Ana Gavrovska

University of Belgrade · School of Electrical Engineering

Modern commercial software tools have the ability to deceive a viewer who is unable to determine whether the image content is authentic or not. Research on visual traces, image modifications as attacks, and possible misleading forensic analysis in practice, led to reexamining common used formats, like JPEG (Joint Photographic Experts Group) compres...

The Image Quality Assessment (IQA) metrics are developed to measure the perceptual quality of images after distortion. In this paper several No-Reference (NR) metrics are tested for different distortion types. In order to see how each type of distortion affects the image quality measures, each image will have only one type of distortion. The experi...

Entire world has been dealing with the number of new Coronavirus 2 or COVID-19 cases. The spread of this severe acute respiratory syndrome has produced many concerns worldwide. Having data related to coronavirus available for tests, novel models for forecasting the number of new cases can be developed. In this paper, a long short-term memory (LSTM)...

The industrial consortium AOMedia has recently developed the AV1 (AOMedia Video 1) format considered as a successor of VP9. One of the main goals with the video coding and compression format is to enable efficient Web based video content delivery, as in the case of OTT (Over-the-top) services. Nowadays, there are different AV1 software implementati...

It is much expected from the relatively novel, open, royalty-free AV1 (Alliance for Open Media (AOMedia) Video 1) standard. At this moment, there are many new variants of AV1 format. It is designed for efficient video internet delivery and high-quality video transmission. AV1 is recognized as Google’s VP9 format successor. One of the reference tool...

Nowadays advantages in face-based modification using DeepFake algorithms made it possible to replace a face of one person with a face of another person. Thus, it is possible to make not only copy-move modifications, but to implement artificial intelligence and deep learning for replacing face movements from one person to another. Still images can b...

Artificial intelligence is making significant changes in industrial internet of things (IIoT). Particularly, machine and deep learning architectures are now used for cybersecurity in smart factories, smart homes, and smart cities. Using advanced mathematical models and algorithms more intelligent protection strategies should be developed. Hacking o...

Digital images and video are the basic media for communication nowadays. They are used as authenticated proofs or corroboratory evidence in different areas like: forensic studies, law enforcement, journalism and others. With development of software for editing digital images, it has become very easy to change image content, add or remove important...

Efficient management of video sequences is based on adequate video content description. This description can be used for various purposes in different applications, telecommunication services, video and multimedia systems. Video hard cut detection represents the foundation of temporal video segmentation. In this paper, a new video hard cut detectio...

Luminance compression is often performed for high dynamic range images (still images and videos). A nonlinear tone mapping is applied for the compression in order to reproduce high dynamic range images using devices with a more limited (low) dynamic range. The images obtained after mapping may provide significant content differences in comparison t...

There has been a sustained effort in the research community over the recent years to develop algorithms that automatically analyze heart sounds. One of the major challenges is identifying primary heart sounds, S1 and S2, as they represent reference events for the analysis. The study presented in this paper analyzes the possibility of improving the...

From the perspective of average viewer, high definition video streams such as HD (High Definition) and UHD (Ultra HD) are increasing their internet presence year over year. This is not surprising, having in mind expansion of HD streaming services, such as YouTube, Netflix etc. Therefore, high definition video streams are starting to challenge netwo...

In this paper two types of physiological signals are analyzed from the music effect standpoint. In particular, we considered the state before (b state) and after listening classical music (p state). The signals belonging to each state of silence are analyzed according to the variability of heartbeats. The made correlation between the states describ...

In this paper we analyze video traffic variability taking into account H.265/HEVC video encoding. In particular, we examine the video trace variability considering high definition and ultra high definition (4k) formats. The preliminary results show the frame based differentiation between the new and common formats. The obtained results seem promisi...

Continued advances in imaging technologies are expected to additionally boost video industry. One of the main driving forces for such evolution is highly correlated with human visual perception. This is proven by the high-end imaging systems and new standards. The deployment of Ultra High-Definition (UHD) television and video does not offer just hi...

High definition video streams increase their Internet presence year over year. Therefore, they are starting to challenge network resource allocation with their bandwidth requirements and statistical characteristics. In this paper, we apply a seasonal autoregressive model for modeling and prediction of 4K video traffic, encoded with H.265 (HEVC) enc...

Healthy versus unhealthy heart sound computer-aided classification tools are very popular for supporting clinical decisions. In this paper a new method is proposed for the classification of heart sound recordings from a statistical standpoint without detection and localization of fundamental heart sounds (S1, S2). This study analyzes the possibilit...

The first step in most computer aided diagnosis systems is an accurate segmentation of breast region, which affects not only the accuracy but also the speed of the analysis because it significantly reduces the area of the image to be examined. The second step usually includes removal of pectoral muscle region, which is seen in mediolateral oblique...

Animation frame content is analyzed in this paper. The animation found in frame/video content is usually characterized by specific color distribution and texture. The existence of animation within a frame can be described by carefully selected features. We analyzed both fully and partially animated frames. Here, we present our initial results relat...

Proper segmentation of breast and pectoral muscle regions is often the first step in systems for automatic analysis of mammographic images. This paper proposes a method for extraction of the breast region using adaptive gamma correction algorithm for image contrast enhancement and k-means algorithm for thresholding the image in order to extract bac...

In this paper we perform animation frame analysis. The animation frame/video content is usually characterized by specific color distribution and texture within frames. The existence of animation within frame can be described by carefully selected features. Furthermore, frames can be partially animated. Here, we present our initial results related t...

New trends in cardiosignal processing show that local fluctuations of the signals may reveal important information for computer-aided analysis. Even though the calculation of variance as a global measure is valuable, variation described locally and in one or more scales can represent a convenient way to describe details found in a signal. In this p...

Echocardiograms are inherently corrupted by the speckle noise. Elimination of the noise is usually treated with low-pass filters which can degrade edges in the image. Adaptive approaches employ masks for edges and restrict low-pass filtering mainly to homogeneous regions. Masks are based on statistical parameters or gradients. In this paper are app...

Heart rate (HR) estimation is of crucial importance in human health evaluation and can be considered as valuable indicator of cardiovascular state. In this paper contact based system for heart rate is presented, where system for HR acquisition is consisted of a laptop and web camera. In such system heart rate variability (HRV) affects the changes i...

The diagnostic process in medicine in many cases depends on readability of medical images. Thus, the image enhancement is very important step in medical imaging. We investigate the potential application of High Dynamic Range (HDR) imaging in echocardiography. HDR image is generated using two Low Dynamic Range (LDR) echocardiograms. Since many devic...

In order to obtain fine contacts over the crystal unit, it is necessary to apply a thin layer of metal electrodes on its surface. The structure of the thin metal film formed by evaporation and condensation under the vacuum depends on many parameters and therefore it is very important to define the terms of steaming. This paper describes a method fo...

In this paper special attention is given to domain selection and local and global signal examination which is important for signal processing, especially in cardiosignals where physiological information is a priority. In the case of phonocardiographic signals, spectral components are analyzed using joint time-frequency and multifractal representati...

Paper introduces a new approach for S1 versus S2 heart sound classification. The blanket fractal dimension is used for the first time for recognizing a heart sound candidate as S1 or S2 without the use of synchronous electrocardiogram. Even though the systole and diastole duration is not calculated for further improvement of the results, the obtain...

Phono- and echocardiogram signals play significant role in describing heart state. In our research we derived and proposed several linear and nonlinear methods for analyzing such signals. Some of our results are described.

In this paper, the cardiosignal analysis using second generation wavelets and lifting structure based on clinically-relevant extracted information is described. The information extraction can be performed by different spectral (such as multifractal) and spectrogram (time-frequency) representations. Here, the novelties are related mainly to the know...

Mitral valve prolapse (MVP) is the most common valve anomaly and the most frequent cause of isolated mitral insuffi-ciency. MVP has a mostly benign course and prognosis in childhood; however, complications, such as severe mitral regurgitation, infectious endocarditis, pulmonary embolism, arrhythmia and sudden death, occur more often in elderly peop...

This paper proposes a face recognition system based on Gabor image representation, principal component analysis and neural networks. Gabor filter bank at 5 scales and 8 orientations is used for face representation. The dimensionality of Gabor representation of face images is reduced by the Principal component analysis. The images are classified usi...

Wavelet-based (WT) denoising has been frequently used as simple and efficient method enabling high signal-to-noise ratio. Moreover, by using second generation WT and lifting scheme, even better performance has been obtained. Besides the WT, empirical mode decomposition (EMD) is another efficient noise reduction technique, which enables adaptive dec...

Phonocardiography has shown a great potential for developing low-cost computer-aided diagnosis systems for cardiovascular monitoring. So far, most of the work reported regarding cardiosignal analysis using multifractals is oriented towards heartbeat dynamics. This paper represents a step towards automatic detection of one of the most common patholo...

Having in mind the availability of electronic stethoscopes, phonocardiograms (PCGs) have become popular for cardiovascular functionality monitoring and signal processing applications. Detection of fundamental heart sounds (HSs), S1s and S2s, is considered to be a crucial step in PCG analysis. Electrocardiogram (ECG), noted as a reference signal, is...

In modern communications, particularly digital one, it is very important to implement quality modules for receiver, transmitter and/or repeater units. One of the most significant components of those modules are crystal filters. This paper presents a computer simulation of crystal filter composed by four crystal units. By changing the filter paramet...

In modern communications, it is of great importance to implement quality modules for signal transmission, transfer and reception. One of the most important components of those modules is crystal filter. This is especially true for digital communications. In this paper, we propose antenna-filter F 106, suitable for high frequencies. This new filter...

This paper describes a new technological process of chemical polishing of SC cut crystal units. Due to the absence of a deformed layer on the surface of chemically polished units, that is always present in mechanically polished units, and the possibility to inspect crystal defects during abrasion, it is possible to select the defect-free units. Che...

Joint time-frequency representations of phonocardiograms enable fundamental characteristics of heart sounds and other physiological changes important for diagnostics to be more visible compared to time and frequency domain, separately. This paper discuss possibilities for relevant component extraction in such representations. Due to improved visual...

In this paper we propose the new multifractal measure inspired by sigmoid activation function usually used in neural networks. By using new measure the Hölder exponent and multifractal spectrum are determined in classical way. New measure is applied to image processing, especially in texture classification. It was shown that by changing the slope o...

This paper considers the implementation of the lifting scheme. It is used as an effective and flexible tool for introducing signal dependence into the problem by improving the wavelet properties. Here, we present the results for simple example of the lazy transform implementation for the signal denoising problem. The results are obtained by testing...

In image processing edges are observed as singularities. Calculating gradient is regular procedure for locating edges within the image. However, there are edges with different order of that given by gradient. Hölder exponent gives mathematical background for different types of edges considering differentiation. In this paper we analyze some well kn...

Tc-DMSA renal scintigraphy is widely accepted as the most sensitive method for evaluating renal parenchym damage like renal scars. Both qualitative and quantitative evaluation standards are used for improving DMSA scintigraphy interpretation. DMSA scan region of interest detection is the first step towards defect revealing and renal morphology eval...

Segmentation of echocardiograms is difficult task since speckle noise embedded in ultrasound images. Standard techniques for image segmentation, such as thresholding, region growing or watershed method, give poor results in heart chambers segmentation. Active contours “without edges” proved to be robust to noise and applicable for ultrasound images...

Kardiosignali, fonokardiogram, ultrazvu~na slika, ehokardiogram, deformacije, tehnike obrade signala i slike, pregled. Abstract The paper considers several techniques for cardiosignal analysis. We made a selection to present the importance of noticeable targeting in electrocardiograms (ECG), phonocar-diograms (PCG), ultrasound (US) imaging and othe...

Texture is defined in acoustics as a subjective measure described with three qualities regarding reflectograms. Here we propose using fractal dimension for texture analysis, thanks to ability of fractals to reveal hidden phenomena within the signal. Further discussion towards quantitative descriptors of texture, based on obtained results, will be p...

Assisting physicians in the auscultation of the patients by providing cost effective, automatic and accurate signal processing module for the heart event detection and recognition is of importance. We suggest that the segmentation of time-frequency representation image of one-dimensional phonocardiogram signals (PCGs) could be effective for observi...

Advancements in digital electronics and signal processing algorithms for various purposes generated the possibility and need for designing artificial neural networks in hardware. The selected platform, FPGA, enables fulfillment of their demands and provides comfortable work and test environment. This paper presents development cycle and specifics f...

Advancements in digital electronics and signal processing algorithms for various purposes generated the possibility and need for designing artificial neural networks in hardware. The selected platform, FPGA, enables fulfillment of their demands and provides comfortable work and test environment. This paper presents development cycle and specifics f...

Optimal selection of decomposition levels in wavelet transform used for both HF and LF filtering of ECG signal is described. By analyzing different types of wavelets and selected decomposition levels, the best parameters are determined and their statistical performances are derived. Three Physionet's ECG databases were used as a test-bed.

U ovom radu razmatrane su mogućnosti upotrebe jednostavnog, linearnog, diskretnog i nelinearnog, neosetljivog Kalmanovog filtra u cilju otklanjanja šuma kod kardiosignala. U odredjenim slučajevima, oni su se pokazali boljim rešenjem od većine standardnih adaptivnih i autoregresivnih tehnika. Osim toga, ove tehnike omogućavaju dobijanje preciznijih...

This paper defines pattern matching using backpropagation neural network for detection of Premature Ventricular Complexes in electrocardiograms. Scalograms using continuous wavelet transformation proved to be a powerful tool in morphological recognition of the segments. Total segment overlapping is also presented as a convenient method for classify...