Segmentazione e categorizzazione di oggetti mediante immagini depth e deep learning

Questo lavoro è iniziato con uno studio teorico delle principali tecniche di classificazione di immagini note in letteratura, con particolare attenzione ai più diffusi modelli di rappresentazione dell’immagine, quali il modello Bag of Visual Words, e ai principali strumenti di Apprendimento Automatico (Machine Learning). In seguito si è focalizzata l’attenzione sulla analisi di ciò che costituisce lo stato dell’arte per la classificazione delle immagini, ovvero il Deep Learning. Per sperimentare i vantaggi dell’insieme di metodologie di Image Classification, si è fatto uso di Torch7, un framework di calcolo numerico, utilizzabile mediante il linguaggio di scripting Lua, open source, con ampio supporto alle metodologie allo stato dell’arte di Deep Learning. Tramite Torch7 è stata implementata la vera e propria classificazione di immagini poiché questo framework, grazie anche al lavoro di analisi portato avanti da alcuni miei colleghi in precedenza, è risultato essere molto efficace nel categorizzare oggetti in immagini. Le immagini su cui si sono basati i test sperimentali, appartengono a un dataset creato ad hoc per il sistema di visione 3D con la finalità di sperimentare il sistema per individui ipovedenti e non vedenti; in esso sono presenti alcuni tra i principali ostacoli che un ipovedente può incontrare nella propria quotidianità. In particolare il dataset si compone di potenziali ostacoli relativi a una ipotetica situazione di utilizzo all’aperto. Dopo avere stabilito dunque che Torch7 fosse il supporto da usare per la classificazione, l’attenzione si è concentrata sulla possibilità di sfruttare la Visione Stereo per aumentare l’accuratezza della classificazione stessa. Infatti, le immagini appartenenti al dataset sopra citato sono state acquisite mediante una Stereo Camera con elaborazione su FPGA sviluppata dal gruppo di ricerca presso il quale è stato svolto questo lavoro. Ciò ha permesso di utilizzare informazioni di tipo 3D, quali il livello di depth (profondità) di ogni oggetto appartenente all’immagine, per segmentare, attraverso un algoritmo realizzato in C++, gli oggetti di interesse, escludendo il resto della scena. L’ultima fase del lavoro è stata quella di testare Torch7 sul dataset di immagini, preventivamente segmentate attraverso l’algoritmo di segmentazione appena delineato, al fine di eseguire il riconoscimento della tipologia di ostacolo individuato dal sistema.

Genome-wide investigation reveals high evolutionary rates in annual model plants

Background: ;Rates of molecular evolution vary widely among species. While significant deviations from molecular clock have been found in many taxa, effects of life histories on molecular evolution are not fully understood. In plants, annual/perennial life history traits have long been suspected to influence the evolutionary rates at the molecular level. To date, however, the number of genes investigated on this subject is limited and the conclusions are mixed. To evaluate the possible heterogeneity in evolutionary rates between annual and perennial plants at the genomic level, we investigated 85 nuclear housekeeping genes, 10 non-housekeeping families, and 34 chloroplast;genes using the genomic data from model plants including Arabidopsis thaliana and Medicago truncatula for annuals and grape (Vitis vinifera) and popular (Populus trichocarpa) for perennials.;Results: ;According to the cross-comparisons among the four species, 74-82% of the nuclear genes and 71-97% of the chloroplast genes suggested higher rates of molecular evolution in the two annuals than those in the two perennials. The significant heterogeneity in evolutionary rate between annuals and perennials was consistently found both in nonsynonymous sites and synonymous sites. While a linear correlation of evolutionary rates in orthologous genes between species was observed in nonsynonymous sites, the correlation was weak or invisible in synonymous sites. This tendency was clearer in nuclear genes than in chloroplast genes, in which the overall;evolutionary rate was small. The slope of the regression line was consistently lower than unity, further confirming the higher evolutionary rate in annuals at the genomic level.;Conclusions: ;The higher evolutionary rate in annuals than in perennials appears to be a universal phenomenon both in nuclear and chloroplast genomes in the four dicot model plants we investigated. Therefore, such heterogeneity in evolutionary rate should result from factors that have genome-wide influence, most likely those associated with annual/perennial life history. Although we acknowledge current limitations of this kind of study, mainly due to a small sample size available and a distant taxonomic relationship of the model organisms, our results indicate that the genome-wide survey is a promising approach toward further understanding of the;mechanism determining the molecular evolutionary rate at the genomic level.

Development of Dynamic Constraint Models for a Model Based Transient Calibration Process

Model based calibration has gained popularity in recent years as a method to optimize increasingly complex engine systems. However virtually all model based techniques are applied to steady state calibration. Transient calibration is by and large an emerging technology. An important piece of any transient calibration process is the ability to constrain the optimizer to treat the problem as a dynamic one and not as a quasi-static process. The optimized air-handling parameters corresponding to any instant of time must be achievable in a transient sense; this in turn depends on the trajectory of the same parameters over previous time instances. In this work dynamic constraint models have been proposed to translate commanded to actually achieved air-handling parameters. These models enable the optimization to be realistic in a transient sense. The air handling system has been treated as a linear second order system with PD control. Parameters for this second order system have been extracted from real transient data. The model has been shown to be the best choice relative to a list of appropriate candidates such as neural networks and first order models. The selected second order model was used in conjunction with transient emission models to predict emissions over the FTP cycle. It has been shown that emission predictions based on air-handing parameters predicted by the dynamic constraint model do not differ significantly from corresponding emissions based on measured air-handling parameters.

Aberrant lipid metabolism in hepatocellular carcinoma revealed by plasma metabolomics and lipid profiling

There has been limited analysis of the effects of hepatocellular carcinoma (HCC) on liver metabolism and circulating endogenous metabolites. Here, we report the findings of a plasma metabolomic investigation of HCC patients by ultraperformance liquid chromatography-electrospray ionization-quadrupole time-of-flight mass spectrometry (UPLC-ESI-QTOFMS), random forests machine learning algorithm, and multivariate data analysis. Control subjects included healthy individuals as well as patients with liver cirrhosis or acute myeloid leukemia. We found that HCC was associated with increased plasma levels of glycodeoxycholate, deoxycholate 3-sulfate, and bilirubin. Accurate mass measurement also indicated upregulation of biliverdin and the fetal bile acids 7α-hydroxy-3-oxochol-4-en-24-oic acid and 3-oxochol-4,6-dien-24-oic acid in HCC patients. A quantitative lipid profiling of patient plasma was also conducted by ultraperformance liquid chromatography-electrospray ionization-triple quadrupole mass spectrometry (UPLC-ESI-TQMS). By this method, we found that HCC was also associated with reduced levels of lysophosphocholines and in 4 of 20 patients with increased levels of lysophosphatidic acid [LPA(16:0)], where it correlated with plasma α-fetoprotein levels. Interestingly, when fatty acids were quantitatively profiled by gas chromatography-mass spectrometry (GC-MS), we found that lignoceric acid (24:0) and nervonic acid (24:1) were virtually absent from HCC plasma. Overall, this investigation illustrates the power of the new discovery technologies represented in the UPLC-ESI-QTOFMS platform combined with the targeted, quantitative platforms of UPLC-ESI-TQMS and GC-MS for conducting metabolomic investigations that can engender new insights into cancer pathobiology.

Metabolomics reveals attenuation of the SLC6A20 kidney transporter in nonhuman primate and mouse models of type 2 diabetes mellitus

To enhance understanding of the metabolic indicators of type 2 diabetes mellitus (T2DM) disease pathogenesis and progression, the urinary metabolomes of well characterized rhesus macaques (normal or spontaneously and naturally diabetic) were examined. High-resolution ultra-performance liquid chromatography coupled with the accurate mass determination of time-of-flight mass spectrometry was used to analyze spot urine samples from normal (n = 10) and T2DM (n = 11) male monkeys. The machine-learning algorithm random forests classified urine samples as either from normal or T2DM monkeys. The metabolites important for developing the classifier were further examined for their biological significance. Random forests models had a misclassification error of less than 5%. Metabolites were identified based on accurate masses (<10 ppm) and confirmed by tandem mass spectrometry of authentic compounds. Urinary compounds significantly increased (p < 0.05) in the T2DM when compared with the normal group included glycine betaine (9-fold), citric acid (2.8-fold), kynurenic acid (1.8-fold), glucose (68-fold), and pipecolic acid (6.5-fold). When compared with the conventional definition of T2DM, the metabolites were also useful in defining the T2DM condition, and the urinary elevations in glycine betaine and pipecolic acid (as well as proline) indicated defective re-absorption in the kidney proximal tubules by SLC6A20, a Na(+)-dependent transporter. The mRNA levels of SLC6A20 were significantly reduced in the kidneys of monkeys with T2DM. These observations were validated in the db/db mouse model of T2DM. This study provides convincing evidence of the power of metabolomics for identifying functional changes at many levels in the omics pipeline.

Distance measures for image segmentation evaluation

The task considered in this paper is performance evaluation of region segmentation algorithms in the ground-truth-based paradigm. Given a machine segmentation and a ground-truth segmentation, performance measures are needed. We propose to consider the image segmentation problem as one of data clustering and, as a consequence, to use measures for comparing clusterings developed in statistics and machine learning. By doing so, we obtain a variety of performance measures which have not been used before in image processing. In particular, some of these measures have the highly desired property of being a metric. Experimental results are reported on both synthetic and real data to validate the measures and compare them with others.

Comparison of the effects of inquiry-based cooperative learning and demonstrations in science education

The reported research project involved studying how teaching science using demonstrations, inquiry-based cooperative learning groups, or a combination of the two methods affected sixth grade students’ understanding of air pressure and density. Three different groups of students were each taught the two units using different teaching methods. Group one learned about the topics through both demonstrations and inquirybased cooperative learning, whereas group two only viewed demonstrations, and group three only participated in inquiry-based learning in cooperative learning groups. The study was designed to answer the following two questions: 1. Which teaching strategy works best for supporting student understanding of air pressure and density: demonstrations, inquirybased labs in cooperative learning groups, or a combination of the two? 2. And what effect does the time spent engaging in a particular learning experience (demonstrations or labs) have on student learning? Overall, the data did not provide sufficient evidence that one method of learning was more effective than the others. The results also suggested that spending more time on a unit does not necessarily equate to a better understanding of the concepts by the students. Implications for science instruction are discussed.

Matching Software and Context in Open Learning Scenarios

We describe the use of log file analysis to investigate whether the use of CSCL applications corresponds to its didactical purposes. Exemplarily we examine the use of the web-based system CommSy as software support for project-oriented university courses. We present two findings: (1) We suggest measures to shape the context of CSCL applications and support their initial and continuous use. (2) We show how log files can be used to analyze how, when and by whom a CSCL system is used and thus help to validate further empirical findings. However, log file analyses can only be interpreted reasonably when additional data concerning the context of use is available.

Prediction of PA12 melt viscosity in Laser Sintering by a Time and Temperature dependent rheological model

Ein auf Basis von Prozessdaten kalibriertes Viskositätsmodell wird vorgeschlagen und zur Vorhersage der Viskosität einer Polyamid 12 (PA12) Kunststoffschmelze als Funktion von Zeit, Temperatur und Schergeschwindigkeit angewandt. Im ersten Schritt wurde das Viskositätsmodell aus experimentellen Daten abgeleitet. Es beruht hauptsächlich auf dem drei-parametrigen Ansatz von Carreau, wobei zwei zusätzliche Verschiebungsfaktoren eingesetzt werden. Die Temperaturabhängigkeit der Viskosität wird mithilfe des Verschiebungsfaktors aT von Arrhenius berücksichtigt. Ein weiterer Verschiebungsfaktor aSC (Structural Change) wird eingeführt, der die Strukturänderung von PA12 als Folge der Prozessbedingungen beim Lasersintern beschreibt. Beobachtet wurde die Strukturänderung in Form einer signifikanten Viskositätserhöhung. Es wurde geschlussfolgert, dass diese Viskositätserhöhung auf einen Molmassenaufbau zurückzuführen ist und als Nachkondensation verstanden werden kann. Abhängig von den Zeit- und Temperaturbedingungen wurde festgestellt, dass die Viskosität als Folge des Molmassenaufbaus exponentiell gegen eine irreversible Grenze strebt. Die Geschwindigkeit dieser Nachkondensation ist zeit- und temperaturabhängig. Es wird angenommen, dass die Pulverbetttemperatur einen Molmassenaufbau verursacht und es damit zur Kettenverlängerung kommt. Dieser fortschreitende Prozess der zunehmenden Kettenlängen setzt molekulare Beweglichkeit herab und unterbindet die weitere Nachkondensation. Der Verschiebungsfaktor aSC drückt diese physikalisch-chemische Modellvorstellung aus und beinhaltet zwei zusätzliche Parameter. Der Parameter aSC,UL entspricht der oberen Viskositätsgrenze, wohingegen k0 die Strukturänderungsrate angibt. Es wurde weiterhin festgestellt, dass es folglich nützlich ist zwischen einer Fließaktivierungsenergie und einer Strukturänderungsaktivierungsenergie für die Berechnung von aT und aSC zu unterscheiden. Die Optimierung der Modellparameter erfolgte mithilfe eines genetischen Algorithmus. Zwischen berechneten und gemessenen Viskositäten wurde eine gute Übereinstimmung gefunden, so dass das Viskositätsmodell in der Lage ist die Viskosität einer PA12 Kunststoffschmelze als Folge eines kombinierten Lasersinter Zeit- und Temperatureinflusses vorherzusagen. Das Modell wurde im zweiten Schritt angewandt, um die Viskosität während des Lasersinter-Prozesses in Abhängigkeit von der Energiedichte zu berechnen. Hierzu wurden Prozessdaten, wie Schmelzetemperatur und Belichtungszeit benutzt, die mithilfe einer High-Speed Thermografiekamera on-line gemessen wurden. Abschließend wurde der Einfluss der Strukturänderung auf das Viskositätsniveau im Prozess aufgezeigt.

NodePM: A Remote Monitoring Alert System for Energy Consumption Using Probabilistic Techniques

In this paper, we propose an intelligent method, named the Novelty Detection Power Meter (NodePM), to detect novelties in electronic equipment monitored by a smart grid. Considering the entropy of each device monitored, which is calculated based on a Markov chain model, the proposed method identifies novelties through a machine learning algorithm. To this end, the NodePM is integrated into a platform for the remote monitoring of energy consumption, which consists of a wireless sensors network (WSN). It thus should be stressed that the experiments were conducted in real environments different from many related works, which are evaluated in simulated environments. In this sense, the results show that the NodePM reduces by 13.7% the power consumption of the equipment we monitored. In addition, the NodePM provides better efficiency to detect novelties when compared to an approach from the literature, surpassing it in different scenarios in all evaluations that were carried out.

Simulation of Drosophila circadian oscillations, mutations, and light responses by a model with VRI, PDP-1, and CLK.

A model of Drosophila circadian rhythm generation was developed to represent feedback loops based on transcriptional regulation of per, Clk (dclock), Pdp-1, and vri (vrille). The model postulates that histone acetylation kinetics make transcriptional activation a nonlinear function of [CLK]. Such a nonlinearity is essential to simulate robust circadian oscillations of transcription in our model and in previous models. Simulations suggest that two positive feedback loops involving Clk are not essential for oscillations, because oscillations of [PER] were preserved when Clk, vri, or Pdp-1 expression was fixed. However, eliminating positive feedback by fixing vri expression altered the oscillation period. Eliminating the negative feedback loop in which PER represses per expression abolished oscillations. Simulations of per or Clk null mutations, of per overexpression, and of vri, Clk, or Pdp-1 heterozygous null mutations altered model behavior in ways similar to experimental data. The model simulated a photic phase-response curve resembling experimental curves, and oscillations entrained to simulated light-dark cycles. Temperature compensation of oscillation period could be simulated if temperature elevation slowed PER nuclear entry or PER phosphorylation. The model makes experimental predictions, some of which could be tested in transgenic Drosophila.

Influence of the Central American Seaway and Drake Passage on ocean circulation and neodymium isotopes: A model study

The sensitivity of the neodymium isotopic composition (ϵNd) to tectonic rearrangements of seaways is investigated using an Earth System Model of Intermediate Complexity. The shoaling and closure of the Central American Seaway (CAS) is simulated, as well as the opening and deepening of Drake Passage (DP). Multiple series of equilibrium simulations with various intermediate depths are performed for both seaways, providing insight into ϵNd and circulation responses to progressive throughflow evolutions. Furthermore, the sensitivity of these responses to the Atlantic Meridional Overturning Circulation (AMOC) and the neodymium boundary source is examined. Modeled ϵNd changes are compared to sediment core and ferromanganese (Fe-Mn) crust data. The model results indicate that the North Atlantic ϵNd response to the CAS shoaling is highly dependent on the AMOC state, i.e., on the AMOC strength before the shoaling to shallow depths (preclosure). Three scenarios based on different AMOC forcings are discussed, of which the model-data agreement favors a shallow preclosure (Miocene) AMOC (∼6 Sv). The DP opening causes a rather complex circulation response, resulting in an initial South Atlantic ϵNd decrease preceding a larger increase. This feature may be specific to our model setup, which induces a vigorous CAS throughflow that is strongly anticorrelated to the DP throughflow. In freshwater experiments following the DP deepening, ODP Site 1090 is mainly influenced by AMOC and DP throughflow changes, while ODP Site 689 is more strongly influenced by Southern Ocean Meridional Overturning Circulation and CAS throughflow changes. The boundary source uncertainty is largest for shallow seaways and at shallow sites.