle over ${\mathcal P}{\mathcal M}^α_s$ obtained by restricting the universal projective bundle over $X\times {\mathcal P}{\mathcal M}^α_s$ . We also prove that there is a universal vector bundle over $X\times {\mathcal P}{\mathcal M}^α_s$ if and only if $m=1$ .

preprint2010arXiv

Derived category of toric varieties with Picard number three

We construct a full, strongly exceptional collection of line bundles on the variety X that is the blow up of the projectivization of the vector bundle O_{P^{n-1}}\oplus O_{P^{n-1}}(b) along a linear space of dimension n-2, where b is a non-negative integer.

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.