CoS2/carbon network flexible film with Co-N bond/π-π interaction enables superior mechanical properties and high-rate sodium ion storage.

Flexible electrodes based on conversion-type materials have potential applications in low-cost and high-performance flexible sodium-ion batteries (FSIBs), owing to their high theoretical capacity and appropriate sodiation potential. However, they suffer from flexible electrodes with poor mechanical properties and sluggish reaction kinetics. In this study, freestanding CoS2 nanoparticles coupled with graphene oxides and carbon nanotubes (CoS2/GO/CNTs) flexible films with robust and interconnected architectures were successfully synthesized. CoS2/GO/CNTs flexible film displays high electronic conductivity and superior mechanical properties (average tensile strength of 21.27 MPa and average toughness of 393.18 KJ m-3) owing to the defect bridge for electron transfer and the formation of the π-π interactions between CNTs and GO. In addition, the close contact between the CoS2 nanoparticles and carbon networks enabled by the Co-N chemical bond prevents the self-aggregation of the CoS2 nanoparticles. As a result, the CoS2/GO/CNTs flexible film delivered superior rate capability (213.5 mAh g-1 at 6 A g-1, better than most reported flexible anode) and long-term cycling stability. Moreover, the conversion reaction that occurred in the CoS2/GO/CNTs flexible film exhibited pseudocapacitive behavior. This study provides meaningful insights into the development of flexible electrodes with superior mechanical properties and electrochemical performance for energy storage.

Enhancing Proton Co-Intercalation in Iron Ion Batteries Cathodes for Increased Capacity.

Recently, aqueous iron ion batteries (AIIBs) using iron metal anodes have gained traction in the battery community as low-cost and sustainable solutions for green energy storage. However, the development of AIIBs is significantly hindered by the limited capacity of existing cathode materials and the poor intercalation kinetic of Fe2+. Herein, we propose a H+ and Fe2+ co-intercalation electrochemistry in AIIBs to boost the capacity and rate capability of cathode materials such as iron hexacyanoferrate (FeHCF) and Na4Fe3(PO4)2(P2O7) (NFPP). This is achieved through an electrochemical activation step during which a FeOOH nanowire layer is formed in situ on the cathode. This layer facilitates H+ co-intercalation in AIIBs, resulting in a high specific capacity of 151 mAh g-1 and 93% capacity retention over 500 cycles for activated FeHCF cathodes. We found that this activation process can also be applied to other cathode chemistries, such as NFPP, where we found that the cathode capacity is doubled as a result of this process. Overall, the proposed H+/Fe2+ co-insertion electrochemistry expands the range of applications for AIBBs, in particular as a sustainable solution for storing renewable energy.

Multi-Person Pose Tracking With Sparse Key-Point Flow Estimation and Hierarchical Graph Distance Minimization.

In this paper, we propose a novel framework for multi-person pose estimation and tracking on challenging scenarios. In view of occlusions and motion blurs which hinder the performance of pose tracking, we proposed to model humans as graphs and perform pose estimation and tracking by concentrating on the visible parts of human bodies which are informative about complete skeletons under incomplete observations. Specifically, the proposed framework involves three parts: (i) A Sparse Key-point Flow Estimating Module (SKFEM) and a Hierarchical Graph Distance Minimizing Module (HGMM) for estimating pixel-level and human-level motion, respectively; (ii) Pixel-level appearance consistency and human-level structural consistency are combined in measuring the visibility scores of body joints. The scores guide the pose estimator to predict complete skeletons by observing high-visibility parts, under the assumption that visible and invisible parts are inherently correlated in human part graphs. The pose estimator is iteratively fine-tuned to achieve this capability; (iii) Multiple historical frames are combined to benefit tracking which is implemented using HGMM. The proposed approach not only achieves state-of-the-art performance on PoseTrack datasets but also contributes to significant improvements in other tasks such as human-related anomaly detection.

Freestanding Ammonium Vanadate Composite Cathodes with Lattice Self-Regulation and Ion Exchange for Long-Lasting Ca-Ion Batteries.

Calcium-ion batteries (CIBs) have emerged as a promising alternative for electrochemical energy storage. The lack of high-performance cathode materials severely limits the development of CIBs. Vanadium oxides are particularly attractive as cathode materials for CIBs, and preinsertion chemistry is often used to improve their calcium storage performance. However, the room temperature cycling lifespan of vanadium oxides in organic electrolytes still falls short of 1000 cycles. Here, based on preinsertion chemistry, the cycling life of vanadium oxides is further improved by integrated electrode and electrolyte engineering. Utilizing a tailored Ca electrolyte, the constructed freestanding (NH4)2V6O16·1.35H2O@graphene oxide@carbon nanotube (NHVO-H@GO@CNT) composite cathode achieves a 305 mAh g-1 high capacity and 10 000 cycles record-long life. Additionally, for the first time, a Ca-ion hybrid capacitor full cell is assembled and delivers a capacity of 62.8 mAh g-1. The calcium storage mechanism of NHVO-H@GO@CNT based on a two-phase reaction and the exchange of NH4 + and Ca2+ during cycling are revealed. The lattice self-regulation of V─O layers is observed and the layered vanadium oxides with Ca2+ pillars formed by ion exchange exhibit higher capacity. This work provides novel strategies to enhance the calcium storage performance of vanadium oxides via integrated structural design of electrodes and electrolyte modification.

Engineering organic polymers as emerging sustainable materials for powerful electrocatalysts.

Cost-effective and high-efficiency catalysts play a central role in various sustainable electrochemical energy conversion technologies that are being developed to generate clean energy while reducing carbon emissions, such as fuel cells, metal-air batteries, water electrolyzers, and carbon dioxide conversion. In this context, a recent climax in the exploitation of advanced earth-abundant catalysts has been witnessed for diverse electrochemical reactions involved in the above mentioned sustainable pathways. In particular, polymer catalysts have garnered considerable interest and achieved substantial progress very recently, mainly owing to their pyrolysis-free synthesis, highly tunable molecular composition and microarchitecture, readily adjustable electrical conductivity, and high stability. In this review, we present a timely and comprehensive overview of the latest advances in organic polymers as emerging materials for powerful electrocatalysts. First, we present the general principles for the design of polymer catalysts in terms of catalytic activity, electrical conductivity, mass transfer, and stability. Then, the state-of-the-art engineering strategies to tailor the polymer catalysts at both molecular (i.e., heteroatom and metal atom engineering) and macromolecular (i.e., chain, topology, and composition engineering) levels are introduced. Particular attention is paid to the insightful understanding of structure-performance correlations and electrocatalytic mechanisms. The fundamentals behind these critical electrochemical reactions, including the oxygen reduction reaction, hydrogen evolution reaction, CO2 reduction reaction, oxygen evolution reaction, and hydrogen oxidation reaction, as well as breakthroughs in polymer catalysts, are outlined as well. Finally, we further discuss the current challenges and suggest new opportunities for the rational design of advanced polymer catalysts. By presenting the progress, engineering strategies, insightful understandings, challenges, and perspectives, we hope this review can provide valuable guidelines for the future development of polymer catalysts.

Extending Cycling Life Beyond 300 000 Cycles in Aqueous Zinc Ion Capacitors Through Additive Interface Engineering.

Developing low-cost and long-cycling-life aqueous zinc (Zn) ion capacitors (AZICs) for large-scale electrochemical energy storage still faces the challenges of dendritic Zn deposition and interfacial side reactions. Here, an interface engineering strategy utilizing a dibenzenesulfonimide (BBI) additive is employed to enhance the stability of the Zn metal anode/electrolyte interface. The first-principles calculation results demonstrate that BBI anions can be chemically adsorbed on Zn metal. Meanwhile, the experimental results confirm that the BBI-Zn interfacial layer converts the original water-richelectric double layer (EDL) into a water-poor EDL, effectively inhibiting the water related parasitic reaction at the electrode/electrolyte interface. In addition, the BBI-Zn interfacial layer introduces an additional Zn ions (Zn2+) migration energy barrier, increasing the Zn2+ de-solvation activation energy, consequently raising the Zn2+ nucleation overpotential, and thus achieving the compact and uniform Zn deposition behavior. Furthermore, the solid electrolyte interphase (SEI) layer derived from the BBI-Zn interfacial layer during cycling can further maintain the interfacial stability of the Zn anode. Owing to the above favorable features, the assembled AZIC exhibits an ultra-long cycling life of over 300 000 cycles based on the additive engineering strategy, which shows application prospects in high-performance AZICs.

Aromatic Organic Small-Molecule Material with (020) Crystal Plane Activation for Wide-Temperature and 68000 Cycle Aqueous Calcium-Ion Batteries.

Calcium-ion batteries are an emerging energy storage device owing to the low redox potential of Ca2+/Ca and the naturally abundant reserves of the Ca element. However, the high charge density and large radius of Ca2+ lead to a low calcium storage capacity or unsatisfactory cycling performance for most electrode materials. Herein, we report the organic crystal 3,4,9,10-perylenetetracarboxylic diimide (PTCDI) as an anode material for aqueous calcium-ion batteries (ACIBs) in a water-in-salt electrolyte. PTCDI delivers a high discharge capacity of 131.8 mAh g-1, excellent rate performance (86.2 mAh g-1@10000 mA g-1), and an ultralong life of 68000 cycles (over 470 days) with a high capacity retention of 72.7%. The calcium storage mechanism of PTCDI is shown to be an enolization reaction by in situ attenuated total reflectance Fourier-transform infrared and ex situ X-ray photoelectron spectroscopy. The activation mechanism of PTCDI microribbon splitting along the (020) crystal plane is studied by in situ X-ray diffraction, 3D tomography reconstruction technologies, and ex situ transmission electron microscopy. In addition, the Ca2+ storage sites and diffusion pathways of PTCDI are studied by density functional theory calculations. Finally, by matching a high-voltage Prussian blue analogue cathode, the assembled aqueous calcium-ion full cells exhibit excellent wide-temperature operating capability (-20 to +50 °C) and an ultralong life of 30000 cycles. Further, an aqueous calcium-ion pouch cell is constructed and exhibits a long lifetime of over 500 cycles.

Serrated lithium fluoride nanofibers-woven interlayer enables uniform lithium deposition for lithium-metal batteries.

The uncontrollable formation of Li dendrites has become the biggest obstacle to the practical application of Li-metal anodes in high-energy rechargeable Li batteries. Herein, a unique LiF interlayer woven by millimeter-level, single-crystal and serrated LiF nanofibers (NFs) was designed to enable dendrite-free and highly efficient Li-metal deposition. This high-conductivity LiF interlayer can increase the Li+ transference number and induce the formation of 'LiF-NFs-rich' solid-electrolyte interface (SEI). In the 'LiF-NFs-rich' SEI, the ultra-long LiF nanofibers provide a continuously interfacial Li+ transport path. Moreover, the formed Li-LiF interface between Li-metal and SEI film renders low Li nucleation and high Li+ migration energy barriers, leading to uniform Li plating and stripping processes. As a result, steady charge-discharge in a Li//Li symmetrical cell for 1600 h under 4 mAh cm-2 and 400 stable cycles under a high area capacity of 5.65 mAh cm-2 in a high-loading Li//rGO-S cell at 17.9 mA cm-2 could be achieved. The free-standing LiF-NFs interlayer exhibits superior advantages for commercial Li batteries and displays significant potential for expanding the applications in solid Li batteries.

Basal Planes Unlocking and Interlayer Engineering Endows Proton Doped-MoO2.8F0.2 with Fast and Stable Magnesium Storage.

Molybdenum trioxide has served as a promising cathode material of rechargeable magnesium batteries (RMBs), because of its rich valence states and high theoretical capacity; yet, it still suffers from sluggish (de)intercalation kinetics and inreversible structure change for highly polarized Mg2+ in the interlayer and intralayer of structure. Herein, F- substitutional and H+ interstitial doping is proposed for α-MoO3 materials (denoted HMoOF) by the intralayer/interlayer engineering strategy to boost the performance of RMBs. F- substitutional doping generates molybdenum vacancies along the Mo-O-□ or Mo-F-□ configurations (where □ represents the cationic vacancy) for unlocking the inactive basal plane of the layered crystal structure, and it further accelerates Mg2+ diffusion along the b-axis. Interstitial-doped H+ can expand interlayer spacing for reducing Mg2+ energy barrier along the ac plane and serve as a "pillar" to stabilize the interlayer structure. Moreover, anion and cation dual doping trigger shallow impurity levels (acceptors levels and donor levels), which helps to easily acquire the electrons from the valence band and donate the electrons to the conduction band. Consequently, the HMoOF electrode exhibits a high reversible capacity (241 mA h g-1 at 0.1 A g-1), an excellent rate capability (137.4 mAh g-1 at 2 A g-1), and a long cycling stability (capacity retention of 98% after 800 cycles at 1 A g-1) in RMBs. This work affords meaningful insights in layered materials for developing high-kinetics and long-life RMBs.

Mo2 C Nanoparticles Embedded in Carbon Nanowires with Surface Pseudocapacitance Enables High-Energy and High-Power Sodium Ion Capacitors.

Electrochemical sodium-ion storage technologies have become an indispensable part in the field of large-scale energy storage systems owing to the widespread and low-cost sodium resources. Molybdenum carbides with high electron conductivity are regarded as potential sodium storage anode materials, but the comprehensive sodium storage mechanism has not been studied in depth. Herein, Mo2 C nanowires (MC-NWs) in which Mo2 C nanoparticles are embedded in carbon substrate are synthesized. The sodium-ion storage mechanism is further systematically studied by in/ex situ experimental characterizations and diffusion kinetics analysis. Briefly, it is discovered that a faradaic redox reaction occurs in the surface amorphous molybdenum oxides on Mo2 C nanoparticles, while the inner Mo2 C is unreactive. Thus, the as-synthesized MC-NWs with surface pseudocapacitance display excellent rate capability (a high specific capacity of 76.5 mAh g-1 at 20 A g-1 ) and long cycling stability (a high specific capacity of 331.2 mAh g-1 at 1 A g-1 over 1500 cycles). The assembled original sodium ion capacitor displays remarkable power density and energy density. This work provides a comprehensive understanding of the sodium storage mechanism of Mo2 C materials, and constructing pseudocapacitive materials is an effective way to achieve sodium-ion storage devices with high power and energy density.

Large-Scale Integration of a Zinc Metasilicate Interface Layer Guiding Well-Regulated Zn Deposition.

Uneven distribution of electric fields at the electrolyte-anode interface and associated Zn dendrite growth is one of the most critical barriers that limit the life span of aqueous zinc-ion batteries. Herein, new-type Zn-A-O (A = Si, Ti) interface layers with thin and uniform thickness, porosity, and hydrophilicity properties are developed to realize homogeneous and smooth Zn plating. For ZnSiO3 nanosheet arrays on Zn foil (Zn@ZSO), their formation follows an "etching-nucleation-growth" mechanism that is confirmed by a well-designed Zn-island-based identical-location microscopy method, the geometric area of which is up to 1000 cm2 in one-pot synthesis based on a low-temperature wet-chemical method. Guided by the structural advantages of the ZSO layer, the Zn2+ flux gets equalized. Besides ultralow polarization, the life spans of symmetric cells and full cells coupled with a high-mass-loading K0.27 MnO2 ·0.54H2 O (8 mg cm-2 ) cathode, are increased by 3-7 times with the Zn@ZSO anode. Moreover, the large-scale preparation of Zn@ZSO foil contributes to a 0.5 Ah multilayer pouch cell with high performance, further confirming its prospects for practical application.

Tiny Vehicle Detection for Mid-to-High Altitude UAV Images Based on Visual Attention and Spatial-Temporal Information.

Mid-to-high altitude Unmanned Aerial Vehicle (UAV) imagery can provide important remote sensing information between satellite and low altitude platforms, and vehicle detection in mid-to-high altitude UAV images plays a crucial role in land monitoring and disaster relief. However, the high background complexity of images and limited pixels of objects challenge the performance of tiny vehicle detection. Traditional methods suffer from poor adaptation ability to complex backgrounds, while deep neural networks (DNNs) have inherent defects in feature extraction of tiny objects with finite pixels. To address the issue above, this paper puts forward a vehicle detection method combining the DNNs-based and traditional methods for mid-to-high altitude UAV images. We first employ the deep segmentation network to exploit the co-occurrence of the road and vehicles, then detect tiny vehicles based on visual attention mechanism with spatial-temporal constraint information. Experimental results show that the proposed method achieves effective detection of tiny vehicles in complex backgrounds. In addition, ablation experiments are performed to inspect the effectiveness of each component, and comparative experiments on tinier objects are carried out to prove the superior generalization performance of our method in detecting vehicles with a limited size of 5 × 5 pixels or less.

Electronic Structure Modulation in MoO2 /MoP Heterostructure to Induce Fast Electronic/Ionic Diffusion Kinetics for Lithium Storage.

Transition metal oxides (TMOs) are considered as the prospective anode materials in lithium-ion batteries (LIBs). Nevertheless, the disadvantages, including large volume variation and poor electrical conductivity, obstruct these materials to meet the needs of practical application. Well-designed mesoporous nanostructures and electronic structure modulation can enhance the electron/Li-ions diffusion kinetics. Herein, a unique mesoporous molybdenum dioxide/molybdenum phosphide heterostructure nanobelts (meso-MoO2 /MoP-NBs) composed of uniform nanoparticles is obtained by one-step phosphorization process. The Mott-Schottky tests and density functional theory calculations demonstrated that meso-MoO2 /MoP-NBs possesses superior electronic conductivity. The detailed lithium storage mechanism (solid solution reaction for MoP and partial conversion for MoO2 ), small change ratio of crystal structure and fast electronic/ionic diffusion behavior of meso-MoO2 /MoP-NBs are systematically investigated by operando X-ray diffraction, ex situ transmission electron microscopy, and kinetic analysis. Benefiting from the synergistic effects, the meso-MoO2 /MoP-NBs displays a remarkable cycling performance (515 mAh g-1 after 1000 cycles at 1 A g-1 ) and excellent rate capability (291 mAh g-1 at 8 A g-1 ). These findings can shed light on the behavior of the electron/ion regulation in heterostructures and provide a potential route to develop high-performance lithium-ion storage materials.

Lightweight Detection Network Based on Sub-Pixel Convolution and Objectness-Aware Structure for UAV Images.

Unmanned Aerial Vehicles (UAVs) can serve as an ideal mobile platform in various situations. Real-time object detection with on-board apparatus provides drones with increased flexibility as well as a higher intelligence level. In order to achieve good detection results in UAV images with complex ground scenes, small object size and high object density, most of the previous work introduced models with higher computational burdens, making deployment on mobile platforms more difficult.This paper puts forward a lightweight object detection framework. Besides being anchor-free, the framework is based on a lightweight backbone and a simultaneous up-sampling and detection module to form a more efficient detection architecture. Meanwhile, we add an objectness branch to assist the multi-class center point prediction, which notably improves the detection accuracy and only takes up very little computing resources. The results of the experiment indicate that the computational cost of this paper is 92.78% lower than the CenterNet with ResNet18 backbone, and the mAP is 2.8 points higher on the Visdrone-2018-VID dataset. A frame rate of about 220 FPS is achieved. Additionally, we perform ablation experiments to check on the validity of each part, and the method we propose is compared with other representative lightweight object detection methods on UAV image datasets.

High-Energy and High-Power Pseudocapacitor-Battery Hybrid Sodium-Ion Capacitor with Na+ Intercalation Pseudocapacitance Anode.

High-performance and low-cost sodium-ion capacitors (SICs) show tremendous potential applications in public transport and grid energy storage. However, conventional SICs are limited by the low specific capacity, poor rate capability, and low initial coulombic efficiency (ICE) of anode materials. Herein, we report layered iron vanadate (Fe5V15O39 (OH)9·9H2O) ultrathin nanosheets with a thickness of ~ 2.2 nm (FeVO UNSs) as a novel anode for rapid and reversible sodium-ion storage. According to in situ synchrotron X-ray diffractions and electrochemical analysis, the storage mechanism of FeVO UNSs anode is Na+ intercalation pseudocapacitance under a safe potential window. The FeVO UNSs anode delivers high ICE (93.86%), high reversible capacity (292 mAh g-1), excellent cycling stability, and remarkable rate capability. Furthermore, a pseudocapacitor-battery hybrid SIC (PBH-SIC) consisting of pseudocapacitor-type FeVO UNSs anode and battery-type Na3(VO)2(PO4)2F cathode is assembled with the elimination of presodiation treatments. The PBH-SIC involves faradaic reaction on both cathode and anode materials, delivering a high energy density of 126 Wh kg-1 at 91 W kg-1, a high power density of 7.6 kW kg-1 with an energy density of 43 Wh kg-1, and 9000 stable cycles. The tunable vanadate materials with high-performance Na+ intercalation pseudocapacitance provide a direction for developing next-generation high-energy capacitors.

3D Nitrogen-Doped Graphene Encapsulated Metallic Nickel-Iron Alloy Nanoparticles for Efficient Bifunctional Oxygen Electrocatalysis.

Invited for the cover of this issue is Liqiang Mai and co-workers at Wuhan University of Technology. The image depicts Ni3 Fe alloy nanoparticles encapsulated in N-doped graphene as an efficient bifunctional oxygen electrocatalyst toward rechargeable Zn-air batteries, which is expected to drive the electric vehicle. Read the full text of the article at 10.1002/chem.201904722.

3D Nitrogen-Doped Graphene Encapsulated Metallic Nickel-Iron Alloy Nanoparticles for Efficient Bifunctional Oxygen Electrocatalysis.

It is extremely desirable to explore high-efficient, affordable and robust oxygen electrocatalysts toward rechargeable Zn-air batteries (ZABs). A 3D porous nitrogen-doped graphene encapsulated metallic Ni3 Fe alloy nanoparticles aerogel (Ni3 Fe-GA1 ) was constructed through a facile hydrothermal assembly and calcination process. Benefiting from 3D porous configuration with great accessibility, high electrical conductivity, abundant active sites, optimal nitrogen content and strong electronic interactions at the Ni3 Fe/N-doped graphene heterointerface, the obtained aerogel showed outstanding catalytic performance toward the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR). Specifically, it exhibited an overpotential of 239 mV to attain 10 mA cm-2 for OER, simultaneously providing a positive onset potential of 0.93 V within a half-wave potential of 0.8 V for ORR. Accordingly, when employed in the aqueous ZABs, Ni3 Fe-GA1 achieved higher power density and superior reversibility than Pt/C-IrO2 catalyst, making it a potential candidate for rechargeable ZABs.

Polyol Solvation Effect on Tuning the Universal Growth of Binary Metal Oxide Nanodots@Graphene Oxide Heterostructures for Electrochemical Applications.

Tuning the uniformity and size of binary metal oxide nanodots on graphene oxide (BMO NDs@GO) is significant but full of challenges in wet-chemistry, owing to the difficulties of controlling the complicated cation/anion co-adsorption, heterogeneous nucleation, and overgrowth processes. Herein, the aim is to tune these processes by understanding the functions of various alcohol solvents for NDs growth on GO. It is found that the polyol solvation effect is beneficial for obtaining highly uniform BMO NDs@GO. Polyol shell capped metal ions exhibit stronger hydrogen-bond interactions with the GO surface, leading to a uniform cation/anion co-adsorption and followed heterogeneous nucleation. The polyol-solvated ions with large diffusion energy barrier drastically limit the ion diffusion kinetics in liquids and at the solid/liquid interface, resulting in a slow and controllable growth. Moreover, the synthesis in polyol systems is highly controllable and universal, thus eleven BMO and polynary metal oxide NDs@GO are obtained by this method. The synthetic strategy provides improved prospects for the manufacture of inorganic NDs and their expanding electrochemical applications.

Strongly Coupled Pyridine-V2 O5 ·nH2 O Nanowires with Intercalation Pseudocapacitance and Stabilized Layer for High Energy Sodium Ion Capacitors.

Developing pseudocapacitive cathodes for sodium ion capacitors (SICs) is very significant for enhancing energy density of SICs. Vanadium oxides cathodes with pseudocapacitive behavior are able to offer high capacity. However, the capacity fading caused by the irreversible collapse of layer structure remains a major issue. Herein, based on the Acid-Base Proton theory, a strongly coupled layered pyridine-V2 O5 ·nH2 O nanowires cathode is reported for highly efficient sodium ion storage. By density functional theory calculations, in situ X-ray diffraction, and ex situ Fourier-transform infrared spectroscopy, a strong interaction between protonated pyridine and VO group is confirmed and stable during cycling. The pyridine-V2 O5 ·nH2 O nanowires deliver long-term cyclability (over 3000 cycles), large pseudocapacitive behavior (78% capacitive contribution at 1 mV s-1 ) and outstanding rate capability. The assembled pyridine-V2 O5 ·nH2 O//graphitic mesocarbon microbead SIC delivers high energy density of ≈96 Wh kg-1 (at 59 W kg-1 ) and power density of 14 kW kg-1 (at 37.5 Wh kg-1 ). The present work highlights the strategy of realizing strong interaction in the interlayer of V2 O5 ·nH2 O to enhance the electrochemical performance of vanadium oxides cathodes. The strategy could be extended for improving the electrochemical performance of other layered materials.

A CNN Model for Semantic Person Part Segmentation with Capacity Optimization.

In this paper, a deep learning model with an optimal capacity is proposed to improve the performance of person part segmentation. Previous efforts in optimizing the capacity of a CNN model suffer from a lack of large datasets as well as the over-dependence on a single-modality CNN which is not effective in learning. We make several efforts in addressing these problems. Firstly, other datasets are utilized to train a CNN module for pre-processing image data and a segmentation performance improvement is achieved without a time-consuming annotation process. Secondly, we propose a novel way of integrating two complementary modules to enrich the feature representations for more reliable inferences. Thirdly, the factors to determine the capacity of a CNN model are studied and two novel methods are proposed to adjust (optimize) the capacity of a CNN to match it to the complexity of a task. The over-fitting and under-fitting problems are eased by using our methods. Experimental results show that our model outperforms the state-of-the-art deep learning models with a better generalization ability and a lower computational complexity.