Recent years have witnessed the popularity of using recurrent neural network (RNN) for action recognition in videos. However, videos are of high dimensionality and contain rich human dynamics with ...various motion scales, which makes the traditional RNNs difficult to capture complex action information. In this paper, we propose a novel recurrent spatial-temporal attention network (RSTAN) to address this challenge, where we introduce a spatial-temporal attention mechanism to adaptively identify key features from the global video context for every time-step prediction of RNN. More specifically, we make three main contributions from the following aspects. First, we reinforce the classical long short-term memory (LSTM) with a novel spatial-temporal attention module. At each time step, our module can automatically learn a spatial-temporal action representation from all sampled video frames, which is compact and highly relevant to the prediction at the current step. Second, we design an attention-driven appearance-motion fusion strategy to integrate appearance and motion LSTMs into a unified framework, where LSTMs with their spatial-temporal attention modules in two streams can be jointly trained in an end-to-end fashion. Third, we develop actor-attention regularization for RSTAN, which can guide our attention mechanism to focus on the important action regions around actors. We evaluate the proposed RSTAN on the benchmark UCF101, HMDB51 and JHMDB data sets. The experimental results show that, our RSTAN outperforms other recent RNN-based approaches on UCF101 and HMDB51 as well as achieves the state-of-the-art on JHMDB.
Recent studies witnessed that context features can significantly improve the performance of deep semantic segmentation networks. Current context based segmentation methods differ with each other in ...how to construct context features and perform differently in practice. This paper firstly introduces three desirable properties of context features in segmentation task. Specially, we find that Global-guided Local Affinity (GLA) can play a vital role in constructing effective context features, while this property has been largely ignored in previous works. Based on this analysis, this paper proposes Adaptive Pyramid Context Network (APCNet) for semantic segmentation. APCNet adaptively constructs multi-scale contextual representations with multiple well-designed Adaptive Context Modules (ACMs). Specifically, each ACM leverages a global image representation as a guidance to estimate the local affinity coefficients for each sub-region, and then calculates a context vector with these affinities. We empirically evaluate our APCNet on three semantic segmentation and scene parsing datasets, including PASCAL VOC 2012, Pascal-Context, and ADE20K dataset. Experimental results show that APCNet achieves state-of-the-art performance on all three benchmarks, and obtains a new record 84.2% on PASCAL VOC 2012 test set without MS COCO pre-trained and any post-processing.
Triclocarban (TCC), one typical antibacterial agent being widely used in various applications, was found to be present in waste activated sludge at significant levels. To date, however, its effect on ...anaerobic fermentation of sludge has not been investigated. This work therefore aims to fill this knowledge gap. Experimental results showed that when TCC content in sludge increased from 26.7 ± 5.3 to 520.5 ± 12.6 mg per kilogram total suspended solids, the maximum concentration of short-chain fatty acids (SCFA) increased from 32.6 ± 2.5 to 228.2 ± 3.6 (without pH control) and from 211.7 ± 2.4 to 378.3 ± 3.2 mg COD/g VSS (initial pH 10), respectively. The large promotion of acetic acid was found to be the major reason for the enhancement of total SCFA production. Although a significant level of TCC was degraded in the fermentation process, SCFA was neither produced from TCC nor affected by its major intermediates at the relevant levels. It was found that TCC facilitated solubilization, acidogenesis, acetogenesis, and homoacetogenesis processes but inhibited methanogenesis process. Microbial analysis revealed that the increase of TCC increased the microbial community diversity, the abundances of SCFA (especially acetic acid) producers, and the activities of key enzymes relevant to acetic acid production.
Display omitted
•Triclocarban affected SCFA production from anaerobic fermentation of sludge.•A significant level of triclocarban was degraded in the fermentation process.•Triclocarban facilitated solubilization, acidogenesis, acetogenesis, and homoacetogenesis processes.•Triclocarban inhibited methanogenesis process.
SARS-CoV-2 has caused tens of thousands of infections and more than one thousand deaths. There are currently no registered therapies for treating coronavirus infections. Because of time consuming ...process of new drug development, drug repositioning may be the only solution to the epidemic of sudden infectious diseases. We systematically analyzed all the proteins encoded by SARS-CoV-2 genes, compared them with proteins from other coronaviruses, predicted their structures, and built 19 structures that could be done by homology modeling. By performing target-based virtual ligand screening, a total of 21 targets (including two human targets) were screened against compound libraries including ZINC drug database and our own database of natural products. Structure and screening results of important targets such as 3-chymotrypsin-like protease (3CLpro), Spike, RNA-dependent RNA polymerase (RdRp), and papain like protease (PLpro) were discussed in detail. In addition, a database of 78 commonly used anti-viral drugs including those currently on the market and undergoing clinical trials for SARS-CoV-2 was constructed. Possible targets of these compounds and potential drugs acting on a certain target were predicted. This study will provide new lead compounds and targets for further in vitro and in vivo studies of SARS-CoV-2, new insights for those drugs currently ongoing clinical studies, and also possible new strategies for drug repositioning to treat SARS-CoV-2 infections.
Twenty structures including 19 SARS-CoV-2 targets and one human target were built by homology modeling. Library of ZINC drug database, natural products, 78 anti-viral drugs were screened against these targets plus human ACE2. This study provides drug repositioning candidates and targets for further in vitro and in vivo studies of SARS-CoV-2. Display omitted
Denitrifying anaerobic methane oxidation (DAMO) can concurrently reduce methane emissions and nitrogen levels in aquatic environments, but how useful is this process? We propose the use of DAMO-based ...technology as a tool for sustainably operating wastewater treatment plants (WWTPs).
Nanotechnology has been extensively studied and exploited for cancer treatment as nanoparticles can play a significant role as a drug delivery system. Compared to conventional drugs, ...nanoparticle-based drug delivery has specific advantages, such as improved stability and biocompatibility, enhanced permeability and retention effect, and precise targeting. The application and development of hybrid nanoparticles, which incorporates the combined properties of different nanoparticles, has led this type of drug-carrier system to the next level. In addition, nanoparticle-based drug delivery systems have been shown to play a role in overcoming cancer-related drug resistance. The mechanisms of cancer drug resistance include overexpression of drug efflux transporters, defective apoptotic pathways, and hypoxic environment. Nanoparticles targeting these mechanisms can lead to an improvement in the reversal of multidrug resistance. Furthermore, as more tumor drug resistance mechanisms are revealed, nanoparticles are increasingly being developed to target these mechanisms. Moreover, scientists have recently started to investigate the role of nanoparticles in immunotherapy, which plays a more important role in cancer treatment. In this review, we discuss the roles of nanoparticles and hybrid nanoparticles for drug delivery in chemotherapy, targeted therapy, and immunotherapy and describe the targeting mechanism of nanoparticle-based drug delivery as well as its function on reversing drug resistance.
Recent studies demonstrate the effectiveness of Recurrent Neural Networks (RNNs) for action recognition in videos. However, previous works mainly utilize video-level category as supervision to train ...RNNs, which may prohibit RNNs to learn complex motion structures along time. In this paper, we propose a recurrent pose-attention network (RPAN) to address this challenge, where we introduce a novel pose-attention mechanism to adaptively learn pose-related features at every time-step action prediction of RNNs. More specifically, we make three main contributions in this paper. Firstly, unlike previous works on pose-related action recognition, our RPAN is an end-to-end recurrent network which can exploit important spatial-temporal evolutions of human pose to assist action recognition in a unified framework. Secondly, instead of learning individual human-joint features separately, our pose-attention mechanism learns robust human-part features by sharing attention parameters partially on the semantically-related human joints. These human-part features are then fed into the human-part pooling layer to construct a highly-discriminative pose-related representation for temporal action modeling. Thirdly, one important byproduct of our RPAN is pose estimation in videos, which can be used for coarse pose annotation in action videos. We evaluate the proposed RPAN quantitatively and qualitatively on two popular benchmarks, i.e., Sub-JHMDB and PennAction. Experimental results show that RPAN outperforms the recent state-of-the-art methods on these challenging datasets.
Real-time detection of frozen meat freshness without thawing is important. This study investigates inspection of frozen pork quality attributes without thawing using fluorescence hyperspectral ...imaging (HSI). Partial least squares regression (PLSR) models were developed based on fluorescence spectra for total volatile basic nitrogen (TVB-N), pH, L*, a*, and b*, and compared with PLSR models based on visible/near-infrared (Vis/NIR) HSI of the same samples. Competitive adaptive reweighted sampling was used to select key fluorescence wavelengths related to each indicator. The correlation coefficients of prediction (Rp) of the models established by fluorescence spectra, with optimal pre-treatment for TVB-N, pH, L*, a*, and b*, were 0.9447, 0.9037, 0.6602, 0.8686, and 0.8699, respectively. Except for L*, fluorescence HSI-based model performance was better than that of Vis-NIR HSI. Model performance was further improved using selected key wavelengths. Results demonstrated that fluorescence HSI could determine freshness indicators of frozen pork without thawing.
•Fluorescence HSI was used for the first time to assess frozen pork freshness.•Relationships between fluorescence peaks and freshness indicators were recognized.•PLSR models were compared based on fluorescence HSI and Vis/NIR HSI.•Key wavelengths were selected for each freshness indicators of frozen pork.
Inspired by the success of dual-targeting drugs, especially bispecific antibodies, we propose to combine the concept of proteolysis targeting chimera (PROTAC) and dual targeting to design and ...synthesize dual PROTAC molecules with the function of degrading two completely different types of targets simultaneously. A library of novel dual-targeting PROTAC molecules has been rationally designed and prepared. A convergent synthetic strategy has been utilized to achieve high synthetic efficiency. These dual PROTAC structures are characterized using trifunctional natural amino acids as star-type core linkers to connect two independent inhibitors and E3 ligands together. In this study, gefitinib, olaparib, and CRBN or VHL E3 ligands were used as substrates to synthesize novel dual PROTACs. They successfully degraded both the epidermal growth factor receptor (EGFR) and poly(ADP-ribose) polymerase (PARP) simultaneously in cancer cells. Being the first successful example of dual PROTACs, this technique will greatly widen the range of application of the PROTAC method and open up a new field for drug discovery.
A novel series of clioquinol-moracin hybrids were designed and synthesized by fusing the pharmacophores of clioquinol and moracin M, and their activities as multitarget-directed ligands against ...Alzheimer’s disease were evaluated. Biological activity results demonstrated that these hybrids possessed significant inhibitory activities against phosphodiesterase 4D (PDE4D) and Aβ aggregation as well as remarkable antioxidant effects and excellent blood–brain barrier permeability. The optimal compound, 18d (WBQ5187), exhibited excellent PDE4D inhibitory potency (IC50 = 0.32 μM), significant antioxidant effects, appropriate biometal chelating functions, and interesting properties that modulated self- and metal-induced Aβ aggregation. Two-dimensional NMR studies revealed that 18d had significant interactions with Aβ1–42 at the R5, H6, H14, Q15, and F20 residues. Furthermore, this typical hybrid possessed preeminent neuroprotective effects against inflammation in microglial cells. Most importantly, oral administration of 18d·HCl demonstrated marked improvements in cognitive and spatial memory in a rat model of Alzheimer’s disease and protected hippocampal neurons from necrosis.