A Lightweight Convolutional Neural Network with Hierarchical Multi-Scale Feature Fusion for Image Classification

Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.

Disparity estimation for multi-scale multi-sensor fusion

The perception module of advanced driver assistance systems plays a vital role.Perception schemes often use a single sensor for data processing and environmental perception or adopt the information processing results of various sensors for the fusion of the detection layer.This paper proposes a multi-scale and multi-sensor data fusion strategy in the front end of perception and accomplishes a multi-sensor function disparity map generation scheme.A binocular stereo vision sensor composed of two cameras and a light deterction and ranging(LiDAR)sensor is used to jointly perceive the environment,and a multi-scale fusion scheme is employed to improve the accuracy of the disparity map.This solution not only has the advantages of dense perception of binocular stereo vision sensors but also considers the perception accuracy of LiDAR sensors.Experiments demonstrate that the multi-scale multi-sensor scheme proposed in this paper significantly improves disparity map estimation.

Attention Guided Multi Scale Feature Fusion Network for Automatic Prostate Segmentation

The precise and automatic segmentation of prostate magnetic resonance imaging(MRI)images is vital for assisting doctors in diagnosing prostate diseases.In recent years,many advanced methods have been applied to prostate segmentation,but due to the variability caused by prostate diseases,automatic segmentation of the prostate presents significant challenges.In this paper,we propose an attention-guided multi-scale feature fusion network(AGMSF-Net)to segment prostate MRI images.We propose an attention mechanism for extracting multi-scale features,and introduce a 3D transformer module to enhance global feature representation by adding it during the transition phase from encoder to decoder.In the decoder stage,a feature fusion module is proposed to obtain global context information.We evaluate our model on MRI images of the prostate acquired from a local hospital.The relative volume difference(RVD)and dice similarity coefficient(DSC)between the results of automatic prostate segmentation and ground truth were 1.21%and 93.68%,respectively.To quantitatively evaluate prostate volume on MRI,which is of significant clinical significance,we propose a unique AGMSF-Net.The essential performance evaluation and validation experiments have demonstrated the effectiveness of our method in automatic prostate segmentation.

Flexible Polydimethylsiloxane Composite with Multi-Scale Conductive Network for Ultra-Strong Electromagnetic Interference Protection

Highly conductive polymer composites(CPCs) with excellent mechanical flexibility are ideal materials for designing excellent electromagnetic interference(EMI) shielding materials,which can be used for the electromagnetic interference protection of flexible electronic devices.It is extremely urgent to fabricate ultra-strong EMI shielding CPCs with efficient conductive networks.In this paper,a novel silver-plated polylactide short fiber(Ag@PL ASF,AAF) was fabricated and was integrated with carbon nanotubes(CNT) to construct a multi-scale conductive network in polydimethylsiloxane(PDMS) matrix.The multi-scale conductive network endowed the flexible PDMS/AAF/CNT composite with excellent electrical conductivity of 440 S m-1and ultra-strong EMI shielding effectiveness(EMI SE) of up to 113 dB,containing only 5.0 vol% of AAF and 3.0 vol% of CNT(11.1wt% conductive filler content).Due to its excellent flexibility,the composite still showed 94% and 90% retention rates of EMI SE even after subjected to a simulated aging strategy(60℃ for 7 days) and 10,000 bending-releasing cycles.This strategy provides an important guidance for designing excellent EMI shielding materials to protect the workspace,environment and sensitive circuits against radiation for flexible electronic devices.

Clothing Parsing Based on Multi-Scale Fusion and Improved Self-Attention Mechanism

Due to the lack of long-range association and spatial location information,fine details and accurate boundaries of complex clothing images cannot always be obtained by using the existing deep learning-based methods.This paper presents a convolutional structure with multi-scale fusion to optimize the step of clothing feature extraction and a self-attention module to capture long-range association information.The structure enables the self-attention mechanism to directly participate in the process of information exchange through the down-scaling projection operation of the multi-scale framework.In addition,the improved self-attention module introduces the extraction of 2-dimensional relative position information to make up for its lack of ability to extract spatial position features from clothing images.The experimental results based on the colorful fashion parsing dataset(CFPD)show that the proposed network structure achieves 53.68%mean intersection over union(mIoU)and has better performance on the clothing parsing task.

Rockburst Intensity Grade Prediction Model Based on Batch Gradient Descent and Multi-Scale Residual Deep Neural Network

Rockburst is a phenomenon in which free surfaces are formed during excavation,which subsequently causes the sudden release of energy in the construction of mines and tunnels.Light rockburst only peels off rock slices without ejection,while severe rockburst causes casualties and property loss.The frequency and degree of rockburst damage increases with the excavation depth.Moreover,rockburst is the leading engineering geological hazard in the excavation process,and thus the prediction of its intensity grade is of great significance to the development of geotechnical engineering.Therefore,the prediction of rockburst intensity grade is one problem that needs to be solved urgently.By comprehensively considering the occurrence mechanism of rockburst,this paper selects the stress index(σθ/σc),brittleness index(σ_(c)/σ_(t)),and rock elastic energy index(Wet)as the rockburst evaluation indexes through the Spearman coefficient method.This overcomes the low accuracy problem of a single evaluation index prediction method.Following this,the BGD-MSR-DNN rockburst intensity grade prediction model based on batch gradient descent and a multi-scale residual deep neural network is proposed.The batch gradient descent(BGD)module is used to replace the gradient descent algorithm,which effectively improves the efficiency of the network and reduces the model training time.Moreover,the multi-scale residual(MSR)module solves the problem of network degradation when there are too many hidden layers of the deep neural network(DNN),thus improving the model prediction accuracy.The experimental results reveal the BGDMSR-DNN model accuracy to reach 97.1%,outperforming other comparable models.Finally,actual projects such as Qinling Tunnel and Daxiangling Tunnel,reached an accuracy of 100%.The model can be applied in mines and tunnel engineering to realize the accurate and rapid prediction of rockburst intensity grade.

MSSTNet:Multi-scale facial videos pulse extraction network based on separable spatiotemporal convolution and dimension separable attention

Background The use of remote photoplethysmography(rPPG)to estimate blood volume pulse in a noncontact manner has been an active research topic in recent years.Existing methods are primarily based on a singlescale region of interest(ROI).However,some noise signals that are not easily separated in a single-scale space can be easily separated in a multi-scale space.Also,existing spatiotemporal networks mainly focus on local spatiotemporal information and do not emphasize temporal information,which is crucial in pulse extraction problems,resulting in insufficient spatiotemporal feature modelling.Methods Here,we propose a multi-scale facial video pulse extraction network based on separable spatiotemporal convolution(SSTC)and dimension separable attention(DSAT).First,to solve the problem of a single-scale ROI,we constructed a multi-scale feature space for initial signal separation.Second,SSTC and DSAT were designed for efficient spatiotemporal correlation modeling,which increased the information interaction between the long-span time and space dimensions;this placed more emphasis on temporal features.Results The signal-to-noise ratio(SNR)of the proposed network reached 9.58dB on the PURE dataset and 6.77dB on the UBFC-rPPG dataset,outperforming state-of-the-art algorithms.Conclusions The results showed that fusing multi-scale signals yielded better results than methods based on only single-scale signals.The proposed SSTC and dimension-separable attention mechanism will contribute to more accurate pulse signal extraction.

Ash Detection of Coal Slime Flotation Tailings Based on Chromatographic Filter Paper Sampling and Multi-Scale Residual Network

The detection of ash content in coal slime flotation tailings using deep learning can be hindered by various factors such as foam,impurities,and changing lighting conditions that disrupt the collection of tailings images.To address this challenge,we present a method for ash content detection in coal slime flotation tailings.This method utilizes chromatographic filter paper sampling and a multi-scale residual network,which we refer to as MRCN.Initially,tailings are sampled using chromatographic filter paper to obtain static tailings images,effectively isolating interference factors at the flotation site.Subsequently,the MRCN,consisting of a multi-scale residual network,is employed to extract image features and compute ash content.Within the MRCN structure,tailings images undergo convolution operations through two parallel branches that utilize convolution kernels of different sizes,enabling the extraction of image features at various scales and capturing a more comprehensive representation of the ash content information.Furthermore,a channel attention mechanism is integrated to enhance the performance of the model.The combination of the multi-scale residual structure and the channel attention mechanism within MRCN results in robust capabilities for image feature extraction and ash content detection.Comparative experiments demonstrate that this proposed approach,based on chromatographic filter paper sampling and the multi-scale residual network,exhibits significantly superior performance in the detection of ash content in coal slime flotation tailings.

Residual Feature Attentional Fusion Network for Lightweight Chest CT Image Super-Resolution

The diagnosis of COVID-19 requires chest computed tomography(CT).High-resolution CT images can provide more diagnostic information to help doctors better diagnose the disease,so it is of clinical importance to study super-resolution(SR)algorithms applied to CT images to improve the reso-lution of CT images.However,most of the existing SR algorithms are studied based on natural images,which are not suitable for medical images;and most of these algorithms improve the reconstruction quality by increasing the network depth,which is not suitable for machines with limited resources.To alleviate these issues,we propose a residual feature attentional fusion network for lightweight chest CT image super-resolution(RFAFN).Specifically,we design a contextual feature extraction block(CFEB)that can extract CT image features more efficiently and accurately than ordinary residual blocks.In addition,we propose a feature-weighted cascading strategy(FWCS)based on attentional feature fusion blocks(AFFB)to utilize the high-frequency detail information extracted by CFEB as much as possible via selectively fusing adjacent level feature information.Finally,we suggest a global hierarchical feature fusion strategy(GHFFS),which can utilize the hierarchical features more effectively than dense concatenation by progressively aggregating the feature information at various levels.Numerous experiments show that our method performs better than most of the state-of-the-art(SOTA)methods on the COVID-19 chest CT dataset.In detail,the peak signal-to-noise ratio(PSNR)is 0.11 dB and 0.47 dB higher on CTtest1 and CTtest2 at×3 SR compared to the suboptimal method,but the number of parameters and multi-adds are reduced by 22K and 0.43G,respectively.Our method can better recover chest CT image quality with fewer computational resources and effectively assist in COVID-19.

A Credit Card Fraud Detection Model Based on Multi-Feature Fusion and Generative Adversarial Network

Credit Card Fraud Detection(CCFD)is an essential technology for banking institutions to control fraud risks and safeguard their reputation.Class imbalance and insufficient representation of feature data relating to credit card transactions are two prevalent issues in the current study field of CCFD,which significantly impact classification models’performance.To address these issues,this research proposes a novel CCFD model based on Multifeature Fusion and Generative Adversarial Networks(MFGAN).The MFGAN model consists of two modules:a multi-feature fusion module for integrating static and dynamic behavior data of cardholders into a unified highdimensional feature space,and a balance module based on the generative adversarial network to decrease the class imbalance ratio.The effectiveness of theMFGAN model is validated on two actual credit card datasets.The impacts of different class balance ratios on the performance of the four resamplingmodels are analyzed,and the contribution of the two different modules to the performance of the MFGAN model is investigated via ablation experiments.Experimental results demonstrate that the proposed model does better than state-of-the-art models in terms of recall,F1,and Area Under the Curve(AUC)metrics,which means that the MFGAN model can help banks find more fraudulent transactions and reduce fraud losses.

Feature Fusion-Based Deep Learning Network to Recognize Table Tennis Actions

A system for classifying four basic table tennis strokes using wearable devices and deep learning networks is proposed in this study.The wearable device consisted of a six-axis sensor,Raspberry Pi 3,and a power bank.Multiple kernel sizes were used in convolutional neural network(CNN)to evaluate their performance for extracting features.Moreover,a multiscale CNN with two kernel sizes was used to perform feature fusion at different scales in a concatenated manner.The CNN achieved recognition of the four table tennis strokes.Experimental data were obtained from20 research participants who wore sensors on the back of their hands while performing the four table tennis strokes in a laboratory environment.The data were collected to verify the performance of the proposed models for wearable devices.Finally,the sensor and multi-scale CNN designed in this study achieved accuracy and F1 scores of 99.58%and 99.16%,respectively,for the four strokes.The accuracy for five-fold cross validation was 99.87%.This result also shows that the multi-scale convolutional neural network has better robustness after fivefold cross validation.

An Intelligent Fault Diagnosis Method of Multi-Scale Deep Feature Fusion Based on Information Entropy

For a single-structure deep learning fault diagnosis model,its disadvantages are an insufficient feature extraction and weak fault classification capability.This paper proposes a multi-scale deep feature fusion intelligent fault diagnosis method based on information entropy.First,a normal autoencoder,denoising autoencoder,sparse autoencoder,and contractive autoencoder are used in parallel to construct a multi-scale deep neural network feature extraction structure.A deep feature fusion strategy based on information entropy is proposed to obtain low-dimensional features and ensure the robustness of the model and the quality of deep features.Finally,the advantage of the deep belief network probability model is used as the fault classifier to identify the faults.The effectiveness of the proposed method was verified by a gearbox test-bed.Experimental results show that,compared with traditional and existing intelligent fault diagnosis methods,the proposed method can obtain representative information and features from the raw data with higher classification accuracy.

Lightweight Image Super-Resolution via Weighted Multi-Scale Residual Network

The tradeoff between efficiency and model size of the convolutional neural network(CNN)is an essential issue for applications of CNN-based algorithms to diverse real-world tasks.Although deep learning-based methods have achieved significant improvements in image super-resolution(SR),current CNNbased techniques mainly contain massive parameters and a high computational complexity,limiting their practical applications.In this paper,we present a fast and lightweight framework,named weighted multi-scale residual network(WMRN),for a better tradeoff between SR performance and computational efficiency.With the modified residual structure,depthwise separable convolutions(DS Convs)are employed to improve convolutional operations’efficiency.Furthermore,several weighted multi-scale residual blocks(WMRBs)are stacked to enhance the multi-scale representation capability.In the reconstruction subnetwork,a group of Conv layers are introduced to filter feature maps to reconstruct the final high-quality image.Extensive experiments were conducted to evaluate the proposed model,and the comparative results with several state-of-the-art algorithms demonstrate the effectiveness of WMRN.

Attention mechanism based multi-scale feature extraction of bearing fault diagnosis

Effective bearing fault diagnosis is vital for the safe and reliable operation of rotating machinery.In practical applications,bearings often work at various rotational speeds as well as load conditions.Yet,the bearing fault diagnosis under multiple conditions is a new subject,which needs to be further explored.Therefore,a multi-scale deep belief network(DBN)method integrated with attention mechanism is proposed for the purpose of extracting the multi-scale core features from vibration signals,containing four primary steps:preprocessing of multi-scale data,feature extraction,feature fusion,and fault classification.The key novelties include multi-scale feature extraction using multi-scale DBN algorithm,and feature fusion using attention mecha-nism.The benchmark dataset from University of Ottawa is applied to validate the effectiveness as well as advantages of this method.Furthermore,the aforementioned method is compared with four classical fault diagnosis methods reported in the literature,and the comparison results show that our pro-posed method has higher diagnostic accuracy and better robustness.

Disease Recognition of Apple Leaf Using Lightweight Multi-Scale Network with ECANet

To solve the problem of difficulty in identifying apple diseases in the natural environment and the low application rate of deep learning recognition networks,a lightweight ResNet(LW-ResNet)model for apple disease recognition is proposed.Based on the deep residual network(ResNet18),the multi-scale feature extraction layer is constructed by group convolution to realize the compression model and improve the extraction ability of different sizes of lesion features.By improving the identity mapping structure to reduce information loss.By introducing the efficient channel attention module(ECANet)to suppress noise from a complex background.The experimental results show that the average precision,recall and F1-score of the LW-ResNet on the test set are 97.80%,97.92%and 97.85%,respectively.The parameter memory is 2.32 MB,which is 94%less than that of ResNet18.Compared with the classic lightweight networks SqueezeNet and MobileNetV2,LW-ResNet has obvious advantages in recognition performance,speed,parameter memory requirement and time complexity.The proposed model has the advantages of low computational cost,low storage cost,strong real-time performance,high identification accuracy,and strong practicability,which can meet the needs of real-time identification task of apple leaf disease on resource-constrained devices.

The deep spatiotemporal network with dual-flow fusion for video-oriented facial expression recognition

The video-oriented facial expression recognition has always been an important issue in emotion perception.At present,the key challenge in most existing methods is how to effectively extract robust features to characterize facial appearance and geometry changes caused by facial motions.On this basis,the video in this paper is divided into multiple segments,each of which is simultaneously described by optical flow and facial landmark trajectory.To deeply delve the emotional information of these two representations,we propose a Deep Spatiotemporal Network with Dual-flow Fusion(defined as DSN-DF),which highlights the region and strength of expressions by spatiotemporal appearance features and the speed of change by spatiotemporal geometry features.Finally,experiments are implemented on CKþand MMI datasets to demonstrate the superiority of the proposed method.

MFF-Net: Multimodal Feature Fusion Network for 3D Object Detection

In complex traffic environment scenarios,it is very important for autonomous vehicles to accurately perceive the dynamic information of other vehicles around the vehicle in advance.The accuracy of 3D object detection will be affected by problems such as illumination changes,object occlusion,and object detection distance.To this purpose,we face these challenges by proposing a multimodal feature fusion network for 3D object detection(MFF-Net).In this research,this paper first uses the spatial transformation projection algorithm to map the image features into the feature space,so that the image features are in the same spatial dimension when fused with the point cloud features.Then,feature channel weighting is performed using an adaptive expression augmentation fusion network to enhance important network features,suppress useless features,and increase the directionality of the network to features.Finally,this paper increases the probability of false detection and missed detection in the non-maximum suppression algo-rithm by increasing the one-dimensional threshold.So far,this paper has constructed a complete 3D target detection network based on multimodal feature fusion.The experimental results show that the proposed achieves an average accuracy of 82.60%on the Karlsruhe Institute of Technology and Toyota Technological Institute(KITTI)dataset,outperforming previous state-of-the-art multimodal fusion networks.In Easy,Moderate,and hard evaluation indicators,the accuracy rate of this paper reaches 90.96%,81.46%,and 75.39%.This shows that the MFF-Net network has good performance in 3D object detection.

Siamese Dense Pixel-Level Fusion Network for Real-Time UAV Tracking

Onboard visual object tracking in unmanned aerial vehicles(UAVs)has attractedmuch interest due to its versatility.Meanwhile,due to high precision,Siamese networks are becoming hot spots in visual object tracking.However,most Siamese trackers fail to balance the tracking accuracy and time within onboard limited computational resources of UAVs.To meet the tracking precision and real-time requirements,this paper proposes a Siamese dense pixel-level network for UAV object tracking named SiamDPL.Specifically,the Siamese network extracts features of the search region and the template region through a parameter-shared backbone network,then performs correlationmatching to obtain the candidate regionwith high similarity.To improve the matching effect of template and search features,this paper designs a dense pixel-level feature fusion module to enhance the matching ability by pixel-wise correlation and enrich the feature diversity by dense connection.An attention module composed of self-attention and channel attention is introduced to learn global context information and selectively emphasize the target feature region in the spatial and channel dimensions.In addition,a target localization module is designed to improve target location accuracy.Compared with other advanced trackers,experiments on two public benchmarks,which are UAV123@10fps and UAV20L fromthe unmanned air vehicle123(UAV123)dataset,show that SiamDPL can achieve superior performance and low complexity with a running speed of 100.1 fps on NVIDIA TITAN RTX.

Multi-Feature Fusion Book Recommendation Model Based on Deep Neural Network

The traditional recommendation algorithm represented by the collaborative filtering algorithm is the most classical and widely recommended algorithm in the practical industry.Most book recommendation systems also use this algorithm.However,the traditional recommendation algorithm represented by the collaborative filtering algorithm cannot deal with the data sparsity well.This algorithm only uses the shallow feature design of the interaction between readers and books,so it fails to achieve the high-level abstract learning of the relevant attribute features of readers and books,leading to a decline in recommendation performance.Given the above problems,this study uses deep learning technology to model readers’book borrowing probability.It builds a recommendation system model through themulti-layer neural network and inputs the features extracted from readers and books into the network,and then profoundly integrates the features of readers and books through the multi-layer neural network.The hidden deep interaction between readers and books is explored accordingly.Thus,the quality of book recommendation performance will be significantly improved.In the experiment,the evaluation indexes ofHR@10,MRR,andNDCGof the deep neural network recommendation model constructed in this paper are higher than those of the traditional recommendation algorithm,which verifies the effectiveness of the model in the book recommendation.

Improved Weather Radar Echo Extrapolation Through Wind Speed Data Fusion Using a New Spatiotemporal Neural Network Model

Weather radar echo extrapolation plays a crucial role in weather forecasting.However,traditional weather radar echo extrapolation methods are not very accurate and do not make full use of historical data.Deep learning algorithms based on Recurrent Neural Networks also have the problem of accumulating errors.Moreover,it is difficult to obtain higher accuracy by relying on a single historical radar echo observation.Therefore,in this study,we constructed the Fusion GRU module,which leverages a cascade structure to effectively combine radar echo data and mean wind data.We also designed the Top Connection so that the model can capture the global spatial relationship to construct constraints on the predictions.Based on the Jiangsu Province dataset,we compared some models.The results show that our proposed model,Cascade Fusion Spatiotemporal Network(CFSN),improved the critical success index(CSI)by 10.7%over the baseline at the threshold of 30 dBZ.Ablation experiments further validated the effectiveness of our model.Similarly,the CSI of the complete CFSN was 0.004 higher than the suboptimal solution without the cross-attention module at the threshold of 30 dBZ.