3284 research outputs found
Sort by
Models and Systems for the Control of Two-Phase Processes in Microfluidics
No full textThe strong point of the microfluidics is the ability to miniaturize and integrate one or several laboratory functions on the same device, to have a portable and user-friendly instrument. Most applications require accurate measures and control within the microfluidic channels. In this thesis, the optical techniques were adopted to monitor, sensing and control the processes, leading to the research area of optofluidics that are based on the integration of fluidics and optics. To reduce the cost to develop these devices, the 3D Printing technology based on the Poly(dimethyl-siloxane) (PDMS) is proposed. All these aspects were addressed considering the two-phase flow (named slug) generated by the interaction of two immiscible fluids, a very common condition in bio-chemical applications. The methodological aspects were discussed in the first part of the thesis, starting from the extraction of parameters for the flow characterization, to their use for the flows real-time modelling and control schemes development; the second part investigates aspects faced for the realization of micro-optical flow detector by using the 3D Printing technology
The Fundamental Photophysics of Fluorescent Carbon Nanodots
No full textCarbon Nanodots (CDs) are a new protagonist of carbon-based nanoscience by which the paradigm of carbon as a black material unable to emit light has been completely revolutionised. They have been emerging as a new frontier in Nanoscience at the beginning of 2000s, and their potential is evident from the explosion of the number of studies, now ranging in the thousands per year. CDs are nanoparticles composed by carbon, oxygen and hydrogen with a size smaller than 10 nm. Their most important hallmark is their strong luminescence, which is combined with many additional benefits as the low cost and ease of synthesis, the high water solubility, the biocompatibility and non-toxicity, the great sensitivity to the external environment, and a marked electron donating and accepting capabilities. The combination of all of these characteristics guarantees the possibility to use CDs in a very broad range of applications which encompasses many different fields as optoelectronics or sensing. Actually, the term CDs includes nanomaterials which display a wide range of possible structures and variable optical properties. Indeed, in the literature, it is common to find different sub-types of CDs: they can be graphitic, amorphous, disks of graphene or with a C3N4 core; they can be hydrophilic or hydrophobic; they can emit blue, green, or red light; their emission can be independent of the excitation wavelength, or more commonly tunable (peak of the emission depends on the excitation wavelength); their fluorescence intensity can be sensitive to one particular ion in solution or they respond to a variety of interactions with other systems, such as carbon nanotubes. Besides, CDs can be synthesized by many different procedures which yield subtypes of CDs capable of emitting fluorescence at different wavelengths. In all the synthesis approaches, their surface is passivated by external agents to get bright emission. All of this leads to the existence in the literature of diverse sub-types of CDs with different core and surface structure, and different specific optical characteristics. Despite of this, there are common characteristics which are recurrent almost in every type of CDs as the small size and the core+corona structure which are found to be crucial to obtain CDs with a visible photoluminescence with a high emission efficiency. Carbon dots research is still in a developing phase despite thousands of studies have already been published on the subject, and several scientific open questions exist about their optical behaviour, the fundamental nature of the electronic states, the key factors determining their bright fluorescence, and the relation between structure and emission. As a consequence, a large effort is in progress to find the most effective ways to tailor them for specific applications.
In this Thesis, we carried out an investigation of the fundamental physics of different families of CDs with the aim to achieve an exhaustive understanding of their entire photocycle from femtosecond to the steady state. To do this, it was necessary to relate the optical properties to the structural ones, and to study the influence of various possible interactions with external agents. Thus, electronic transitions of CDs having different structures or exposed to different environments have been studied by the combined use of several experimental techniques. In particular, as an important novelty in the field, a variety of new methods which involve ultrafast spectroscopic techniques have been used. Ultrafast spectroscopy is a powerful tool to investigate in real time electronic and vibrational processes with picosecond and femtosecond time resolution. With the use of these methods it was possible to map the photocycle of CDs revealing several dynamics which occur on short time scale as solvation, charge transfer, and fluorescence quenching
Egocentric Vision Based Localization of Shopping Cart
No full textIndoor camera localization from egocentric images is a challenge computer vision problem which has been strongly investigated in the last years. Localizing a camera in a 3D space can open many useful applications in different domains. In this work, we analyse this challenge to localize shopping cart in stores. Three main contributions are given with this thesis. As first, we propose a new dataset for shopping cart localization which includes both RGB and depth images together with the 3-DOF data corresponding to the cart position and orientation in the store. The dataset is also labelled with respect to 16 different classes associated to different areas of the considered retail. A second contribution is related to a benchmark study where different methods are compared for both, cart pose estimation and retail area classification. Last contribution is related to the computational analysis of the considered approaches
Chemotassonomia e attività biologica degli oli essenziali di taxa del genere Thymus.
No full textIl lavoro di ricerca prende spunto dalle problematiche tassonomiche inerenti al genere Thymus correlate alla notevole variabilità dei tratti morfologici spesso non sufficientemente discriminanti.
Lo studio ha riguardato la caratterizzazione dei componenti chimici degli oli essenziali di differenti taxa del genere Thymus, quali validi marcatori chemotassonomici, per contribuire all identificazione delle entità indagate.
La ricerca è stata finalizzata anche alla valutazione dell attività biologica, in vitro e in vivo, degli oli essenziali di Thymus su germinazione, sopravvivenza e crescita di diverse specie infestanti, per un loro potenziale impiego come bio-erbicidi.
L analisi dei componenti chimici degli oli essenziali è stata condotta utilizzando la gas cromatografia accoppiata alla spettrometria di massa (GC/FID, GC/MS), in differenti taxa del genere Thymus: T. longicaulis C. Presl subsp. longicaulis; T. paronychioides Celak., T. praecox Opiz subsp. parvulus (Lojac.) Bartolucci, Peruzzi & N.G. Passal., T. spinulosus Ten., T. richardii Pers. subsp. nitidus (Guss.) Jalas, T. striatus Vahl subsp. striatus, T. vulgaris L. subsp. vulgaris, Thymus. sp.
Per il confronto in termini composizionali tra i taxa, a livello intra ed interspecifico, è stata applicata l analisi cluster con il metodo agglomerativo UPGMA (Unweighted pair group method using arithmetic averages), basato su distanze Euclidee.
I saggi di fitotossicità sono stati realizzati, in vitro e in vivo, su differenti specie target (Amaranthus blitum L., Amaranthus retroflexus L., Portulaca oleracea s.l., Avena fatua L., Echinocloa crus-galli (L.) P. Beauv. e Lepidium sativum L.) impiegando oli essenziali di Thymus tal quali, nanoemulsionati con diverse sostanze naturali (es. gomma arabica e proteine del siero del latte) o microincapsulati in particelle di zeina.
L effetto in vitro degli oli essenziali è stato stimato calcolando le percentuali di inibizione della germinazione (GI%) e di fitotossicità.
L azione erbicida, in vivo, è stata valutata adottando una scala di valori che commisura il grado di danneggiamento delle piante.
La caratterizzazione chimica degli oli essenziali ha evidenziato una variabilità intra e interspecifica tra i taxa indagati e ha contribuito a consolidare e ampliare, per le entità distribuite in Sicilia, le relative conoscenze tassonomiche.
I saggi di fitotossicità hanno evidenziato una significativa azione inibente degli oli essenziali di T. longicaulis subsp. longicaulis e T. praecox subsp. parvulus, T. spinulosus, T. vulgaris subsp. vulgaris. Essi, si configurano, quindi, potenzialmente interessanti per applicazioni in ambito agronomico
A-priori estimates for some classes of elliptic problems
No full textL'obiettivo di questa tesi è di studiare alcuni aspetti di un potente strumento ampiamente utilizzato in analisi matematica, che è rappresentato dalle stime a priori. Infatti, le stime a priori hanno un ruolo chiave nella teoria delle equazioni differenziali a derivate parziali e nel calcolo delle variazioni, perché sono intimamente legate all'esistenza di soluzione per un dato problema. Nella tesi vengono presentati tre lavori scritti durante il periodo del dottorato, in ciascuno dei quali vengono utilizzate le stime a priori.
Il primo lavoro, scritto in collaborazione con il Prof. S. Mosconi, riguarda l'esistenza di soluzione per la seguente equazione differenziale ordinaria del quarto ordine (equazione di Swift-Hohenberg), , dove è un parametro reale e è una funzione , coerciva e quasi-convessa.
Il secondo lavoro, scritto in collaborazione con il prof. P. Winkert, riguarda stime a priori per un problema ellittico in cui gli operatori hanno crescita critica, sia nel dominio che sulla frontiera.
Il terzo lavoro, scritto in collaborazione con i Prof. S.A. Marano e A. Moussaoui, riguarda l'esistenza di soluzione per un sistema ellittico definito in tutto lo spazio , in cui le nonlinearità contengono termini singolari, cioè che possono tendere a quando la variabile tende a zero
Deeply Incorporating Human Capabilities into Machine Learning Models for Fine-Grained Visual Categorization
No full textArtificial intelligence and machine learning have long attempted to emulate human visual system.
With the recent advances in deep neural networks, which take inspiration from the architecture of the primate visual hierarchy, human-level visual abilities are now coming within reach of artificial systems. However, the existing computational models are designed with engineering goals, loosely emulating computations and connections of biological neurons, especially in terms of intermediate visual representations.
In this thesis we aim at investigating how human skills can be integrated into computational models in order to perform fine-grained image categorization, a task which requires the application of specific perceptive and cognitive abilities to be solved. In particular, our goal is to develop systems which, either implicitly or explicitly, combine human reasoning processes with deep classification models. Our claims is that by the emulation of the process carried out by humans while performing a recognition task it is possible to yield improved classification performance.
To this end, we first attempt to replicate human visual attention by modeling a saliency detection system able to emulate the integration of the top-down (task-controlled, classification-driven) and bottom-up (sensory information) processes; thus, the generated saliency maps are able to represent implicitly the way humans perceive and focus their attention while performing recognition, and, therefore, a useful supervision for the automatic classification system. We then investigate if and to what extent the learned saliency maps can support visual classification in nontrivial cases. To achieve this, we propose SalClassNet, a CNN framework consisting of two networks jointly trained: a) the first one computing top-down saliency maps from input images, and b) the second one exploiting the computed saliency maps for visual classification.
Gaze shifts change in relation to a task is not the only process when performing classification in specific domains, but humans also leverage a-priori specialized knowledge to perform recognition. For example, distinguishing between different dog breeds or fruit varieties requires skills that not all human possess but only domain experts. Of course, one may argue that the typical learning-by-example approach can be applied by asking domain experts to collect enough annotations from which machine learning methods can derive the features necessary for the classification. Nevertheless, this is a really costly process and often infeasible. Thus, the second part of this thesis aim at explicitly modeling and exploiting domain-specific knowledge to perform recognition.
To this end, we introduce and demonstrate that computational ontologies can explicitly encode human knowledge and that it can be used to support multiple tasks from data annotation to classification. In particular, we propose an ontology-based annotation tool, able to reduce significantly the efforts to collect highly-specialized labels and demonstrate its effectiveness building the VegImage dataset, a collection of about 4,000 images belonging to 24 fruit varieties, annotated with over 65,000 bounding boxes and enriched with a large knowledge base consisting of more than 1,000,000 OWL triples.
We then exploit this ontology-structured knowledge by combining a semantic-classifier, which performs inference based on the information encoded in the domain ontology, with a visual convolutional neural network, showing that the integration of semantics into automatic classification models can represents the key to solve a complex task such as the fine-grained recognition of fruit varieties, a task which requires the contribution of domain expert to be completely solved.
Performance evaluation of the proposed approaches provides a basis to assess the validity of our claim along with the scientific soundness of developed models
Electrochemical sensors for environmental and clinical analyses
No full textNowadays, the concepts of smart cities, smart houses and Homo Deus (potential next stage in human evolution) are taking more and more attention. Sensors have a key role in this context: recording different parameters it will be possible to design cities able to manage efficiently all the resources or to warn people if a particular kind of disease, such as cancer or diabetes, is taking place. Current way of detecting these parameters or molecules are often laborious and expensive and cannot be used as in situ and real time. Electrochemical sensors, especially nano-sized sensors, are perfect candidates to address these challenges. Indeed, these sensors do not require special instrumentations to work but just an usual battery, so that this technology is cheap and suitable for in situ action. Furthermore, the electrical signal can be recorded over time and can be acquired and managed in remote. The main challenge of this technology is to achieve a Limit Of Detection (LOD) low enough to make the use of these sensors competitive with other analytical techniques, such as Inductively Coupled Plasma Mass Spectrometry (ICP-MS) or Enzyme Linked ImmunoSorbent Assay (ELISA). These goals can be achieved by nano-sized materials because they enhance mass transport and electron transfer rate. In addition, to once found the right sensing material, it is possible to select the electrochemical detection technique giving the best performances in dependence on the analyte that has to be detected. During my Ph.D I studied different electrodes to detect, electrochemically, 3 analytes: i) heavy metals, ii) proteins, and iii) H2O2. Heavy metals are the main source of water pollution. They have been extensively used in various applications due to their specific properties. The main problem using these chemicals is that they are non-biodegradable and thus they can accumulate in the human body through the food chain. Among heavy metals, one of the most dangerous is mercury because just the exposure to some ppb (µg/l) can cause several problems to human body. Lead, cadmium, zinc, arsenic and copper are considered toxic and dangerous as well, therefore, their monitoring is really important. H2O2 is a widely used chemical, employed as bleaching agent in textile and paper industry, for medical and pharmaceutical applications and to remove organic compounds from waste water and contaminated soil. Furthermore, H2O2 has a key role in the human body as well. For instance, its detection can be useful because can give indications about the glucose concentration and so it could be useful for diabetic patients. Furthermore, it is a biomarker of oxidative stress that is a pathological condition due to breakdown of the antioxidant defense system.. Detection of proteins was also investigated during my Ph.D. In order to detect these bio-compounds it is mandatory to use some bio recognition elements (such as antibodies, DNA or aptamers) so that the sensors are usually named biosensors. During my studies, I developed a biosensors towards Human ImmunoGlobulin G (H-IgG) and ParaThyroid Hormone Like Hormone (PTHLH). H-IgG is a protein always present in the human fluids (blood, urine, sweat) and its detection has not any particular relevance. It can be used as a model because is a cheap protein that has the bio-chemical properties of many other proteins. Instead, PTHLH is overproduced owing to different kind of cancer, consequently it can be used as a biomarker. This kind of application is really of great value because PTHLH starts to be produced at the beginning of the disease, so that its detection is useful for early diagnosis. Summarizing, the main goals of this Ph.D work are:
1. The development of new and innovative ways to fabricate electrodes with high surface area;
2. To find new, cheap, robust electrochemical sensors for detecting H2O2, heavy metals, and proteins;
3. Validate these sensors using real sample
Codimension two ACM varieties in P1xP1xP1 and regularity of bicyclic graphs and their powers
No full textIn this PhD thesis, we discuss several different results about some homological invariants (e.g., graded Betti numbers, Hilbert function, regularity) of some special varieties. In particular, we focus on the codimension two ACM varieties in P1×P1×P1 (called varieties of lines), and the edge ideals of bicyclic graphs. We study the Hilbert function of Ferrers varieties of lines, a special case of ACM variety of lines, and we describe the trigraded minimal free resolution of the defining ideal of a variety of lines arising from a complete intersection of points. We also compute the Castelnuovo-Mumford regularity of the defining ideal of grids of lines and complete intersections of lines in P1×P1×P1. Then we study the regularity of another special variety, i.e., the edge ideal of a bicyclic graph and its powers. Specifically, we compute the regularity of the edge ideal of a dumbbell graph, and then we give a combinatorial characterization of the regularity of the edge ideal of an arbitrary bicyclic graph in terms of its induced matching number. Finally we study the regularity of powers of edge ideals of some specific bicyclic graphs, i.e., dumbbell graphs with path having at most two vertices
RNA sequencing, bioinformatics and CRISPR system studies on Italian MDR Staphylococci
No full textAntibiotic resistance is worldwide an increasing problem for the public health. Scientists are working to better know the features of Multi-Drug Resistant (MDR) bacteria to develop new strategies to overcome the antibiotic resistance. Among the most common bacteria capable to infect humans and become resistant to antibiotics, the members of the genus Staphylococcus are the etiological agents of some infectious diseases that can be lethal if the pharmacological treatment fails. S. aureus can cause a wide spectrum of diseases, ranging from mild skin forms to systemic forms, whilst S. epidermidis has been frequently implicated in endocarditis and infections of surgical implants and it seems to be the source for antibiotic resistance related gene acquisition by S. aureus.
A comparative transcriptomic analysis by RNA-seq and bioinformatic of two Methicillin-Resistant S. aureus isogenic pairs, the characterization of the CRISPR-Cas system and phage infection sensitivity tests of ten Italian pathogenic clinical MDR S. epidermidis strains were performed to deepen knowledge the antibiotic resistance related traits in Italian MDR Staphylococci.
Results of transcriptomic analysis showed the contribution of Staphylococcal mRNAs and small-RNAs in daptomycin resistance, with multiple pathways associated, including the cell-wall biosynthesis and organization, metabolism, nucleic acid metabolism, stress response and transport, confirming the role of transcriptome in developing antibiotic resistance.
CRISPR-Cas system presence seems to not be a common trait in MDR S. epidermidis (33%) although they showed a high resistance to phage infection, results that should be considered in the perspective of developing a phage therapy.
Finally, recent discoveries suggest the possibility to use CRISPR, in the future, to target not only the genes, but also RNAs (messenger and small), linking transcriptome analysis and CRISPR system studies as parts of a possible strategy to cure the antibiotic resistance
Smoothed Particle Hydrodynamics method and flow dynamics: the case of lava numerical modeling and simulation
No full textLo Smoothed Particle Hydrodynamics è un metodo Lagrangiano mesh-free che sta riscuotendo interesse nel campo della fluidodinamica computazionale. Grazie alla sua natura, il metodo SPH è in grado di gestire fluidi complessi caratterizzati da reolgie non-Newtoniane, superfici libere, dipendenze dalla temperatura, transizioni di fase, grandi deformazioni, e così via. Il metodo SPH è caratterizzato da una natura intrinsecamente parallela, che ne rende possibile l'esecuzione su hardware per il calcolo parallelo ad alte prestazioni, come ad esempio le moderne schede grafiche (GPU), ottenendo così tempi di simulazione vantaggiosi. In questa tesi lavoreremo su GPUSPH, un'implementazione su GPU del metodo SPH. Studieremo la simulazione di un fluido molto complesso: la lava. La combinazione di superficie libera, topografia naturale, transizioni di fase e la formazione di strutture come argini e tunnel, rende la modellazione e la simulazione della lava un aspetto di interesse nel campo della fluidodinamica computazionale, con un notevole impatto in numerosi campi dell'Ingegneria e della ricerca scientifica. Vedremo l'introduzione in GPUSPH di modelli e strategie che permetteranno di trattare le diverse caratteristiche delle colate laviche, includendo lo sviluppo di uno schema di integrazione semi-implicito, che permetterà la simulazione di fluidi altamente viscosi, assicurando robustezza e riducendo i tempi di simulazione.
La nuova implementazione sarà testata per verificarne la correttezza e studiarne i livelli accuratezza e performance raggiunti