Computer Science Journal (AGH University of Science and Technology, Krakow)
Not a member yet
476 research outputs found
Sort by
Tree-based Control Space Structures for Discrete Metric Sources in 3D Meshing
Get PDFThis article compares the different variations of the octree and kd-tree structures used to create a control space based on a set of discrete metric point-sources. The control space thus created supervises the generation of the mesh providing efficient access to the required information on the desired shape and size of the mesh elements at each point of the discretized domain. Structures are compared in terms of computational and memory complexity as well as regarding the accuracy of the approximation of the set of discrete metric sources in the created control space structure
A ROUTING ALGORITHM AND A ROUTER ARCHITECTURE FOR 3D NOC
Get PDFIn recent years, the enhancement of microchip technologies has enabled large scale Systems-on-Chip (SoC). Due to sharp increase in number of processing elements, SoC faces various challenges in design and testing. Network on Chip (NoC) is an alternative technology to overcome the challenges in SoC design and testing. NoC emerged as a key architecture that allows one to optimize the parameters like power and area. In spite of its applications, NoC faces some real time challenges like designing an optimum topology, routing scheme and application mappings. In this paper, we address the main three issues on NoC, namely, designing of an optimal topology, routing algorithm and a router design for the topology. First, we propose a topology and a routing algorithm. We prove that our recursive network topology is Hamiltonian connected and we propose an algorithm for data packet transmissions, which is free from cyclic deadlocks and the algorithm maximizes the congestion factor. Our experimental results show that the proposed topology gives better performance in terms of average latency and power than the other topologies. Finally, we propose a router architecture for our 3D-NoC. The router architecture is based on shared buffers. Also, our experimental results indicate that the proposed router architecture consumes less area and power than the Virtual Channel architecture
Computational intelligence for prediction of biological effects of drugs absorption in lungs
No full textLungs are recently extensively examined as a route for delivering drugs (active pharmaceutical ingredients, API) into the bloodstream mainly due to the possibility of noninvasive administration of macromolecules, such as proteins and peptides. Absorption mechanisms of chemical compounds in lungs are still not fully understood, what makes pulmonary formulation composition development challenging. The manuscript presents development of an empirical model capable of predicting the excipients’ influence on the absorption of the drug in lungs. Due to complexity of the problem and not fully understood mechanisms of absorption, computational intelligence tools were applied. As a result, a mathematical formula was established and analyzed. The normalized root-mean-squared error (NRMSE) and R2 of the model were 4.57%, and 0.83, respectively. The presented approach is beneficial both practically by developing \textit{in sillico} predictive model and theoretically by gaining knowledge about the influence of API and excipient structure on absorption in lungs
An approach to classification of data with highly localized unmarked features using neural networks
No full textTo face the increasing demand on quality healthcare, cutting edge automation technology is being applied in such demanding areas as medical imaging. This paper proposes a novel approach to classification problems on datasets with sparse, highly localized features. It is based on the use of saliency map in amplification of features. Unlike previous efforts, this approach does not use any prior information about feature localization. We present an experimental study based on Diabetic Retinopathy classification problem, in which our method has shown to achieve over 20\% higher accuracy in solving a two-class Diabetic Retinopathy classification problem than a naive approach based solely on residual neural networks. The dataset consists of 35120 images of various quality, inconsistent resolution and aspect ratio
Current research opportunities of image processing and computer vision
Get PDFImage processing and computer vision is an important and essential area in today’s scenario. Several problems can be solved through computer vision techniques. There are a large number of challenges and opportunities which require skills in the field of computer vision to address them. Computer vision applications cover each band of the electromagnetic spectrum and there are numerous applications in every band. This article is targeted to the research students, scholars and researchers who are interested to solve the problems in the field of image processing and computer vision. It addresses the opportunities and current trends of computer vision applications in all emerging domains. The research needs are identified through available literature survey and classified in the corresponding domains. The possible exemplary images are collected from the different repositories available for research and shown in this paper. The opportunities mentioned in this paper are explained through the images so that a naive researcher can understand it well before proceeding to solve the corresponding problems. The databases mentioned in this article could be useful for researchers who are interested in further solving the problem. The motivation of the article is to expose the current opportunities in the field of image processing and computer vision along with corresponding repositories. Interested researchers who are working in the field can choose a problem through this article and can get the experimental images through the cited references for working further.
Certificateless Digital Signature Technology for e-Governance Solutions
Get PDF. In spite of the fact that digital signing is an essential requirement for implementation of e-governance solutions in any organization, its use in large scale Government ICT implementation is negligible in India. In order to understand the reasons for low-level acceptance of the technology, authors performed a detailed study of a famous e-governance initiative of India. The outcome of the study revealed that the reasons are related to the challenges concerning the use of cryptographic devices carrying private key and the complicated process of generation, maintenance and disposal of Digital Signature Certificates (DSC).The solution, for the challenges understood from the case study, required implementation of a certificateless technology where private keys should be generated as and when required rather than storing them on cryptographic devices. Although many solutions which provide certificateless technology exist, to date there have been no practical implementation for using biometrics for implementing the solution. This paper presents the first realistic architecture to implement Identity Based Cryptography with biometrics using RSA algorithm. The solution presented in the paper is capable of providing a certificateless digital signature technology to the users, where public and private keys are generated on-the-fly
Installation and testing the LOFAR software on ACC Cyfronet cluster Prometheus
Get PDFWe present the actual status and the most important issues related to the installation of the data reduction LOFAR software on high power computer Prometheus located in ACC Cyfronet. We refer to the software itself as well its practical use cases in the context of the scientific tool and the detailed installation/testing methodology. We address most typical challenges and problems that occured during our attempts to set up the complete and ready-to-use LOFAR environment (including not only programs, but also libraries, scripts and other additional tools) on non-standard (cluster-type) computing system. The result of these works is then briefly summarized. We also discuss the issues related to LOFAR documentation, maintenance, distribution and further development. Finally, we propose some future improvements
Adapting Text Categorization for Manifest based Android Malware Detection
Get PDFThere are mainly three different approaches to detect malwares: i) static, ii) dynamic, and iii) hybrid. Static approach uses static source of the program without executing it. Dynamic approach, on the other hand, executes the program in a controlled environment and obtains information from operating system during runtime. Hybrid approach, as its name implies, is the combination of these two approaches. Although static approach may seem to have some disadvantages, it is highly preferred because of its lower cost. In this paper, we assume that obfuscated malware is processed by dynamic analysis and perform static malware detection based on text categorization methods. To reach our goal, we apply text mining techniques like feature extraction by using bag-of-words, n-grams, etc. from \texttt{manifest content} of programs to investigate the effectiveness of the malware detection. Our experimental results revealed that our approach is capable of detecting malicious applications with an accuracy between 94.0% and 99.3%
Analysis of data pre-processing methods for the sentiment analysis of reviews
Get PDFThe aim of this study is to analyse the effects of data pre-processing methods for sentiment analysis and determine which of these pre-processing methods and their combinations are effective for English and an agglutinative language like Turkish. We also try to answer the research question “is there any difference between agglutinative and non-agglutinative languages in terms of pre-processing methods for sentiment analysis?” We find that the performance results for the English reviews are generally higher than for the Turkish reviews related to the differences between the two languages in terms of vocabularies, writing styles, and agglutinative property of the Turkish language
Track Finding with Deep Neural Networks
Get PDFHigh Energy Physics experiments require fast and efficient methods toreconstruct the tracks of charged particles. Commonly used algorithms aresequential and the CPU required increases rapidly with a number of tracks.Neural networks can speed up the process due to their capability to modelcomplex non-linear data dependencies and finding all tracks in parallel.In this paper we describe the application of the Deep Neural Networkto the reconstruction of straight tracks in a toy two-dimensional model. It isplanned to apply this method to the experimental data taken by the MUonEexperiment at CERN