Lanzhou Jiaotong University

Abstract Deep learning has been widely used for medical image segmentation and a large number of papers has been presented recording the success of deep learning in the field. A comprehensive thematic survey on medical image segmentation using deep learning techniques is presented. This paper makes two original contributions. Firstly, compared to traditional surveys that directly divide literatures of deep learning on medical image segmentation into many groups and introduce literatures in detail for each group, we classify currently popular literatures according to a multi‐level structure from coarse to fine. Secondly, this paper focuses on supervised and weakly supervised learning approaches, without including unsupervised approaches since they have been introduced in many old surveys and they are not popular currently. For supervised learning approaches, we analyse literatures in three aspects: the selection of backbone networks, the design of network blocks, and the improvement of loss functions. For weakly supervised learning approaches, we investigate literature according to data augmentation, transfer learning, and interactive segmentation, separately. Compared to existing surveys, this survey classifies the literatures very differently from before and is more convenient for readers to understand the relevant rationale and will guide them to think of appropriate improvements in medical image segmentation based on deep learning approaches.

Agriculture in the 21st century is facing multiple challenges, such as those related to soil fertility, climatic fluctuations, environmental degradation, urbanization, and the increase in food demand for the increasing world population. In the meanwhile, the scientific community is facing key challenges in increasing crop production from the existing land base. In this regard, traditional farming has witnessed enhanced per acre crop yields due to irregular and injudicious use of agrochemicals, including pesticides and synthetic fertilizers, but at a substantial environmental cost. Another major concern in modern agriculture is that crop pests are developing pesticide resistance. Therefore, the future of sustainable crop production requires the use of alternative strategies that can enhance crop yields in an environmentally sound manner. The application of rhizobacteria, specifically, plant growth-promoting rhizobacteria (PGPR), as an alternative to chemical pesticides has gained much attention from the scientific community. These rhizobacteria harbor a number of mechanisms through which they promote plant growth, control plant pests, and induce resistance to various abiotic stresses. This review presents a comprehensive overview of the mechanisms of rhizobacteria involved in plant growth promotion, biocontrol of pests, and bioremediation of contaminated soils. It also focuses on the effects of PGPR inoculation on plant growth survival under environmental stress. Furthermore, the pros and cons of rhizobacterial application along with future directions for the sustainable use of rhizobacteria in agriculture are discussed in depth.

Abstract Involving eight electron transfer process and multiple intermediates of nitrate (NO 3 − ) reduction reaction leads to a sluggish kinetic and low Faradaic efficiency, therefore, it is essential to get an insight into the reaction mechanism to develop highly efficient electrocatalyst. Herein, a series of reduced‐graphene‐oxide‐supported RuCu alloy catalysts (Ru x Cu x /rGO) are fabricated and used for the direct reduction of NO 3 − to NH 3 . It is found that the Ru 1 Cu 10 /rGO shows the ammonia formation rate of 0.38 mmol cm −2 h −1 (loading 1 mg cm −2 ) and the ammonia Faradaic efficiency of 98% under an ultralow potential of −0.05 V versus Reversible Hydrogen Electode (RHE), which is comparable to Ru catalyst. The highly efficient activity of Ru 1 Cu 10 /rGO can be attributed to the synergetic effect between Ru and Cu sites via a relay catalysis, in which the Cu shows the exclusively efficient activity for the reduction of NO 3 − to NO 2 − and Ru exhibits the superior activity for NO 2 − to NH 3 . In addition, the doping of Ru into Cu tunes the d‐band center of alloy and effectively modulates the adsorption energy of the NO 3 − and NO 2 − , which promotes the direct reduction of NO 3 − to NH 3 . This synergetic electrocatalysis strategy opens a new avenue for developing highly efficient multifunctional catalysts.

Abstract We demonstrate the great feasibility of MBenes as a new class of tandem catalysts for electrocatalytic nitrate reduction to ammonia (NO 3 RR). As a proof of concept, FeB 2 is first employed as a model MBene catalyst for the NO 3 RR, showing a maximum NH 3 ‐Faradaic efficiency of 96.8 % with a corresponding NH 3 yield of 25.5 mg h −1 cm −2 at −0.6 V vs. RHE. Mechanistic studies reveal that the exceptional NO 3 RR activity of FeB 2 arises from the tandem catalysis mechanism, that is, B sites activate NO 3 − to form intermediates, while Fe sites dissociate H 2 O and increase *H supply on B sites to promote the intermediate hydrogenation and enhance the NO 3 − ‐to‐NH 3 conversion.

The real-time performance and accuracy of traffic flow prediction directly affect the efficiency of traffic flow guidance systems, and traffic flow prediction is a hotspot in the field of intelligent transportation. To further improve the accuracy of short-term traffic flow prediction, a short-term traffic flow prediction model based on traffic flow time series analysis, and an improved long short-term memory network (LSTM) is proposed. First, perform time series analysis on traffic flow data and perform smoothing and standardization processing to obtain a stable time series as model input data, which can improve the accuracy of model training and eliminate the impact of a wide range of feature values. Then, an improved LSTM model based on LSTM and bidirectional LSTM networks are established. Combining the advantages of sequential data and the long-term dependence of forwarding LSTM and reverse LSTM, the bidirectional long-term memory network (BILSTM) is integrated into the prediction model. The first layer of the LSTM network learns and predicts the input time series and further learns and trains through the bidirectional LSTM network to effectively overcome the large prediction errors. Finally, the performance of the proposed method is evaluated by comparing the predicted results with actual traffic data. The model that is proposed in this paper is compared with the long short-term memory network (LSTM) model and the bidirectional long-term memory network (BILSTM) model. The results demonstrate that the proposed method outperforms both compared methods in terms of accuracy and stability.

This paper provides an updated review on the subjects, the available alternative to produce biochar from biomass, quantification and characterization of biochar, the adsorptive capacity for the adsorption of contaminants, and the effect of biochar addition to agricultural soils on contaminant bioavailability. The property of biochar produced is much dependent upon the composition and type of biomass and the conditions at which biomass is carbonized. The physical and chemical characterizations are necessary to identify the basic structure and property of biochar and to predict its potential in various environmental application. Biochar is a promising alternative to remedy the soils contaminated with heavy metals and organic compounds through adsorption and immobilization due to its large surface area, charged surface, and functional groups. Overall, the bioavailability of heavy metals and organic compounds decreases when biochar is amended into soils.

The electrochemical nitrate reduction to ammonia reaction (NO3RR) has emerged as an appealing route for achieving both wastewater treatment and ammonia production. Herein, sub-nm RuOx clusters anchored on a Pd metallene (RuOx/Pd) are reported as a highly effective NO3RR catalyst, delivering a maximum NH3-Faradaic efficiency of 98.6% with a corresponding NH3 yield rate of 23.5 mg h–1 cm–2 and partial a current density of 296.3 mA cm–2 at −0.5 V vs RHE. Operando spectroscopic characterizations combined with theoretical computations unveil the synergy of RuOx and Pd to enhance the NO3RR energetics through a mechanism of hydrogen spillover and hydrogen-bond interactions. In detail, RuOx activates NO3– to form intermediates, while Pd dissociates H2O to generate *H, which spontaneously migrates to the RuOx/Pd interface via a hydrogen spillover process. Further hydrogen-bond interactions between spillovered *H and intermediates makes spillovered *H desorb from the RuOx/Pd interface and participate in the intermediate hydrogenation, contributing to the enhanced activity of RuOx/Pd for NO3–-to-NH3 conversion.

Abstract Electrochemical reduction of nitrate to ammonia (NO 3 RR) holds a great promise for attaining both NH 3 electrosynthesis and wastewater purification. Herein, single‐atom Bi alloyed Pd metallene (Bi 1 Pd) is reported as a highly effective NO 3 RR catalyst, showing a near 100% NH 3 ‐Faradaic efficiency with the corresponding NH 3 yield of 33.8 mg h −1 cm −2 at −0.6 V versus RHE, surpassing those of almost all ever reported NO 3 RR catalysts. In‐depth theoretical and operando spectroscopic investigations unveil that single‐atom Bi electronically couples with its neighboring Pd atoms to synergistically activate NO 3 − and destabilize *NO on Bi 1 Pd, leading to the reduced energy barrier of the potential‐determining step (*NO→*NOH) and enhanced protonation energetics of NO 3 − ‐to‐NH 3 pathway.

Abstract Graphene nanoplatelets (GNPs) exhibit ultra‐high strength and elastic modulus. Therefore, they are potential ideal reinforcements in metal–matrix composites (MMCs). In this work, we report the use of GNPs to strengthen the bulk Cu‐matrix composites. GNP reinforced Cu‐matrix (GNP/Cu) composites were prepared by a combination of the ball milling and hot‐pressing processing, and their mechanical properties were investigated. Microstructure studies indicated that the GNPs with 0–8 vol.% contents were well dispersed in the Cu matrix by ball milling. Compared to unreinforced Cu, the GNP/Cu composites showed a remarkable increase in yield strength and Young's modulus up to 114 and 37% at 8 vol.% GNP content, respectively. The extraordinary reinforcement is attributed to the homogeneous dispersion of GNPs and grain refinement. However, the mechanical improvement of GNP/Cu composites was still below the theoretical value. The possible reasons for this deviation were discussed and the methods for further mechanical improvement of GNP/Cu composites were proposed.

Abstract The electrochemical N 2 reduction reaction (NRR) offers a promising approach for sustainable NH 3 production, and modulating the structural/electronic configurations of the catalyst materials with optimized electrocatalytic properties is pivotal for achieving high‐efficiency NRR electrocatalysis. Herein, vacancy and heterostructure engineering are rationally integrated to explore O‐vacancy‐rich MoO 3‐ x anchored on Ti 3 C 2 T x ‐MXene (MoO 3‐ x /MXene) as a highly active and selective NRR electrocatalyst, achieving an exceptional NRR activity with an NH 3 yield of 95.8 µg h −1 mg −1 (−0.4 V) and a Faradaic efficiency of 22.3% (−0.3 V). A combination of in situ spectroscopy, molecular dynamics simulations and density functional theory computations is employed to unveil the synergistic effect of O‐vacancies and heterostructures for the NRR, which demonstrates that O‐vacancies on MoO 3‐ x serve as the active sites for N 2 chemisorption and activation, while the MXene substrate can further regulate the O‐vacancy sites to break the scaling relation to effectively stabilize *N 2 /*N 2 H while destabilizing *NH 2 /*NH 3 , resulting in more optimized binding affinity of NRR intermediates toward reduced energy barriers and an enhanced NRR activity for MoO 3‐ x /MXene.

The electrochemical nitrogen reduction reaction (NRR) is a very efficient method for sustainable NH3 production, but it requires effective catalysts to expedite the NRR kinetics and inhibit the concomitant hydrogen evolution reaction (HER). Two-dimensional (2D)/2D interface engineering is an effective method to design powerful catalysts due to intimate face-to-face contact of two 2D materials that facilitates the strong interfacial electronic interactions. Herein, we explored a 2D/2D MoS2/C3N4 heterostructure as an active and stable NRR catalyst. MoS2/C3N4 exhibited a conspicuously improved NRR performance with an NH3 yield of 18.5 μg h–1 mg–1 and a high Faradaic efficiency (FE) of 17.8% at −0.3 V, far better than those of the individual MoS2 or C3N4 component. Density functional theory calculations revealed that the interfacial charge transport from C3N4 to MoS2 could enhance the NRR activity of MoS2/C3N4 by promoting the stabilization of the key intermediate *N2H on Mo edge sites of MoS2 and concurrently decreasing the reaction energy barrier. Meanwhile, MoS2/C3N4 rendered a more favorable *H adsorption free energy on S edge sites than on Mo edge sites of MoS2, thereby protecting the NRR-active Mo edge sites from the competing HER and leading to a high FE.

BACKGROUND: Reactive microglia are associated with β-amyloid (Aβ) deposit and clearance in Alzhiemer's Disease (AD). Paradoxically, entocranial resident microglia fail to trigger an effective phagocytic response to clear Aβ deposits although they mainly exist in an "activated" state. Oligomeric Aβ (oAβ), a recent target in the pathogenesis of AD, can induce more potent neurotoxicity when compared with fibrillar Aβ (fAβ). However, the role of the different Aβ forms in microglial phagocytosis, induction of inflammation and oxidation, and subsequent regulation of phagocytic receptor system, remain unclear. RESULTS: We demonstrated that Aβ(1-42) fibrils, not Aβ(1-42) oligomers, increased the microglial phagocytosis. Intriguingly, the pretreatment of microglia with oAβ(1-42) not only attenuated fAβ(1-42)-triggered classical phagocytic response to fluorescent microspheres but also significantly inhibited phagocytosis of fluorescent labeled fAβ(1-42). Compared with the fAβ(1-42) treatment, the oAβ(1-42) treatment resulted in a rapid and transient increase in interleukin 1β (IL-1β) level and produced higher levels of tumor necrosis factor-α (TNF-α), nitric oxide (NO), prostaglandin E2 (PGE2) and intracellular superoxide anion (SOA). The further results demonstrated that microglial phagocytosis was negatively correlated with inflammatory mediators in this process and that the capacity of phagocytosis in fAβ(1-42)-induced microglia was decreased by IL-1β, lippolysaccharide (LPS) and tert-butyl hydroperoxide (t-BHP). The decreased phagocytosis could be relieved by pyrrolidone dithiocarbamate (PDTC), a nuclear factor-κB (NF-κB) inhibitor, and N-acetyl-L-cysteine (NAC), a free radical scavenger. These results suggest that the oAβ-impaired phagocytosis is mediated through inflammation and oxidative stress-mediated mechanism in microglial cells. Furthermore, oAβ(1-42) stimulation reduced the mRNA expression of CD36, integrin β1 (Itgb1), and Ig receptor FcγRIII, and significantly increased that of formyl peptide receptor 2 (FPR2) and scavenger receptor class B1 (SRB1), compared with the basal level. Interestingly, the pre-stimulation with oAβ(1-42) or the inflammatory and oxidative milieu (IL-1β, LPS or t-BHP) significantly downregulated the fAβ(1-42)-induced mRNA over-expression of CD36, CD47 and Itgb1 receptors in microglial cells. CONCLUSION: These results imply that Aβ oligomers induce a potent inflammatory response and subsequently disturb microglial phagocytosis and clearance of Aβ fibrils, thereby contributing to an initial neurodegenerative characteristic of AD. Antiinflammatory and antioxidative therapies may indeed prove beneficial to delay the progression of AD.

Abstract Urea electrosynthesis from co–electrolysis of NO 3 − and CO 2 (UENC) offers a promising technology for achieving sustainable and efficient urea production. Herein, a diatomic alloy catalyst (CuPd 1 Rh 1 –DAA), with mutually isolated Pd and Rh atoms alloyed on Cu substrate, is theoretically designed and experimentally confirmed to be a highly active and selective UENC catalyst. Combining theoretical computations and operando spectroscopic characterizations reveals the synergistic effect of Pd 1 –Cu and Rh 1 –Cu active sites to promote the UENC via a tandem catalysis mechanism, where Pd 1 –Cu site triggers the early C–N coupling and promotes *CO 2 NO 2 –to–*CO 2 NH steps, while Rh 1 –Cu site facilitates the subsequent protonation step of *CO 2 NH 2 to *COOHNH 2 toward the urea formation. Impressively, CuPd 1 Rh 1 –DAA assembled in a flow cell presents the highest urea Faradaic efficiency of 72.1% and urea yield rate of 53.2 mmol h −1 g cat −1 at −0.5 V versus RHE, representing nearly the highest performance among all reported UENC catalysts.

Microplastics (MPs) alter soil aggregation stability. However, studies have yet to determine whether these alterations further affect microbial community structures and diversities within different soil aggregates and whether they influence the responses of soil microbial structures and diversities to MPs in different aggregate fractions. In this study, long-term soil incubation experiments and soil fractionation were combined to investigate the effects of polyethylene microplastics (PE-MPs) on soil aggregate properties and microbial communities in soil aggregates with different particle sizes. Results showed that the existence of PE-MPs significantly reduced the physicochemical properties of soil aggregates, inhibited the activities of soil enzymes, and changed the richness and diversity of bacterial and fungal communities. Such variations exerted notable differences in soil aggregate levels. The response sensitivity of bacteria in the silt and clay fraction was higher than that in the macroaggregate fraction, but the response sensitivity of fungi in the macroaggregate fraction was higher than that in the silt and clay fraction. Relationships and path analysis between soil aggregate properties and microbial communities after PE-MPs addition were proposed. PE-MPs affected microbial community structures by directly and indirectly influencing soil microenvironmental conditions. The relative abundances of Acidobacteria, Gemmatimonadetes, Bacteroides, Basidiomycota, Chtridiomyota, and Glomeromycota were significantly correlated with physicochemical properties and soil enzyme activities. Enzyme activities were direct factors influencing soil microbial community structures, and physicochemical properties (i.e., dissolved organic carbon, soil available phosphorus) could indirectly affect these structures by acting on soil enzyme activities. Our findings helped improve our understanding of the responses of soil microbial structures and diversities to MPs through the perspective of different soil aggregates.

After going through the deep network, there will be some loss of pedestrian information, which will cause the disappearance of gradients, causing inaccurate pedestrian detection. This paper improves the network structure of YOLO algorithm and proposes a new network structure YOLO-R. First, three Passthrough layers were added to the original YOLO network. The Passthrough layer consists of the Route layer and the Reorg layer. Its role is to connect the shallow layer pedestrian features to the deep layer pedestrian features and link the high and low resolution pedestrian features. The role of the Route layer is to pass the pedestrian characteristic information of the specified layer to the current layer, and then use the Reorg layer to reorganize the feature map so that the currently-introduced Route layer feature can be matched with the feature map of the next layer. The three Passthrough layers added in this algorithm can well transfer the network's shallow pedestrian fine-grained features to the deep network, enabling the network to better learn shallow pedestrian feature information. This paper also changes the layer number of the Passthrough layer connection in the original YOLO algorithm from Layer 16 to Layer 12 to increase the ability of the network to extract the information of the shallow pedestrian features. The improvement was tested on the INRIA pedestrian dataset. The experimental results show that this method can effectively improve the detection accuracy of pedestrians, while reducing the false detection rate and the missed detection rate, and the detection speed can reach 25 frames per second.

Density functional theory calculations revealed that CoO possessed poor HER activity but favorable NRR activity. CoO quantum dots (2–5 nm) supported on graphene exhibited a high NH₃ yield of 21.5 μg h⁻¹ mg⁻¹ and a faradaic efficiency of 8.3% at −0.6 V <italic>vs.</italic> RHE under ambient conditions, superior to most of the reported NRR catalysts.

Abstract Single‐atom alloys hold great promise for electrocatalytic nitrogen reduction reaction (NRR), while the comprehensive experimental/theoretical investigations of SAAs for the NRR are still missing. Herein, PdFe 1 single‐atom alloy metallene, in which the Fe single atoms are confined on a Pd metallene support, is first developed as an effective and robust NRR electrocatalyst, delivering exceptional NRR performance with an NH 3 yield of 111.9 μg h −1 mg −1 , a Faradaic efficiency of 37.8 % at −0.2 V (RHE), as well as a long‐term stability for 100 h electrolysis. In‐depth mechanistic investigations by theoretical computations and operando X‐ray absorption/Raman spectroscopy indentify Pd‐coordinated Fe single atoms as active centers to enable efficient N 2 activation via N 2 ‐to‐Fe σ‐donation, reduced protonation energy barriers, suppressed hydrogen evolution and excellent thermodynamic stability, thus accounting for the high activity, selectivity and stability of PdFe 1 for the NRR.

Fe-doping induced synergetic effects, including the morphological change of crystalline CeO₂ to partial-amorphous nanosheets, enriched O-vacancies and active Ce³⁺–Ce³⁺ pairs, were all responsible for the significantly enhanced NRR activity of Fe-CeO₂.

Abstract Seawater electrolysis under alkaline conditions presents an attractive alternative to traditional freshwater electrolysis for mass sustainable high‐purity hydrogen production. However, the lack of active and robust electrocatalysts severely impedes the industrial application of this technology. Herein, carbon‐doped nanoporous cobalt phosphide (C‐Co 2 P) prepared by electrochemical dealloying is reported as an electrocatalyst for hydrogen evolution reaction (HER). The C‐Co 2 P achieves an overpotential of 30 mV at a current density of 10 mA cm −2 in 1 m KOH, along with impressive catalytic activity and stability at large current densities in artificial alkaline seawater electrolyte containing mixed chlorides of NaCl, MgCl 2 , and CaCl 2 . Experimental analysis and density functional theory calculations reveal that the C atom with strong electronegativity and small atomic radius can tailor the electronic structure of Co 2 P, leading to weakened Co–H bonding toward promoted HER kinetics. Moreover, the C doping introduces a two‐stepped H delivery pathway by forming C–H ad intermediate, thus reducing the energy barrier of water dissociation. This study offers a new vision toward the development of seawater electrolysis for large‐scale hydrogen production.

Sewage sludge pyrolysis into biochar is an effective way to realize sludge harmless and resourceful utilization. The study aimed to investigate the relationship between pyrolysis temperatures (300 °C, 500 °C and 700 °C) and biochar properties and environmental risks associated with heavy metals (Cu, Zn, Pb, Cd, Ni and Cr) and to discuss possible mechanisms of heavy metal immobilization. The results showed that the biochar pH value increased from 5.87 to 10.50 and the specific area grew from 5.26 m2/g to a maximum of 15.23 m2/g with pyrolysis temperature rose. Pyrolysis enhanced aromatization and increased stability of biochar, and part of oxygen-containing functional groups decomposed. Although heavy metals enriched in pyrolysis, high temperatures promoted the bioavailable fractions converted to more stable ones, and the heavy metals leaching rate was reduced by 5.5% on average. This may be related to the functional group decomposition, pore collapse and wrapping, and the new crystalline phases formation. The environmental risk was significantly reduced (RI value below 51.00), indicating the relatively safe and reliable performance of sludge biochar. The study can provide a theoretical basis for sludge management and the biochar safe use in soil.