972 resultados para Stamp collecting
Resumo:
Die vorliegende Dissertation betrachtet institutionsinterne lokale (Critical-)Incident-Reporting-Systeme ((C)IRS) als eine Möglichkeit zum Lernen aus Fehlern und unerwünschten kritischen Ereignissen (sogenannte Incidents) im Krankenhaus. Die Notwendigkeit aus Incidents zu lernen, wird im Gesundheitswesen bereits seit den 1990er Jahren verstärkt diskutiert. Insbesondere risikoreichen Organisationen, in denen Incidents fatale Konsequenzen haben können, sollten umfassende Strategien erarbeiten, die sie vor Fehlern und unerwünschten Ereignissen schützen und diese als Lernpotenzial nutzen können. Dabei können lokale IRS als ein zentraler Bestandteil des Risikomanagements und freiwillige Dokumentationssysteme im Krankenhaus ein Teil dieser Strategie sein. Sie können eine Ausgangslage für die systematische Erfassung und Auswertung von individuellen Lerngelegenheiten und den Transfer zurück in die Organisation schaffen. Hierfür sind eine lernförderliche Gestaltung, Implementierung und Einbettung lokaler IRS eine wichtige Voraussetzung. Untersuchungen über geeignete lerntheoretisch fundierte und wirkungsvolle IRS-Modelle und empirische Daten fehlen bisher im deutschsprachigen Raum. Einen entsprechenden Beitrag leistet die vorliegende Fallstudie in einem Schweizer Universitätsspital (800 Betten, 6.100 Mitarbeitende). Zu diesem Zweck wurde zuerst ein Anforderungsprofil an lernförderliche IRS aus der Literatur abgeleitet. Dieses berücksichtigt zum einen literaturbasierte Kriterien für die Gestaltung und Nutzung aus der IRS-Literatur, zum anderen die aus der Erziehungswissenschaft und Arbeitspsychologie entlehnten Gestaltungsbedingungen und Erfolgskriterien an organisationales Lernen. Das Anforderungsprofil wurde in drei empirischen Teilstudien validiert und entsprechend adaptiert. In der ersten empirischen Teilstudie erfolgte eine Standortbestimmung der lokalen IRS. Die Erhebung erfolgte in vier Kliniken mittels Dokumentenanalyse, leitfadengestützter Interviews (N=18), sieben strukturierter Gruppendiskussionen und teilnehmender Beobachtungen über einen Zeitraum von 22 Monaten. Erfolgskritische IRS-Merkmale wurden identifiziert mit dem Ziel einer praxisgerechten lernförderlichen Systemgestaltung und Umsetzung von Incident Reporting unter Betrachtung von organisationalen Rahmenbedingungen, Lernpotenzialen und Barrieren. Die zweite Teilstudie untersuchte zwei Fallbeispiele organisationalen Lernens mittels Prozessbegleitung, welche zu einem verwechslungssicheren Design bei einem Medizinalprodukt und einer verbesserten Patientenidentifikation in Zusammenhang mit Blutentnahmen führten. Für das organisationale Lernen im Spital wurden dabei Chancen, Barrieren und Gestaltungsansätze abgeleitet, wie erwünschte Veränderungen und Lernen unter Nutzung von IRS initiiert werden können und dabei ein besseres Gesundheitsresultat erreicht werden kann. Die dritte Teilstudie überprüfte, inwiefern die Nutzung und Implementierung lokaler IRS mittels einer Mitarbeitervollbefragung zur Sicherheitskultur gefördert werden kann. Hierfür wurde eine positive Interaktion, zwischen einer starken Sicherheitskultur und der Bereitschaft ein IRS zu implementieren und Incidents zu berichten, angenommen. Zum Einsatz kam eine deutschsprachige Version des Hospital Survey on Patient Safety Culture (Patientensicherheitsklimainventar) mit einem Rücklauf von 46.8% (2.897 gültige Fragebogen). In 23 von 37 Kliniken führte laut einer Nachbefragung die Sicherheitskulturbefragung zum Implementierungsentscheid. Dies konnte durch Monitoring der IRS-Nutzung bestätigt werden. Erstmals liegen mit diesen Studien empirische Daten für eine wirkungsvolle und lernförderliche Gestaltung und Umsetzung von lokalen IRS am Beispiel einer Schweizer Gesundheitsorganisation vor. Die Ergebnisse der Arbeit zeigen Chancen und Barrieren für IRS als Berichts- und Lernsysteme im Krankenhaus auf. Als Resultat unsachgemäss gestalteter und implementierter IRS konnte dabei vor allem Lernverhinderung infolge IRS aufgezeigt werden. Blinder Aktionismus und eine fehlende Priorisierung von Patientensicherheit, unzureichende Kompetenzen, Qualifikationen und Ressourcen führten dabei zur Schaffung neuer Fehlerquellen mit einer Verstärkung des Lernens erster Ordnung. Eine lernförderliche Gestaltung und Unterhaltung der lokalen IRS, eingebettet in eine klinikumsweite Qualitäts- und Patientensicherheitsstrategie, erwiesen sich hingegen als wirkungsvoll im Sinne eines organisationalen Lernens und eines kontinuierlichen Verbesserungsprozesses. Patientensicherheitskulturbefragungen erwiesen sich zudem bei entsprechender Einbettung als effektives Instrument, um die Implementierung von IRS zu fördern. Zwölf Thesen zeigen in verdichteter Form auf, welche Gestaltungsprinzipien für IRS als Instrument des organisationalen Lernens im Rahmen des klinischen Risikomanagements und zur Förderung einer starken Patientensicherheitskultur zu berücksichtigen sind. Die Erkenntnisse aus den empirischen Studien münden in ein dialogorientiertes Rahmenmodell organisationalen Lernens unter Nutzung lokaler IRS. Die Arbeit zeigt damit zum einen Möglichkeiten für ein Lernen auf den verschiedenen Ebenen der Organisation auf und weist auf die Notwendigkeit einer (Re-)Strukturierung der aktuellen IRS-Diskussion hin.
Resumo:
Summary: Productivity, botanical composition and forage quality of legume-grass swards are important factors for successful arable farming in both organic and conventional farming systems. As these attributes can vary considerably within a field, a non-destructive method of detection while doing other tasks would facilitate a more targeted management of crops, forage and nutrients in the soil-plant-animal system. This study was undertaken to explore the potential of field spectral measurements for a non destructive prediction of dry matter (DM) yield, legume proportion in the sward, metabolizable energy (ME), ash content, crude protein (CP) and acid detergent fiber (ADF) of legume-grass mixtures. Two experiments were conducted in a greenhouse under controlled conditions which allowed collecting spectral measurements which were free from interferences such as wind, passing clouds and changing angles of solar irradiation. In a second step this initial investigation was evaluated in the field by a two year experiment with the same legume-grass swards. Several techniques for analysis of the hyperspectral data set were examined in this study: four vegetation indices (VIs): simple ratio (SR), normalized difference vegetation index (NDVI), enhanced vegetation index (EVI) and red edge position (REP), two-waveband reflectance ratios, modified partial least squares (MPLS) regression and stepwise multiple linear regression (SMLR). The results showed the potential of field spectroscopy and proved its usefulness for the prediction of DM yield, ash content and CP across a wide range of legume proportion and growth stage. In all investigations prediction accuracy of DM yield, ash content and CP could be improved by legume-specific calibrations which included mixtures and pure swards of perennial ryegrass and of the respective legume species. The comparison between the greenhouse and the field experiments showed that the interaction between spectral reflectance and weather conditions as well as incidence angle of light interfered with an accurate determination of DM yield. Further research is hence needed to improve the validity of spectral measurements in the field. Furthermore, the developed models should be tested on varying sites and vegetation periods to enhance the robustness and portability of the models to other environmental conditions.
Resumo:
Geographically, Taiwan is an Island and situated in the northeast of Asia, on the western side of the Pacific Basin, at the southeast of main China, south of Japan, and north of the Philippines. The main topographic character is the longitudinally oriented mountainous area. More than 200 peaks rise above 3000 m. They departed Taiwan into two lowland areas, an eastern and western plain. Taiwan is departed into subtropical (north area) and tropical zone (south area), which have a warm and humid climate, due to the Tropic of Cancer passing through. The average annual temperature in the lowland amounts to 28°C (7~38°C). The temperate climate also presents in the mountainous areas. The tropical typhoons usually come in summer and bring heavy rain, while the monsoon seasons have an important effect on the regional rainfall distribution. The mean annual rainfall of Taiwan is about 2600 mm (1000~6700 mm); the mountainous areas receive more rain than the lowlands. In Taiwan, according to different temperature and vegetation, the ecological environments were given rise to vertical biotic zonations, and form five major types: highland snowfield, highland meadow, coniferous forest, deciduous forest, and tropical forest. Six National Parks in Taiwan are located in the mountainous areas, in the north, the south, and on Jinmen Island. The National Parks represent about 8.4% of the country area. In this study, the collection sites are situated in Yangmingshan, Shei-Pa, Yushan, and Kenting National Park. Due to the island isolation, the proportions of endemic species are great in Taiwan, which also presents a high biodiversity. There are 4255 species of vascular plants including 1133 endemic. 5936 species in 1276 genera of fungi are hitherto reported in Taiwan. Among them, 233 Corticiaceae species were recorded, over one third (79 species) of them are known only from Taiwan. The first fungal report in Taiwan is about Phytophthora cyperi, published by the Japanese researcher T. Kawakami in 1904. Therefore, the history of research about fungi in Taiwan is more than one hundred years old. An eminent Japanese mycologist K. Sawada made an intensive survey from 1919 to 1959, and reported 2464 fungi species in his eleven volumes of “Descriptive Catalogue of Formosan Fungi”. However, only a few species (21 species in 9 genera) of Corticiaceae were recorded. From 1973, Chen and Lin resumed the study on Corticiaceae, and also some other foreign mycologists contributed for this field after 1980. The German research group lead by Franz Oberwinkler from Tübingen University collected in Taiwan several times. They published a number of new species and new records. Since 1989, S. H. Wu, a Taiwanese mycologist, has published a great amount of reports on corticioid fungi from Taiwan. Corticioid fungi were made up by the large and heterogeneous unnatural family Corticiaceae and other resupinate fungi belonging to other natural families in the Agaricomycetes. Molecular studies have shown that corticioid genera are distributed across all major clades of Agaricomycetes indicating that the corticioid fungi represent a polyphyletic group. They have resupinate fruitbodies and similar habitats. Species are characterized by simple fruitbody, more or less effused, and present smooth, porioid, grandinioid to odontioid hymenial surface. The fruitbodies are differently colored and usually soft to tough. Most of the Corticiaceae species are wood-saprobic organisms and gain the energy from the decomposing of wood-substrate such as cellulose or lignin. Materials for this study were collected by the author and other mycologists in Taiwan during surveys in April and May 1996, and March 2007, using the spring season with its high humidity and warm climate which are optimal conditions for the development of fungi. For assembling, the convenience sampling method was used in this study. This approach was chosen because it enables to detect a high biodiversity in a short time, and also to find species with rare or patchy distribution. The collecting sites from the North to the South include four National Parks and some preserved forests. They cover many different habitats such as low lands and high mountains. Fresh specimens were dried and analysed with a light microscope. 265 specimens belonging to Corticiaceae were studied in this research. Among them, 50 species in 21 genera including 11 new records and 10 new species were described with text and drawing. Four new species are belonging to Hyphodontia (H. sp. nov. 1, H. sp. nov. 2, H. sp. nov. 3, and H. sp. nov. 4), four to Schizopora (Sch. sp. nov. 1, Sch. sp. nov. 2, Sch. sp. nov. 3, and Sch. sp. nov. 4), one in Trechispora (T. sp. nov. 1), and one in Tubulicrinis (T. sp. nov. 1). Species recorded as new are Aleurodiscus amorphus, Botryohypochnus isabellinus, Hyphodontia cineracea, Hyphodontia palmae, Hypochnicium vellereum, Merulius tremellosus, Metulodontia nivea, Paullicorticium ansatum, Phlebia radiata, Phlebiella ardosiaca, and Xylobolus frustulatus. Besides, Botryohypochnus, Merulius, Metulodontia, Paullicorticium, and Xylobolus are also newly recorded genera in Taiwan. The genus Hyphodontia presents the highest diversity with 20 out of 50 species recorded. The second important genus is Hyphoderma, however with only 5 species. This indicates that Hyphodontia and Hyphoderma have a higher ability to develop in variable environments and approximately shows the predominance of these two genera in Taiwanese Corticiaceae. There are 11 new records out of the 50 species recorded, representing 22%. Some species, e.g. Hypochnicium vellereum and Paullicorticium ansatum were in the past recorded only in Europe and North America with cold and temperate climate. The samples of them are for the first time found in the subtropical belt, and display some difference from those of temperate regions. These collections should be molecularly investigated to clarify if they represent the same species of temperate areas. Patchily distributed species, for example Phlebiella ardosiaca, previously known only in Europe, and Hyphodontia palmae collected only in Brazil, were first recorded in different continents. Two possibilities are indicated by these new records: they are worldwide species but very rare to be found, or the Taiwanese specimens are taxonomically different. More survey from other continents and molecular study for these collections should be done in the future to solve this question. The distribution of Corticiaceae in Taiwan presents the variations in the north, central, and south areas and shows the diversity in lowlands and high mountains. The results of this study provide the evidence that the temperate Corticiaceae species displays a wider distribution. Subtropical and tropical taxa probably have also high dispersal capacities, and could possibly be found in the future in neighboring areas such as China, Japan, Korea or South Asia, but this needs further researches. In the total of 50 species, 10 new taxa were described in this study, giving about 20%. Some new species (e.g. Hyphodontia sp. 1, Hyphodontia sp. 2, and Hyphodontia sp. 3) are very similar to known species (Hyphodontia sambuci and Hyphodontia formosana), and the distinctive characters of Schizopora sp. nov. 1 are intermediate between those of Schizopora paradoxa and Hyphodontia flavipora. Thus, these small differences between the new and known species, suggest that the speciation occurred when the fungi migrated into Taiwan, due to the high diversity of environment, and amounts of the endemic plants. Taiwan is an intermediate place for the south (tropical) fungal species to migrate and adapt to north (temperate) regions. The middle and high altitude environments in Taiwan offer good conditions for the fungal speciation and possibly the occurrence of physiological changes to adapt to the temperate climate. Thus Taiwan has an important position for the biogeography of Asia mycobiota. 5936 known species in Taiwan represent about only 20% of the estimated number (24000) of Taiwanese fungal taxa. In this study, the findings (22% new records and 20% new species) indicate that amounts of unknown fungi species are expected in Taiwan. The lack of knowledge indicates that many new species are awaiting description, and fungal survey in Taiwan remains in a Pioneer phase. The last three wide surveys of Corticiaceae researches took place 20 years before this study (Chen & Lin 1977, Lin & Chen 1989, Wu 1990). After previous important contributions, the present taxonomic study comprising 21 genera is the most extensive on Corticiaceae of Taiwan.
Resumo:
Der Europäische Markt für ökologische Lebensmittel ist seit den 1990er Jahren stark gewachsen. Begünstigt wurde dies durch die Einführung der EU-Richtlinie 2092/91 zur Zertifizierung ökologischer Produkte und durch die Zahlung von Subventionen an umstellungswillige Landwirte. Diese Maßnahmen führten am Ende der 1990er Jahre für einige ökologische Produkte zu einem Überangebot auf europäischer Ebene. Die Verbrauchernachfrage stieg nicht in gleichem Maße wie das Angebot, und die Notwendigkeit für eine Verbesserung des Marktgleichgewichts wurde offensichtlich. Dieser Bedarf wurde im Jahr 2004 von der Europäischen Kommission im ersten „Europäischen Aktionsplan für ökologisch erzeugte Lebensmittel und den ökologischen Landbau“ formuliert. Als Voraussetzung für ein gleichmäßigeres Marktwachstum wird in diesem Aktionsplan die Schaffung eines transparenteren Marktes durch die Erhebung statistischer Daten über Produktion und Verbrauch ökologischer Produkte gefordert. Die Umsetzung dieses Aktionsplans ist jedoch bislang nicht befriedigend, da es auf EU-Ebene noch immer keine einheitliche Datenerfassung für den Öko-Sektor gibt. Ziel dieser Studie ist es, angemessene Methoden für die Erhebung, Verarbeitung und Analyse von Öko-Marktdaten zu finden. Geeignete Datenquellen werden identifiziert und es wird untersucht, wie die erhobenen Daten auf Plausibilität untersucht werden können. Hierzu wird ein umfangreicher Datensatz zum Öko-Markt analysiert, der im Rahmen des EU-Forschungsprojektes „Organic Marketing Initiatives and Rural Development” (OMIaRD) erhoben wurde und alle EU-15-Länder sowie Tschechien, Slowenien, Norwegen und die Schweiz abdeckt. Daten für folgende Öko-Produktgruppen werden untersucht: Getreide, Kartoffeln, Gemüse, Obst, Milch, Rindfleisch, Schaf- und Ziegenfleisch, Schweinefleisch, Geflügelfleisch und Eier. Ein zentraler Ansatz dieser Studie ist das Aufstellen von Öko-Versorgungsbilanzen, die einen zusammenfassenden Überblick von Angebot und Nachfrage der jeweiligen Produktgruppen liefern. Folgende Schlüsselvariablen werden untersucht: Öko-Produktion, Öko-Verkäufe, Öko-Verbrauch, Öko-Außenhandel, Öko-Erzeugerpreise und Öko-Verbraucherpreise. Zudem werden die Öko-Marktdaten in Relation zu den entsprechenden Zahlen für den Gesamtmarkt (öko plus konventionell) gesetzt, um die Bedeutung des Öko-Sektors auf Produkt- und Länderebene beurteilen zu können. Für die Datenerhebung werden Primär- und Sekundärforschung eingesetzt. Als Sekundärquellen werden Publikationen von Marktforschungsinstituten, Öko-Erzeugerverbänden und wissenschaftlichen Instituten ausgewertet. Empirische Daten zum Öko-Markt werden im Rahmen von umfangreichen Interviews mit Marktexperten in allen beteiligten Ländern erhoben. Die Daten werden mit Korrelations- und Regressionsanalysen untersucht, und es werden Hypothesen über vermutete Zusammenhänge zwischen Schlüsselvariablen des Öko-Marktes getestet. Die Datenbasis dieser Studie bezieht sich auf ein einzelnes Jahr und stellt damit einen Schnappschuss der Öko-Marktsituation der EU dar. Um die Marktakteure in die Lage zu versetzen, zukünftige Markttrends voraussagen zu können, wird der Aufbau eines EU-weiten Öko-Marktdaten-Erfassungssystems gefordert. Hierzu wird eine harmonisierte Datenerfassung in allen EU-Ländern gemäß einheitlicher Standards benötigt. Die Zusammenstellung der Marktdaten für den Öko-Sektor sollte kompatibel sein mit den Methoden und Variablen der bereits existierenden Eurostat-Datenbank für den gesamten Agrarmarkt (öko plus konventionell). Eine jährlich aktualisierte Öko-Markt-Datenbank würde die Transparenz des Öko-Marktes erhöhen und die zukünftige Entwicklung des Öko-Sektors erleichtern. ---------------------------
Resumo:
Internationalization of higher education has become one of the most important policies for institutions of higher education worldwide. Though universities are international by nature, the need for intensified quality activities of international nature has promoted internationalization to be under spotlight of researchers, administrators and policy makers and to be an area for research. Each institution follows its certain way to govern its international affairs. Most Universities, especially in the 'Developed World' started to plan it strategically. This study explores the meanings and importance of internationalization especially that it means different things to different people. It also studies the rationales behind internationalizing higher education. It focuses on the four main prevailing rationales; political, cultural/social, economic/financial, and academic on both national and institutional levels. With the increasing need to strategically plan, the study explores internationalization strategies in terms of how to develop them, what are their approaches and types, and their components and dimensions. Damascus University has witnessed an overwhelming development of its international relations and activities. Therefore, it started to face a problem of how to deal with this increasing load especially that its International Office is the only unit that deals with the international issues. In order to study the internationalization phenomenon at Damascus University, the 2WH approach, which asks the what, why, and how questions, is used and in order to define the International Office's role in the internationalization process of the University, it studies it and the international offices of Kassel University, and Humboldt University in Germany, The University of Jordan, and Al Baath University in Syria using the 'SOCIAL' approach that studies and analyses the situation, organization, challenges, involvement, ambitions, and limitations of these offices. The internationalization process at the above-mentioned Universities is studied and compared in terms of its meaning, rationales for both the institution and its academic staff, challenges and strategic planning. Then a comparison is made among the international offices of the Universities to identify their approaches, what led to their success and what led to their failure in their practices. The aim is to provide Damascus University and its International Office with some good practices and, depending on the experiences of the professionals of the case-studies, a suggested guidance to the work of this Office and the University in general is given. The study uses the interviews with the different officials and stakeholders of the case-studies as the main method of collecting the information in addition to site visits, studying their official documents and their websites. The study belongs to qualitative research that has an action dimension in it since the recommendations will be applied in the International Office. The study concludes with few learned lessons for Damascus University and its International Office depending on the comparison that was done according to a set of dimensions. Finally a reflection on the relationship between internationalization of higher education and politics, the impact of politics on Middle Eastern Universities, and institutional internationalization strategies are presented.
Resumo:
Proposes a behavior-based scheme for high-level control of autonomous underwater vehicles (AUVs). Two main characteristics can be highlighted in the control scheme. Behavior coordination is done through a hybrid methodology, which takes in advantages of the robustness and modularity in competitive approaches, as well as optimized trajectories
Resumo:
In order to successfully deploy multicast services in QoS-aware networks, pricing architectures must take into account the particular characteristics of multicast sessions. With this objective, we propose a charging scheme for QoS multicast services, assuming that the unicast cost of each interconnecting link is determined and that such cost is expressed in terms of quality of service (QoS) parameters. Our scheme allows determining the cost distribution of a multicast session along a cost distribution tree (CDT), and basing such distribution in those pre-existing unicast cost functions. The paper discusses in detail the main characteristics of the problem in a realistic interdomain scenario and how the proposed scheme would contribute to its solution
Resumo:
The estimation of camera egomotion is a well established problem in computer vision. Many approaches have been proposed based on both the discrete and the differential epipolar constraint. The discrete case is mainly used in self-calibrated stereoscopic systems, whereas the differential case deals with a unique moving camera. The article surveys several methods for mobile robot egomotion estimation covering more than 0.5 million samples using synthetic data. Results from real data are also given
Resumo:
A technique for simultaneous localisation and mapping (SLAM) for large scale scenarios is presented. This solution is based on the use of independent submaps of a limited size to map large areas. In addition, a global stochastic map, containing the links between adjacent submaps, is built. The information in both levels is corrected every time a loop is closed: local maps are updated with the information from overlapping maps, and the global stochastic map is optimised by means of constrained minimisation
Resumo:
Our work is focused on alleviating the workload for designers of adaptive courses on the complexity task of authoring adaptive learning designs adjusted to specific user characteristics and the user context. We propose an adaptation platform that consists in a set of intelligent agents where each agent carries out an independent adaptation task. The agents apply machine learning techniques to support the user modelling for the adaptation process
Resumo:
In the context of the round table the following topics related to image colour processing will be discussed: historical point of view. Studies of Aguilonius, Gerritsen, Newton and Maxwell. CIE standard (Commission International de lpsilaEclaraige). Colour models. RGB, HIS, etc. Colour segmentation based on HSI model. Industrial applications. Summary and discussion. At the end, video images showing the robustness of colour in front of B/W images will be presented
Resumo:
We propose a probabilistic object classifier for outdoor scene analysis as a first step in solving the problem of scene context generation. The method begins with a top-down control, which uses the previously learned models (appearance and absolute location) to obtain an initial pixel-level classification. This information provides us the core of objects, which is used to acquire a more accurate object model. Therefore, their growing by specific active regions allows us to obtain an accurate recognition of known regions. Next, a stage of general segmentation provides the segmentation of unknown regions by a bottom-strategy. Finally, the last stage tries to perform a region fusion of known and unknown segmented objects. The result is both a segmentation of the image and a recognition of each segment as a given object class or as an unknown segmented object. Furthermore, experimental results are shown and evaluated to prove the validity of our proposal
Resumo:
We present a new approach to model and classify breast parenchymal tissue. Given a mammogram, first, we will discover the distribution of the different tissue densities in an unsupervised manner, and second, we will use this tissue distribution to perform the classification. We achieve this using a classifier based on local descriptors and probabilistic Latent Semantic Analysis (pLSA), a generative model from the statistical text literature. We studied the influence of different descriptors like texture and SIFT features at the classification stage showing that textons outperform SIFT in all cases. Moreover we demonstrate that pLSA automatically extracts meaningful latent aspects generating a compact tissue representation based on their densities, useful for discriminating on mammogram classification. We show the results of tissue classification over the MIAS and DDSM datasets. We compare our method with approaches that classified these same datasets showing a better performance of our proposal
Resumo:
Given a set of images of scenes containing different object categories (e.g. grass, roads) our objective is to discover these objects in each image, and to use this object occurrences to perform a scene classification (e.g. beach scene, mountain scene). We achieve this by using a supervised learning algorithm able to learn with few images to facilitate the user task. We use a probabilistic model to recognise the objects and further we classify the scene based on their object occurrences. Experimental results are shown and evaluated to prove the validity of our proposal. Object recognition performance is compared to the approaches of He et al. (2004) and Marti et al. (2001) using their own datasets. Furthermore an unsupervised method is implemented in order to evaluate the advantages and disadvantages of our supervised classification approach versus an unsupervised one
Resumo:
We investigate whether dimensionality reduction using a latent generative model is beneficial for the task of weakly supervised scene classification. In detail, we are given a set of labeled images of scenes (for example, coast, forest, city, river, etc.), and our objective is to classify a new image into one of these categories. Our approach consists of first discovering latent ";topics"; using probabilistic Latent Semantic Analysis (pLSA), a generative model from the statistical text literature here applied to a bag of visual words representation for each image, and subsequently, training a multiway classifier on the topic distribution vector for each image. We compare this approach to that of representing each image by a bag of visual words vector directly and training a multiway classifier on these vectors. To this end, we introduce a novel vocabulary using dense color SIFT descriptors and then investigate the classification performance under changes in the size of the visual vocabulary, the number of latent topics learned, and the type of discriminative classifier used (k-nearest neighbor or SVM). We achieve superior classification performance to recent publications that have used a bag of visual word representation, in all cases, using the authors' own data sets and testing protocols. We also investigate the gain in adding spatial information. We show applications to image retrieval with relevance feedback and to scene classification in videos