Trojachanec, Katarina

Preferred name

Official Name

Trojachanec, Katarina

Main Affiliation

katarina.trojacanec@finki.ukim.mk

Now showing 1 - 10 of 20

Some of the metrics are blocked by your
consent settings
Longitudinal Brain MRI Retrieval for Alzheimer’s Disease Using Different Temporal Information
(Institute of Electrical and Electronics Engineers (IEEE), 2018)
;
;
;
Some of the metrics are blocked by your
consent settings
Comparison of Different Methods for Estimation of Arterial Blood Pressure Using PPG Signals
(Springer, Cham, 2024-06-25)
Mladenovska, Teodora
;
;
;
;
The use of photoplethysmography (PPG) signals to predict the arterial blood pressure (ABP) waveform has gained popularity in recent years. Currently, there is a limited number of studies investigating this approach. This chapter elaborates a comparative analysis of two methodologies: a deep neural network approach and an encoder–decoder model for ABP waveform estimation with different window sizes expressed in seconds: 1s (175 signal points) and 4s (512 signal points). By estimating the systolic blood pressure (SBP) and diastolic blood pressure (DBP) scalars, this approach differs from conventional regression models that predict blood pressure through direct estimation; and it also enables another feature—evaluation of cardiovascular anomalies by analyzing the waveform patterns derived from the input PPG signal, which enables further medical analysis. The best obtained results are an R2 score of 0.76 for ABP, an MAE of 6.52 mmHg for DBP, using an encoder–decoder model on a sequence of 4s, and an MAE of 10.48 mmHg for SBP using GRU neural network on a sequence of 1s.
Some of the metrics are blocked by your
consent settings
Content based retrieval of MRI based on brain structure changes in Alzheimer’s disease
(SCITEPRESS, 2015-01-12)
;
;
Loshkovska, Suzana
;
The aim of the paper is to present Content Based Retrieval of MRI based on the brain structure changes characteristic for Alzheimer’s Disease (AD). The approach used in this paper aims to improve the retrieval performance while using smaller number of features in comparison to the descriptor dimensionality generated by the traditional feature extraction techniques. The feature vector consists of the measurements of cortical and subcortical brain structures, including volumes of the brain structures and cortical thickness. Two main stages are required to obtain these features: segmentation and calculation of the quantitative measurements. The feature subset selection is additionally applied using Correlation-based Feature Selection (CFS) method. Euclidean distance is used as a similarity measurement. The retrieval performance is evaluated using MRIs provided by the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Experimental results show that the strategy used in this research outperforms the traditional one despite its simplicity and small number of features used for representation.
Some of the metrics are blocked by your
consent settings
Mammography image classification using texture features
(Faculty of Computer Science and Engineering, Ss. Cyril and Methodius University in Skopje, Macedonia, 2012)
;
;
;
Jankulovski, Blagojche
Mammography image classification is a very important research field due to its implementation domain. The aim of this paper is propose techniques for automation of the mammography image classification process. This requires the images to be described using feature extraction algorithms and then classified using machine learning algorithms. In that context, the goal is to find which combination of feature extraction algorithm and classification algorithm yield the best results for mammography image classification. The following feature extraction methods were used LBP, GLDM, GLRLM, Haralick, Gabor filters and a combined descriptor. The images were classified using several machine learning algorithms i.e. support vector machines, random forests and k-nearest neighbour classifier. The best results were obtained when the images were described using GLDM together with the support vector machines as a classification technique.
Some of the metrics are blocked by your
consent settings
Пребарување на медицински слики базирано на лонгитудинални податоци за Алцхајмерова болест
(ФИНКИ, УКИМ, Скопје, 2018)
In this thesis, we performed research towards finding an efficient way to organize and represent the data from the medical cases with the aim to improve their retrieval from the large medical databases. The ultimate goal is to provide semantically and clinically relevant answer from the retrieval system. Particularly, we focused our research on improving the medical case retrieval on the bases of the information extracted from Magnetic Resonance Images (MRI) applied to Alzheimer’s disease. We represented the images with descriptors generated by using the domain knowledge to improve the semantic relevance, precision and efficiency. Considering the nature of the application domain, i.e. the fact that AD is a neurodegenerative disease that progresses over time, we also approached the problem in a longitudinal manner and we represented the medical cases by using longitudinal information. Additionally, we addressed the problem of incomplete data, which is the main challenge regarding the longitudinal data. Considering the limitations of the traditional approach for image representation, we investigated alternative approaches towards getting relevant information and optimal image description by using the domain knowledge. In that direction, we used imaging markers extracted from the regions of interest (ROI) that reflect the brain anatomy, such as volume of the brain structures and cortical thickness to generate the descriptor. Additionally, we explored a representation based on the special pattern of abnormality extracted from MRI. We also included the time component, i.e. we used longitudinal data to represent the patients’ cases. For that purpose, we evaluated feature vectors based on static and dynamic features. Descriptors based on the static measures contain a combination of the volumetric measures of the cortical and sub-cortical regions as well as cortical thickness, extracted from the available images acquired at multiple consecutive time points. On the other hand, dynamic measures, such as rate of change, percent change and symmetrized percent change reflect the severity of the disease and the advance of the degeneration. Additionally, we evaluated a representation comprised of a combination of static and dynamic measures. The experiments were performed with and without quality control (QC) to determine the influence of the errors caused by the automated processing to the results relevance. We also applied feature subset selection to reduce the feature vector dimensionality and to select the most relevant features. Considering the longitudinal data, a key problem that arises is incomplete data, i.e. lack of data for one or more time points. In this PhD thesis, we also proposed a possible solution for this problem. Namely, we suggested to represent the incomplete data with the dynamic features extracted from the available time points because they provide information about the disease progression. Additionally, this strategy provides the same dimension of the descriptors for all patients, regardless the number of the available time points. The proposed techniques were evaluated on a publicly available dataset, provided by the Alzheimer’s Disease Neuroimaging Initiative (ADNI). The evaluation of the proposed strategies showed improvements in comparison to the research already made in the area. Additionally, the obtained results provided answers to some key questions within this area. Thus, the strategies that we used for image representation during the retrieval process in this PhD thesis provided more precise and clinically more relevant, as well as more efficient image retrieval. In fact, the approaches proposed and used in this thesis, directed the retrieval from answering the question “find all cases with similar visual content” (characteristic for the traditional approaches) towards “find all cases with similar structural brain characteristics”, and even more “find all cases with similar brain changes”. This is very important for this research domain. Our future work is directed towards feature learning using deep neural networks. Additionally, we will explore this approach to address the incomplete data as well. Moreover, we will investigate in the direction of dealing with converters, i.e. patients who converted the diagnosis during the period of examination.
Some of the metrics are blocked by your
consent settings
Online Presence for Learning
(2012)
;
;
This paper focuses on presenting the objectives of the “Online Presence for Learning” (OP4L) project. The development of different kinds of services for the benefit of online learners based on their online presence and providing support for advanced, context-aware Learning Process Management (LPM) within Personal Learning Environments (PLEs) is the goal of the project. OP4L project uses an innovative technological approach. It offers new ways of combining Social Web presence data and Semantic Web technologies to provide advanced support for managing online learning resources and processes within personal learning environments (PLEs).
Some of the metrics are blocked by your
consent settings
Application of edge histogram descriptor and region shape descriptor to MRIs
(2009)
;
Davcev, Dancho
;
Medical images have a crucial place in the medical diagnosis. As a specific part of medicine, Magnetic Resonance Imaging (MRI) has become a useful modality since it provides plentiful information and high sensitivity. The rapid growth of MRI technology results in increasingly amount of MRIs which have to be efficiently stored, analyzed and described. By improving these processes, the decision making process for the clinicians is getting easier. In this paper we apply the Edge Histogram and the Region-based Shape descriptors standardized by MPEG-7 standard to MRIs. We have used 13917 MRIs provided by ImageCLEF 2008. Experimental results showed the Edge Histogram Descriptor achieves higher precision. We have also combined both descriptors by averaging their output ranks. Analysis showed that averaging the output ranks leads to higher precision. However, the obtained results and the fact that MRIs have specific characteristics suggest that additional image processing techniques have to be used.
Some of the metrics are blocked by your
consent settings
Hierarchical classification architectures applied to Magnetic Resonance Images
(IEEE, 2011-06-27)
;
;
Madjarov, Gjorgji
;
Gjorgjevikj, Dejan
The main goal of the paper is to explore hierarchical classification. The investigation is performed on the dataset of Magnetic Resonance Images (MRI) which is hierarchically organized. Generalized top-down hierarchical classification architecture is proposed in the paper. Additionally, two specific cases of the generalized architecture are explored: three-stage hierarchical architecture based on SVM and three-stage hierarchical architecture based on ANN. From the performed experiments, it is concluded that the SVM based scheme outperforms the ANN based scheme. Moreover, the gain of the investigation conducted in this paper becomes bigger with the possibilities given by the proposed generalized architecture for further investigations.
Some of the metrics are blocked by your
consent settings
Multi-query Content Based Medical Image Retrieval
(Faculty of Computer Science and Engineering, Ss. Cyril and Methodius University in Skopje, Macedonia, 2013)
;
Stojanova, Elena
;
;
This paper focuses on applying multi-query single-group methods to improve the content based image retrieval performance. The Multi-query-Max and Multi-query-Avg methods were applied using different numbers of query examples, namely three, five, and ten. The dataset contained medical images. The results obtained from the multi-query methods are compared to the single-query approach. The multi-query outperformed the single-query approach in all cases, meaning three, five, and ten queries based retrieval. Additionally, the Multi-query-Max method gives the best results on the bases of MAP (Mean Average Precision) value, when for the feature extraction purposes the Edge Histogram Descriptor (EHD) is used.
Some of the metrics are blocked by your
consent settings
Influence of Segmentation over Magnetic Resonance Image Classification
(2010-09)
;
;
Magnetic resonance imaging is an image based diagnostic technique which is widely used in medical environment. Thus, the efficient automated analysis of this kind of images is of great importance for both, scientific and clinical environment. In this paper, analysis of evaluation results of the classification of magnetic resonance images with different classifiers is conducted. This analysis is provided in both cases, with and without application of graph-based segmentation technique. The aim of the paper is to investigate whether or not this kind of segmentation technique induces improvements in the classification of MRIs. Seven descriptors are used for feature extraction in our paper, and the classification is analyzed in all seven cases. The ultimate goal of the paper is to signify in which combination of classification technique and feature extraction algorithm, the examined segmentation technique is the most appropriate for magnetic resonance images. For the overall investigation in this paper, a specific hierarchical organized dataset of magnetic resonance images is used.

Trojachanec, Katarina

Filters

Settings

Sort By

Results per page