Robertas Damasevicius - Academia.edu (original) (raw)

Papers by Robertas Damasevicius

Cancers

Laryngeal carcinoma is the most common malignant tumor of the upper respiratory tract. Total lary... more Laryngeal carcinoma is the most common malignant tumor of the upper respiratory tract. Total laryngectomy provides complete and permanent detachment of the upper and lower airways that causes the loss of voice, leading to a patient’s inability to verbally communicate in the postoperative period. This paper aims to exploit modern areas of deep learning research to objectively classify, extract and measure the substitution voicing after laryngeal oncosurgery from the audio signal. We propose using well-known convolutional neural networks (CNNs) applied for image classification for the analysis of voice audio signal. Our approach takes an input of Mel-frequency spectrogram (MFCC) as an input of deep neural network architecture. A database of digital speech recordings of 367 male subjects (279 normal speech samples and 88 pathological speech samples) was used. Our approach has shown the best true-positive rate of any of the compared state-of-the-art approaches, achieving an overall accu...

Sensors

The identification of human activities from videos is important for many applications. For such a... more The identification of human activities from videos is important for many applications. For such a task, three-dimensional (3D) depth images or image sequences (videos) can be used, which represent the positioning information of the objects in a 3D scene obtained from depth sensors. This paper presents a framework to create foreground–background masks from depth images for human body segmentation. The framework can be used to speed up the manual depth image annotation process with no semantics known beforehand and can apply segmentation using a performant algorithm while the user only adjusts the parameters, or corrects the automatic segmentation results, or gives it hints by drawing a boundary of the desired object. The approach has been tested using two different datasets with a human in a real-world closed environment. The solution has provided promising results in terms of reducing the manual segmentation time from the perspective of the processing time as well as the human input...

Informatics

Low back pain is a leading cause of disability worldwide, putting a significant strain on individ... more Low back pain is a leading cause of disability worldwide, putting a significant strain on individual sufferers, their families, and the economy as a whole. It has a significant economic impact on the global economy because of the costs associated with healthcare, lost productivity, activity limitation, and work absence. Self-management, education, and adopting healthy lifestyle behaviors, such as increasing physical activity, are all widely recommended treatments. Access to services provided by healthcare professionals who provide these treatments can be limited and costly. This evaluation study focuses on the application of the MyRelief serious game, with the goal of addressing such challenges by providing an accessible, interactive, and fun platform that incorporates self-management, behavior change strategies, and educational information consistent with recommendations for managing low-back pain, based on self-assessment models implemented through ontology-based mechanics. Functi...

Journal of Personalized Medicine

Currently, most mask extraction techniques are based on convolutional neural networks (CNNs). How... more Currently, most mask extraction techniques are based on convolutional neural networks (CNNs). However, there are still numerous problems that mask extraction techniques need to solve. Thus, the most advanced methods to deploy artificial intelligence (AI) techniques are necessary. The use of cooperative agents in mask extraction increases the efficiency of automatic image segmentation. Hence, we introduce a new mask extraction method that is based on multi-agent deep reinforcement learning (DRL) to minimize the long-term manual mask extraction and to enhance medical image segmentation frameworks. A DRL-based method is introduced to deal with mask extraction issues. This new method utilizes a modified version of the Deep Q-Network to enable the mask detector to select masks from the image studied. Based on COVID-19 computed tomography (CT) images, we used DRL mask extraction-based techniques to extract visual features of COVID-19 infected areas and provide an accurate clinical diagnos...

Advances in Intelligent Systems and Computing, 2018

The paper describes the results of experiments in applying the Random Projection (RP) method for ... more The paper describes the results of experiments in applying the Random Projection (RP) method for authorship identification of online texts. We propose using RP for feature dimensionality reduction to low-dimensional feature subspace combined with probability density function (PDF) estimation for identification of the features of each author. In our experiments, we use the dataset of Internet comments posted on a web news site in Lithuanian language, and we have achieved 92% accuracy of author identification.

2020 International Conference on Intelligent Computing and Human-Computer Interaction (ICHCI), 2020

Computer numerical simulation is used to study the propagation law of seismic waves in undergroun... more Computer numerical simulation is used to study the propagation law of seismic waves in underground media, which is one of the most practical and convenient research methods in the field of seismic exploration in modern geophysics. Due to the many anisotropy parameters of elastic wave, the calculation in the simulation process is large and time-consuming. This paper propose the idea of using acoustic waves for simulation. Acoustic anisotropy theory sets the velocity of the shear wave part of the elastic wave to zero and eliminates many of its parameters. This paper propose the idea of staggered mesh finite difference to discretize the acoustic wave equation. The purpose is to convert the differential equation into a difference equation. The homogeneous medium model and layered medium model are used in the forward modeling process. This paper use the perfectly matched layer(PML) boundary absorption algorithm to process the wave propagating to the model boundary. The experimental results show that our algorithm is correct and of great significance to the study of seismic wave propagation in earth media.

Applied Sciences, 2022

Breast cancer is a major research area in the medical image analysis field; it is a dangerous dis... more Breast cancer is a major research area in the medical image analysis field; it is a dangerous disease and a major cause of death among women. Early and accurate diagnosis of breast cancer based on digital mammograms can enhance disease detection accuracy. Medical imagery must be detected, segmented, and classified for computer-aided diagnosis (CAD) systems to help the radiologists for accurate diagnosis of breast lesions. Therefore, an accurate breast cancer detection and classification approach is proposed for screening of mammograms. In this paper, we present a deep learning system that can identify breast cancer in mammogram screening images using an “end-to-end” training strategy that efficiently uses mammography images for computer-aided breast cancer recognition in the early stages. First, the proposed approach implements the modified contrast enhancement method in order to refine the detail of edges from the source mammogram images. Next, the transferable texture convolutiona...

Computing and Informatics, 2021

Sensors, 2022

Current research endeavors in the application of artificial intelligence (AI) methods in the diag... more Current research endeavors in the application of artificial intelligence (AI) methods in the diagnosis of the COVID-19 disease has proven indispensable with very promising results. Despite these promising results, there are still limitations in real-time detection of COVID-19 using reverse transcription polymerase chain reaction (RT-PCR) test data, such as limited datasets, imbalance classes, a high misclassification rate of models, and the need for specialized research in identifying the best features and thus improving prediction rates. This study aims to investigate and apply the ensemble learning approach to develop prediction models for effective detection of COVID-19 using routine laboratory blood test results. Hence, an ensemble machine learning-based COVID-19 detection system is presented, aiming to aid clinicians to diagnose this virus effectively. The experiment was conducted using custom convolutional neural network (CNN) models as a first-stage classifier and 15 supervis...

Educational robots can serve as smart, mobile and tangible learning objects which can explicitly ... more Educational robots can serve as smart, mobile and tangible learning objects which can explicitly represent knowledge through actions and engage students through immersion and instant feedback. As pedagogical background, we employ the elements of Norman's foundational theory of action and the Internet-of-Things Supported Collaborative Learning (IoTSCL) paradigm, which is based on constructivism. We describe the application of educational robots in project-based teaching at the university course. Positive relationship between successful implementation of robotics project and assimilation of theoretical knowledge has been established.

Higher-level programming such as metaprogramming introduces a layer of abstraction above the doma... more Higher-level programming such as metaprogramming introduces a layer of abstraction above the domain language programs. Metaprogramming allows describing generic components and managing variability in a domain. It is especially useful for developing program generators for domains, where a great deal of commonalties exists. It allows increasing the level of abstraction and hiding details that are unnecessary to the designer. Information abstraction and hiding reduces the amount of "user-visible" information. In this paper, we estimate the increase of abstraction by evaluating the information content at the lower (domain) and higher (meta) layers of abstraction. The estimation method is based on the Kolmogorov complexity and uses a common compression algorithm. The method is evaluated experimentally on families of DSP components.

Computational Science and Its Applications – ICCSA 2017, 2017

We explore the relationship between attention and action, and focus on human reaction to stress i... more We explore the relationship between attention and action, and focus on human reaction to stress in the Supervisory Control and Data Acquisition (SCADA) based Human Computer Interface (HCI) environment aiming to measure the reaction time and warn against attention deficit. To provoke human reaction we simulate several provocative situations mimicking real-world accidents while working on the industrial production line. During the simulation of the industrial line control, the subjects are presented on screen with affective visual stimuli imitating the possible accident and the reaction of subjects is tracked with a gaze tracker. We measure a subjects’ response time from stimuli onset to the eye fixation (gaze time) and to the pressing of “line stop” button (press time). The reaction time patterns are analysed with respect to subject’s gender, age, colour and position of stop sign. The results confirm the significance of gender, age, sign colour and position factors.

Computers, 2021

Despite the attractive benefits of cloud-based business processes, security issues, cloud attacks... more Despite the attractive benefits of cloud-based business processes, security issues, cloud attacks, and privacy are some of the challenges that prevent many organizations from using this technology. This review seeks to know the level of integration of security risk management process at each phase of the Business Process Life Cycle (BPLC) for securing cloud-based business processes; usage of an existing risk analysis technique as the basis of risk assessment model, usage of security risk standard, and the classification of cloud security risks in a cloud-based business process. In light of these objectives, this study presented an exhaustive review of the current state-of-the-art methodology for managing cloud-based business process security risk. Eleven electronic databases (ACM, IEEE, Science Direct, Google Scholar, Springer, Wiley, Taylor and Francis, IEEE cloud computing Conference, ICSE conference, COMPSAC conference, ICCSA conference, Computer Standards and Interfaces Journal)...

Pattern Recognition Letters, 2021

Informatica, 2002

In this paper, we analyze the abstractions used for developing component-based domain generators.... more In this paper, we analyze the abstractions used for developing component-based domain generators. These include programming paradigms, programming languages, component models, and generator architecture models. On the basis of the analysis, we present a unified relationship model between the domain content, technological factors (structuring, composition, and generalization), and domain architecture. We argue that this model is manifested in the known software generator models, too.

Informatica, 2000

In this paper we suggest a three-language (3L) paradigm for building the program gen- erator mode... more In this paper we suggest a three-language (3L) paradigm for building the program gen- erator models. The basis of the paradigm is a relationship modelof the specification, scripting and target languages. It is not necessary that all three languages would be the separate ones. We con- sider some internal relationship (roles) between the capabilities of a given la nguage for specifying, scripting (gluing) and describing the domain functionality. We also assume that a target la nguage is basic. We introduce domain architecture (functionality) with the generic componentsusually com- posed using the scripting and target languages. The specification language is for describing user's needs for the domain functionality to be extracted from the system. We present the framework for implementing the 3L paradigm and some results from the experimental systems developed for a validation of the approach.

Informatica, 2000

Data, Engineering and Applications, 2019

Crime and criminal activities have impacted negatively on the socioeconomic development of Nigeri... more Crime and criminal activities have impacted negatively on the socioeconomic development of Nigeria over the years. As a result, life and property of citizens are threatened continually. Putting in place a secured and fear-free environment is thus a critical role of any government. In recent times, Information and Communication Technologies are being leveraged to overcome this challenge. In particular, the advent of Internet of things shows great potential in this regard. Therefore, this study proposes a framework of Internet of things and Big Data technologies to track and monitor crime and criminalities real-time online in order to reduce crime rate in Nigeria. A prototype of the framework is also developed.

Recognition of specific functionally-important DNA sequence fragments is one of the most importan... more Recognition of specific functionally-important DNA sequence fragments is one of the most important problems in bioinformatics. Common sequence analysis methods such as pattern search can not solve this problem because of noisy data and variability of consensus sequences across different species. Machine learning methods such as Support Vector Machine (SVM) can be used for sequence classification, because they can learn useful descriptions of genetic concepts from data instances only rather than explicit definitions. Before applying SVM classification one must define a mapping of classified sequences into a feature space. We analyze binary and kmer frequency-based feature mapping rules. The efficiency of such rules is demonstrated for the recognition of promoters and splice-junction sites.

Cancers

Sensors

Informatics

Journal of Personalized Medicine

Advances in Intelligent Systems and Computing, 2018

2020 International Conference on Intelligent Computing and Human-Computer Interaction (ICHCI), 2020

Applied Sciences, 2022

Computing and Informatics, 2021

Sensors, 2022

Computational Science and Its Applications – ICCSA 2017, 2017

Computers, 2021

Pattern Recognition Letters, 2021

Informatica, 2002

Informatica, 2000

Data, Engineering and Applications, 2019

Physiological computing is a paradigm of computing that treats users’ physiological data as input... more Physiological computing is a paradigm of computing that treats users’ physiological data as input during computing tasks in an Ambient Assisted Living (AAL) environment. By monitoring, analyzing and responding to such inputs, Physiological Computing Systems (PCS) are able to respond to the users’ cognitive, emotional and physical states. A specific case of PCS is Neural Computer Interface (NCI), which uses electrical signals governing users’ muscular activity (EMG data) to establish a direct communication pathway between the user and a computer. We present taxonomy of speller application parameters, propose a model of PCS, and describe the development of the EMG-based speller as a benchmark application. We analyze and develop an EMG-based speller application with a traditional letter-based as well as visual concept-based interface. Finally, we evaluate the performance and usability of the developed speller using empirical (accuracy, information transfer speed, input speed) metrics.