Esta pesquisa visa analisar a transferência de tensões de compressão entre elementos de concreto pré-moldado através de almofadas de argamassa modificada. A argamassa foi modificada com as adições de fibras de polipropileno, látex estireno-butadieno e vermiculita termo-expandida. A caracterização da argamassa foi realizada com ensaios de compressão, de tração por compressão diametral e de módulo de elasticidade, além de ensaios de rigidez de almofada, que medem a capacidade da peça de se deformar sob tensão de compressão uniforme. As análises de transferência de tensões foram realizadas intercalando uma almofada entre dois blocos de concreto e submetendo o corpo de prova formado a ensaios de compressão. Os parâmetros variados nesses ensaios foram: rugosidade superficial e resistência mecânica dos blocos ligados, excentricidade de carregamento, solidarização ou não da almofada ao bloco inferior de concreto e carregamento monotônico e cíclico. Em relação à massa do cimento, a argamassa das almofadas possuía 30% de agregados, dos quais 95% eram areia fina e 5% eram vermiculita, 10% de látex, 2% de taxa volumétrica de fibras de polipropileno e relação água / cimento de 0,25, além da água incorporada no látex. A argamassa apresentou resistência à compressão média de 27,65 MPa, resistência à tração de 3,62 MPa e módulo de elasticidade igual a 11,53 GPa. Os principais resultados dos ensaios de ligação foram: a) para compressão centrada, a presença da almofada aumentou a resistência em 20% para espessuras de saliências de 0,5 mm e 30% para saliências de 1,0 mm; b) ganho de efetividade da almofada à medida que se reduziu a resistência do concreto; c) em ligações excêntricas, a dispersão dos resultados não permitiu uma correta avaliação dos resultados; d) para carregamento cíclico, a almofada de apoio aumentou em 45% a resistência da ligação. / This research aims to analyze the compressive stresses transfer between precast concrete elements by modified mortar pads. The mortar was modified with additions of polypropylene fibers, styrene-butadiene latex and heat-expanded vermiculite. Mortar characterization was performed in compression, tension and elastic modulus tests, further pad rigidity tests, that measure the pad capacity to deform under uniform compressive stress. Stress transfer analyzes were performed interleaving a bearing pad between two concrete blocks and subjecting the specimen formed in the compression tests. Varied parameters in these tests were: surface roughness and mechanical strength of the connected blocks, loading eccentricity, solidarization or not of the bearing pad to the bottom concrete block and monotonic and cyclic loading. In relation to the cement weight, the mortar of the pads showed 30% of aggregates, which 95% were fine sand and 5% were vermiculite, 10% of latex, 2% of volumetric rate of polypropylene fibers and water / cement ratio of 0,25, in addition of water incorporated into the latex. The mortar showed compression strength of 27,65 MPa, tension strength of 3,62 MPa and elastic modulus equal to 11,53 GPa. The main results of connection test were: a) for centered compression, presence of pad increased the strength in 20% for thicknesses of imperfections of 0,5 mm and 30% for imperfections of 1,0 mm; b) gain of effectiveness of the bearing pad when the concrete strength was reduced; c) in eccentrical connections, the dispersion of results did not allowed a correct evaluation of results; d) for cyclic loading, the bearing pad increased in 45% the connections strength.

In der vorliegenden Arbeit wurde das Pad-Protein von Legionella pneumophila mit zell- bzw. molekularbiologischen sowie immunochemischen Methoden charakterisiert. Mittels Einbau des mutierten Pad-Locus, kodiert auf dem Plasmid pCS-25, in das Genom der L. pneumophila Stämme WH88 (Serogruppen 6) bzw. WH89 (Serogruppe 10) wurden je drei Pad-negative Mutanten gewonnen. Mit Hilfe der PCR, des Southernblot bzw. ELISA wurde der korrekte Austausch der Wildtypsequenz gegen den mutierten Bereich des Proteins nachgewiesen. Die Infektionsversuche mit den Mutanten bestätigten die Vermutung, dass das Protein eine Rolle an der initialen Kontaktaufnahme von L. pneumophila in den natürlichen Wirt, A. castellanii, hat. Dabei wurden signifikant niedrigere Aufnahmeraten der Mutanten im Vergleich zu den beiden Wildtypen beobachtet. Durch Kultivierung und Infektion des Ursprungstammes L. pneumophila Corby, der pad-negativen Mutante CP7 und der Pad-komplementierten Mutante bei 25°C, 30°C bzw. 37°C wurde der Einfluss der Temperatur auf die Funktion des Proteins untersucht. Die Ergebnisse zeigten dabei einen signifikanten Anstieg der Aufnahmeraten bei Vergleich von 25°C und 30°C bzw. 25°C und 37°C. Keine Signifikanz trat zwischen 30°C und 37°C auf. Ein zweiter Aspekt dieses Versuchs war die Prüfung der Infektionsrate in Abhängigkeit von der Proteinexpression unter dem Einfluss der Wachstumsphase. Dabei wurden durch vergleichende Infektion von A. castellanii mit exponentiellen (EP) bzw. stationäre Phase-Kulturen (SP) in diesem Versuch signifikante Unterschiede zwischen dem Wildtyp Corby und der Mutante CP7 bzw. zwischen der Mutante CP7 und der Pad+-Komplementante bei den jeweils untersuchten Temperaturen beobachtet. Keine signifikanten Unterschiede waren zwischen Wildtyp und Pad-Komplementante feststellbar. Zur Speziesübergreifenden Charakterisierung des Proteins Pad wurde das Wildtyplocus-kodierende Plasmid (pCS31) in die fünf non-pneumophila Stämme L. parisiensis (H962); L. bozemanii (L99-343); L. bozemanii (Frankreich5); L. longbeachae (A46) und L. tauriniensis (Koper11) durch Elektroporation übertragen. Im anschließenden ELISA wurde die Expression des Proteins Pad in den komplementierten Transformanten nachgewiesen. Die Infektionsversuche ergaben überwiegend signifikant höhere Aufnahmeraten der Pad-komplementierten Klone im Vergleich zu den Pad-negativen Wildtypen. Die Infektionsversuche zeigen, dass Pad nicht an der intrazellulären Replikation beteiligt ist. Die Ergebnisse der Infektionsversuche wurden parallel dazu durch mikroskopische Untersuchung intrazellulärer Bakterien bestätigt. Die Vermutung, dass die Expression von Pad als infektionsbeteiligtes Protein beim Übergang in die virulente Phase induziert wird, wurde durch die Kultivierung über 6 Tage bis in die späte stationäre Phase bestätigt. Die Kulturen des L. pneumophila Stammes JR32 zeigen im ELISA mit dem Pad-spezifischen Antikörper 61-1 ab der spätexponentiellen Phase (nach 24 Stunden) eine ansteigende Extinktion. Im Gegensatz dazu wurde für die GacA (LetA)-negative Mutante JR32gac- eine gleichbleibend hohe Expression des Proteins Pad über den gesamten Versuchszeitraum gemessen, was erste Hinweise liefert, dass Pad gacA bzw. letA-reguliert ist. Ein zellschädigender Einfluss der Bakterien auf die Atmungskette der Amöbenzelle zu frühen Zeitpunkten der Infektion (5, 30, 120 min) wurde mittels des Cell Titer Blue Cell Viability Assays festgestellt. Die Infektion mit hitzeabgetöteten Legionellen ergab dabei eine Toxizität unter 10 % bei allen drei untersuchten Zeitpunkten. Nach Infektion mit EP- als auch mit SP-Kulturen des Wildtyps Corby wurde eine maximale Toxizität zwischen 30 % und 40 % (120 min) gemessen. Bei der Mutante wurde eine Toxizität von ca. 28% der SP-Kultur nach 120 min Infektion beobachtet. Für die weiterführende Charakterisierung des Pad-Proteins wurde eine Maus mit dem rekombinanten Protein immunisiert. Dabei wurden zusätzlich zum vorhandenen Antikörper 61-1 die Antikörper 83-1 und 83-2 gewonnen. Bei diesen handelt es sich um IgG-Antikörper. Die durchgeführten Tests lassen vermuten, dass es sich um, in der Reaktion von Mak 61-1, ähnliche Antikörper handelt. Zur strukturellen Analyse des Pad-Proteins wurde mit zwei unterschiedlichen Phagenbibliotheken nach möglichen Epitopen für den Pad-spezifischen Antikörper 61-1 gesucht. Dabei wurde unabhängig voneinander mit beiden Phagen-Bibliotheken ein Epitop (L2) im C-terminalen Bereich ermittelt. Das Einfügen von Zufallsmutationen ließ keine Eingrenzung des Epitops zu. Da es jedoch dem, mit Mak 61-7 im Peptidspotting von C. Steudel (2001) ermittelten, Epitop C3 entspricht, ist davon auszugehen, dass es sich tatsächlich um das zweite putative Epitop für den Antikörper handelt. Des Weiteren wurde mit beiden Phagen-Bibliotheken das Epitop (L1) lokalisiert. Eine Eingrenzung dieses Epitops mittels Einfügen von Zufallsmutanten war nicht möglich. Durch site-spezifische Mutation wurden alle Aminosäuren des Epitops C1 (STEUDEL, 2001) in der putativen Signalsequenz ausgetauscht. Bei Testung der Mutanten auf ihre Bindungsfähigkeit an den Antikörper 61-1 im ELISA konnte eine Beteiligung der Aminosäuren Leucin11, Phenylalanin15, Glycin18 und Prolin20 beobachtet werden. Die Ergebnisse bestätigen, dass Pad an frühen Phasen der Infektion von A.castellanii durch L.pneumophila beteiligt ist. Es ist zu vermuten, dass dieses Pad-Protein einen Selektionsvorteil für den am häufigsten nachweisbaren Vertreter dieser Familie darstellt.

Background: As a company it is more important than ever to consider the environmental sustainability aspect within the whole supply chain and all the stakeholders. Studies show that people's concerns about environmental sustainability increases and will continue to grow in the future. The wet dyeing process is the part of the textile production that consists of liquor and contains three main steps which are pretreatment, coloration and finishing. This process is presented as the most water, energy and chemicals consuming process within the textile industry and because of that, it is important to investigate it to be able to improve the environmental sustainability.  Purpose: The purpose of the thesis is to increase the understanding of IKEA’s suppliers' current usage of water, energy and chemicals within the wet dyeing process. Followed by an identification of opportunities for improvements in the wet dyeing process, aiming to become more environmentally sustainable.  Method: Through interviews with two suppliers, information regarding IKEA’s current wet dyeing processes will be collected. After that, data from relevant theoretical sources will be collected to analyze possible improvements of methods and colorants that can be relevant for IKEA as well as other companies within the textile industry to implement.  Conclusion: For research question 1, the researchers reach the conclusion that IKEA’s supplier number one has a more environmentally sustainable wet dyeing process compared to the second supplier.. A conclusion is nevertheless drawn that improvements within both suppliers wet dyeing processes should be considered to become more environmentally sustainable. For research question 2, the researchers come to the conclusion that Spray dyeing is the best alternative method to implement for improving environmental sustainability within the wet dyeing process. The researchers also recommend companies to actively become a part of the development of bacterial colorants, the future of coloration.

Este trabalho descreve um ataque que detecta as teclas pressionadas em teclados mecânicos pela análise das vibrações geradas quando as mesmas são pressionadas. Dois equipamentos foram experimentados no ataque: um teclado genérico de automação comercial e um terminal de ponto de venda (POS / PIN-pad). Acelerômetros são utilizados como sensores de vibração. Propositalmente, o equipamento necessário para a execução do ataque é de baixíssimo custo, de modo a ressaltar o risco das vulnerabilidades encontradas. Obtivemos taxas de sucesso médio de 69% no reconhecimento das teclas pressionadas para o terminal PIN-pad em repouso e 75% para o mesmo sendo segurado na mão. No caso de teclado de automação comercial, as taxas médias de acerto ficaram em torno de 99%. / This work describes an attack that identifies the sequence of keystrokes analyzing mechanical vibrations generated by the act of pressing keys. We use accelerometers as vibration sensors. The apparatus necessary for this attack is inexpensive and can be unobtrusively embedded within the target equipment. We tested the proposed attack on an ATM keypad and a PIN-pad. We achieved the key recognition rates of 99% in ATM keypad, 69% in PIN-pad resting on a hard surface and 75% in PIN-pad hold in hand.

Este trabalho descreve um ataque que detecta as teclas pressionadas em teclados mecânicos pela análise das vibrações geradas quando as mesmas são pressionadas. Dois equipamentos foram experimentados no ataque: um teclado genérico de automação comercial e um terminal de ponto de venda (POS / PIN-pad). Acelerômetros são utilizados como sensores de vibração. Propositalmente, o equipamento necessário para a execução do ataque é de baixíssimo custo, de modo a ressaltar o risco das vulnerabilidades encontradas. Obtivemos taxas de sucesso médio de 69% no reconhecimento das teclas pressionadas para o terminal PIN-pad em repouso e 75% para o mesmo sendo segurado na mão. No caso de teclado de automação comercial, as taxas médias de acerto ficaram em torno de 99%. / This work describes an attack that identifies the sequence of keystrokes analyzing mechanical vibrations generated by the act of pressing keys. We use accelerometers as vibration sensors. The apparatus necessary for this attack is inexpensive and can be unobtrusively embedded within the target equipment. We tested the proposed attack on an ATM keypad and a PIN-pad. We achieved the key recognition rates of 99% in ATM keypad, 69% in PIN-pad resting on a hard surface and 75% in PIN-pad hold in hand.

Static characteristics and rotordynamic coefficients were experimentally determined for a four-pad tilting-pad journal bearing with ball-in-socket pivots in loadbetween- pad configuration. A frequency-independent [M]-[C]-[K] model fit the measurements reasonably well, except for the cross-coupled damping coefficients. Test conditions included speeds from 4,000 to 12,000 rpm and unit loads from 0 to 1896 kPa (0 to 275 psi). The test bearing was manufactured by Rotating Machinery Technology (RMT), Inc. Though it has a nominal diameter of 101.78 mm (4.0070 in.), measurements indicated significant bearing crush with radial bearing clearances of 99.6 μm (3.92 mils) and 54.6 μm (2.15 mils) in the axes 45º counterclockwise and 45º clockwise from the loaded axis, respectively. The pad length is 101.6 mm (4.00 in.), giving L/D = 1.00. The pad arc angle is 73º, and the pivot offset ratio is 65%. The preloads of the loaded and unloaded pads are 0.37 and 0.58, respectively. A bulk-flow Navier-Stokes model was used for predictions, using adiabatic conditions for the bearing fluid. Because the model assumes constant nominal clearances at all pads, the average of the measured clearances was used as an estimate. Eccentricities and attitude angles were markedly under predicted while power loss was under predicted at low speeds and very well predicted at high speeds. The maximum detected pad temperature was 71ºC (160ºF) and the rise from inlet to maximum bearing temperature was over predicted by 10-40%. Multiple-frequency force inputs were used to excite the bearing. Direct stiffness and damping coefficients were significantly over predicted, but addition of a simple stiffness-in-series model substantially improved the agreement between theory and experiment. Direct added masses were zero or negative at low speeds and increased with speed up to a maximum of about 50 kg; they were normally greater in the unloaded direction. Although significant cross-coupled stiffness terms were present, they always had the same sign. The bearing had zero whirl frequency ratio netting unconditional stability over all test conditions. Static stiffness in the y direction (obtained from steadystate loading) matched the rotordynamic stiffness Kyy (obtained from multiple-frequency excitation) reasonably at low loads but poorly at the maximum test load.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Most of the free radicals generated in the body is derived from oxygen and are associated with the development of many diseases. Medicinal plants are a natural resource in the treatment prevention of diseases and have been used for the first healthcare. These properties of the plants are related to the presence of phenolic compounds which possess several pharmacological and antioxidant properties. The high performance liquid chromatography has been used to identify and quantify these compounds in plants. Therefore, this study developed and validated an HPLC - PAD method using gold as working electrode to characterize 12 phenolic compounds in herbal extracts from Amazon. The extracts from Connarus perrotettii var angustifolius, Mansoa alliacea, Cecropia palmata, Cecropia obtusa, Bauhinia variegata and Bauhinia alboflava showed the presence of at least one of the antioxidants compounds under study. Furthermore, the species showed in vitro antioxidant activity against reactive oxygen species. The acetate fraction of the species M. alliacea showed the highest in vitro antioxidant activity (70.85 ± 4.06%) against the superoxide radical. Front of the peroxyl radical, the species C. palmata was responsible for the hidhest antioxidant activity (95.33 ± 0.95%) and against hydroxyl radical species C. obtusa showed the highest scavenging activity (98.19 ± 5.93). The study of the in vitro antioxidant activity of these species coupled with chemical characterization of their extracts comprises an important key to potential in vivo studies. Moreover, the compounds identified and quantified in these species extracts have pharmacological properties and may be linked to popular use of these medicinal plants / A maioria dos radicais livres gerados no organismo é derivada do oxigênio e estão associados ao desenvolvimento de diversas doenças. As plantas medicinais são um recurso terapêutico natural na prevenção de patologias e têm sido utilizadas a milênios pela humanidade nos primeiros cuidados com a saúde. Muitas dessas propriedades das plantas estão relacionadas com a presença de compostos fenólicos os quais possuem diversas propriedades farmacológicas e antioxidantes. A cromatografia líquida de alta eficiência tem sido utilizada na identificação e quantificação desses compostos em plantas. Dessa forma, o presente trabalho desenvolveu e validou um método de HPLC PAD utilizando eletrodo de ouro como eletrodo de trabalho para a caracterização de 12 compostos fenólicos em extratos de plantas medicinais da Amazônia. Os extratos das espécies Connarus perrotettii var angustifolius, Mansoa alliacea, Cecropia palmata, Cecropia obtusa, Bauhinia variegata e Bauhinia alboflava demonstraram a presença de pelo menos um dos compostos antioxidantes em estudo. Além disso, as espécies apresentaram atividade antioxidante in vitro frente à espécies reativas de oxigênio. A fração acetato da espécie M. alliacea demonstrou a maior atividade antioxidante in vitro (70,85 ± 4,06 %) frente ao radical superóxido. Frente ao radical peroxil, a espécie C. palmata foi responsável pela maior atividade antioxidante (95,33 ± 0,95 %) e frente ao radical hidroxil a espécie C. obtusa demonstrou a maior atividade sequestradora (98,19 ± 5,93). O estudo da atividade antioxidante in vitro dessas espécies aliada a caracterização química dos seus extratos compreende uma importante chave para possíveis estudos in vivo. Ademais, os compostos identificados e quantificados nos extratos dessas espécies apresentam propriedades farmacológicas comprovadas e podem estar associados ao uso popular dessas plantas medicinais.

Background: Despite the fact that peripheral arterial disease (PAD) significantly increases the risk of cardiovascular mortality, it is significantly underdiagnosed and underrated. The purpose of this study was to evaluate the practice at a northeast Tennessee university primary care clinic regarding the diagnosis and treatment of PAD. Methods: A retrospective medical record survey was conducted to evaluate practice patterns in diagnosing and treating PAD in a university primary care clinic. A clinic population of 711 patients was selected using International Classification of Diseases-9 codes for coronary artery disease (CAD), cerebovascular disease (CVD), and/ or PAD. A sample of 180 patients (25.3%) was randomly selected using a systematic statistical method. Of these, 125 patients met the diagnostic criteria for CAD, CVD, and/or PAD. The study covered a 3-year period, from July 2001 until June 2004. Demographic and other data, including the use of antiplatelet therapy, were collected. Results: One hundred ten patients met all of the inclusion and exclusion criteria. Thirty-nine percent were males, and 61% were females. Overall, 79% had CAD, 53% had CVD, and 25% had PAD. Almost half of the patients had some combination of these. Only about 2% had PAD only compared with 36% with CAD only and 17% with CVD only. Although the prevalence of CAD and CVD (among other atherosclerotic vascular diseases) in our clinic was comparable to national figures, the prevalence of PAD was significantly lower (p = .004). The overall use of any antiplatelet agent was 84.2% for patients with only CAD and 80% for only CVD. There was not an adequate number of patients with only PAD to evaluate the use of antiplatelet therapy in this group. Conclusion: The low prevalence of PAD only (most PAD patients had coexisting CAD and CVD) indicates that PAD is underdiagnosed at our clinic. There was suboptimal use of aspirin and other antiplatelet drugs among patients with atherosclerotic vascular disease.

Skin cancer is one of the most common forms of cancer, of which melanoma is the most lethal. Early detection is critical to long term survival rates. The use of machine learning to detect melanoma shows promising results in detecting malignant forms. The use of melanoma detecting models could potentially lead to faster diagnosis, and ensure treatment is applied before the melanoma has spread to other organs. In recent years, many studies have been made on the topic, especially on larger datasets such as ISIC, but there is a lack of studies on how machine learning perform on datasets containing smartphone images of skin lesions. PAD-UFES-20 is a dataset containing 2298 smartphone images of skin lesions, where the malignant forms are all clinically confirmed. The purpose of this thesis was to produce a comparison of the accuracy, precision, recall, and Area under the Curve metrics for several convolutional neural network models that were trained on the PAD-UFES-20 and ISIC datasets to differentiate between melanoma and benign lesions. The goal of this project is to contribute to the knowledge surrounding how neural networks perform in terms of melanoma detection. This is of interest for anyone getting started in the domain of melanoma detection using machine learning, building a melanoma detection application, and for anyone benefiting from earlier diagnosis. A qualitative and inductive approach with the support of quantitative data was adopted. Eight models were selected. They were designed using transfer learning, where the base model acted as a feature extractor, providing input to a dropout layer, and finally to the classification layer, which was retrained for 50 epochs, using the ADAM optimizer with a learning rate of 0.0003. These models were then evaluated against a test set for accuracy, precision, recall and AUC. The results of this study showed that the best performing models in terms of detecting melanoma on the PAD-UFES-20 dataset were Resnet_v2_50 and Mobilenet_v3_Large. Resnet_v2_50 achieved an accuracy of 0.9062, precision of 0.8, recall of 0.6667 and AUC of 0.8846. Mobilenet_v3_Large achieved accuracy of 0.9062, precision of 0.8, recall of 0.6667 and AUC of 0.9295. The best performing model on the ISIC dataset was the Inception_Resnet_v2 model with accuracy of 0.9342, precision of 0.9627, recall of 0.9018 and AUC of 0.9675. The conclusions of this study show that several models achieve promising results on the PAD-UFES-20 dataset. Overall, models that extracted a larger number of features from the base model performed better on the PAD-UFES-20 dataset, whereas this was not the general case on the larger ISIC dataset. It is not too futuristic to claim that machine learning can be applied as complement to standard diagnosis, but the study highlights the need for similar studies on larger and more diverse datasets of smartphone images, from which more transferable results can be obtained. / Hudcancer är en av de vanligaste formerna av cancer, där melanom är den mest dödliga. Tidig upptäckt är avgörande för långsiktiga överlevnadschanser. Användning av maskininlärning för att upptäcka melanom visar lovande resultat i att upptäcka maligna former av hudskador. Användning av modeller för att upptäcka melanom kan potentiellt leda till snabbare diagnos och säkerställa att behandling tillämpas innan melanomet har spridit sig till andra organ. De senaste åren har många studier gjorts inom ämnet, särskilt på större datamängder såsom ISIC datamängden, men det saknas studier om hur maskininlärning presterar på datamängder som innehåller smarttelefonbilder av hudskador. PAD-UFES-20 är en datamängd som innehåller 2298 smarttelefonbilder av hudskador, där de maligna formerna alla är kliniskt bekräftade. Syftet med denna studie var att producera en jämförelse av noggrannhet, precision, recall och Area under the Curve måtten för flera faltnings neuronnätverksmodeller som tränades på PAD-UFES-20 och ISIC-dataseten för att skilja mellan melanom och godartade hudskador. Målet med denna studie är att bidra till kunskapen om hur neurala nätverk presterar när det gäller upptäckt av melanom. Detta är av intresse för alla som börjar inom området för upptäckt av melanom med hjälp av maskininlärning och för alla som gynnas av en tidigare diagnos. En kvalitativ och induktiv ansats med stöd av kvantitativa data antogs. Åtta modeller valdes. De designades med hjälp av överföringsinlärning, där basmodellen fungerade som en producent av funktioner, som gav inmatning till ett Dropout lager och slutligen till klassificeringslagret, som tränades i 50 epoker, med hjälp av ADAM-optimeraren med en inlärningshastighet på 0.0003. Dessa modeller utvärderades sedan mot en test mängd för noggrannhet, precision, recall och AUC. Resultaten av denna studie visade att de bästa presterande modellerna när det gäller att upptäcka melanom på PAD-UFES-20-datasetet var Resnet_v2_50 och Mobilenet_v3_Large. Resnet_v2_50 uppnådde en noggrannhet på 0.9062, precision på 0.8, recall på 0.6667 och AUC på 0.8846. Mobilenet_v3_Large uppnådde noggrannhet på 0.9062, precision på 0.8, recall på 0.6667 och AUC på 0.9295. Den bäst presterande modellen på ISIC-datasetet var Inception_Resnet_v2-modellen med noggrannhet på 0.9342, precision på 0.9627, recall på 0.9018 och AUC på 0.9675. Slutsatserna av denna studie visar att flera modeller uppnår lovande resultat på PAD-UFES-20-datasetet. Sammantaget, presterade modeller som extraherade ett större antal funktioner från basmodellen, bättre på PAD-UFES-20-datasetet, medan detta inte var det generella fallet på det större ISIC-datasetet. Det är inte för långsökt att hävda att maskininlärning kan tillämpas som komplement till standarddiagnos metoder, men denna studie framhäver behovet av liknande studier på större och mer diversifierade datamängder av smarttelefonbilder, från vilka mer överförbara resultat kan erhållas.

No description available.