robust method Dotaz Zobrazit nápovědu
- Publikační typ
- abstrakt z konference MeSH
The aim of this study was to develop a simple extraction procedure and a multiresidual liquid chromatography-tandem mass spectrometry method for determination of a wide range of pharmaceuticals from various soil types. An extraction procedure for 91 pharmaceuticals from 13 soil types, followed by liquid chromatography-tandem mass spectrometry analysis, was optimized. The extraction efficiencies of three solvent mixtures for ultrasonic extraction were evaluated for 91 pharmaceuticals. The best results were obtained using acetonitrile/water (1/1 v/v with 0.1 % formic acid) followed by acetonitrile/2-propanol/water (3/3/4 v/v/v with 0.1 % formic acid) for extracting 63 pharmaceuticals. The method was validated at three fortification levels (10, 100, and 1000 ng/g) in all types of representative soils; recovery of 44 pharmaceuticals ranged between 55 and 135 % across all tested soils. The method was applied to analyze actual environmental samples of sediments, soils, and sludge, and 24 pharmaceuticals were found above limit of quantification with concentrations ranging between 0.83 ng/g (fexofenadine) and 223 ng/g (citalopram).
- MeSH
- 2-propanol MeSH
- acetonitrily MeSH
- chromatografie kapalinová metody MeSH
- látky znečišťující půdu analýza MeSH
- léčivé přípravky analýza MeSH
- monitorování životního prostředí metody MeSH
- odpadní vody chemie MeSH
- půda chemie MeSH
- tandemová hmotnostní spektrometrie metody MeSH
- Publikační typ
- časopisecké články MeSH
- validační studie MeSH
Image analysis methods commonly used in forensic anthropology do not have desirable robustness properties, which can be ensured by robust statistical methods. In this paper, the face localization in images is carried out by detecting symmetric areas in the images. Symmetry is measured between two neighboring rectangular areas in the images using a new robust correlation coefficient, which down-weights regions in the face violating the symmetry. Raw images of faces without usual preliminary transformations are considered. The robust correlation coefficient based on the least weighted squares regression yields very promising results also in the localization of such faces, which are not entirely symmetric. Standard methods of statistical machine learning are applied for comparison. The robust correlation analysis can be applicable to other problems of forensic anthropology.
- MeSH
- dospělí MeSH
- lidé MeSH
- lineární modely MeSH
- mladiství MeSH
- mladý dospělý MeSH
- obličej anatomie a histologie MeSH
- počítačové zpracování obrazu MeSH
- soudní antropologie MeSH
- statistické modely MeSH
- Check Tag
- dospělí MeSH
- lidé MeSH
- mladiství MeSH
- mladý dospělý MeSH
- mužské pohlaví MeSH
- ženské pohlaví MeSH
- Publikační typ
- časopisecké články MeSH
- práce podpořená grantem MeSH
Acute respiratory distress syndrome (ARDS) is a disease that has a high reported mortality rate. The treatment for ARDS typically involves mechanical ventilation that is tailored to each patient's needs. A crucial aspect of this treatment is maintaining adequate oxygen saturation of haemoglobin by setting the fraction of inspired oxygen. This paper proposes a design method of robust proportional-integral-derivative (PID) controllers using a gas exchange model during ARDS. Several PID controllers were synthesized for different sub-operational ranges defined by measurable quantities of the mechanical ventilator and the patient using a mixed sensitivity H∞ approach. In simulations, the controller demonstrated high robustness to external changes and changes in the patient's condition, with saturation always above 88%. Although further validation of the controller is required, the results indicate that the presented robust control method has the potential to be clinically relevant.
A method is presented in which conventional speech algorithms are applied, with no modifications, to improve their performance in extremely noisy environments. It has been demonstrated that, for eigen-channel algorithms, pre-training multiple speaker identification (SID) models at a lattice of signal-to-noise-ratio (SNR) levels and then performing SID using the appropriate SNR dependent model was successful in mitigating noise at all SNR levels. In those tests, it was found that SID performance was optimized when the SNR of the testing and training data were close or identical. In this current effort multiple i-vector algorithms were used, greatly improving both processing throughput and equal error rate classification accuracy. Using identical approaches in the same noisy environment, performance of SID, language identification, gender identification, and diarization were significantly improved. A critical factor in this improvement is speech activity detection (SAD) that performs reliably in extremely noisy environments, where the speech itself is barely audible. To optimize SAD operation at all SNR levels, two algorithms were employed. The first maximized detection probability at low levels (-10 dB ≤ SNR < +10 dB) using just the voiced speech envelope, and the second exploited features extracted from the original speech to improve overall accuracy at higher quality levels (SNR ≥ +10 dB).
- MeSH
- algoritmy * MeSH
- hluk * MeSH
- lidé MeSH
- percepce řeči fyziologie MeSH
- počítačové zpracování signálu MeSH
- poměr signál - šum * MeSH
- řeč * MeSH
- Check Tag
- lidé MeSH
- Publikační typ
- časopisecké články MeSH
Tento článek se zabývá automatickou lokalizací objektů (očí, úst) ve dvourozměrných (2D) černobílých obrazech obličejů. Je motivován praktickým problémem v genetice člověka a výstup lokalizace objektů v dané databázi obrazů je zapotřebí pro řešení dalších úloh v genetickém výzkumu. V článku se aplikuje robustní filtr na obrazy s cílem odstranit šum. Hlavní metodou jsou šablony. Ústa a obě oči se lokalizují současně za použití váženého Pearsonova korelačního koeficientu nebo jeho robustní analogie založené na robustních regresních metodách. V databázi s 212 obrazy obličejů tato metoda správně nalezne ústa a oči ve 100 % případů. Také robustní korelační koeficient založený na regresní metodě nejmenších vážených čtverců lokalizuje ústa a oči ve 100 % obrazů uvažované databáze. Článek studuje robustní aspekty této metody vzhledem k otočení, šumu, okluzi a asymetrii v obraze. Současná lokalizace úst i obou očí je invariantní vůči libovolnému otočení obličeje. Tato studie využívá speciální vlastnosti daných obrazů obličejů vzhledem k očekávanému použití v genetických aplikacích.
This paper is devoted to automatic localization of objects (eyes, mouth) in two-dimensional (2D) grey scale images of faces. Motivated by a practical problem in human genetics, the output of the localization of objects in the given database of images is needed for further tasks in the genetic research. A robust filter is applied on the image to ensure denoising. Templates are used as the main method. The mouth and both eyes are localized jointly using the weighted Pearson product-moment correlation coefficient or its robust analogy based on robust regression methods. In the database with 212 images of faces the method allows to locate the mouth and eyes correctly in 100 % of cases. Also the robust correlation coefficient based on the least weighted squares regression localizes the mouth and both eyes in 100 % of images of the given database. Robustness aspects of the method are examined with respect to rotation, noise, occlusion and asymmetry in the image. The joint localization of the mouth and both eyes produces the method invariant to rotation of any degree. This work is tailor made for the given images with expected usage of the methods in genetic applications.
- Klíčová slova
- lokalizace objektů, šablony, detekce oči a úst, robustní korelační analýza, redukce šumu,
- MeSH
- biometrie metody MeSH
- citlivost na kontrast fyziologie MeSH
- databáze jako téma normy MeSH
- fotografování metody MeSH
- genetický výzkum MeSH
- interpretace obrazu počítačem metody MeSH
- lidé MeSH
- obličej MeSH
- oči MeSH
- počítačové zpracování obrazu metody MeSH
- regresní analýza MeSH
- reprodukovatelnost výsledků MeSH
- rozpoznávání fyziologické fyziologie MeSH
- subtrakční technika normy MeSH
- ústa MeSH
- vylepšení obrazu metody MeSH
- Check Tag
- lidé MeSH
Resource specialization is a key concept in ecology, but it is unexpectedly difficult to parameterize. Differences in resource availability, sampling effort and abundances preclude comparisons of incompletely sampled biotic interaction webs. Here, we extend the distance-based specialization index (DSI) that measures trophic specialization by taking resource phylogenetic relatedness and availability into account into a rescaled version, DSI*. It is a versatile metric of specialization that expands considerably the scope and applicability, hence the usefulness, of DSI. The new metric also accounts for differences in abundance and sampling effort of consumers, which enables robust comparisons among distinct guilds of consumers. It also provides an abundance threshold for the reliability of the metric for rare species, a very desirable property given the difficulty of assessing any aspect of rare species accurately. We apply DSI* to an extensive dataset on interactions between insect herbivores from four folivorous guilds and their host plants in Papua New Guinean rainforests. We demonstrate that DSI*, contrary to the original DSI, is largely independent of sample size and weakly and non-linearly related with several host specificity measures that do not adjust for plant phylogeny. Thus, DSI* provides further insights into host specificity patterns; moreover, it is robust to the number and phylogenetic diversity of plant species selected to be sampled for herbivores. DSI* can be used for a broad range of comparisons of distinct feeding guilds, geographical locations and ecological conditions. This is a key advance in elucidating the interaction structure and evolution of highly diversified systems.
- MeSH
- býložravci * MeSH
- fylogeneze * MeSH
- hmyz klasifikace genetika MeSH
- nutriční stav MeSH
- potravní řetězec MeSH
- reprodukovatelnost výsledků MeSH
- rostliny klasifikace MeSH
- zvířata MeSH
- Check Tag
- zvířata MeSH
- Publikační typ
- časopisecké články MeSH
Detection of grapes in real-life images is a serious task solved by researchers dealing with precision viticulture. In the case of white wine varieties, grape detectors based on SVMs classifiers, in combination with a HOG descriptor, have proven to be very efficient. Simplified versions of the detectors seem to be the best solution for practical applications. They offer the best known performance versus time-complexity ratio. As our research showed, a conversion of RGB images to grayscale format, which is implemented at an image preprocessing level, is ideal means for further improvement of performance of the detectors. In order to enhance the ratio, we explored relevance of the conversion in a context of a detector potential sensitivity to a rotation of berries. For this purpose, we proposed a modification of the conversion, and we designed an appropriate method for a tuning of such modified detectors. To evaluate the effect of the new parameter space on their performance, we developed a specialized visualization method. In order to provide accurate results, we formed new datasets for both tuning and evaluation of the detectors. Our effort resulted in a robust grape detector which is less sensitive to image distortion.
The Minimum Redundancy Maximum Relevance (MRMR) approach to supervised variable selection represents a successful methodology for dimensionality reduction, which is suitable for high-dimensional data observed in two or more different groups. Various available versions of the MRMR approach have been designed to search for variables with the largest relevance for a classification task while controlling for redundancy of the selected set of variables. However, usual relevance and redundancy criteria have the disadvantages of being too sensitive to the presence of outlying measurements and/or being inefficient. We propose a novel approach called Minimum Regularized Redundancy Maximum Robust Relevance (MRRMRR), suitable for noisy high-dimensional data observed in two groups. It combines principles of regularization and robust statistics. Particularly, redundancy is measured by a new regularized version of the coefficient of multiple correlation and relevance is measured by a highly robust correlation coefficient based on the least weighted squares regression with data-adaptive weights. We compare various dimensionality reduction methods on three real data sets. To investigate the influence of noise or outliers on the data, we perform the computations also for data artificially contaminated by severe noise of various forms. The experimental results confirm the robustness of the method with respect to outliers.
