Radovan Obradovic

Fully fractional anisotropic diffusion for image denoising

This paper introduces a novel Fully Fractional Anisotropic Diffusion Equation for noise removal which contains spatial as well as time fractional derivatives. It is a generalization of a method proposed by Cuesta which interpolates between the heat and the wave equation by the use of time fractional derivatives, and the method proposed by Bai and Feng, which interpolates between the second and the fourth order anisotropic diffusion equation by the use of spatial fractional derivatives. This equation has the benefits of both of these methods. For the construction of a numerical scheme, the proposed partial differential equation (PDE) has been treated as a spatially discretized Fractional Ordinary Differential Equation (FODE) model, and then the Fractional Linear Multistep Method (FLMM) combined with the discrete Fourier transform (DFT) is used. We prove that the analytical solution to the proposed FODE has certain regularity properties which are sufficient to apply a convergent and stable fractional numerical procedure. Experimental results confirm that our model manages to preserve edges, especially highly oscillatory regions, more efficiently than the baseline parabolic diffusion models.

Speech Technologies for Serbian and Kindred South Slavic Languages

This chapter will present the results of the research and development of speech technologies for Serbian and other kindred South Slavic languages used in five countries of the Western Balkans, carried out by the University of Novi Sad, Serbia in cooperation with the company AlfaNum. The first section will describe particularities of highly inflected languages (such as Serbian and other languages dealt with in this chapter) from the point of view of speech technologies. The following sections will describe the existing speech and language resources for these languages, the automatic speech recognition (ASR) and text-to-speech synthesis (TTS) systems developed on the basis of these resources as well as auxiliary software components designed in order to aid this development. It will be explained how the resources originally built for the Serbian language facilitated the development of speech technologies in Croatian, Bosnian, and Macedonian as well. The paper is concluded by the directions of further research aimed at development of multimodal dialogue systems in South Slavic languages.

AlfaNum System for Speech Synthesis in Serbian Language

This paper presents some basic criteria for conception of a concatenative text-to-speech synthesizer in Serbian language. The paper describes the prosody generator which was used and reflects upon several peculiarities of Serbian language which led to its adoption. Within the paper, the results of an experiment showing the influence of natural-sounding prosody on human speech recognition are discussed. The paper also describes criteria for on-line selection of appropriate segments from a large speech corpus, as well as criteria for off-line preparations of the speech database for synthesis.

Image denoising by a direct variational minimization

In this article we introduce a novel method for the image de-noising which combines a mathematically well-posdenes of the variational modeling with the efficiency of a patch-based approach in the field of image processing. It based on a direct minimization of an energy functional containing a minimal surface regularizer that uses fractional gradient. The minimization is obtained on every predefined patch of the image, independently. By doing so, we avoid the use of an artificial time PDE model with its inherent problems of finding optimal stopping time, as well as the optimal time step. Moreover, we control the level of image smoothing on each patch (and thus on the whole image) by adapting the Lagrange multiplier using the information on the level of discontinuities on a particular patch, which we obtain by pre-processing. In order to reduce the average number of vectors in the approximation generator and still to obtain the minimal degradation, we combine a Ritz variational method for the actual minimization on a patch, and a complementary fractional variational principle. Thus, the proposed method becomes computationally feasible and applicable for practical purposes. We confirm our claims with experimental results, by comparing the proposed method with a couple of PDE-based methods, where we get significantly better denoising results specially on the oscillatory regions.

An approach to instruction set compiled simulator development based on a target processor C compiler back-end design

Many instruction set simulation approaches place the retargetability and/or cycle-accuracy as the key features for easier architectural exploration and performance estimation early in the hardware development phase. This paper describes an approach in which importance of speed and controllability is placed above the cycle-accuracy and retargetability, thus providing a better platform for software development. The main idea behind this work is to associate the compiled simulator effort with the development of the C language compiler for the target embedded processor, using the knowledge related to compilers and reusing some common software elements. Through the prototype design of a compiled simulator for the Cirrus Logic Coyote DSP architecture, many implementation aspects are presented showing that this approach has great potential.

Robust sparse image denoising

In this paper, we propose a novel method for denoising images corrupted by the mixture of the additive white Gaussian noise and the heavy tailed noise. The proposed method is based on robust statistical approach, i.e. application of M-estimators in the combination with ℓ1 sparse regularization technique. Thus, we perform the denoising by numerical solving of the convex optimization problem. Additionally, the developed method performs denoising adaptively in order to preserve the image details, by adjusting regularization coefficients according to the local variance. The proposed de-noising scheme produces excellent results, both objectively (in terms of PSNR and MSSIM) and subjectively and outperforms state-of-the-art filters for denoising heavy tailed noise in images.

An Approach to Instruction Set Compiled Simulator Development Based on a Target Processor C Compiler Back-End Design

Many instruction set simulation approaches place the retargetability and/or cycle-accuracy as the key features for easier architectural exploration and performance estimation early in the hardware development phase. This paper describes an approach in which importance of speed and controllability is placed above the cycle-accuracy and retargetability, thus providing a better platform for software development. The main idea behind this work is to try to associate the compiled simulator effort with the development of the C language compiler for the target embedded processor, using the knowledge from that field of work and reusing some common software elements. Through the prototype design of a compiled simulator for the Cirrus Logic Coyote DSP architecture, many implementation aspects are presented proving that this approach has a great potential.

The Hough transformation of rectangle

Hough transformation for finding straight line segments is well known in digital image processing. In this paper we have used Hough, houghline and houghpeaks functions for creating a new 4- dimensional parameter space named h_space for extracting rectangles in digital image. These parameters are coordinates of the center of a rectangle (intersection of diagonals), angle that the shorter edge makes with x-axis and length of the longer edge. In our opinion this result might be useful for the shape descriptors associated with rectangles because the parameters of h-space are suitable for calculating descriptors of a rectangle (like area and perimeter) and they give a good idea about the position of a rectangle.

Sparse representation of precision matrices used in GMMs

The paper presents a novel precision matrix modeling technique for Gaussian Mixture Models (GMMs), which is based on the concept of sparse representation. Representation coefficients of each precision matrix (inverse covariance), as well as an accompanying overcomplete matrix dictionary, are learned by minimizing an appropriate functional, the first component of which corresponds to the sum of Kullback-Leibler (KL) divergences between the initial and the target GMM, and the second represents the sparse regularizer of the coefficients. Compared to the existing, alternative approaches for approximate GMM modeling, like popular subspace-based representation methods, the proposed model results in notably better trade-off between the representation error and the computational (memory) complexity. This is achieved under assumption that the training data in the recognition system utilizing GMM have an inherent sparseness property, which enables application of the proposed model and approximate representation using only one dictionary and a significantly smaller number of coefficients. Proposed model is experimentally compared with the Subspace Precision and Mean (SPAM) model, a state of the art instance of subspace-based representation models, using both the data from a real Automatic Speech Recognition (ASR) system, and specially designed sets of artificially created/synthetic data.