No description available.

Orientador: Sandra Regina Gimenez-Paschoal / Banca: Elisabete de Fátima Polo de Almeida Nunes / Banca: Olga Maria Piazentini Rolim Rodrigues / Banca: Maria Angélica Pagliarini Waidman / Banca: Maria de Lourdes Morales Horiguela / Resumo: Analisa comportamentos de segurança e de risco para acidente com material biológico (AMB) durante preparo e descarte de materiais utilizados na administração de medicações. Os participantes são 19 discentes de enfermagem, escolhidos aleatoriamente, que haviam terminado o módulo que aborda de forma teórica o tema biossegurança e realizavam estágio em pronto socorro, e sete docentes supervisores de estágio, todos de uma Universidade Pública do Estado do Paraná. Utiliza filmagem e observação ao vivo dos participantes durante o estágio curricular para identificar e registrar comportamentos e situações de risco e de segurança para AMB. Seleciona cenas das fitas gravadas e registros obtidos da observação ao vivo, os encaminha para juízes e calcula o índice de concordância. Utiliza questionário com discentes e docentes, agrupa os recortes das opiniões mais freqüentes em categorias e sub-categorias, submete à avaliação de juízes e calcula o índice de concordância entre os instrumentos. Entrega recortes de cenas e registros aos discentes e docentes para classificação quanto ao risco ou segurança para AMB, analisa e compara as classificações com aspectos ligados ao processo de ensino-aprendizagem. Verifica 2,86 vezes mais comportamentos de risco (109) para acidente com material biológico que de segurança (38). Identifica 38 comportamentos de segurança: 12 pela filmagem e 26 pela técnica de observação ao vivo, e 109 comportamentos de risco (76 pela filmagem e 33 pela observação ao vivo). Identifica pela observação ao vivo que: dos 30 comportamentos de risco apresentados por discentes, 20 (67%) relacionam-se ao preparo de medicações, estando o docente supervisionando a atividade em 13 (65%) situações; 26 (86,66%) comportamentos/situações envolvem o recipiente de descarte que não atende... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Safety and risky behaviors for accidents with biological material (ABM) are subjected to analysis during preparation and discard of materials used in administration of medication. The participants are 19 undergraduate nursing students, randomly chosen, who had finished the module about biosafety, approached theoretically, and were doing internship in the emergency room, and seven faculty internship supervisors, all from a public university in Paraná State, Brazil. Video recordings and live observation of the participants during internship are used to identify and register safety and risky behaviors for ABM. It selects footage and registers obtained from live observation, submitting them to judges and calculates the agreement rate. It carries out surveys among undergraduate students and faculty members, it groups clippings of most recurrent opinions in categories and subcategories, which is submitted to judges for evaluation and the agreement rate among the instruments is calculated. Footages and registers are given in to undergraduate students and teachers for classification regarding risk and safety in ABM, it analyses and compares the classifications connected to aspects of the teaching and learning process. It verifies 2,86 times more risky behaviors (109) for accidents with biological material than safety (38). It identifies 38 safety behaviors: 12 by the video recordings and 26 by the live observation technique, and 109 risky behaviors (76 by video recording and 33 by live observation). It identifies by live observation that: out of 30 risky behaviors presented by undergraduate nursing students, 20 (67%) are related to the preparation of medication, with the professor supervising the activity in 13 (65%) situations; 26 (86,66%) behaviors/situations involving the discarded container which does not comply with... (Complete abstract click electronic access below) / Doutor

PROBLÉMATIQUE: L’Organisation Mondiale de la Santé (OMS) considère les maladies cardiovasculaires (MCVs) comme l'hypertension, la maladie coronarienne (par exemple, infarctus du myocarde), l'insuffisance cardiaque ainsi que les accidents cérébrovasculaires, parmi les principales causes de mortalité dans le monde. Les MCVs sont des maladies multifactorielles caractérisées par des interactions complexes entre le génome et l'environnement et dont la prévalence augmente rapidement dans toutes les populations du globe, ce qui vient compliquer d'autant l'étude de leurs bases héréditaires. Nos études précédentes sur la population fondatrice des familles Canadiennes-françaises de la région du Saguenay-Lac-Saint-Jean (SLSJ) au Québec ont permis d’obtenir une carte des loci significativement liés à des déterminants qualitatifs et quantitatifs de l’hypertension et ses déterminants métaboliques [1, 2]. HYPOTHÈSE ET OBJECTIF: Puisque nos données préliminaires nous suggèrent que la mort prématurée consécutive aux MCVs possède des composantes génétique et environnementale, notre hypothèse de départ est que les maladies avec occurrences fatales et non fatales (OF et ONF, respectivement) ont des caractéristiques distinctes, surtout lorsqu’en lien avec le système CV. Pour réaliser ce projet, nos objectifs sont d’analyser les causes de morbidité/mortalité d’hypertendus avec ou sans obésité chez des familles de la région du SLSJ. Nous accomplirons ceci en interrogeant les registres des hôpitaux et de l'état civil de même que les données généalogiques de 1950 jusqu'à maintenant. Nous voulons décrire et étudier les OF pour les comparer aux NFO. RÉSULTATS: Nous avons identifié un total de 3,654 diagnostiques appartenant aux OF et ONF chez les 343 sujets étudiés. Pour les OF, nous avons trouvé que: (1) un grand total de 1,103 diagnostiques du système circulatoire ont affecté 299 sujets avec 555 occurrences et 247 premières occurrences; (2) 333 des sujets participants ont reçu 1,536 diagnostiques non-CV avec 195 occurrences et 107 premières occurrences; (3) 62 diagnostiques de toutes autres causes chez 62 des sujets participants avec 81 occurrences et 11 premières occurrences. Pour les ONF: (1) 156 diagnostiques du système circulatoire ont affecté 105 sujets; (2) 60 diagnostiques de causes non-CV chez 53 des sujets; (3) et 718 diagnostiques de toutes autres causes chez 252 des sujets. Pour les OF, 109 des 333 sujets affectés par les maladies non-CV et 58 des 62 par toutes autres maladies étaient atteints simultanément par des MCV. Nous avons décrit les caractéristiques des maladies avec occurrences fatales et non fatales. Les MCVs prédominaient dans les résultats des premières occurrences et occurrences totales tandis que les maladies non-CV étaient les plus élevées pour les diagnostiques. De plus, les OF CV ont affecté 67.1% de notre échantillon de population, incluant les sujets co-affectés par les maladies non-CV ou de toutes autres causes. En fait, nos sujets ont un risque trois fois plus élevé de développer des MCVs (p<0.0001; χ2=1,575.348), tandis qu’il diminue de moitié pour les maladies non-CV comparativement au reste de la population du SLSJ (p=0.0006; χ2=11.834). Enfin, le risque de développer des tumeurs malignes est diminué de moitié dans notre échantillon comparativement à l’incidence régionale. CONCLUSION: Cette étude a apporté une nouvelle perspective sur les OF et ONF chez nos sujets de la région SLSJ du Québec après 11 ans. Quand on observe ces résultats en conjonction avec les MCVs, ce risque double. / BACKGROUND: The World Health Organization (WHO) recognizes cardiovascular diseases (CVD), such as high blood pressure (hypertension; HBP), coronary heart disease (e.g. myocardial infarction), cerebrovascular disease (stroke) and heart failure, as the major cause of death globally. Furthermore, CVD are multifactorial traits with a steeply rising prevalence worldwide, determined by a complex interplay of interactions between genome and environment, which renders their hereditary analysis a complicated task. Our previous studies in the founder population of French-Canadian families from the Saguenay-Lac-Saint-Jean (SLSJ) region of Quebec revealed the largest set of significantly linked loci to hypertension and its metabolic components [1, 2]. HYPOTHESIS AND OBJECTIVE: Since our preliminary findings suggest that early mortality from CVD has genetic and environmental factors, we hypothesize that our population will have distinct characteristics from diseases with and without fatal outcomes (FO and NFO, respectfully), especially in relation to CV cause. Our objectives are to analyze morbidity causes, their temporal characteristics and their clustering in the hypertensive families with or without obesity from the SLSJ region, using hospital, civic death registries and genealogical records from 1950 to the present. We want to describe and study FO and compare them to NFO. RESULTS: In total, we identified 3,654 diagnosis from FO and NFO in 343 participating subjects. For FO, we report: (1) 299 participants suffered from diseases of the circulatory system with a grand total of 1,103 diagnosis, 555 outcomes and 247 first outcomes; (2) 333 of the participating subjects were affected by non-CV systems with a total of 1,536 diagnosis, 195 outcomes and 107 first outcomes; (3) all other diseases were responsible for a total of 81 diagnosis, 30 outcomes and 11 first outcomes in 62 of the participating subjects; and for NFO: (1) the circulatory system affected 105 of the participating subjects with a total of 156 diagnosis; (2) non-CVD, 53 of participating subjects with a total of 60 diagnosis; (3) all other diseases were found in 252 of participating subjects with a total of 718 diagnosis. For FO, 109 of 333 affected subjects by non-CVD and 58 of 62 by all other diseases had also a concomitant CVD. We were able to show characteristics from both FO and NFO. In both first outcomes and total outcomes, CVD predominated over non-CV and all other diseases. When examining CV co-affection with non-CV or all other diseases, 67.1% of our sample population was affected by CV FO. In fact, our sample showed a threefold risk increase in developing CVD (p<0.0001; χ2=1,575.348) when compared to the general population of SLSJ, while it diminished by half for non-CVD (p=0.0006; χ2=11.834). Finally, the relative risk for developing malignancies decreased by half in our sample in contrast to the same region. CONCLUSION: This 11 year follow-up study provided a unique insight into affections for both FO and NFO. When looking at malignancies in conjunction with CVD, this risk grew twofold in our sample.

PROBLÉMATIQUE: L’Organisation Mondiale de la Santé (OMS) considère les maladies cardiovasculaires (MCVs) comme l'hypertension, la maladie coronarienne (par exemple, infarctus du myocarde), l'insuffisance cardiaque ainsi que les accidents cérébrovasculaires, parmi les principales causes de mortalité dans le monde. Les MCVs sont des maladies multifactorielles caractérisées par des interactions complexes entre le génome et l'environnement et dont la prévalence augmente rapidement dans toutes les populations du globe, ce qui vient compliquer d'autant l'étude de leurs bases héréditaires. Nos études précédentes sur la population fondatrice des familles Canadiennes-françaises de la région du Saguenay-Lac-Saint-Jean (SLSJ) au Québec ont permis d’obtenir une carte des loci significativement liés à des déterminants qualitatifs et quantitatifs de l’hypertension et ses déterminants métaboliques [1, 2]. HYPOTHÈSE ET OBJECTIF: Puisque nos données préliminaires nous suggèrent que la mort prématurée consécutive aux MCVs possède des composantes génétique et environnementale, notre hypothèse de départ est que les maladies avec occurrences fatales et non fatales (OF et ONF, respectivement) ont des caractéristiques distinctes, surtout lorsqu’en lien avec le système CV. Pour réaliser ce projet, nos objectifs sont d’analyser les causes de morbidité/mortalité d’hypertendus avec ou sans obésité chez des familles de la région du SLSJ. Nous accomplirons ceci en interrogeant les registres des hôpitaux et de l'état civil de même que les données généalogiques de 1950 jusqu'à maintenant. Nous voulons décrire et étudier les OF pour les comparer aux NFO. RÉSULTATS: Nous avons identifié un total de 3,654 diagnostiques appartenant aux OF et ONF chez les 343 sujets étudiés. Pour les OF, nous avons trouvé que: (1) un grand total de 1,103 diagnostiques du système circulatoire ont affecté 299 sujets avec 555 occurrences et 247 premières occurrences; (2) 333 des sujets participants ont reçu 1,536 diagnostiques non-CV avec 195 occurrences et 107 premières occurrences; (3) 62 diagnostiques de toutes autres causes chez 62 des sujets participants avec 81 occurrences et 11 premières occurrences. Pour les ONF: (1) 156 diagnostiques du système circulatoire ont affecté 105 sujets; (2) 60 diagnostiques de causes non-CV chez 53 des sujets; (3) et 718 diagnostiques de toutes autres causes chez 252 des sujets. Pour les OF, 109 des 333 sujets affectés par les maladies non-CV et 58 des 62 par toutes autres maladies étaient atteints simultanément par des MCV. Nous avons décrit les caractéristiques des maladies avec occurrences fatales et non fatales. Les MCVs prédominaient dans les résultats des premières occurrences et occurrences totales tandis que les maladies non-CV étaient les plus élevées pour les diagnostiques. De plus, les OF CV ont affecté 67.1% de notre échantillon de population, incluant les sujets co-affectés par les maladies non-CV ou de toutes autres causes. En fait, nos sujets ont un risque trois fois plus élevé de développer des MCVs (p<0.0001; χ2=1,575.348), tandis qu’il diminue de moitié pour les maladies non-CV comparativement au reste de la population du SLSJ (p=0.0006; χ2=11.834). Enfin, le risque de développer des tumeurs malignes est diminué de moitié dans notre échantillon comparativement à l’incidence régionale. CONCLUSION: Cette étude a apporté une nouvelle perspective sur les OF et ONF chez nos sujets de la région SLSJ du Québec après 11 ans. Quand on observe ces résultats en conjonction avec les MCVs, ce risque double. / BACKGROUND: The World Health Organization (WHO) recognizes cardiovascular diseases (CVD), such as high blood pressure (hypertension; HBP), coronary heart disease (e.g. myocardial infarction), cerebrovascular disease (stroke) and heart failure, as the major cause of death globally. Furthermore, CVD are multifactorial traits with a steeply rising prevalence worldwide, determined by a complex interplay of interactions between genome and environment, which renders their hereditary analysis a complicated task. Our previous studies in the founder population of French-Canadian families from the Saguenay-Lac-Saint-Jean (SLSJ) region of Quebec revealed the largest set of significantly linked loci to hypertension and its metabolic components [1, 2]. HYPOTHESIS AND OBJECTIVE: Since our preliminary findings suggest that early mortality from CVD has genetic and environmental factors, we hypothesize that our population will have distinct characteristics from diseases with and without fatal outcomes (FO and NFO, respectfully), especially in relation to CV cause. Our objectives are to analyze morbidity causes, their temporal characteristics and their clustering in the hypertensive families with or without obesity from the SLSJ region, using hospital, civic death registries and genealogical records from 1950 to the present. We want to describe and study FO and compare them to NFO. RESULTS: In total, we identified 3,654 diagnosis from FO and NFO in 343 participating subjects. For FO, we report: (1) 299 participants suffered from diseases of the circulatory system with a grand total of 1,103 diagnosis, 555 outcomes and 247 first outcomes; (2) 333 of the participating subjects were affected by non-CV systems with a total of 1,536 diagnosis, 195 outcomes and 107 first outcomes; (3) all other diseases were responsible for a total of 81 diagnosis, 30 outcomes and 11 first outcomes in 62 of the participating subjects; and for NFO: (1) the circulatory system affected 105 of the participating subjects with a total of 156 diagnosis; (2) non-CVD, 53 of participating subjects with a total of 60 diagnosis; (3) all other diseases were found in 252 of participating subjects with a total of 718 diagnosis. For FO, 109 of 333 affected subjects by non-CVD and 58 of 62 by all other diseases had also a concomitant CVD. We were able to show characteristics from both FO and NFO. In both first outcomes and total outcomes, CVD predominated over non-CV and all other diseases. When examining CV co-affection with non-CV or all other diseases, 67.1% of our sample population was affected by CV FO. In fact, our sample showed a threefold risk increase in developing CVD (p<0.0001; χ2=1,575.348) when compared to the general population of SLSJ, while it diminished by half for non-CVD (p=0.0006; χ2=11.834). Finally, the relative risk for developing malignancies decreased by half in our sample in contrast to the same region. CONCLUSION: This 11 year follow-up study provided a unique insight into affections for both FO and NFO. When looking at malignancies in conjunction with CVD, this risk grew twofold in our sample.

Název práce: Quantifying Determiners from the Distributional Semantics View Autor: Maria Ximena Gutierrez Vasques Katedra: Ústav formální a aplikované lingvistiky Vedoucí diplomové práce: doc. RNDr. Markéta Lopatková, Ph.D. Abstrakt: Distribuční sémanika představuje moderní přístup k zachycení sémantiky přirozeného jazyka. Jedním z témat, kterým zatím v rámci tohoto přístupu nebyla věnována dostatečná pozornost, je možnost automatické detekce logických relací jako vyplývání. Tato diplomová práce navazuje na práci autorů Baroni, Bernar- di, Do and Shan (2012), kteří se zabývají relací vyplývání mezi kvantifikujícími výrazy. Citovaná práce využívá detekce pomocí SVN klasifikátorů natrénavaných na sémantických vektorech reprezentujících relaci vyplývání. Popisované exper- imenty se nezaměřovaly na nastaveni parametrů SVN klasifikátoru, proto se v této práci vracíme k původním experimentům popisujícím relaci vyplývání mezi kvantifikovanýmo jmennými konstrukcemi, navrhujeme nové konfigurace klasi- fikátoru a optimalizujeme nastavení parametrů. Dosaženou přesnost predikce porovnáváme s původními výsledky a ukazujeme, že SVM klasifikátor s kvadrat- ickým polynomiálním jádrem dosahuje lepších výsledků....

Made available in DSpace on 2014-06-11T19:31:44Z (GMT). No. of bitstreams: 0 Previous issue date: 2008-02-21Bitstream added on 2014-06-13T19:21:07Z : No. of bitstreams: 1 belei_ra_dr_mar.pdf: 1070474 bytes, checksum: a94e8409a324752559085447d8390b23 (MD5) / Analisa comportamentos de segurança e de risco para acidente com material biológico (AMB) durante preparo e descarte de materiais utilizados na administração de medicações. Os participantes são 19 discentes de enfermagem, escolhidos aleatoriamente, que haviam terminado o módulo que aborda de forma teórica o tema biossegurança e realizavam estágio em pronto socorro, e sete docentes supervisores de estágio, todos de uma Universidade Pública do Estado do Paraná. Utiliza filmagem e observação ao vivo dos participantes durante o estágio curricular para identificar e registrar comportamentos e situações de risco e de segurança para AMB. Seleciona cenas das fitas gravadas e registros obtidos da observação ao vivo, os encaminha para juízes e calcula o índice de concordância. Utiliza questionário com discentes e docentes, agrupa os recortes das opiniões mais freqüentes em categorias e sub-categorias, submete à avaliação de juízes e calcula o índice de concordância entre os instrumentos. Entrega recortes de cenas e registros aos discentes e docentes para classificação quanto ao risco ou segurança para AMB, analisa e compara as classificações com aspectos ligados ao processo de ensino-aprendizagem. Verifica 2,86 vezes mais comportamentos de risco (109) para acidente com material biológico que de segurança (38). Identifica 38 comportamentos de segurança: 12 pela filmagem e 26 pela técnica de observação ao vivo, e 109 comportamentos de risco (76 pela filmagem e 33 pela observação ao vivo). Identifica pela observação ao vivo que: dos 30 comportamentos de risco apresentados por discentes, 20 (67%) relacionam-se ao preparo de medicações, estando o docente supervisionando a atividade em 13 (65%) situações; 26 (86,66%) comportamentos/situações envolvem o recipiente de descarte que não atende... / Safety and risky behaviors for accidents with biological material (ABM) are subjected to analysis during preparation and discard of materials used in administration of medication. The participants are 19 undergraduate nursing students, randomly chosen, who had finished the module about biosafety, approached theoretically, and were doing internship in the emergency room, and seven faculty internship supervisors, all from a public university in Paraná State, Brazil. Video recordings and live observation of the participants during internship are used to identify and register safety and risky behaviors for ABM. It selects footage and registers obtained from live observation, submitting them to judges and calculates the agreement rate. It carries out surveys among undergraduate students and faculty members, it groups clippings of most recurrent opinions in categories and subcategories, which is submitted to judges for evaluation and the agreement rate among the instruments is calculated. Footages and registers are given in to undergraduate students and teachers for classification regarding risk and safety in ABM, it analyses and compares the classifications connected to aspects of the teaching and learning process. It verifies 2,86 times more risky behaviors (109) for accidents with biological material than safety (38). It identifies 38 safety behaviors: 12 by the video recordings and 26 by the live observation technique, and 109 risky behaviors (76 by video recording and 33 by live observation). It identifies by live observation that: out of 30 risky behaviors presented by undergraduate nursing students, 20 (67%) are related to the preparation of medication, with the professor supervising the activity in 13 (65%) situations; 26 (86,66%) behaviors/situations involving the discarded container which does not comply with... (Complete abstract click electronic access below)

Book entries are vehicles used in accounting to accommodate non-cash transactions, timing differences and provisions. The use of book entries is a normal activity in accounting and may have their origin in accrual accounting. The management of a company may apply creative accounting techniques in the form of earnings management, in particular, adopting the practices of income smoothing and taking the so-called ‘big bath’. These practices may result in the financial manager or accountant misusing book entries. This could then lead to information of a different integrity to that which would have resulted had these creative accounting practices not been performed in the company. The question addressed in this dissertation and for which an answer is sought, is whether there is any notable difference in the integrity issues of information supplied through the accounting process and created by real transactions (real events) as opposed to information created by book entries (artificial events). The hypothesis underlying this dissertation is: The integrity of information created by book entries is based on subjective opinions because it is based on future events therefore it is not the same as integrity of information created by real transactions that is based on historical events. The new science is concerned with new guidelines, amongst other things, regarding reality, observation, objectivity, predictions and relationships among events. These new guidelines could be seen as explaining certain aspects which is relevant to the field of accounting. The attributes of a book entry are not based on reality, but are based on subjective predictions of future transactions etc. Another similarity is that a book entry is often not objective but is based on subjective observation. Notable differences were observed in the integrity of the information emerging from a real, historical event and a future event. These differences were established through the application of two research methods, namely, the use of a questionnaire and the analysis of the financial statements of 30 companies listed on The JSE Securities Exchange South Africa (JSE). The influence of book entries on certain ratios was considered, and the ratios influenced by two major book entries, namely, depreciation provision and deferred taxation, differ substantially in interpretation when the two book entries are reclassified. The results of the questionnaire also indicate that a large proportion of the financial managers in practice believe that book entries substantially influence the integrity of information. / Dissertation (MCom Financial Management Sciences)--University of Pretoria, 2003. / Financial Management / unrestricted

L'identification d'appareils photos a récemment fait l'objet d'une grande attention en raison de son apport en terme sécurité et juridique. Établir l'origine d'un médias numériques, obtenus par un appareil d'imagerie est important à chaque fois que le contenu numériques est présente et utilise comme preuve devant un tribunal.L'identification d'appareils photos consiste à déterminer la marque, le modèle, ou le dispositif qui a été utilisé pour prendre une image.Notre première contribution pour l'identification du modèle d'appareil photo numérique est basée sur l'extraction de trois ensembles de caractéristiques puis l'utilisation d'apprentissage automatique. Ces caractéristiques sont la matrice de cooccurrences,des corrélations inter-canaux mesurant la trace laissée par l'interpolation CFA, et les probabilités conditionnelles calculées dans le domaine JPEG. Ces caractéristiques donnent des statistiques d'ordre élevées qui complètent et améliorent le taux d'identification. Les expériences prouvent la force de notre proposition, car la précision obtenue est supérieure à celle des méthodes basées sur la corrélation.La deuxième contribution est basée sur l'utilisation des CNNs. Contrairement aux méthodes traditionnelles, les CNNs apprennent simultanément les caractéristiques et la classification. Nous proposons d'ajouter une couche de pré-traitement (filtre passe-haut applique à l'image d’entrée) au CNN. Le CNN obtenu donne de très bonnes performances pour une faible complexité d'apprentissage. La méthode proposée donne des résultats équivalent à ceux obtenu par une approche en deux étapes (extraction de caractéristiques + SVM). Par ailleurs nous avons également examines les CNNs : AlexNet et GoogleNet. GoogleNet donne actuellement les meilleurs taux d'identification pour une complexité d'apprentissage plus grande / Source camera identification has recently received a wide attention due to its importantrole in security and legal issue. The problem of establishing the origin ofdigital media obtained through an imaging device is important whenever digitalcontent is presented and is used as evidence in the court. Source camera identification is the process of determining which camera device or model has been used to capture an image.Our first contribution for digital camera model identification is based on the extractionof three sets of features in a machine learning scheme. These featuresare the co-occurrences matrix, some features related to CFA interpolation arrangement,and conditional probability statistics computed in the JPEG domain.These features give high order statistics which supplement and enhance the identification rate. The experiments prove the strength of our proposition since it achieves higher accuracy than the correlation-based method.The second contribution is based on using the deep convolutional neural networks(CNNs). Unlike traditional methods, CNNs can automatically and simultaneouslyextract features and learn to classify during the learning process. A layer ofpreprocessing is added to the CNN model, and consists of a high pass filter which isapplied to the input image. The obtained CNN gives very good performance for avery small learning complexity. Experimental comparison with a classical two stepsmachine learning approach shows that the proposed method can achieve significantdetection performance. The well known object recognition CNN models, AlexNetand GoogleNet, are also examined.

Le district de Litsa est une région de la Péninsule de Kola particulièrement prometteuse pour la métallogénie de l'uranium. Présentant toutes les caractéristiques d'une province métallifère, ce district a enregistré une histoire géologique complexe avec des enrichissements successifs en uranium. Trente anomalies uranifères y sont observées, estimées à 102 000 tonnes (IAEA nomenclature), pour une teneur moyenne de 0,01%. Elles sont encaissées dans différentes formations de niveau structural et d'âge différents, depuis l'Archéen jusqu'au Calédonien. La minéralisation présentant les teneurs en uranium les plus économiques sont des veines de pechblende observées autour du complexe granitique de Litsa-Araguba, où sont présents des protolites, des structures et des phénomènes épigénétiques favorables / The Litsa district is one of the most interesting areas of the Kola Peninsula with respect to uranium metallogenesis. Presenting all the features of uraniferous province, it endured a protracted geological history with several successive uranium enrichments. There are over 30 uranium occurrences and manifestations, totally estimated at 102,000 tons in Speculative Resources (IAEA nomenclature), with an average grade of Uranium 0.01%. They occur at different stratigraphic levels of the geological succession in the district: Archean, Paleoproterozoic, Mesoproterozoic and Paleozoic. Uranium content in the geological formations gradually increases during the geological evolution of the Litsa district from Archean to Paleozoic. The most abundant and probably prospective for the economic-grade uranium mineralization is Paleozoic pitchblende vein type, located around the Litsa-Araguba granitoid complex, where favorable protoliths, structures and epigenetic processes are encountered

Poliquetas são anelídeos em sua maioria marinhos, apresentando grande diversidade de formas e hábitos de vida. Há tanto animais sedentários quanto errantes, incluindo alguns táxons planctônicos, mas a maioria vive em ambientes bentônicos. A família Syllidae Grube, 1850 é uma das mais complexas, diversificadas, abundantes e amplamente distribuídas famílias de poliqueta. É composta por poliquetas errantes, atualmente distribuídos em mais de 70 gêneros e aproximadamente 700 espécies, distribuídas em 5 subfamílias (Anoplosyllinae, Autolytinae, Eusyllinae, Exogoninae e Syllinae). Os silídeos são facilmente reconhecidos por possuírem uma estrutura muscular geralmente robusta, posteriormente à faringe, denominada proventrículo. A família atualmente é considerada monofilética e a presença desse órgão tem sido considerada como uma sinapomorfia do grupo. Além do proventrículo, os silídeos caracterizam-se por apresentarem corpos relativamente pequenos, com segmentação homônoma e parapódios unirremes, excepcionalmente adquirindo condição birreme nas formas epítocas. Reproduzem-se principalmente por epitoquia, embora incubação externa de ovos, viviparidade, paternogênese e arquitomia também já tenham sido registradas no grupo. O principal objetivo deste estudo foi realizar um levantamento taxonômico das espécies da família Syllidae presentes em substratos consolidados do litoral raso dos estados da Paraíba e de Pernambuco, fornecendo descrições e ilustrações detalhadas dos táxons encontrados. Amostragens foram realizadas no entremarés e infralitoral raso ao longo desses dois estados, totalizando 21 localidades amostradas. Foram analisados morfologicamente 4.233 indivíduos, distribuídos em 20 gêneros e 57 espécies. Dessas, 19 espécies são novas para a ciência, duas são novas ocorrências de gêneros para o Atlântico Sul (Basidiosyllis e Amblyosyllis), e 14 e 18 são novas ocorrências de espécies para o Brasil e para o nordeste brasileiro, respectivamente. Dentre os táxons identificados, estão representadas as cinco subfamílias de Syllidae, assim como os gêneros independentes Amblyosyllis e Basidiosyllis / Polychaetes are usually marine annelids extremely variable in form and lifestyle. Among them, there are examples of sedentary and errant animals, most living in benthic environments, but also including some planktonic taxa. The Syllidae Grube, 1850 is one of the most complex, diverse, abundant and widespread families of polychaetes. It is represented by errant animals, currently comprising more than 70 genera and about 700 species, divided in 5 subfamilies (Anoplosyllinae, Autolytinae, Eusyllinae, Exogoninae e Syllinae). Syllids are recognized by the presence of the proventricle, a muscular structure, generally strong, posteriorly to the pharynx. The family is considered monophyletic, the proventricle considered as a synapomorphy of the group. Besides the proventricle, syllids are also characterized by having relatively small-sized bodies, with homonomous segmentation and uniramous parapodia, exceptionally acquiring a biramous condition in epitokous forms. The reproduction is usually by epitoky, but external brooding of eggs, viviparity, parthenogenesis and architomy are all also found among syllids. The aim of this study was to make a taxonomic inventory of the syllid species occurring on hard bottoms of shallow waters off the states of Paraíba and Pernambuco, providing descriptions and detailed illustrations of the taxa found. Samples were taken from the intertidal and shallow sublitoral zones of both states, totalizing 21 localities sampled. In total, 4,233 individuals were morphologically analysed, belonging to 20 genera and 57 species; 19 of these species are new to science, two are new records for the genera to Brazil (Amblyosyllis and Basidiosyllis), and 14 and 18 are new records for the species to Brazil and to the northestern Brazilian coast, respectively. Among the taxa, all five subfamilies of Syllidae are represented, besides two independent genera, Amblyosyllis and Basidiosyllis