Search

Recent studies suggest that plant disease identification via machine learning approach is vital for preventing the spread of diseases. Identifying multiple diseases simultaneous on a single leaf is one of the most irritating issues in agricultural production. However, the existing approaches are difficult to meet the requirements of production practice in accuracy or interpretability. Here, we present residual attention based multi-label learning framework (RAMDI), a method for predicting apple leaf diseases in natural environment. Built upon an attention based multi-label learning framework, the channel and spatial attention mechanisms are investigated and embedded in residual network for multi-label disease prediction, which takes advantage of channel-wise and spatial-wise attention weights. Experimental results indicate that the RAMDI achieves 0.976 accuracy, 0.986 F-score, and 0.979 mAPs, outperforms the existing state-of-the-art apple leaf disease identification models. RAMDI not only predicts multi-disease on a single leaf simultaneously, but also reveals the interpretability among positive predictions that contribute most to identify the key features that are significant for the leaf diseases. This method achieves the following two achievements. Firstly, it provides a solution for detecting multiple diseases on a single leaf. Secondly, this approach gains an interpretable understanding for apple leaf disease identification.

As the trajectory of transistor scaling defined by Moore’s law encounters challenges, the paradigm of ever-evolving integrated circuit technology shifts to explore unconventional materials and architectures to sustain progress. Two-dimensional (2D) semiconductors, characterized by their atomic-scale thickness and exceptional electronic properties, have emerged as a beacon of promise in this quest for the continued advancement of field-effect transistor (FET) technology. The energy-efficient complementary circuit integration necessitates strategic engineering of both n-channel and p-channel 2D FETs to achieve symmetrical high performance. This intricate process mandates the realization of demanding device characteristics, including low contact resistance, precisely controlled doping schemes, high mobility, and seamless incorporation of high- κ dielectrics. Furthermore, the uniform growth of wafer-scale 2D film is imperative to mitigate defect density, minimize device-to-device variation, and establish pristine interfaces within the integrated circuits. This review examines the latest breakthroughs with a focus on the preparation of 2D channel materials and device engineering in advanced FET structures. It also extensively summarizes critical aspects such as the scalability and compatibility of 2D FET devices with existing manufacturing technologies, elucidating the synergistic relationships crucial for realizing efficient and high-performance 2D FETs. These findings extend to potential integrated circuit applications in diverse functionalities.

For decades, the fundamental driving force behind energy-efficient and cost-effective electronic components has been the downward scaling of electronic devices. However, due to approaching the fundamental limits of silicon-based complementary metal-oxide-semiconductor (CMOS) devices, various emerging materials and device structures are considered alternative aspirants, such as negative-capacitance field-effect transistors (NCFETs), for their promising advantages in terms of scaling, speed, and power consumption. In this article, we present a brief overview of the progress made on NCFETs, including theoretical and experimental approaches, a current understanding of NCFET device physics, possible physical mechanisms for NC, and future functionalization prospects. In addition, in the context of recent findings, critical technological difficulties that must be addressed in the NCFET development are also discussed.

An iridium-catalysed hydrogen transfer strategy, enabling straightforward access to tetrahydro pyridine derivatives from aryl-1,8-naphthyridines and indolines, was developed. This method proceeds with unprecedented synthetic effectiveness including high step-economic fashion together with the advantages of having no by-product and no need for external high-pressure H2 gas, offering an important basis for the transformation of 1,8-naphthyridines and indolines into functionalized products.

As one of the most common forms of waste, waste PET is a serious pollutant in natural and human living environments. There is an urgent need to recycle PET. For this study, the complete degradation of PET was realized at a low temperature. A lipophilic hydrophobic membrane was formed on the surface of a stainless steel mesh (SSM) using a simple dip coating method, and an oil–water separation material was successfully prepared. After loading with degradation products, the surface roughness of SSM increased from 19.09 μm to 62.33 μm. The surface changed from hydrophilic to hydrophobic, and the water contact angle increased to 123°. The oil–water separation flux of the modified SSM was 9825 L/(m2·h), and the separation efficiency was 98.99%. The modified SSM had good reuse performance. This hydrophobic modification method can also be used to modify other porous substrates, such as activated carbon, filter paper, foam, and other materials. The porous substrate modified by the degradation product of waste PET was used to prepare oil–water separation materials, not only solving the problem of white pollution but also reducing the dependence on non-renewable resources in the conventional methods used for the preparation of oil–water separation materials. This study provides new raw materials and methods for the industrial production of oil–water separation materials, which have important application prospects.

The increase in carbon dioxide emissions has significantly impacted human society and the global environment. As carbon dioxide is the most abundant and cheap C1 resource, the conversion and utilization of carbon dioxide have received extensive attention from researchers. Among the many carbon dioxide conversion and utilization methods, the reverse water–gas conversion (RWGS) reaction is considered one of the most effective. This review discusses the research progress made in RWGS with various heterogeneous metal catalyst types, covering topics such as catalyst performance, thermodynamic analysis, kinetics and reaction mechanisms, and catalyst design and preparation, and suggests future research on RWGS heterogeneous catalysts.

Abstract This paper develops an active magnetic near field probe (H‐field probe) by using a four‐layer printed circuit board (PCB) technique. Two‐turn detection structure and a low noise amplifier are used to improve probe's frequency response. The two‐turn detection structure can maximize the use of PCB stack resources, and a 14.5 dB gain amplifier can increase signal output capability. The probe can then be used for electromagnetic compatibility (EMC) test from 150 kHz to 3 GHz. Compared with traditional active shielded‐loop H‐field probe (TAHP) with loop area of 500 μm2, the frequency response of the proposed probe can improve 20 dB at 0.5 GHz. Different from the method of comparing the sensitivity of the probe only through the frequency response, this paper measures and analyzes the real sensitivity of the probing device composed of the proposed probe. Under the condition of the same receiver parameter setting, the sensitivity at 0.5 GHz is increased by 11.7 dB compared with commercial passive probing device and 20.2 dB at 0.5 GHz compared with the probing device with TAHP device. It reaches −32.4 dB μA. In addition, the proposed probe has acceptable spatial resolution of 1.28 mm at 1 GHz, which is more suitable for printed circuit board level EMI analysis.

As COVID-19 spread worldwide, many major grain-producing countries have adopted measures to restrict their grain exports; food security has aroused great concern from various parties. How to improve grain production has become one of the most important issues facing all countries. However, crop diseases are a difficult problem for many farmers so it is important to master the severity of crop diseases timely and accurately to help staff take further intervention measures to minimize plants being further infected. In this paper, a restructured residual dense network was proposed for tomato leaf disease identification; this hybrid deep learning model combines the advantages of deep residual networks and dense networks, which can reduce the number of training process parameters to improve calculation accuracy as well as enhance the flow of information and gradients. The original RDN model was first used in image super resolution, so we need to restructure the network architecture for classification tasks through adjusted input image features and hyper parameters. Experimental results show that this model can achieve a top-1 average identification accuracy of 95% on the Tomato test dataset in AI Challenger 2018 datasets, which verifies its satisfactory performance. The restructured residual dense network model can obtain significant improvements over most of the state-of-the-art models in crop leaf identification, as well as requiring less computation to achieve high performance.

In practice, early detection of disease is of high importance to practical value, corresponding measures can be taken at the early stage of plant disease. However, in the early stage of disease or when a rare disease occurs, there are limited training samples in practice, which makes machine learning especially deep learning models hardly work well while the stronger representative ability needs large-scale training data. To solve this problem, a fine grained-GAN based grape leaf spot identification method was proposed for local spot area image data augmentation to the generated local spot area images which were added and fed them into deep learning models for training to further strengthen the generalization ability of the classification models, which can effectively improve the accuracy and robustness of the prediction. Including 500 early-stage grape leaf spot images every category were fed into the proposed fine grained-GAN for local spot area data augmentation, 1000 local spot area sub images every category were generated in this study. The improved faster R-CNN was integrated in fine grained-GAN as local spot area detector. After that, the segmented sub-images and generated sub-images were mixed as training data input into the deep learning models, while the original segmented sub-images were used for testing. Experimental results showed that the proposed method had achieved higher identification accuracy on five state-of-art deep learning models; especially ResNet-50 had got 96.27% accuracy, which obtained significant improvements than other data augmentation methods and verified its satisfactory performance. This is of great practical significance for rare diseases or limited training samples.

As the COVID-19 spread worldwide, countries around the world are actively taking measures to fight against the epidemic. To prevent the spread of it, a high sensitivity and efficient method for COVID-19 detection is necessary. By analyzing the COVID-19 chest X-ray images, a combination method of image regrouping and ResNet-SVM was proposed in this study. The lung region was segmented from the original chest X-ray images and divided into small pieces, and then the small pieces of lung region were regrouped into a regular image randomly. Furthermore the regrouped images were fed into the deep residual encoder block for feature extraction. Finally the extracted features were as input into support vector machine for recognition. The visual attention was introduced in the novel method, which paid more attention to the features of COVID-19 without the interference of shapes, rib and other related noises. The experimental results showed that the proposed method achieved 93% accuracy without large number of training data, outperformed the existing COVID-19 detection models.

The paper presents the design and fabrication of lateral-field-excited (LFE) resonators based on 42° Y-cut single-crystal LiTaO3 (LT) on silicon dioxide (SiO2). A simple coplanar top electrode is defined to excite the bulk acoustic wave modes in the suspended LT/SiO2 structure, and the fabrication process that only involves two lithography steps is more simplified compared to that of commercial film bulk acoustic wave resonators. For a model structure consisting of LT(670 nm)/SiO2 (1500 nm) thin film, two types of acoustic modes are both piezoelectrically active in the LT film: the first one is the thickness-shear mode with a resonance frequency of 2.46 GHz, an electromechanical coupling ( $k_{eff}^{2}$ ) of 1.4%, a high quality factor ( ${Q}$ ) of 1690, and the second one corresponds to longitudinal mode with a resonance frequency of 4.39 GHz, ${k} _{eff}^{2}$ of 1.2%, a high ${Q}$ of 1590, which is among the highest reported for piezoelectric MEMS resonators operating at this frequency range. The excellent performance would enable application scenarios including high-resolution sensors, low-phase-noise oscillators, and low-loss, high selectivity filters in the sub-6 GHz range for the fifth-generation (5G) wireless communication.

Abstract Self-driven photodetectors that can detect light without any external voltage bias are important for low-power applications, including future internet of things, wearable electronics, and flexible electronics. While two-dimensional (2D) materials exhibit good optoelectronic properties, the extraordinary properties have not been fully exploited to realize high-performance self-driven photodetectors. In this paper, a metal–semiconductor–metal (MSM) photodetector with graphene and Au as the two contacts have been proposed to realize the self-driven photodetector. Van der Waals contacts are formed by dry-transfer methods, which is important in constructing the asymmetrical MSM photodetector to avoid the Fermi-level pinning effect. By choosing graphene and Au as the two contact electrodes, a pronounced photovoltaic effect is obtained. Without any external bias, the self-driven photodetector exhibits a high responsivity of 7.55 A W−1 and an ultrahigh photocurrent-to-dark current ratio of ~108. The photodetector also shows gate-tunable characteristics due to the field-induced Fermi-level shift in the constituent 2D materials. What is more, the high linearity of the photodetector over almost 60 dB suggests the easy integration with processing circuits for practical applications.

In this letter, a compact model for charge and drain current in molybdenum disulfide (MoS2) field-effect transistors (FETs) is developed, which is valid from ballistic to quasi-ballistic to drift-diffusion electronic transport regimes. Considering the influence of trap charges in MoS2 transistors, a physical-based and analytical charge model is derived. Based on the virtual source model which applies to both ballistic and quasi-ballistic transports, the carrier number density and current expressions are combined to yield the current-voltage (I-V) characteristics. Furthermore, the presented model is validated by experimental data as well as recently reported simulations for MoS2 FETs with different gate lengths. It shows that our model is accurate, straight-forward, scalable and compatible for short- and long-channel devices.

Two-dimensional MoS2 exhibits excellent optical and electrical properties, which have aroused great research interest in a wide range of interdisciplinary fields. Chemical vapor deposition (CVD) is considered to be an effective and simple technique for the synthesis of monolayer MoS2 thin films. However, the controlled growth of large size single-crystalline monolayer MoS2 is still a challenge. This work presents controlled growth of large-area monolayer MoS2 using evaporated Mo films pretreated with NaCl solution. The maximum size of the monolayer MoS2 flake can reach 300 μm, which could be obtained by optimizing the mass concentration of NaCl solution. In addition, the study also reveals that an appropriate increase of NaCl concentration can affect the edge growth rate of sulfur and molybdenum, thus affecting the relative stability of the growth mode of the crystal. This provides new insight and research direction for the specific mechanism of fractal growth, which can also be used for the controllable growth of MoS2 with high-density Mo-terminated edges for catalytic applications.

A stable, earth-abundant, reusable cobalt-based heterogeneous catalyst is developed for the oxidative esterification of alcohols under ambient conditions, featuring broad substrate scope, providing good to excellent product yields. This protocol enables easy recyclability of the catalyst, measured up to five times without significant loss of efficiency. The active sites of Co-N-Si/AC are proposed to be Co-N species.

Synthesis of the vertically aligned carbon nanotubes (CNTs) using complementary metal-oxide-semiconductor (CMOS)-compatible methods is essential to integrate the CNT contact and interconnect to nanoscale devices and ultra-dense integrated nanoelectronics. However, the synthesis of high-density CNT array at low-temperature remains a challenging task. The advances in the low-temperature synthesis of high-density vertical CNT structures using CMOS-compatible methods are reviewed. Primarily, recent works on theoretical simulations and experimental characterizations of CNT growth emphasized the critical roles of catalyst design in reducing synthesis temperature and increasing CNT density. In particular, the approach of using multilayer catalyst film to generate the alloyed catalyst nanoparticle was found competent to improve the active catalyst nanoparticle formation and reduce the CNT growth temperature. With the multilayer catalyst, CNT arrays were directly grown on metals, oxides, and 2D materials. Moreover, the relations among the catalyst film thickness, CNT diameter, and wall number were surveyed, which provided potential strategies to control the tube density and the wall density of synthesized CNT array.

A method to synthesize high-density, vertically-aligned, multi-wall carbon nanotubes (MWCNTs) on an insulating substrate at low temperature using a complementary metal–oxide–semiconductor (CMOS) compatible process is presented. Two factors are identified to be important in the carbon nanotube (CNT) growth, which are the catalyst design and the substrate material. By using a Ni–Al–Ni multilayer catalyst film and a ZrO2 substrate, vertically-aligned CNTs can be synthesized at 340 °C using plasma-enhanced chemical vapor deposition (PECVD). Both the quality and density of the CNTs can be enhanced by increasing the synthesis temperature. The function of the aluminum interlayer in reducing the activation energy of the CNT formation is studied. The nanoparticle sintering and quick accumulation of amorphous carbon covering the catalyst can prematurely stop CNT synthesis. Both effects can be suppressed by using a substrate with a high surface energy such as ZrO2.

Traditional manual identification of crop leaf diseases is challenging. Owing to the limitations in manpower and resources, it is challenging to explore crop diseases on a large scale. The emergence of artificial intelligence technologies, particularly the extensive application of deep learning technologies, is expected to overcome these challenges and greatly improve the accuracy and efficiency of crop disease identification. Crop leaf disease identification models have been designed and trained using large-scale training data, enabling them to predict different categories of diseases from unlabeled crop leaves. However, these models, which possess strong feature representation capabilities, require substantial training data, and there is often a shortage of such datasets in practical farming scenarios. To address this issue and improve the feature learning abilities of models, this study proposes a deep transfer learning adaptation strategy. The novel proposed method aims to transfer the weights and parameters from pre-trained models in similar large-scale training datasets, such as ImageNet. ImageNet pre-trained weights are adopted and fine-tuned with the features of crop leaf diseases to improve prediction ability. In this study, we collected 16,060 crop leaf disease images, spanning 12 categories, for training. The experimental results demonstrate that an impressive accuracy of 98% is achieved using the proposed method on the transferred ResNet-50 model, thereby confirming the effectiveness of our transfer learning approach. [ABSTRACT FROM AUTHOR]