Preferencje help
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 25

Liczba wyników na stronie
first rewind previous Strona / 2 next fast forward last
Wyniki wyszukiwania
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 2 next fast forward last
1
Content available remote Rule extraction from active contour classifiers
EN
In this paper, the idea of rule extraciton from active contour classifiers is presented. The concepts are new in relation to active contour approach. The problem is illustrated by examples having roots in technical diagnosis and in analysis of content of images.
EN
A paper web produced under industrial conditions always exhibits an anisotropy in its mechanical properties measured in the machine and cross directions. Proper and fast information about current paper anisotropy would significantly increase the efficiency of web quality control systems and allow to produce paper with greater precision. The main objective of the work presented was to determine the possibility of measurement of paper anisotropy. The method proposed is based on the image analysis of the orientation and location of special markers – luminescent fibres – which were introduced to the structure of the paper. Studies have shown that the method of image analysis provides a detection of fibres with a satisfactory level of efficiency, thus allowing accurate examination of the anisotropy of the paper samples analyzed. On the basis of experiments, it was also found that the method of image analysis proposed allows to detect fibres with an accuracy appropriate for the problem under consideration. The invention presented has already been registered in a Patent Office as submission no. P.411294.
PL
Celem pracy było określenie możliwości wyznaczenia zorientowania włókien celulozowych w papierze co pozwoliłoby określić anizotropię wytwarzanego materiału. Parametr ten jest niezwykle istotny w technologii papieru ze względu na ścisłą zależność między anizotropią a właściwościami mechanicznymi wytworów papierowych. Pomiar orientacji włókien realizowano drogą analizy obrazu, gdzie analizowanymi obiektami były wprowadzane do struktury papieru znaczniki – sztuczne włókna celulozowe o właściwościach luminescencyjnych. Cechą charakterystyczną tych włókien były ich właściwości zbliżone do właściwości naturalnych włókien celulozowych. Dzięki temu włókna luminescencyjne nie zmieniały właściwości papieru a przy tym zachowywały się w sposób podobny do pozostałych włókien występujących w masie papierniczej. Efekt luminescencji pojawiał się w chwili oświetlenia tych włókien światłem UV, co pozwalało na uzyskanie wyraźnego obrazu badanych obiektów. Na podstawie uzyskanych wyników stwierdzono, że zaproponowane: metoda pomiarowa oraz zastosowany algorytm analizy obrazu umożliwiły określenie zorientowania włókien w badanych strukturach papierniczych oraz pozwoliły na wyznaczenie anizotropii tych struktur. Wyniki zostały potwierdzone laboratoryjnymi badaniami wytrzymałościowymi badanych papierów. Na tej podstawie zaproponowano również sposób realizacji układu pomiarowego dla maszyny papierniczej, w którym byłaby zastosowana ta metoda. Rozwiązanie zostało już zgłoszone w Urzędzie Patentowym RP i zarejestrowane pod numerem P.411294.
3
Content available remote Neural Models of Demands for Electricity - Prediction and Risk Assessment
EN
Two neural systems for forecasting the electricity demand by the group of retail consumers are presented along with two methods for risk assessment of demand prediction models. The first forecasting system is composed of series-connected local neural predictors in the form of multilayer perceptron (MLP) networks. The system is mainly formed on the basis of expert knowledge and statistical tests. The second forecasting system has two levels. The first contains a neural classifier and the second consists of a set of local neural predictors. The classifier is built on the basis of a self-organising neural network (SOM). MLP or radial basis function (RBF) networks are used as predictors. Finally, two methods for assessing the risk of forecasting models are proposed. These consider financial risk measures such as value at risk (VaR) and conditional value at risk (CVaR). Possible economic losses posed by the application of predictions from a forecasting model are calculated using these risk measures. The risk analysis facilitates the selection of the forecasting model that generates the smallest risk of losses when selling energy contracts. The proposed methods are tested using data from the Polish electricity market.
PL
W pracy zostały przedstawione dwa neuronowe systemy przeznaczone do prognozowania zapotrzebowania na energię elektryczną grupy konsumentów detalicznych. Ponadto zaprenzetowano dwie metody oceny ryzyka modeli prognozowania.
EN
To a large extent, the chromatographic data obtained by measurements on power transformers reflect the state of a power transformer and allow the assessment of possible faults. The distribution of real learning data is not even approximately uniform and makes the partitioning of decision space difficult. The purpose of this paper is to present the results of the application of an EC-based classifier and a number of novel methods.
PL
Jak wiadomo, wyniki analizy chromatograficznej gazów rozpuszczonych w oleju transformatorowym (Dissofoed Gas Analysis - DGA) mogą być użyte do diagnostyki transformatorów. Zwykle rozmieszczenie tych danych (ściśle ilorazów koncentracji wybranych gazów) w przestrzeni jest bardzo nierównomierne, a ponadto jednoznaczny podział tej przestrzeni na obszary decyzyjne o rozsądnej wielkości i liczbie jest bardzo trudny. Celem pracy jest dokonanie przeglądu zastosowań nowych metod, wśród nich tych mających korzenie w obliczeniach inteligentnych i odniesienie się do wyników uzyskiwanych za pomocą standardu EC (International Electrotechnical Commission).
EN
Active region models are methods for automatic image segmentation. The models are able to detect shapes of irregular borders. In the present paper, the method is examined using medical images of liver changed locally by cancer cells.
EN
In the paper, the roles of intelligence, knowledge, learning and wisdom are discussed in the context of image content understanding. The known model of automatic image understanding is extended by the role of learning. References to example implementations are also given.
7
Content available remote Spatch Based Active Partitions with Linguistically Formulated Energy
EN
The present paper shows the method of cognitive hierarchical active partitions that can be applied to creation of automatic image understanding systems. The approach, which stems from active contours techniques, allows one to use not only the knowledge contained in an image, but also any additional expert knowledge. Special emphasis is put on the effcient way of knowledge retrieval, which could minimise the necessity to render information expressed in a natural language into a description convenient for recognition algorithms and machine learning.
EN
Medical Image Understanding is a recently defined semantic oriented image recognition task. Its specific requirements, highlighting complex characteristics of recognised objects as well as indispensable use of human-level expert knowledge almost every step of data processing sets new requirements for implemented algorithms. This paper focuses on linguistic image description method, designed to segment low level, semantically coherent image regions and mine adjacency relations among them. Example method results on medical images are presented to specify some methods properties.
9
Content available remote Notes on a linguistic description as the basis for automatic image understanding
EN
The main paradigm of image understanding and a concept for its practical machine realisation are presented. The crucial elements of the presented approach are the formalisation of human knowledge about the class of images that are to be automatically interpreted, a linguistic description and the realization of cognitive resonance.
EN
Potential contours are methods for automatic image analysis. In the present paper, potential contours adapted in the supervised way are used for segmentation of disjoint objects and examined using medical images.
EN
In the paper, selected visualization methods are described such as: surface rendering, volume rendering and the simplest approach basing on texture mapping. This work, however, does not aim at a detailed description of those methods but at a friendly presentation of possible ways of visualization and their features which can be of importance for a physician who has some expectations or needs, such as gaining insight into a 3-D object. Those expectations can be satisfied for example by making incisions or using surface transparency option which significantly improves visualization effect and can be of use in diagnostic process.
EN
Medical objects are based on different formats, where graphics are the most representative. They provide us with tremendous volumes of the data, very troublesome in management processes. For simplifying these processes, many works for simple and fast descriptors finding were undertaken. Fashionable platform for solving these problems one can be found in XML environment. This technology enables ordering and simple organisation of the database, with friendly linguistic description of the discussed items.
13
Content available remote Price Prediction of the Electric Energy - Regression versus Neural Approach
EN
In this paper, two models for price prediction of the day ahead market are presented and evaluated. The work consists of two parts. The first part includes short description the day ahead market of electric energy exchange. In the second one, the regression and neural models applied. As an example, the polish power exchange market is used.
14
Content available remote Interval-valued linguistic summaries of databases
EN
The so-called linguistic summaries of databases are the semi-natural language sentences that enable distilling the most relevant information from large numbers of tuples, and present it in the human consistent forms. Recently, the methods of constructing and evaluating linguistic summaries have been based on Zadeh's fuzzy sets, which represent uncertain data. The main aim of the paper is to enhance and generalize the Yager's approach to linguistic summarization of data. This enhancement is based on interval-valued fuzzy sets. The newly presented methods enable handling fuzzy concepts, whose membership degrees are not given by real values explicitly, but are approximated by intervals in [0,1]. From now on, the Yager's approach can be viewed as a special case of the method presented in this paper. Finally, illustrative examples are presented.
PL
W pracy przedstawiono możliwość zastosowania uczących się sieci logicznych (adaptative logic networks) do diagnostyki transformatorów w oparciu o wyniki analizy chromatograficznej rozpuszczonych w oleju gazów (Dissolved Gas Analysis - DGA). Zestawiono wyniki w postaci reguł logicznych uzyskanych tą metodą obliczeń inteligentnych (soft computing) z regułami zbudowanymi według międzynarodowego kodu IEC (International Electrotechnical Commision).
EN
In the paper, application of adaptive logic networks to the diagnosis of power transformers on (he basis dissolved gas analysis of is presented. The results in the form of logical rules obtained using the proposed method are compared to those given by the IEC code.
16
EN
In this paper, the problem of unified analysis of data and descriptions of objects is discussed. Basic concept for measure of similarity between features described in diverse manner is presented as well as the method for unification of different types of description.
EN
In the paper, evaluation of two approaches to modelling of hemodialysis is performed. Results obtained by regression are compared to those generated by neural models. Differences in the modelling quality are small. Both models shown the same qualitative dependencies between analyzed parameters.
EN
Dealing with the problem of co-ordination of subjects involved in kidney transplantation the paper displays new possibilities of using complex information systems for integration of dialysis centres and specialised health care units. Growing requirements of information system customers and development of object modeling methods stimulate creation of complex information systems designed to assist the work of medical staff. In the system presented in the paper information needs of customers are modeled by means of UML diagrams and the system itself involves a number of computer technologies, like inference in uncertainty conditions, artificial intelligence algorithms, knowledge bases, distributed processing, and Internet techniques. The system implementation covers numerous functional modules which are necessary to ensure efficient assistance of medical staff in their work, to improve the quality of patients’ life, and to carry out research in the fields of medicine and computer science.
EN
The paper presents a general review of two approaches to modeling of urea concentration in serum after hemodialysis. The first approach is a well-established classic method, widely accepted by medical staff dealing with practical nephrology. The other one utilises the power of artificial neural networks, which is a novel application to that field. Unlike classic models that base on theoretical investigation, this technique uses only a number of data from previous treatment to construct a model through so called learning from examples.
EN
The paper focuses on textual semi-structured data processing and mining. An original composition of a simple method of textual data mining [1] and Yager's linguistic summaries of databased [2] proposed in the paper makes it possible to summarize not only numerical or string data, but – even and especially – the textual noncrisp and semi-structured information as well. As a result of application and implementation of the presented method to a real medical database a user-friendly and easy-to-operate system is achieved.
first rewind previous Strona / 2 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.