Optimizing 3D Convolution Kernels on Stereo Matching for Resource Efficient Computations.

Despite recent stereo matching algorithms achieving significant results on public benchmarks, the problem of requiring heavy computation remains unsolved. Most works focus on designing an architecture to reduce the computational complexity, while we take aim at optimizing 3D convolution kernels on the Pyramid Stereo Matching Network (PSMNet) for solving the problem. In this paper, we design a series of comparative experiments exploring the performance of well-known convolution kernels on PSMNet. Our model saves the computational complexity from 256.66 G MAdd (Multiply-Add operations) to 69.03 G MAdd (198.47 G MAdd to 10.84 G MAdd for only considering 3D convolutional neural networks) without losing accuracy. On Scene Flow and KITTI 2015 datasets, our model achieves results comparable to the state-of-the-art with a low computational cost.

[Impact of Respiratory Waveform on Gate Signal Generation on Respiratory Gated Irradiation and Evaluation Method of Respiratory Waveform to Be Used].

PURPOSE: The respiratory gated irradiation using the real-time position management system (RPM) was used to clarify the generation of the gated signal when the respiration waveform changed, and also the evaluation method of the respiration waveform was also examined. METHODS: The respiratory waveform was changed using a moving phantom. Respiratory waveform was analyzed from the data recorded in RPM, and the out-of-phase gated rate was examined. Analysis was made by focusing on the coefficient of variation of the respiratory wavelength in the evaluation of respiratory waveform. RESULTS: Immediately after the change of respiratory wavelength from the short cycle to the long cycle, a gated signal was generated at a phase before the set gated phase, and a maximum advance of 1.259 ± 0.212 s occurred. Immediately after the change of respiratory wavelength from the long cycle to the short cycle, the gated signal was generated at the phase exceeding the set gated phase, and a delay of 0.997 ± 0.180 s occurred at the maximum. As the value of the coefficient of variation increased, the gated rate which was out of setting also increased. CONCLUSION: In respiratory gated irradiation using RPM, it became clear that the gated signal is generated out of the phase set by the respiratory waveform change. Coefficient of variation of the respiratory wavelength is considered to be an indicator for evaluating the respiratory waveform to be used in the respiratory gated irradiation.

The Pseudomonas aeruginosa dnaK gene is involved in bacterial translocation across the intestinal epithelial cell barrier.

Pseudomonas aeruginosa can penetrate through polarized epithelial cell monolayers produced by the human adenocarcinoma cell line Caco-2. We previously identified genes associated with bacterial translocation through Caco-2 cell monolayers by analysing transposon insertion mutants with dramatically reduced penetration activity relative to that of the wild-type P. aeruginosa PAO1 strain. In this study, we focused on the dnaK mutant because the association between this gene and penetration activity is unknown. Inactivation of dnaK caused significant repression of bacterial penetration through Caco-2 cell monolayers, with decreased swimming, swarming and twitching motilities; bacterial adherence; and fly mortality rate; as well as dramatic repression of type III effector secretion and production of elastase and exotoxin A. However, type IV pilus protein PilA expression was not affected. These results suggest that dnaK is associated with bacterial motility and adherence, which are mediated by flagella and pili, and with toxin secretion, which plays a key role in the penetration of P. aeruginosa through Caco-2 cell monolayers. Inactivation of P. aeruginosa dnaK function may interfere with bacterial translocation and prevent septicaemia caused by P. aeruginosa.

Dysfunction of neurogenic VIP-mediated relaxation in mouse distal colon with dextran sulfate sodium-induced colitis.

Vasoactive intestinal peptide (VIP) regulates various functions including motility and immune homeostasis in colon. The VIP system including its receptors has been established to control the development of ulcerative colitis, but the functional changes of the system-regulated motility in colon with ulcerative colitis are not well understood. In this study, we investigated VIP-related contractile responses in distal colon from mice with dextran sulfate sodium (DSS)-induced acute colitis. Electrical stimulation (ES) under our conditions caused relaxation during ES and contraction after withdrawal of ES in a tetrodotoxin-sensitive manner. Pharmacological analyses showed two phases of ES-induced relaxation: a transient neuronal nitric oxide (NO) synthase-dependent phase (I), and a continued VIP receptor-mediated phase (II). Inhibition of VIP receptors and protein kinase A decreased both phases. In colon from DSS-treated mice, ES-induced phase II (also phase I) and VIP-induced, but not cyclic AMP analog-induced, relaxation were decreased. Stimulation with VIP significantly increased cyclic AMP formation in colon preparations from control but not DSS-treated mice. In colon from DSS-treated mice, the basal cyclic AMP level was markedly greater without changes in the level of VIP receptor VPAC(2). Isoprenaline- and forskolin-induced relaxation and cyclic AMP formation were not changed by DSS treatment. These findings suggest that dysfunction of VIP receptors in muscles, in addition to loss of the neuronal VIP and NO pathways, are involved in abnormal motility in mouse colon with DSS-induced colitis.

Neurogenic contraction of mouse rectum via the cyclooxygenase pathway: Changes of PGE2-induced contraction with dextran sulfate sodium-induced colitis.

Recent reports suggest that cyclooxygenases (COXs) including COX-2 are constitutively expressed, and prostaglandins (PGs) regulate motility and/or contraction in the colon and rectum. This study examines the role of COXs in the regulation of neuromuscular function in longitudinal preparations of isolated rectum and distal colon (Side A, close to the transverse colon; and Side B, close to the rectum) in normal mice and after the induction of colitis by dextran sulfate sodium (DSS). In control rectum, electrical stimulation (ES)-induced contractions were inhibited by atropine and by COX inhibitors, in an independent manner. PGE(2) at 3microM caused a marked contraction, but the secondary response at 20min after the first application was 60% desensitized. In rectum from DSS-treated mice, spontaneous and ES-induced contractions were significantly less intense than in the control preparations, and the response to PGE(2) was abolished but that to 3microM acetylcholine was not. In control distal colon, the responses to PGE(2) in neither side were desensitized by the repeated application. In DSS-treated distal colon, PGE(2) response was impaired in the two regions, and was desensitized on Side B more than Side A. DSS treatment impaired contractions by 40mM KCl in rectum and on Side B but not Side A. DSS treatment increased COX-2 expression in rectum, but not in distal colon. These findings suggest that the induction of colitis by DSS affects ES- and PGE(2)-regulated motility in the order rectum>distal colon close to the rectum>distal colon in mice.

Pseudomonas aeruginosa serA Gene Is Required for Bacterial Translocation through Caco-2 Cell Monolayers.

To specify critical factors responsible for Pseudomonas aeruginosa penetration through the Caco-2 cell epithelial barrier, we analyzed transposon insertion mutants that demonstrated a dramatic reduction in penetration activity relative to P. aeruginosa PAO1 strain. From these strains, mutations could be grouped into five classes, specifically flagellin-associated genes, pili-associated genes, heat-shock protein genes, genes related to the glycolytic pathway, and biosynthesis-related genes. Of these mutants, we here focused on the serA mutant, as the association between this gene and penetration activity is yet unknown. Inactivation of the serA gene caused significant repression of bacterial penetration through Caco-2 cell monolayers with decreased swimming and swarming motilities, bacterial adherence, and fly mortality rate, as well as repression of ExoS secretion; however, twitching motility was not affected. Furthermore, L-serine, which is known to inhibit the D-3-phosphoglycerate dehydrogenase activity of the SerA protein, caused significant reductions in penetration through Caco-2 cell monolayers, swarming and swimming motilities, bacterial adherence to Caco-2 cells, and virulence in flies in the wild-type P. aeruginosa PAO1 strain. Together, these results suggest that serA is associated with bacterial motility and adherence, which are mediated by flagella that play a key role in the penetration of P. aeruginosa through Caco-2 cell monolayers. Oral administration of L-serine to compromised hosts might have the potential to interfere with bacterial translocation and prevent septicemia caused by P. aeruginosa through inhibition of serA function.

L-Cysteine/D,L-homocysteine-regulated ileum motility via system L and B°(,+) transporter: Modification by inhibitors of hydrogen sulfide synthesis and dietary treatments.

Previous studies including ours demonstrated that L-cysteine treatments decreased motility in gastrointestinal tissues including the ileum via hydrogen sulfide (H2S), which is formed from sulfur-containing amino acids such as L-cysteine and L-homocysteine. However, the amino acid transport systems involved in L-cysteine/L-homocysteine-induced responses have not yet been elucidated in detail; therefore, we investigated these systems pharmacologically by measuring electrical stimulation (ES)-induced contractions with amino acids in mouse ileum preparations. The treatments with L-cysteine and D,L-homocysteine inhibited ES-induced contractions in ileum preparations from fasted mice, and these responses were decreased by the treatment with 2-aminobicyclo[2.2.1]heptane-2-carboxylate (BCH), an inhibitor of systems L and B°(,+). The results obtained using ileum preparations and a model cell line (PC12 cells) with various amino acids and BCH showed that not only L-cysteine, but also aminooxyacetic acid and D,L-propargylglycine, which act as H2S synthesis inhibitors, appeared to be taken up by these preparations/cells in L and B°(,+) system-dependent manners. The L-cysteine and D,L-homocysteine responses were delayed and abolished, respectively, in ileum preparations from fed mice. Our results suggested that the regulation of ileum motility by L-cysteine and D,L-homocysteine was dependent on BCH-sensitive systems, and varied depending on feeding in mice. Therefore, the effects of aminooxyacetic acid and D,L-propargylglycine on transport systems need to be considered in pharmacological analyses.

Hydrogen sulfide-mediated regulation of contractility in the mouse ileum with electrical stimulation: roles of L-cysteine, cystathionine ß-synthase, and K+ channels.

Hydrogen sulfide (H2S) is considered to be a signaling molecule. The precise mechanisms underlying H2S-related events, including the producing enzymes and target molecules in gastrointestinal tissues, have not been elucidated in detail. We herein examined the involvement of H2S in contractions induced by repeated electrical stimulations (ES). ES-induced contractions were neurotoxin-sensitive and increased by aminooxyacetic acid, an inhibitor of cystathionine ß-synthase (CBS) and cystathionine γ-lyase, but not by D,L-propargylglycine, a selective inhibitor of cystathionine γ-lyase, in an ES trial-dependent manner. ES-induced contractions were markedly decreased in the presence of L-cysteine. This response was inhibited by aminooxyacetic acid and an antioxidant, and accelerated by L-methionine, an activator of CBS. The existence of CBS was confirmed. NaHS transiently inhibited ES- and acetylcholine-induced contractions, and sustainably decreased basal tone for at least 20 min after its addition. The treatment with glibenclamide, an ATP-sensitive K+ channel blocker, reduced both the L-cysteine response and NaHS-induced inhibition of contractions. The NaHS-induced decrease in basal tone was inhibited by apamin, a small conductance Ca2+-activated K+ channel blocker. These results suggest that H2S may be endogenously produced via CBS in ES-activated enteric neurons, and regulates contractility via multiple K+ channels in the ileum.

Nicotine-induced neurogenic relaxation in the mouse colon: changes with dextran sodium sulfate-induced colitis.

Nicotine has been shown to reduce both tone and muscular activity in the human colon by releasing nitric oxide (NO) from nerves. To our knowledge, however, the effect of nicotine on mouse colon has not been elucidated, and the response in tissue from ulcerative colitis (UC) has not been investigated. We examined nicotine-induced responses in colon from control mice and mice with dextran sodium sulfate (DSS)-induced UC. In controls, bath application of nicotine caused a transient relaxation in longitudinal preparations from the transverse and distal colons but not from the rectum. The response was observed in the presence of bethanechol, abolished by treatment with tetrodotoxin and hexamethonium, and mediated partially (>50%) by the NO pathway. In longitudinal preparations of the distal colon from DSS-treated mice, spontaneous contractions decreased markedly, and nicotine caused contraction without relaxation in half of the preparations tested. Nicotine-induced relaxation in the presence of bethanechol was significantly decreased in the DSS-treated distal colon without changing bethanechol-induced contractions. These data suggest that 1) responses to nicotine differ dependent on colon regions, 2) DSS treatment predominantly caused nicotine-sensitive neurogenic changes in distal colon, and 3) DSS treatment may reverse the direction of nicotine-evoked responses in the colon, in mice.

Si(111) surface modified with alpha,beta-unsaturated carboxyl groups studied by MIR-FTIR.

A Si(111) surface modified with alpha,beta-unsaturated carboxyl groups was fabricated using activated alkynes such as propiolic acid and propiolic acid methyl ester via hydrosilylation reaction. The obtained coverage of carboxyl groups was roughly estimated to be 55-60% in both cases from the Si-2p and C-1s X-ray photoelectron specroscopy (XPS) peak intensities. The detailed surface structures were investigated by multiple internal reflection Fourier transform infrared (MIR-FTIR) measurement. It was revealed that this reaction was promoted by visible light irradiation at room temperature. The Si surface modified with functional groups prepared under such a moderate condition is adaptable to functional devices which are easily damaged under UV irradiation or high temperature conditions.