This article presents the problem of improving the classifier of handwritten letters from historical alphabets, using letter classification algorithms and transliterating them to Latin. We apply it on Palmyrene alphabet, which is a complex alphabet with letters, some of which are very similar to each other. We created a mobile application for Palmyrene alphabet that is able to transliterate hand-written letters or letters that are given as photograph images. At first, the core of the application was based on MobileNet, but the classification results were not suitable enough. In this article, we suggest an improved, better performing convolutional neural network architecture for hand-written letter classifier used in our mobile application. Our suggested new convolutional neural network architecture shows an improvement in accuracy from 0.6893 to 0.9821 by 142% for hand-written model in comparison with the original MobileNet. Future plans are to improve the photographic model as well.
Voice over Internet Protocol (VoIP) networks are an increasingly important field in the world of telecommunication due to many involved advantages and potential revenue. Measuring speech quality in VoIP networks is an important aspect of such networks for legal, commercial and technical reasons. The E-model is a widely used objective approach for measuring the quality as it is applicable to monitoring live-traffic, automatically and non-intrusively. The E-model suffers from several drawbacks. Firstly, it considers the effect of packet loss on the speech quality collectively without looking at the content of the speech signal to check whether the loss occurred in voiced or unvoiced parts of the signal. Secondly, it depends on subjective tests to calibrate its parameters, which makes it applicable to limited conditions corresponding to specific subjective experiments. In this paper, a solution is proposed to overcome these two problems. The proposed solution improves the accuracy of the E-model by differentiating between packet loss during speech and silence periods. It also avoids the need for subjective tests, which makes it extendable to new network conditions. The proposed solution is based on an Artificial Neural Networks (ANN) approach and is compared with the accurate Perceptual Evaluation of Speech Quality (PESQ) model and the original E-model to confirm its accuracy. Several experiments are conducted to test the effectiveness of the proposed solution on two well-known ITU-T speech codecs; namely, G.723.1 and G.729.
The photosynthetic gas-exchange has been assessed traditionally either as O2 evolution or CO2 consumption. In this study, we used a liquid-phase O2 electrode combined with CO2 optodes to examine simultaneously photosynthesis in intact leaves of mangrove Rhizophora mucronata. We verified suitable conditions for leaf photosynthetic rates by assessing pH levels and NaHCO3 concentrations and compared these to the gas-exchange method at various PAR levels. The photosynthetic rate in response to pH exhibited a similar pattern both for O2 evolution and CO2 consumption, and higher rates were associated with intermediate pH compared with low and high pH values. The net photosynthetic quotient (PQ) of R. mucronata leaves ranged from 1.04-1.28. The PQ values, which were never lesser than 1, suggested that photorespiration did not occur in R. mucronata leaves under aqueous conditions. The similar maximum photosynthetic rates suggested that all measurements had a high capacity to adjust the photosynthetic apparatus under a light saturation condition. The simultaneous measurements of O2 evolution and CO2 consumption using the Clark oxygen electrode polarographic sensor with the CO2 optode sensor provided a simple, stable, and precise measurement of PQ under aqueous and saturated light conditions., T. Z. Ulqodry, A. Nose, S.-H. Zheng., and Obsahuje seznam literatury
A new pre-processing algorithm for improved discrimination of odor samples is proposed. The pre-processed odor sample outputs from two sensors are input using a learning-vector quantization (LVQ) classifier as a means of odor recognition to be employed within electronic nose applications. The proposed algorithm brings out highly scattered classes while minimizing the within-class scatter of the samples given an odor class. LVQ is observed to operate robustly and reliably in terms of variation of parameters of interest, mainly a learning parameter. Due to the increased performance along with computational simplicity and robustness, the scheme is suitable to sample-by-sample identification of olfactory sensory data and can be easily adapted to hierarchical processing with other sensory data in real-time.
In the general tropospheric tomography model, the tomographic area is divided into a large number of voxels, which provides convenience for reconstructing tomographic observation equations. However, due to the defect of GNSS acquisition geometry, there are plenty of empty voxels for any tomographic epoch. Moreover, an unreasonable assumption that water vapor density is constant within a voxel was imposed on the tomographic model. In this study, we proposed an improved method based on the dynamic node parameterized algorithm to solve both key problems. The proposed approach first tries to select effective GNSS signals and determines the dynamic scope of the tomographic area using the dynamic algorithm. The parameterization of the tomography model is performed by a cubic spline formula and Gauss weighted function. Additionally, a piecewise linear fitting method based on Newton-Cotes interpolation is introduced to estimate the tomographic observation of slant water vapor (SWV). The experimental results show that the average number of effective signals increased by 32.33 % and the mean RMSE of the tomographic results is decreased by 45 % with the proposed method. Further, compared with the tomographic results of the general method, the improved solutions have a more centralized distribution and a smaller bias., Wenyuan Zhang, Shubi Zhang, Nan Ding and Pengxu Ma., and Obsahuje bibliografii
Suppose that $A$ is an $n\times n$ nonnegative matrix whose eigenvalues are $\lambda = \rho (A), \lambda _2,\ldots , \lambda _n$. Fiedler and others have shown that $\det (\lambda I - A) \le \lambda ^n - \rho ^n$, for all $\lambda > \rho $, with equality for any such $\lambda $ if and only if $A$ is the simple cycle matrix. Let $a_i$ be the signed sum of the determinants of the principal submatrices of $A$ of order $i\times i$, $i = 1,\ldots ,n - 1$. We use similar techniques to Fiedler to show that Fiedler’s inequality can be strengthened to: $\det (\lambda I - A) + \sum _{i = 1}^{n - 1} \rho ^{n - 2i}|a_i|(\lambda - \rho )^i \le \lambda ^n -\rho ^n$, for all $\lambda \ge \rho $. We use this inequality to derive the inequality that: $\prod _{2}^{n}(\rho - \lambda _i) \le \rho ^{n - 2}\sum _{i = 2}^{n}(\rho - \lambda _i)$. In the spirit of a celebrated conjecture due to Boyle-Handelman, this inequality inspires us to conjecture the following inequality on the nonzero eigenvalues of $A$: If $\lambda _1 = \rho (A),\lambda _2,\ldots , \lambda _k$ are (all) the nonzero eigenvalues of $A$, then $\prod _{2}^{k}(\rho - \lambda _i) \le \rho ^{k-2}\sum _{i = 2}^{k}(\rho -\lambda )$. We prove this conjecture for the case when the spectrum of $A$ is real.
In this paper we proposed a fuzzy neural network model which can
einbody a fuzzy Takagi-Sugeno model and carry out fuzzy inference and support structure of fuzzy rules. The algorithm of model properties improvement consists of several new procedures námely input space partition, fuzzy terms number and rule number extending, low-effective fuzzy terms and rules extraction and consequent structure Identification. A fuzzy neural network is constructed based on fuzzy model. By learning of the neural network we can tuně of embedded initial fuzzy model. To show the applicability of new method and to niake a possibility to reál systerns rnodelling, we designed the fuzzy-neural network prograrnrne tool FUZNET. Next, we perforrned numerical experiment to do fuzzy rnodelling for an artifical tirne series and reál non-linear complex systém.
We tried to determine as accurately as possible the geocentric coordinates of the Hvar Doppler station in the coordinate system of broadcast and in one of precise ephemerides, on the basis of Doppler observations from the project IDOC-82 and of broadcast ephemerides alone. With this aim in view it was necessary in the first place to calculate for the project IDOC-82 two new variations of multipoint solutions by means of broadcast ephemerides (MPBE), taking into account 11 stations, i.e. 10 suitable stations. Coordinates X, Y, Z for the Hvar station contained in this way were thereupon converted from BE-system by means of
three-dimensional Helmert transformation and by using available identical stations from previous projects EDOC-2, ERIDOC and ALGEDOP-82 for all of which multipoint solutions with precise ephemerides (MPPE) are dsposable.
The goal of this paper is to present and defend an inferentialist account of the meaning of fictional names on the basis of Sellars-Brandom’s inferentialist semantics and a Brandomian anaphoric theory of reference. On this inferentialist account, the meaning of a fictional name is constituted by the relevant language norms which provide the correctness conditions for its use. In addition, the Brandomian anaphoric theory of reference allows us to understand reference in terms of anaphoric word-word relations, rather than substantial word-world relations. In this paper I argue that this inferentialist account has many important merits over its rival theories. One important merit is that it explains why we can use fictional names to make true statements, even if they lack bearers. As a consequence, this theory allows us to use fictional names without committing ourselves to an implausible ontology of fictional entities. Another important merit is that it provides a uniform semantic account of fictional names across different types of statements in which fictional names are involved.
An influence of changing the humic acids content on soil water repellency and saturated hydraulic conductivity was studied on soil samples of Mollic Gleysol from Cilizska Radvan in the Danubian Lowland. Water repellency was measured with the water drop penetration time (WDPT) test on original soil samples and on soil samples with increased humic acids content. Saturated hydraulic conductivity coefficient was measured on the above-mentioned samples with falling head permeameter. From the results of measuring it follows that an increasing of humic acids content in soil resulted in an decreasing of the coefficient of saturated hydraulic conductivity of the soil under study. Original soil was non-water reppelent soil. Already a small increasing of humic acids content in soil (in 17.9 % at original amount) caused that the soil became slightly or strongly water repellent in the average of soil moisture 15-30 %. At soil moisture less than 15 % time of penetration decreased probably as a result of shrinking and cracking of the soil. Water repellency of soil samples from horizon 0 - 5 cm was usually higher than water repellency of soil samples from horizon 5 - 10 cm both in case of humic acids extracted from peat and in case of humic acids extracted from the same soil from Cilizska Radvan. and Na pôdnej vzorke čiernice glejovej (ČA G) (MKSP, 2000) z lokality Čiližská Radvaň v Podunajskej nížině bol skúmaný pomocou testu času vsaku kvapky vody (WDPT test) vplyv zmeny obsahu humínových kyselin na vodoodpudivosť a nasýtenú hydraulickú vodivosť pôdy. Vodoodpudivosť bola meraná na pôvodných pôdnych vzorkách a vzorkách zo zvýšeným obsahom humínových kyselín. Koeficient nasýtenej hydraulickej vodivosti bol meraný na týchto pôdnych vzorkách metódou premenlivého hydraulického sklonu. Z nameraných výsledkov vyplýva, že nárast obsahu humínových kyselín v pôde mal za následok pokles koeficientu nasýtenej hydraulickej vodivosti študovanej pôdy. Už malé zvýšenie obsahu humínových kyselín v pôde (o 17,9 % pôvodnej hodnoty ) spôsobilo, že pôda sa stala slabo až silne vodoodpudivou vo vlhkostnom rozsahu 15-30 %. Pri pôdnej vlhkosti nižšej ako 15 % sa čas vsakovania zmenšil pravdepodobne v dôsledku zmršťovania pôdy a vzniku puklín. Vodoodpudivosť pôdnych vzoriek z horizontu 0 - 5 cm vo väčšine prípadov bola vyššia než vodoodpudivosť pôdnych vzoriek z horizontu 5 - 10 cm aj v prípade pridania humínových kyselín, extrahovaných z rašeliny, aj v prípade pridania humínových kyselín, extrahovaných z tej istej pôdy.