LiViT-Net: A U-Net-like, lightweight Transformer network for retinal vessel segmentation

The intricate task of precisely segmenting retinal vessels from images, which is critical for diagnosing various eye diseases, presents significant challenges for models due to factors such as scale variation, complex anatomical patterns, low contrast, and limitations in training data. Building on these challenges, we offer novel contributions spanning model architecture, loss function design, robustness, and real-time efficacy. To comprehensively address these challenges, a new U-Net-like, lightweight Transformer network for retinal vessel segmentation is presented. By integrating MobileViT+ and a novel local representation in the encoder, our design emphasizes lightweight processing while capturing intricate image structures, enhancing vessel edge precision. A novel joint loss is designed, leveraging the characteristics of weighted cross-entropy and Dice loss to effectively guide the model through the task's challenges, such as foreground-background imbalance and intricate vascular structures. Exhaustive experiments were performed on three prominent retinal image databases. The results underscore the robustness and generalizability of the proposed LiViT-Net, which outperforms other methods in complex scenarios, especially in intricate environments with fine vessels or vessel edges. Importantly, optimized for efficiency, LiViT-Net excels on devices with constrained computational power, as evidenced by its fast performance. To demonstrate the model proposed in this study, a freely accessible and interactive website was established (https://hz-t3.matpool.com:28765?token=aQjYR4hqMI), revealing real-time performance with no login requirements.

GMAlignNet: multi-scale lightweight brain tumor image segmentation with enhanced semantic information consistency

Although the U-shaped architecture, represented by UNet, has become a major network model for brain tumor segmentation, the repeated convolution and sampling operations can easily lead to the loss of crucial information. Additionally, directly fusing features from different levels without distinction can easily result in feature misalignment, affecting segmentation accuracy. On the other hand, traditional convolutional blocks used for feature extraction cannot capture the abundant multi-scale information present in brain tumor images. This paper proposes a multi-scale feature-aligned segmentation model called GMAlignNet that fully utilizes Ghost convolution to solve these problems. Ghost Hierarchical Decoupled Fusion Unit and Ghost Hierarchical Decoupled Unit are used instead of standard convolutions in the encoding and decoding paths. This transformation replaces the holistic learning of volume structures by traditional convolutional blocks with multi-level learning on a specific view, facilitating the acquisition of abundant multi-scale contextual information through low-cost operations. Furthermore, a feature alignment unit is proposed that can utilize semantic information flow to guide the recovery of upsampled features. It performs pixel-level semantic information correction on misaligned features due to feature fusion. The proposed method is also employed to optimize three classic networks, namely DMFNet, HDCNet, and 3D UNet, demonstrating its effectiveness in automatic brain tumor segmentation. The proposed network model was applied to the BraTS 2018 dataset, and the results indicate that the proposed GMAlignNet achieved Dice coefficients of 81.65%, 90.07%, and 85.16% for enhancing tumor, whole tumor, and tumor core segmentation, respectively. Moreover, with only 0.29M parameters and 26.88G FLOPs, it demonstrates better potential in terms of computational efficiency and possesses the advantages of lightweight. Extensive experiments on the BraTS 2018, BraTS 2019, and BraTS 2020 datasets suggest that the proposed model exhibits better potential in handling edge details and contour recognition.

Light3DHS: A lightweight 3D hippocampus segmentation method using multiscale convolution attention and vision transformer

The morphological analysis and volume measurement of the hippocampus are crucial to the study of many brain diseases. Therefore, an accurate hippocampal segmentation method is beneficial for the development of clinical research in brain diseases. U-Net and its variants have become prevalent in hippocampus segmentation of Magnetic Resonance Imaging (MRI) due to their effectiveness, and the architecture based on Transformer has also received some attention. However, some existing methods focus too much on the shape and volume of the hippocampus rather than its spatial information, and the extracted information is independent of each other, ignoring the correlation between local and global features. In addition, many methods cannot be effectively applied to practical medical image segmentation due to many parameters and high computational complexity. To this end, we combined the advantages of CNNs and ViTs (Vision Transformer) and proposed a simple and lightweight model: Light3DHS for the segmentation of the 3D hippocampus. In order to obtain richer local contextual features, the encoder first utilizes a multi-scale convolutional attention module (MCA) to learn the spatial information of the hippocampus. Considering the importance of local features and global semantics for 3D segmentation, we used a lightweight ViT to learn high-level features of scale invariance and further fuse local-to-global representation. To evaluate the effectiveness of encoder feature representation, we designed three decoders of different complexity to generate segmentation maps. Experiments on three common hippocampal datasets demonstrate that the network achieves more accurate hippocampus segmentation with fewer parameters. Light3DHS performs better than other state-of-the-art algorithms.

Lightweight neural network for Alzheimer's disease classification using multi-slice sMRI

Alzheimer's disease (AD) is a progressive neurodegenerative disease. Early detection and intervention are crucial in preventing the progression of AD. To achieve efficient and scalable AD auto-detection based on structural Magnetic Resonance Imaging (sMRI), a lightweight neural network using multi-slice sMRI is proposed in this paper. The backbone for feature extraction is based on ShuffleNet V1 architecture, which is effective for overcoming the limitations posed by limited sMRI data and resource-restricted devices. In addition, we incorporate Efficient Channel Attention (ECA) to capture cross-channel interaction information, enabling us to effectively enhance features of disease associated brain regions. To optimize the model, we employ both cross entropy loss and triplet loss functions to constrain the predicted probabilities to the ground-truth labels, and to ensure appropriate representation of distances between different classes in the learned features. Experimental results show that the classification accuracies of our method for AD vs. CN, AD vs. MCI, and MCI vs. CN classification tasks are 95.00%, 87.50%, and 85.62% respectively. Our method utilizes only 3.42 M parameters and 6.08G FLOPs, while maintaining a comparable level of performance compared to the other 5 latest lightweight methods. This model design is computationally efficient, allowing it to process large amounts of data quickly and accurately in a timely manner. Additionally, it has the potential to advance the intelligent detection of Alzheimer's disease on devices with limited computing capabilities.

Characterization of lightweight aerated mortars using waste-to-energy bottom ash (WtE-BA) as aerating agent

The management of Waste-to-Energy Bottom Ash (WtE-BA), generated during the incineration of waste, poses a global challenge. Presently, the majority of WtE-BA is disposed of in landfills due to the lack of alternatives. Meanwhile, the construction industry remains the primary consumer of raw materials and significantly contributes to Greenhouse Gas Emissions. This study attempts to address these issues by utilizing the fine fraction of WtE-BA (<2 mm) as a raw material for aerated mortar production. Thanks to its metallic aluminum content, WtE-BA is utilized as an aerating agent. The study investigates how the quantities of water and WtE-BA, as well as its granulometric sub-fractions, impact the properties of the final product. An analysis of properties such as density, compressive strength, and thermal conductivity was conducted. Additionally, the environmental impact of each raw material (i.e. WtE-BA, cement and sand) was assessed through leaching tests and elemental content analysis enabling the determination of their individual contribution to the presence of trace elements in the produced mortars. The aforementioned properties are discussed using microstructure and porosity analyses. The findings demonstrate that the quantity of water is a crucial factor in controlling the aeration of mortars, whereas the granulometry of the WtE-BA particles did not significantly affect their macro-properties. Furthermore, this study highlights that WtE-BA based mortars has the potential to exhibit better environmental and insulating performances than standard aerated mortar of equal density and strength. The differences in pore size and type between WtE-BA and aerated mortars can account for the variation in performance. Thus, WtE-BA proves to be an effective substitute for aerating agent in the production of aerated mortars.

Pomegranate plasma heterostructure regulated 1D biomass derived microtube networks for lightweight broadband microwave absorber

The excessive aggregation of magnetic metal particles and the resulting skin effect tend to cause a serious imbalance in impedance matching, which hinders its application in aerospace and military wave absorption fields. Obviously, effective dispersion configuration and network construction are two practical measures to develop broadband lightweight absorbers. Based on the recycling theme, pomegranate plasma heterostructure regulated one-dimensional (1D) biomass derived microtube networks are achieved through the conversion and utilization of waste Platanus ball fibers. The metal-organic framework strategy successfully avoids the hard agglomeration of metal particles. The pomegranate seed-like heterostructure effectively modulated the impedance of carbon microtubes, resulting in coordinated dielectric and magnetic losses. Such composites exhibited an effective absorbing bandwidth of 6.08 GHz and a minimum reflection loss of -29.8 dB. This work provides a new approach for constructing sustainable ultralight electromagnetic wave absorbers using plasmon modification and a 1D built-up network structure.

SC-Unext: A Lightweight Image Segmentation Model with Cellular Mechanism for Breast Ultrasound Tumor Diagnosis

Automatic breast ultrasound image segmentation plays an important role in medical image processing. However, current methods for breast ultrasound segmentation suffer from high computational complexity and large model parameters, particularly when dealing with complex images. In this paper, we take the Unext network as a basis and utilize its encoder-decoder features. And taking inspiration from the mechanisms of cellular apoptosis and division, we design apoptosis and division algorithms to improve model performance. We propose a novel segmentation model which integrates the division and apoptosis algorithms and introduces spatial and channel convolution blocks into the model. Our proposed model not only improves the segmentation performance of breast ultrasound tumors, but also reduces the model parameters and computational resource consumption time. The model was evaluated on the breast ultrasound image dataset and our collected dataset. The experiments show that the SC-Unext model achieved Dice scores of 75.29% and accuracy of 97.09% on the BUSI dataset, and on the collected dataset, it reached Dice scores of 90.62% and accuracy of 98.37%. Meanwhile, we conducted a comparison of the model's inference speed on CPUs to verify its efficiency in resource-constrained environments. The results indicated that the SC-Unext model achieved an inference speed of 92.72 ms per instance on devices equipped only with CPUs. The model's number of parameters and computational resource consumption are 1.46M and 2.13 GFlops, respectively, which are lower compared to other network models. Due to its lightweight nature, the model holds significant value for various practical applications in the medical field.

A lightweight network for traffic sign recognition based on multi-scale feature and attention mechanism

Traffic sign recognition is an important part of intelligent transportation system. It uses computer vision and traffic sign recognition technology to detect and recognize traffic signs on the road automatically. In this paper, we propose a lightweight model for traffic sign recognition based on convolutional neural networks called ConvNeSe. Firstly, the feature extraction module of the model is constructed using the Depthwise Separable Convolution and Inverted Residuals structures. The model extracts multi-scale features with strong representation ability by optimizing the structure of convolutional neural networks and fusing of features. Then, the model introduces Squeeze and Excitation Block (SE Block) to improve the attention to important features, which can capture key information of traffic sign images. Finally, the accuracy of the model in the German Traffic Sign Recognition Benchmark Database (GTSRB) is 99.85%. At the same time, the model has good robustness according to the results of ablation experiments.

OnDev-LCT: On-Device Lightweight Convolutional Transformers towards federated learning

Federated learning (FL) has emerged as a promising approach to collaboratively train machine learning models across multiple edge devices while preserving privacy. The success of FL hinges on the efficiency of participating models and their ability to handle the unique challenges of distributed learning. While several variants of Vision Transformer (ViT) have shown great potential as alternatives to modern convolutional neural networks (CNNs) for centralized training, the unprecedented size and higher computational demands hinder their deployment on resource-constrained edge devices, challenging their widespread application in FL. Since client devices in FL typically have limited computing resources and communication bandwidth, models intended for such devices must strike a balance between model size, computational efficiency, and the ability to adapt to the diverse and non-IID data distributions encountered in FL. To address these challenges, we propose OnDev-LCT: Lightweight Convolutional Transformers for On-Device vision tasks with limited training data and resources. Our models incorporate image-specific inductive biases through the LCT tokenizer by leveraging efficient depthwise separable convolutions in residual linear bottleneck blocks to extract local features, while the multi-head self-attention (MHSA) mechanism in the LCT encoder implicitly facilitates capturing global representations of images. Extensive experiments on benchmark image datasets indicate that our models outperform existing lightweight vision models while having fewer parameters and lower computational demands, making them suitable for FL scenarios with data heterogeneity and communication bottlenecks.

Dust detection and cleanliness assessment based on S-YOLOv5s for NPP reactor containment wall-climbing cleaning robot

NPP reactor containment dust can easily turn into radioactive dust, endangering staff health and the environment. However, the nuclear reactor containment wall-climbing cleaning robot cleans blindly without the ability to clean the dust in a timely and thoroughly. In this paper, ShuffleNetV2-YOLOv5s (S-YOLOv5s) model is proposed to solve the problem of wall-climbing robots unable to detect different categories of dust in time. The use of ShuffleNetV2 in the backbone of the network not only ensures a large number of characterized channels and a large network capacity, but also reduces the complexity of the model; SIoU is chosen for the loss function to improve the model accuracy. Then, planar cleaning index (PCI) is proposed by combining the results of S-YOLOv5s to evaluate whether the wall-climbing cleaning robot cleans thoroughly. Compared to other methods, PCI considers amount and area occupation of different classes of dust. The dust data set is collected to train the designed model, and the model size is reduced to 14 % of the original model, and the FPS is 7.313 higher than the original model. Especially when compared with the state-of-the-art lightweight methods, our model has smaller model size and higher recognition speed. Experimental results have shown that our dust detection and cleanliness assessment method can be used on a wall-climbing cleaning robot to clean dust in time and thoroughly.

Ultrastrong lightweight nanocellulose-based composite aerogels with robust superhydrophobicity and durable thermal insulation under extremely environment

Ultra-lightweight porous aerogels based on nanocellulose (NC) have promising applications in various fields such as building insulation, sewage treatment, energy storage, and aerospace. One of the key advantages of these aerogels is their exceptionally low thermal conductivity. Nevertheless, the thermal insulation of NC aerogel (NCA) can deteriorate with changes in temperature and humidity conditions, making it crucial to develop a bulk aerogel that can maintain exceptional thermal insulating properties in harsh environmental conditions. A sustainable and user-friendly approach to synthesizing cellulose/poly(vinyl alcohol) aerogel (CellPA) materials has been developed, which are lightweight, possess good insulating properties, and demonstrate robust superhydrophobicity even in harsh environmental conditions. The CellPA are both exceptionally lightweight and robust, boasting outstanding resistance to combustion while also displaying a thermal conductivity of 36.1 mW m-1 K-1, suggesting they hold great promise for insulation applications. Furthermore, CellPA also exhibits robust superhydrophobicity even under harsh conditions, confirming the homogenous superhydrophobic modification of the biodegradable PVA through chemical methods. The manufacturing of bio-based composite materials with enhanced mechanical and thermal insulation features has gained immense popularity in a broad spectrum of contemporary engineering applications. These composite materials are particularly valuable as a robust, energy-efficient, lightweight, waterproof and flameproof for construction materials.

Multi-scale feature selection network for lightweight image super-resolution

Recently, many super-resolution (SR) methods based on convolutional neural networks (CNNs) have achieved superior performance by utilizing deep and heavy models, which may not be suitable for real-world low-budget devices. To address this issue, we propose a novel lightweight SR network called a multi-scale feature selection network (MFSN). As the basic building block of MFSN, the multi-scale feature selection block (MFSB) is presented to extract the rich multi-scale features from a coarse-to-fine receptive field level. For a better representation ability, a wide-activated residual unit is adopted in each branch of MFSB except the last one. In MFSB, the scale selection module (SSM) is designed to effectively fuse the features from two adjacent branches by adjusting receptive field sizes adaptively. Further, a comprehensive channel attention mechanism (CCAM) is integrated into SSM to learn the dynamic selection weight by considering the local and global inter-channel dependencies. Extensive experimental results illustrate that the proposed MFSN is superior to other lightweight methods.

QMLS: quaternion mutual learning strategy for multi-modal brain tumor segmentation

Objective.Due to non-invasive imaging and the multimodality of magnetic resonance imaging (MRI) images, MRI-based multi-modal brain tumor segmentation (MBTS) studies have attracted more and more attention in recent years. With the great success of convolutional neural networks in various computer vision tasks, lots of MBTS models have been proposed to address the technical challenges of MBTS. However, the problem of limited data collection usually exists in MBTS tasks, making existing studies typically have difficulty in fully exploring the multi-modal MRI images to mine complementary information among different modalities.Approach.We propose a novel quaternion mutual learning strategy (QMLS), which consists of a voxel-wise lesion knowledge mutual learning mechanism (VLKML mechanism) and a quaternion multi-modal feature learning module (QMFL module). Specifically, the VLKML mechanism allows the networks to converge to a robust minimum so that aggressive data augmentation techniques can be applied to expand the limited data fully. In particular, the quaternion-valued QMFL module treats different modalities as components of quaternions to sufficiently learn complementary information among different modalities on the hypercomplex domain while significantly reducing the number of parameters by about 75%.Main results.Extensive experiments on the dataset BraTS 2020 and BraTS 2019 indicate that QMLS achieves superior results to current popular methods with less computational cost.Significance.We propose a novel algorithm for brain tumor segmentation task that achieves better performance with fewer parameters, which helps the clinical application of automatic brain tumor segmentation.

A facile strategy for constructing lightweight, fire safety and compression resistance poly(vinylalcohol) aerogels with highly-efficient expansible graphene oxide/layered double hydroxides hybrid synergistic flame retardant

Poly(vinyl alcohol) (PVA) aerogels with excellent environmentally friendly properties have been considered to replace undegradable polymer foams. However, due to highly flammable, hydrophilic, and worse compression resistance performance, PVA aerogels have always been excluded from practical. Herein, a fire safety and compression resistance PVA/expansible graphene oxide (EGO)/Layered double hydroxides (LDHs) (PGL) aerogel was prepared via the freeze-drying method and electrostatic adsorption of flame retardant. The ice crystals from aerogels were sublimated and left a mass of tree-like pore tunnel structures. Meantime, the compound of EGO and LDHs rendered PGL aerogels high compressive strength of 6.0917 MPa (at 80% of strains), a high specific modulus of 19.16 m2/s2, and an ultra-low density of 0.059 g/cm3. Especially, the as-prepared PGL aerogels showed heat release reduced by 55.4%, smoke release reduced by 54.3%, and the limiting oxygen index reaching up to 31%. Moreover, LDHs also enhanced the interface with PVA/EGO resulting in hydrophobic performance improvement. The proposed enhancements mechanism suggested that (i) chemical reactions between EGO and PVA matrix; (ii) a mass of negative potential sites from the interface of PVA/EGO composites made LDHs sheets easily adsorbing; (iii) oxygen-containing groups from EGO and LDHs absorbed mass of heat during combustion; (iv) the compact char residues on the surface of aerogels acting as barriers suppressed smoke and prevented PVA matrix from further combustion. Therefore, electrostatic adsorption as a facile production process was paved for meeting the compression resistance, flame-retardant, heat-insulating, and smoke-suppressed requirements of PVA aerogels in this work.

Focus-RCNet: a lightweight recyclable waste classification algorithm based on focus and knowledge distillation

Waste pollution is a significant environmental problem worldwide. With the continuous improvement in the living standards of the population and increasing richness of the consumption structure, the amount of domestic waste generated has increased dramatically, and there is an urgent need for further treatment. The rapid development of artificial intelligence has provided an effective solution for automated waste classification. However, the high computational power and complexity of algorithms make convolutional neural networks unsuitable for real-time embedded applications. In this paper, we propose a lightweight network architecture called Focus-RCNet, designed with reference to the sandglass structure of MobileNetV2, which uses deeply separable convolution to extract features from images. The Focus module is introduced to the field of recyclable waste image classification to reduce the dimensionality of features while retaining relevant information. To make the model focus more on waste image features while keeping the number of parameters small, we introduce the SimAM attention mechanism. In addition, knowledge distillation was used to further compress the number of parameters in the model. By training and testing on the TrashNet dataset, the Focus-RCNet model not only achieved an accuracy of 92[Formula: see text] but also showed high deployment mobility.

U-MLP: MLP-based ultralight refinement network for medical image segmentation

The convolutional neural network (CNN) and Transformer play an important role in computer-aided diagnosis and intelligent medicine. However, CNN cannot obtain long-range dependence, and Transformer has shortcomings in computational complexity and a large number of parameters. Recently, compared with CNN and Transformer, the Multi-Layer Perceptron (MLP)-based medical image processing network can achieve higher accuracy with smaller computational and parametric quantities. Hence, in this work, we propose an encoder-decoder network, U-MLP, based on the ReMLP block. The ReMLP block contains an overlapping sliding window mechanism and a Multi-head Gate Self-Attention (MGSA) module, where the overlapping sliding window can extract local features of the image like convolution, then combines MGSA to fuse the information extracted from multiple dimensions to obtain more contextual semantic information. Meanwhile, to increase the generalization ability of the model, we design the Vague Region Refinement (VRRE) module, which uses the primary features generated by network inference to create local reference features, thus determining the pixel class by inferring the proximity between local features and labeled features. Extensive experimental evaluation shows U-MLP boosts the performance of segmentation. In the skin lesions, spleen, and left atrium segmentation on three benchmark datasets, our U-MLP method achieved a dice similarity coefficient of 88.27%, 97.61%, and 95.91% on the test set, respectively, outperforming 7 state-of-the-art methods.

Improved MobileNetV2 crop disease identification model for intelligent agriculture

Using intelligent agriculture is an important way for the industry to achieve high-quality development. To improve the accuracy of the identification of crop diseases under conditions of limited computing resources, such as in mobile and edge computing, we propose an improved lightweight MobileNetV2 crop disease identification model. In this study, MobileNetV2 is used as the backbone network for the application of an improved Bottleneck structure. First, the number of operation channels is reduced using point-by-point convolution, the number of parameters of the model is reduced, and the re-parameterized multilayer perceptron (RepMLP) module is introduced; the latter can capture long-distance dependencies between features and obtain local a priori information to enhance the global perception of the model. Second, the efficient channel-attention mechanism is added to adjust the image-feature channel weights so as to improve the recognition accuracy of the model, and the Hardswish activation function is introduced instead of the ReLU6 activation function to further improve performance. The final experimental results show that the improved MobilNetV2 model achieves 99.53% accuracy in the PlantVillage crop disease dataset, which is 0.3% higher than the original model, and the number of covariates is only 0.9M, which is 59% less than the original model. Also, the inference speed is improved by 8.5% over the original model. The crop disease identification method proposed in this article provides a reference for deployment and application on edge and mobile devices.

SLWE-Net: An improved lightweight U-Net for Sargassum extraction from GOCI images

The Sargassum bloom has severely impacted the ecological environment of the East China Sea and the Yellow Sea, causing significant economic losses. In recent years, deep learning has seen extensive development due to its outstanding feature extraction capabilities. However, the deep learning process typically involves a large number of parameters and computations. To address this issue, this paper proposes a lightweight deep learning network based on the U-Net framework, called SLWE-NET, which uses lightweight modules to replace the feature extraction modules in U-Net. In this experiment, SLWE-Net performed the best in both extraction accuracy and model lightweight. Compared to the formal U-Net, the number of parameters decreased by 65.83 %, the model size reduced from 94.97 MB to 32.51 MB, and the mIoU increased to 93.81 %. Therefore, the method proposed in this paper is beneficial for Sargassum extraction and provides a basis for operational monitoring.

LightR-YOLOv5: A compact rotating detector for SARS-CoV-2 antigen-detection rapid diagnostic test results

Nucleic acid testing is currently the golden reference for coronaviruses (SARS-CoV-2) detection, while the SARS-CoV-2 antigen-detection rapid diagnostic tests (RDT) is an important adjunct. RDT can be widely used in the community or regional screening management as self-test tools and may need to be verified by healthcare authorities. However, manual verification of RDT results is a time-consuming task, and existing object detection algorithms usually suffer from high model complexity and computational effort, making them difficult to deploy. We propose LightR-YOLOv5, a compact rotating SARS-CoV-2 antigen-detection RDT results detector. Firstly, we employ an extremely light-weight L-ShuffleNetV2 network as a feature extraction network with a slight reduction in recognition accuracy. Secondly, we combine semantic and texture features in different layers by judiciously combining and employing GSConv, depth-wise convolution, and other modules, and further employ the NAM attention to locate the RDT result detection region. Furthermore, we propose a new data augmentation approach, Single-Copy-Paste, for increasing data samples for the specific task of RDT result detection while achieving a small improvement in model accuracy. Compared with some mainstream rotating object detection networks, the model size of our LightR-YOLOv5 is only 2.03MB, and it is 12.6%, 6.4%, and 7.3% higher in mAP@.5:.95 metrics compared to RetianNet, FCOS, and R³Det, respectively.

A comparative life cycle assessment of graphene nanoplatelets- and glass fibre-reinforced poly(propylene) composites for automotive applications

The circulation of motor vehicles with an exhaust system is one of the leading causes of pollution. Due to the risk factor for human health and contribution to climate change, countries have been growing efforts to achieve a sustainable level of air quality and limit carbon dioxide (CO₂) emissions in recent years. Hence, there has been significant interest in developing technological innovations such as alternative fuels and lighter thermoplastic composites for auto applications. Thermoplastic nanocomposites show substantial properties improvements, incorporating much lower nanofiller percentages than fibre-reinforced thermoplastic composites, allowing for a reduction in the weight of automotive parts (AP). For these motivations, this study presents a comparative life cycle assessment (LCA) of poly(propylene) (PP)-based graphene nanoplatelets (GnPs) nanocomposite and PP/glass fibre (GF) composite for automotive applications. The AP selected as the functional unit was the front end part of 1.5 tons weight car. The LCA included preparing raw materials, AP manufacturing, AP use, and AP end-of-life (EoL). Three different EoL scenarios were considered in this analysis. Several midpoint environmental impact indicators were evaluated. Overall, the results suggest that the different EoL scenarios do not affect the global environmental impact of both AP systems. The environmental impacts of the AP depend on the type of material. The AP of the PP-based GnPs nanocomposite exhibited a weight 28 % lower than the AP of the PP/GF composite. This lightweight resulted in energy and emissions savings, especially during the AP use stage, which was the stage with the most significant contribution in most environmental impact categories. Using nanocomposite reduced the AP environmental impact from 17 % to 75 %, depending on the impact category. The study suggested that substituting traditional composite with a new lighter nanocomposite can decrease the global environmental impacts caused by AP.

Real-time airplane detection using multi-dimensional attention and feature fusion

The remote sensing image airplane object detection tasks remain a challenge such as missed detection and misdetection, and that is due to the low resolution occupied by airplane objects and large background noise. To address the problems above, we propose an AE-YOLO (Accurate and Efficient Yolov4-tiny) algorithm and thus obtain higher detection precision for airplane detection in remote sensing images. A multi-dimensional channel and spatial attention module is designed to filter out background noise information, and we also adopt a local cross-channel interaction strategy without dimensionality reduction so as to reduce the loss of local information caused by the scaling of the fully connected layer. The weighted two-way feature pyramid operation is used to fuse features and the correlation between different channels is learned to improve the utilization of features. A lightweight convolution module is exploited to reconstruct the network, which effectively reduce the parameters and computations while improving the accuracy of the detection model. Extensive experiments validate that the proposed algorithm is more lightweight and efficient for airplane detection. Moreover, experimental results on the airplane dataset show that the proposed algorithm meets real-time requirements, and its detection accuracy is 7.76% higher than the original algorithm.

Lightweight similarity checking for English literatures in mobile edge computing

With the advent of information age, mobile devices have become one of the major convenient equipment that aids people's daily office activities such as academic research, one of whose major tasks is to check the repetition rate or similarity among different English literatures. Traditional literature similarity checking solutions in cloud paradigm often call for intensive computational cost and long waiting time. To tackle this issue, in this paper, we modify the traditional literature similarity checking solution in cloud paradigm to make it suitable for the light-weight mobile edge environment. Furthermore, we put forward a lightweight similarity checking approach SC MEC for English literatures in mobile edge computing environment. To validate the advantages of SC MEC , we have designed massive experiments on a dataset. The reported experimental results show that SC MEC can deliver a satisfactory similarity checking result of literatures compared to other existing approaches.

Nanocellulose interface enhanced all-cellulose foam with controllable strength via a facile liquid phase exchange route

The development of sustainable, biodegradable, non-toxic biomass foams with outstanding physical properties to replace traditional petroleum-based foams is urgent. In this work, we proposed a simple, efficient, and scalable approach to fabricate nanocellulose (NC) interface enhanced all-cellulose foam through ethanol liquid phase exchange and subsequent ambient drying. In this process, NCs served as reinforcer and binder were integrated with pulp fiber to improve cellulose interfibrillar bonding and interface adhesion between NCs and pulp microfibrils. The resultant all-cellulose foam displayed stable microcellular structure (porosity of 91.7-94.5 %), low apparent density (0.08-0.12 g/cm³), and high compression modulus (0.49-2.96 MPa) by regulating the content and size of NCs. Further, the strengthening mechanism of the structure and property of all-cellulose foam were investigated in detail. This proposed process enabled ambient drying, and is simple and feasible for low-cost, practicable, and scalable production of biodegradable, green bio-based foam without special apparatuses and other chemicals.

Lightweight and Strong Ceramic Network with Exceptional Damage Tolerance

Lightweight materials such as porous ceramics have attracted increasing attention for applications in energy conservation, aerospace and automobile industries. However, porous ceramics are usually weak and brittle; in particular, tiny defects could cause catastrophic failure, which affects their reliability and limits the potential use greatly. Here we report a SiC/SiO2 nanowire network constructed from numerous well-bonded SiC nanowires coated by a biphasic structure consisting of amorphous SiO2 and nanocrystal SiC. The as-obtained SiC/SiO2 nanowire network is lightweight (360 ± 10 mg cm-3), mechanically strong (compressive strength of 16 MPa), and damage-tolerant. The high strength of the network is attributed to the biphasic mixed structure of the binding coating which can restrict the deformation of nanowires upon compression. The lightweight and strong SiC/SiO2 nanowire network shows potential for engineering applications in harsh environments.

Printable Aligned Single-Walled Carbon Nanotube Film with Outstanding Thermal Conductivity and Electromagnetic Interference Shielding Performance

Ultrathin, lightweight, and flexible aligned single-walled carbon nanotube (SWCNT) films are fabricated by a facile, environmentally friendly, and scalable printing methodology. The aligned pattern and outstanding intrinsic properties render "metal-like" thermal conductivity of the SWCNT films, as well as excellent mechanical strength, flexibility, and hydrophobicity. Further, the aligned cellular microstructure promotes the electromagnetic interference (EMI) shielding ability of the SWCNTs, leading to excellent shielding effectiveness (SE) of ~ 39 to 90 dB despite a density of only ~ 0.6 g cm^-3 at thicknesses of merely 1.5-24 µm, respectively. An ultrahigh thickness-specific SE of 25 693 dB mm^-1 and an unprecedented normalized specific SE of 428 222 dB cm² g^-1 are accomplished by the freestanding SWCNT films, significantly surpassing previously reported shielding materials. In addition to an EMI SE greater than 54 dB in an ultra-broadband frequency range of around 400 GHz, the films demonstrate excellent EMI shielding stability and reliability when subjected to mechanical deformation, chemical (acid/alkali/organic solvent) corrosion, and high-/low-temperature environments. The novel printed SWCNT films offer significant potential for practical applications in the aerospace, defense, precision components, and smart wearable electronics industries.

A Perspective for Developing Polymer-Based Electromagnetic Interference Shielding Composites

The rapid development of aerospace weapons and equipment, wireless base stations and 5G communication technologies has put forward newer and higher requirements for the comprehensive performances of polymer-based electromagnetic interference (EMI) shielding composites. However, most of currently prepared polymer-based EMI shielding composites are still difficult to combine high performance and multi-functionality. In response to this, based on the research works of relevant researchers as well as our research group, three possible directions to break through the above bottlenecks are proposed, including construction of efficient conductive networks, optimization of multi-interfaces for lightweight and multifunction compatibility design. The future development trends in three directions are prospected, and it is hoped to provide certain theoretical basis and technical guidance for the preparation, research and development of polymer-based EMI shielding composites.

Lightweight, amphipathic and fire-resistant prGO/MXene spherical beads for rapid elimination of hazardous chemicals

Frequent leaks of hazardous chemicals have a huge impact on human lives, property and the ecological environment. Therefore, the three-dimensional functional porous materials with high absorption efficiency and special wettability for the disposal of hazardous chemical spills is an urgent demand. In this work, a series of spherical beads consisting of partially reduced graphene oxide (prGO) and MXene (Ti₃C₂T_x) nanosheets were constructed by hydrogen bond induced self-assembly along with freeze-drying and thermal treatment. The lightweight and amphipathic prGO/MXene spherical beads (prGMSBDs) had millimeter-level size, spherical morphology and highly porous internal structure, which were especially suitable for eliminating hazardous chemicals. Because of their excellent thermal stability and fire retardance, the prGMSBDs could be used to absorb flammable organic liquids, reducing the fire risk of the flammable hazardous chemical spills. Indeed, the prGMSBDs exhibited outstanding absorption performances for various hazardous chemicals, including organic solvents and water-based concentrated acid and alkali. Moreover, the prGMSBDs showed relatively stable absorption performance after five absorption-drying cycles. Due to meeting the requirements of both amphipathic characteristic and flame retardancy, the prGMSBDs reported in this work may offer a promising strategy for rapidly cleaning up various hazardous chemicals and open a feasible route to protecting the combustible hazardous chemical spills from fire.

Lightweight and efficient neural network with SPSA attention for wheat ear detection

Advancements in deep neural networks have made remarkable leap-forwards in crop detection. However, the detection of wheat ears is an important yet challenging task due to the complex background, dense targets, and overlaps between wheat ears. Currently, many detectors have made significant progress in improving detection accuracy. However, some of them are not able to make a good balance between computational cost and precision to meet the needs of deployment in real world. To address these issues, a lightweight and efficient wheat ear detector with Shuffle Polarized Self-Attention (SPSA) is proposed in this paper. Specifically, we first utilize a lightweight backbone network with asymmetric convolution for effective feature extraction. Next, SPSA attention is given to adaptively select focused positions and produce a more discriminative representation of the features. This strategy introduces polarized self-attention to spatial dimension and channel dimension and adopts Shuffle Units to combine those two types of attention mechanisms effectively. Finally, the TanhExp activation function is adopted to accelerate the inference speed and reduce the training time, and CIOU loss is used as the border regression loss function to enhance the detection ability of occlusion and overlaps between targets. Experimental results on the Global Wheat Head Detection dataset show that our method achieves superior detection performance compared with other state-of-the-art approaches.

Chest compressions become deeper when pushing with forward lean: A simulation study

Aim

Chest compression depth (CCD) in cardiopulmonary resuscitation is important. However, lightweight rescuers have difficulty achieving an appropriate depth. Chest compression force (CCFORCE) can be increased by placing the arms at 100° to the patient's frontal plane. In a simulation manikin study, we compared the CCD at 90° and 100° among lightweight Asian females and hypothesized that the CCD would be greater when the arms were placed at 100°.

Methods

We included 35 lightweight female students from Shimane University who performed compressions 30 times each at 90° and 100°. The CCFORCE and CCD and the residual force on the chest wall during decompression for each chest compression were compared using CPRmeter-2.

Results

Of the 35 participants, 3 were excluded because their angles deviated from the prescribed angle. Thirty-two participants were categorized according to CCD at 90°: ≤40 mm (group 1), 41-49 mm (group 2), and ≥ 50 mm (group 3). The overall mean CCD increased from 90° to 100° (44.3 ± 8.2 mm vs. 48.1 ± 7.2 mm; p < 0.05). The mean CCD changes between 90° and 100° were 34.4 ± 4.7 mm vs. 42.9 ± 4.8 mm (p < 0.05) in group 1, 44.9 ± 2.5 mm vs. 47.0 ± 4.2 mm (p = 0.17) in group 2, and 53.0 ± 2.7 mm vs. 55.4 ± 5.6 mm (p < 0.05) in group 3. The residual force was greater when the chest compression angle was 100°.

Conclusion

CCD can be increased for lightweight rescuers when using a forward leaning position of 100° to the frontal plane of the patient. Further research is required to obtain more realistic situations.

Efficient microwave absorber and supercapacitors derived from puffed-rice-based biomass carbon: Effects of activating temperature

Biomass-based carbon is gaining increasing attention because it presents a promising prospect for economic growth and social sustainable development. Moreover, it is an excellent medium for application in electromagnetic and electronic devices. Here, puffed-rice-based carbon is obtained at various activating temperatures, and when the hollow bulges on the carbon disappear, the morphology of the carbon changes into sheet-like structures. The R-800 sample displays the highest I_D/I_G value and demonstrates the best performance when used as both a microwave absorber and an electrode material. The minimum reflection loss (RL) and bandwidth for RL < -10 dB of the R-800 sample reach -72.1 dB and 13.2 GHz, respectively, and the bandwidth for RL < -20 dB is as large as 7.0 GHz, illustrating the widest bandwidth among the five carbon specimens. The multiple reflection effects and scattering, good impedance matching, and interfacial polarization synergistically enhance the microwave absorption performances of the sample. At 1 A g^-1, the specific capacitance of the R-800 sample reaches 117.2 F g^-1 and the capacitance retention remains at 85.3%. Moreover, a hybrid supercapacitor R-800//R-800 demonstrates an outstanding energy density of 15.23 Wh kg^-1, power density of 5739.43 W kg^-1, and high cycle stability (94.5% after 5000 cycles). This functionalized biomass carbon provides a promising media for constructing a bridge between sustainable development and biomass materials.

Hierarchically maze-like structured nanofiber aerogels for effective low-frequency sound absorption

Noise has been regarded as an environmental pollutant that greatly affects people's physical and psychiatric health. Fibrous sound absorption materials are widely used to release the annoyance that brought by noise pollution, however, the fibrous materials are limited by poor sound absorption ability in low-frequency, heavyweight, and excessive thickness. Herein, composite nanofiber aerogels are designed with a hierarchical maze-like microstructure, which is fabricated by interweaving the cellulose nanocrystal lamellas with polyacrylonitrile electrospun nanofiber networks through the freeze-casting technique. The designed maze-like structure shows obvious enhancement in the low-frequency sound absorption band compared to the fiber aerogels made by the network structure. Moreover, through carefully regulating the maze structure, composite nanofiber aerogels with excellent sound absorption performance (with an NRC of 0.58) and lightweight property (11.05 mg cm^-3) can be fabricated. In addition to the superior sound absorption ability, the hierarchical nature of the maze-like structure also guarantees the nanofiber aerogels with robust mechanical properties, which can be tailored to various shaped objects on a large scale. These favorable characters present that the composite nanofiber aerogels potential choice for sound absorption in the fields of vehicles, buildings, and indoor reverberation.

Lightweight, compression-resistant cellular structures inspired from the infructescence of Liquidambar formosana

In order to adapt to the environment, plants have evolved many structural designs to improve material utilization. The head infructescence can be described as the Fibonacci sequence, in consistent with plant developmental biology. The lignified framework inside the head infructescence possesses idiographic structural designs that optimize maximum energy efficiency, growing space, seed spreading probability, and enhance the mechanical behavior of the infructescences. In this study, the hierarchical structure and mechanical properties of the infructescence of Liquidambar formosana, commonly called Formosan gum, were investigated. Liquidambar formosana has maple-like leaves and burr-like infructescences. The buckyball-like framework inside infructescence consists of chambers (cells), which support the whole structure under compression. Inspired by the framework, we proposed three models: Thomson model based on the lowest potential energy state, Poisson disc model indicated random distribution, and spherical Fibonacci model represented plant development. Three-dimensional physical entities of these models were fabricated by additive manufacturing. We discovered that under compression testing, these models appear different mechanical properties and deformation mechanisms based on their structures. Spherical Fibonacci model provides superior mechanical properties compared to Thomson and Poisson disc models due to its unique structural design. It is the first time that spherical Fibonacci model brought into the bio-inspired mechanics models through structural analysis and finite element method. The unique construction of Liquidambar formosana has great potential in the designs of novel lightweight, anti-buckling composites, and bio-inspired architectures.

A review on nanocellulose as a lightweight filler of polyolefin composites

Nanocellulose (NC) possesses low density, high aspect ratio, impressive mechanical properties, nanoscale dimensions, which shows huge potential applications as a reinforced filler. Polyolefin (PO), represented by polyethylene (PE) and polypropylene (PP), has been widely used in industries. Recently nanocellulose/polyolefin nanocomposites (NC/PO nanocomposites) have caught more attention from the application of automotive components, aerospace, furniture, building, home appliances, and sport. In this review, the surface modifications of nanocellulose and polyolefin are summarized respectively, such as surface adsorption modification, small molecule modification, and graft copolymerization modification. The common preparations of NC/PO nanocomposites are discussed, including the melting compounding, the solvent casting, and the in-situ polymerization. The lightweight, mechanical properties, and aging-resistant properties of NC/PO nanocomposites are highlighted. Finally, the potentials and challenges for industrial production development of NC/PO nanocomposites are discussed.

Recognition and Clinical Diagnosis of Cervical Cancer Cells Based on our Improved Lightweight Deep Network for Pathological Image

Accurate recognition of cervical cancer cells is of great significance to clinical diagnosis, but these existing algorithms are designed by low-level manual feature, and their performance improvements are limited an improved algorithm based on residual neural network is proposed to improve the accuracy of diagnosis. Firstly, momentum parameters are introduced into the training model; secondly, by changing the number of training samples, the recognition rate of the algorithm can be improved. Therefore, aiming at the task of object recognition under resource constrained condition, we optimize the design method of the network structure such as convolution operation, model parameter compression and enhancement of feature expression depth, and design and implement the lightweight network model structure for embedded platform. Our proposed deep network model can reduce the parameters of the model and the resources needed for operation under the condition of guaranteeing the precision. The experimental results show that the lightweight deep model has better performance than that of the existing comparison models, and it can achieve the model accuracy of 94.1% under the condition that the model with fewer parameters on cervical cells data set.

Recycling waste thermoplastic for energy efficient construction materials: An experimental investigation

A large stream of research has studied the performance of waste plastics impregnated concrete, reporting multiple benefits and advocating its use in construction works. But no study has reported the merits of bricks impregnated with waste plastics. The present paper reports the results of experiments done on bricks made up of varying percentages of waste thermoplastics (0 - 10% by weight) and sand (60 - 70% by weight), holding percentages of fly ash and ordinary Portland cement constant at 15% (by weight) each. Three types of waste thermoplastics were used, forming three separate batches of bricks. The plastics were polycarbonates, polystyrenes, and mixed plastics. The bricks were cured under water for 28 days. Some of the batches were baked at temperatures ranging from 90 °C to 110 °C for 2 hours in order to melt the plastics to form voids. The bricks made with the above-stated compositions were found to possess low thermal conductivity and adequately high compressive strength. The compressive strength of these bricks is observed to be more than 17 MPa, which lies within the upper half of the range of strengths specified for bricks in the IS 1077:1992 standard. The waste plastics impregnated bricks display high thermal resistance, a feature that can add economic value to the brick manufacturers, motivating them to establish the necessary logistics for collection and use of all types of waste thermoplastics. The paper also presents a regression model to predict the compressive strength of bricks at varying plastic contents. The study, thus, introduces a new strand of research on sustainable recycling of waste thermoplastics in the context of the circular economy.

Biomechanical compatibility of high strength nickel free stainless steel bone plate under lightweight design

High nitrogen nickel-free stainless steel (HNNFSS) has excellent mechanical properties, corrosion resistance and biocompatibility, but its strength advantage is not fully used even though with one time higher than that of the conventional 316 L stainless steel. In this work, the lightweight design of HNNFSS bone plate was studied using finite element analysis, and the effect of lightweight plate fixation on histological and biomechanical behavior of healing bone were also researched on fractured rabbit femur. The finite element analysis results showed that the lightweight plate within 18.2% thickness reduction had higher bending strength and more homogeneous stress distribution compared with 316 L stainless steel plate. There was no obvious difference in radiography, histology analysis of callus and expression pattern of insulin like growth factor-1(IGF-1) of callus between the lightweight HNNFSS plate group and 316 L stainless steel plate group in animal test, and the IGF-1 concentrations of callus and the biomechanical bending test results also showed no statistical significance (p > 0.05), even though the data of the lightweight HNNFSS plate group were relatively better than that of 316 L stainless steel plate group. Therefore, the high nitrogen nickel-free stainless steel has the lightweight potential to keep good fixing function and improve bone healing compared with 316 L stainless steel plate.

A method to optimize the shield compact and lightweight combining the structure with components together by genetic algorithm and MCNP code

To optimize the shield for neutrons and gamma rays compact and lightweight, a method combining the structure and components together was established employing genetic algorithms and MCNP code. As a typical case, the fission energy spectrum of ²³⁵U which mixed neutrons and gamma rays was adopted in this study. Six types of materials were presented and optimized by the method. Spherical geometry was adopted in the optimization after checking the geometry effect. Simulations have made to verify the reliability of the optimization method and the efficiency of the optimized materials. To compare the materials visually and conveniently, the volume and weight needed to build a shield are employed. The results showed that, the composite multilayer material has the best performance.

Outperforming hummingbirds' load-lifting capability with a lightweight hummingbird-like flapping-wing mechanism

The stroke-cam flapping mechanism presented in this paper closely mimics the wing motion of a hovering Rufous hummingbird. It is the only lightweight hummingbird-sized flapping mechanism which generates a harmonic wing stroke with both a high flapping frequency and a large stroke amplitude. Experiments on a lightweight prototype of this stroke-cam mechanism on a 50 mm-long wing demonstrate that a harmonic stroke motion is generated with a peak-to-peak stroke amplitude of 175° at a flapping frequency of 40 Hz. It generated a mass lifting capability of 5.1 g, which is largely sufficient to lift the prototype's mass of 3.39 g and larger than the mass-lifting capability of a Rufous hummingbird. The motor mass of a hummingbird-like robot which drives the stroke-cam mechanism is considerably larger (about five times) than the muscle mass of a hummingbird with comparable load-lifting capability. This paper presents a flapping wing nano aerial vehicle which is designed to possess the same lift- and thrust-generating principles of the Rufous hummingbird. The application is indoor flight. We give an overview of the wing kinematics and some specifications which should be met to develop an artificial wing, and also describe the applications of these in the mechanism which has been developed in this work.

Summary: The stroke-cam flapping mechanism closely mimics the wing motion of a hummingbird. It is the only lightweight hummingbird-sized flapping mechanism which generates a harmonic wing stroke with both a high flapping frequency and a large stroke amplitude.