Multi-scale stacked sequential learning

Sequential learning is the discipline of machine learning that deals with dependent data such that neighboring labels exhibit some kind of relationship. The paper main contribution is two-fold: first, we generalize the stacked sequential learning, highlighting the key role of neighboring interactions modeling. Second, we propose an effective and efficient way of capturing and exploiting sequential correlations that takes into account long-range interactions. We tested the method on two tasks: text lines classification and image pixel classification. Results on these tasks clearly show that our approach outperforms the standard stacked sequential learning as well as state-of-the-art conditional random fields.     

Analytical prediction of abnormal temporary overvoltages due to ground faults in MV networks

The paper deals with the temporary overvoltages that build up in radial MV distribution networks following the inception of a 1-phase-to-ground fault. It will be demonstrated that, for extended cable/overhead MV distribution networks with ungrounded neutral, in case of low resistance faults at critical locations along overhead lines, the neutral voltage can reach very large values, significantly higher than 1 p.u. (up to 2.5 p.u.). The attendant temporary overvoltages on healthy phases can be very large, much higher than √3 p.u. (up to 3.5 p.u.). Fault currents are also affected, reaching twice the value calculated with simplified methods, i.e. neglecting series impedances. Quick yet accurate analytical formulas for the prediction of maximum overvoltages, fault currents and critical fault distances are presented, together with their validation by detailed ATP-EMTP simulations. The results of a study on an existing Italian 15 kV–50 Hz extended mixed cable/overhead distribution network are finally presented and discussed.     

A fully integrated 0.18-/spl mu/m CMOS direct conversion receiver front-end with on-chip LO for UMTS

This paper presents a 0.18-/spl mu/m CMOS direct-conversion IC realized for the Universal Mobile Telecommunication System (UMTS). The chip comprises a variable gain low-noise amplifier, quadrature mixers, variable gain amplifiers, and local oscillator generation circuits. The solution is based on very high dynamic range front-end blocks, a low-power superharmonic injection-locking technique for quadrature generation and continuous-time dc offset removal. Measured performances are an overall gain variable between 21 and 47 dB, 5.6 dB noise figure, -2 dBm out-of-band IIP3, -10 dBm in-band IIP3, 44.8-dBm minimum IIP2, and -155-dBc/Hz phase noise at 135 MHz from carrier frequency, while drawing 21 mA from a 1.8-V supply.     

Quantitative analysis of mycotoxins in cereal foods by collision cell fragmentation-high-resolution mass spectrometry: performance and comparison with triple-stage quadrupole detection

A liquid chromatography-high-resolution mass spectrometry (LC-HRMS) method for the simultaneous determination of aflatoxins (B(1), B(2), G(1), G(2)), ochratoxin A, deoxynivalenol, zearalenone, T-2 and HT-2 toxins in wheat flour, barley flour and crisp bread was developed. Mycotoxin fragmentation patterns obtained by high-energy collision dissociation (HCD) were investigated to obtain quantitative and confirmatory information (two characteristic masses per mycotoxin) using Orbitrap?-based high-resolution mass spectrometry. LC-HRMS (full-scan) detection carried out by HCD allows the monitoring of the pseudo-molecular ion and an additional characteristic fragment (for each mycotoxin) with mass accuracy in the range 0.1-3.9?ppm, meeting current European regulatory requirements for LC-MS confirmatory analysis. A sample preparation procedure based on polymeric solid-phase extraction cartridges was applied, allowing recoveries higher than 74% for nine mycotoxins, with a relative standard deviation lower than 13%. Detection limits in the range 0.5-3.4?μg?kg(-1) were obtained for three cereal matrices. A critical comparison between the proposed method and a validated method based on triple quadrupole mass spectrometry showed similar performance in terms of detection limits, recoveries and repeatability, and matrix effects. Based on an efficient sample extraction and clean-up, the LC-HCD-HRMS method reported here represents a reliable and robust alternative tool for mycotoxin analysis in food matrices as compared with well-established triple quadrupole-based approaches.     

Speed-up Technique for a Local Automatic Colour Equalization Model

In this paper we propose a speed‐up technique for a local automatic colour equalization operator derived from a model of the human vision system. This method is characterized by local and global filtering effects, that simultaneously achieve different equalization tasks e.g. performing colour and lightness constancy, realizing dynamic image data driven stretching, controlling the contrast. We describe a way to quickly create a filtering mapping function to perform the global component of the mapping. This method is based on singular value decomposition (SVD) applied to sampled and filtered points in the input image. Then, the local information is added computing the basic algorithm on a neighbourhood of each input pixel. A slight quality loss is the price that we have to pay for a speed‐up of more than two orders of magnitude of the basic algorithm. We present the results on several images and discuss the efficiency and the drawbacks of the speed‐up technique.     

A multiscale framework for spatial gamut mapping.

Image reproduction devices, such as displays or printers, can reproduce only a limited set of colors, denoted the color gamut. The gamut depends on both theoretical and technical limitations. Reproduction device gamuts are significantly different from acquisition device gamuts. These facts raise the problem of reproducing similar color images across different devices. This is well known as the gamut mapping problem. Gamut mapping algorithms have been developed mainly using colorimetric pixel-wise principles, without considering the spatial properties of the image. The recently proposed multilevel gamut mapping approach takes spatial properties into account and has been demonstrated to outperform spatially invariant approaches. However, they have some important drawbacks. To analyze these drawbacks, we build a common framework that encompasses at least two important previous multilevel gamut mapping algorithms. Then, when the causes of the drawbacks are understood, we solve the typical problem of possible hue shifts. Next, we design appropriate operators and functions to strongly reduce both haloing and possible undesired over compression. We use challenging synthetic images, as well as real photographs, to practically show that the improvements give the expected results.     

Short-term effects of transcatheter arterial chemoembolisation on metabolic activity of the liver of cirrhotic patients with hepatocellular carcinoma

BACKGROUND: Transcatheter arterial chemoembolisation, a procedure for the treatment of hepatocellular carcinoma, provokes a pronounced but transient increase in hepatic cytolysis parameters. A definite evaluation of the impairment of liver function after this treatment, performed by adequate techniques, is still lacking. AIMS: To assess and quantify the impairment of liver metabolic activity after arterial chemoembolisation in patients with cirrhosis. The variations of hepatic vein pressure gradient provoked by this procedure were evaluated. PATIENTS: 15 patients with cirrhosis (Child's class A and B) and hepatocellular carcinoma. METHODS: 17 transcatheter arterial chemoembolisations with epirubicin, iodised oil, and gelfoam were performed; liver function was assessed before, the following day, and after seven days measuring galactose elimination capacity; aminopyrine breath test was also performed in six patients before the procedure and seven days after. In 10 patients intrinsic hepatic clearance of indocyanine green and hepatic vein pressure gradient were measured by hepatic vein catheterisation before and 30 minutes after chemoembolisation. RESULTS: Intrinsic hepatic clearance of indocyanine green decreased significantly from (mean (SEM)) 355 (140) ml/min to 277 (98) ml/min after the procedure (p = 0.0007). Galactose elimination capacity did not show significant changes, being 4.00 (0.90) mg/min/kg body weight at baseline, 4.20 (0.90) mg/min/kg body weight after one day, and 3.95 (0.87) mg/min/kg body weight seven days after chemoembolisation. Aminopyrine breath test was 2.31 (1.09)% and remained unchanged after treatment, being 2.39 (2.04)% at day 7. Baseline hepatic vein pressure gradient was 17.0 (5.5) mm Hg, and 14.4 (3.7) mm Hg 30 minutes after chemoembolisation (p = 0.09). CONCLUSIONS: A single transcatheter chemoembolisation in cirrhotic patients was detected by galactose elimination capacity and aminopyrine breath test one and seven days after the procedure. Therefore it can be considered a safe therapeutic tool for hepatocellular carcinoma in Child's class A and B cirrhotic patients.