FASSVid: Fast and Accurate Semantic Segmentation for Video Sequences

Most of the methods for real-time semantic segmentation do not take into account temporal information when working with video sequences. This is counter-intuitive in real-world scenarios where the main application of such methods is, precisely, being able to process frame sequences as quickly and accurately as possible. In this paper, we address this problem by exploiting the temporal information provided by previous frames of the video stream. Our method leverages a previous input frame as well as the previous output of the network to enhance the prediction accuracy of the current input frame. We develop a module that obtains feature maps rich in change information. Additionally, we incorporate the previous output of the network into all the decoder stages as a way of increasing the attention given to relevant features. Finally, to properly train and evaluate our methods, we introduce CityscapesVid, a dataset specifically designed to benchmark semantic video segmentation networks. Our proposed network, entitled FASSVid improves the mIoU accuracy performance over a standard non-sequential baseline model. Moreover, FASSVid obtains state-of-the-art inference speed and competitive mIoU results compared to other state-of-the-art lightweight networks, with significantly lower number of computations. Specifically, we obtain 71% of mIoU in our CityscapesVid dataset, running at 114.9 FPS on a single NVIDIA GTX 1080Ti and 31 FPS on the NVIDIA Jetson Nano embedded board with images of size 1024×2048 and 512×1024, respectively.     

Identification of a quenching species in ruthenium tris-bipyridine electroluminescent devices

We have used matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry and micro-Raman spectroscopy to identify a quenching species that is formed during operation of [Ru(bpy)3]2+ electroluminescent devices. We identify this performance-degrading product to be the oxo-bridged dimer [(bpy)2(H2O)RuORu(OH2)(bpy)2]4+ and show this dimer to be an effective quencher of device luminescence. This work is the first to detect a specific chemical degradation product formed during iTMC OLED operation.     

Evaluation of damage induced by high irradiation levels on α-Ni-Ni3Si eutectic structure

Diluted alloys of the binary system Ni–Si have been used as target of beam of ions, electrons, neutrons and so on because in this kind of alloy occurs transformations order–disorder, when the temperature is raised. This fact has permitted to evaluate the phenomena associated with the damage induced by irradiation (DII). The results of these works have been employed to understand the behavior under irradiation of complex alloys and to evaluate the reliability of the results of mathematical simulation of the evolution of the DII. The interest in the alloy system Ni-Si has been reborn due to the necessity of developing materials, which have better resistance against the corrosion on more aggressive environments such as those generated on the nuclear power plants or those that exist out of the Earth's atmosphere. Now, a growing interest to use concentrated alloys of this binary system on diverse fields of the materials science has been taking place because up to determined concentration of silicon, a regular eutectic is formed, and this fact opens the possibility to develop lamellar composite material by directional solidification. However, nowadays, there is a lack of fundamental knowledge about the behavior of this type of lamellar structure under aggressive environments, like those mentioned before. Hence, the task of this work is to evaluate the effect that has the irradiation over the microstructure of the concentrated alloy Ni22at%Si. The dendritic region of the hypereutectic alloy consists of an intermetallic phase Ni₃Si, whereas the interdendritic region is formed by the alternation of lamellas of solid solution α-Ni and intermetallic phase Ni₃Si. Such kind of microstructure has the advantage to get information of the DII over different phases individually, and at the same time, about of the microstructure influence over the global damage in the alloy. The hypereutectic Ni22at%Si alloy was irradiated perpendicularly to its surface, with 3.66?MeV – Ni ions up to 380?dpa at 650°C in a Tandetron linear accelerator. The level of irradiation dose was chosen similar to the irradiation conditions of the next-generation nuclear reactors. The theoretical maximum depth of the DII (maximum depth of damage (MDD)) was calculated as 1.35?µm using the SRIM-2013 program; the laminar microstructure of the eutectic was simulated using the lattice parameters of the eutectic before irradiation. The experimental MDD was 1.47?µm, as determined through transmission electron microscope (TEM) images and the DII was characterized using µX-ray diffraction and TEM. The elimination of cubic phase of the intermetallic Ni₃Si, the suppression of lamellae of the α-Ni phase, the generation of dislocation loops and lines, all of these changes generated by the irradiation are clear evidences that the DII was severe. Based on theoretical and experimental evidence, we propose that the amount of phases, alternate of lamellae with different chemical concentrations of silicon and lamellae spatial distribution have a direct relation with the severe evolution of the DII.     

W6S8] octahedral tungsten clusters functionalized with thiophene derivatives: toward polymerizable building blocks

The functionalization of octahedral [W(6)S(8)] clusters with a family of phosphino-thiophene ligands has been investigated with the goal of synthesizing extended networks of [W(6)S(8)] units covalently linked to one another through thiophene-conjugated bridges. In addition to new phosphino-thiophene ligands, eight clusters were synthesized and characterized by (1)H and (31)P NMR spectroscopies, elemental analysis, and UV-vis absorption. These clusters are formulated [W(6)S(8)(T-PPh(2))6] (1a), [W(6)S(8)(T-PEt(2))(6)] (1b), [W(6)S(8)(2T-PPh(2))(6)] (2a), [W(6)S(8)(2T-PEt(2))(6)] (2b), [W(6)S(8)(3T-PPh(2))(6)] (3a), [W(6)S(8)(3T-PEt(2))(6)] (3b), [W(6)S(8)((2T)(3)P)(6)] (4), and [W(6)S(8)(2EDOT-PEt(2))(6)] (5) (T = thiophene and EDOT = 3,4-ethylenedioxythiophene). The molecular structure of six of them has been obtained by single-crystal X-ray diffraction analysis. All of them crystallize in the P1 triclinic space group except 3b, which has the P2(1)/c monoclinic symmetry. The redox behavior of both the ligands and the corresponding functionalized clusters has been investigated by cyclic voltammetry. An attempt to electropolymerize these species is also reported.     

A Versatile Precursor System for Supercritical Fluid Electrodeposition of Main‐Group Materials

For the first time, a versatile electrolyte bath is described that can be used to electrodeposit a wide range of p‐block elements from supercritical difluoromethane (scCH₂F₂). The bath comprises the tetrabutylammonium chlorometallate complex of the element in an electrolyte of 50×10^?3 mol dm^?3 tetrabutylammonium chloride at 17.2 MPa and 358 K. Through the use of anionic ([GaCl₄]^?, [InCl₄]^?, [GeCl₃]^?, [SnCl₃]^?, [SbCl₄]^?, and [BiCl₄]^?) and dianionic ([SeCl₆]^2? and [TeCl₆]^2?) chlorometallate salts, the deposition of elemental Ga, In, Ge, Sn, Sb, Bi, Se, and Te is demonstrated. In all cases, with the exception of gallium, which is a liquid under the deposition conditions, the resulting deposits are characterised by SEM, energy‐dispersive X‐ray analysis, XRD and Raman spectroscopy. An advantage of this electrolyte system is that the reagents are all crystalline solids, reasonably easy to handle and not highly water or oxygen sensitive. The results presented herein significantly broaden the range of materials accessible by electrodeposition from supercritical fluid and open up the future possibility of utilising the full scope of these unique fluids to electrodeposit functional binary or ternary alloys and compounds of these elements.     

Novel rearrangement of N-enoyl oxazolidinethiones to N-substituted 1,3-thiazine-2,4-diones promoted by NbCl5

NbCl₅ has been employed as promoter of a novel rearrangement to afford chiral N-substituted 1,3-thiazine-2,4-diones with one or two new stereogenic centers from di- and trisubstituted N-enoyl oxazolidinethiones. The trisubstituted E-isomers provide the anti-diastereomers mainly.     

Stereoselective Michael addition of 6-amino-1,3-dimethyl-2,4-pyrimidinedione to the exocyclic methylene of three sesquiterpene lactones. 1H and 13C NMR evidence of a new C-C bond and lactam formation

The Stereoselective addition of the pyrimidine derivative 1 to the exocyclic methylene of the α,β unsaturated dehydrocostus lactone 2, Ivalin acetate 3 and Zaluzanin A diacetate 4, was achieved resulting in a new C-C bond formation. In the cases of compounds 3 and 4, after the addition, the lactone was cleaved followed by reclosure into a lactam ring system.     

Machine Learning Models and Statistical Complexity to Analyze the Effects of Posture on Cerebral Hemodynamics

The mechanism of cerebral blood flow autoregulation can be of great importance in diagnosing and controlling a diversity of cerebrovascular pathologies such as vascular dementia, brain injury, and neurodegenerative diseases. To assess it, there are several methods that use changing postures, such as sit-stand or squat-stand maneuvers. However, the evaluation of the dynamic cerebral blood flow autoregulation (dCA) in these postures has not been adequately studied using more complex models, such as non-linear ones. Moreover, dCA can be considered part of a more complex mechanism called cerebral hemodynamics, where others (CO₂ reactivity and neurovascular-coupling) that affect cerebral blood flow (BF) are included. In this work, we analyzed postural influences using non-linear machine learning models of dCA and studied characteristics of cerebral hemodynamics under statistical complexity using eighteen young adult subjects, aged 27 ± 6.29 years, who took the systemic or arterial blood pressure (BP) and cerebral blood flow velocity (BFV) for five minutes in three different postures: stand, sit, and lay. With models of a Support Vector Machine (SVM) through time, we used an AutoRegulatory Index (ARI) to compare the dCA in different postures. Using wavelet entropy, we estimated the statistical complexity of BFV for three postures. Repeated measures ANOVA showed that only the complexity of lay-sit had significant differences.