Vehicle Abnormal Behavior Detection Based on Dense Block and Soft Thresholding

With the rapid advancement of social economies,intelligent transportation systems are gaining increasing atten-tion.Central to these systems is the detection of abnormal vehicle behavior,which remains a critical challenge due to the complexity of urban roadways and the variability of external conditions.Current research on detecting abnormal traffic behaviors is still nascent,with significant room for improvement in recognition accuracy.To address this,this research has developed a new model for recognizing abnormal traffic behaviors.This model employs the R3D network as its core architecture,incorporating a dense block to facilitate feature reuse.This approach not only enhances performance with fewer parameters and reduced computational demands but also allows for the acquisition of new features while simplifying the overall network structure.Additionally,this research integrates a self-attentive method that dynamically adjusts to the prevailing traffic conditions,optimizing the relevance of features for the task at hand.For temporal analysis,a Bi-LSTM layer is utilized to extract and learn from time-based data nuances.This research conducted a series of comparative experiments using the UCF-Crime dataset,achieving a notable accuracy of 89.30%on our test set.Our results demonstrate that our model not only operates with fewer parameters but also achieves superior recognition accuracy compared to previous models.

MRI Brain Tumor Segmentation Using 3D U-Net with Dense Encoder Blocks and Residual Decoder Blocks

The main task of magnetic resonance imaging (MRI) automatic brain tumor segmentation is to automaticallysegment the brain tumor edema, peritumoral edema, endoscopic core, enhancing tumor core and nonenhancingtumor core from 3D MR images. Because the location, size, shape and intensity of brain tumors vary greatly, itis very difficult to segment these brain tumor regions automatically. In this paper, by combining the advantagesof DenseNet and ResNet, we proposed a new 3D U-Net with dense encoder blocks and residual decoder blocks.We used dense blocks in the encoder part and residual blocks in the decoder part. The number of output featuremaps increases with the network layers in contracting path of encoder, which is consistent with the characteristicsof dense blocks. Using dense blocks can decrease the number of network parameters, deepen network layers,strengthen feature propagation, alleviate vanishing-gradient and enlarge receptive fields. The residual blockswere used in the decoder to replace the convolution neural block of original U-Net, which made the networkperformance better. Our proposed approach was trained and validated on the BraTS2019 training and validationdata set. We obtained dice scores of 0.901, 0.815 and 0.766 for whole tumor, tumor core and enhancing tumorcore respectively on the BraTS2019 validation data set. Our method has the better performance than the original3D U-Net. The results of our experiment demonstrate that compared with some state-of-the-art methods, ourapproach is a competitive automatic brain tumor segmentation method.

Dense Spatial-Temporal Graph Convolutional Network Based on Lightweight OpenPose for Detecting Falls

Fall behavior is closely related to high mortality in the elderly,so fall detection becomes an important and urgent research area.However,the existing fall detection methods are difficult to be applied in daily life due to a large amount of calculation and poor detection accuracy.To solve the above problems,this paper proposes a dense spatial-temporal graph convolutional network based on lightweight OpenPose.Lightweight OpenPose uses MobileNet as a feature extraction network,and the prediction layer uses bottleneck-asymmetric structure,thus reducing the amount of the network.The bottleneck-asymmetrical structure compresses the number of input channels of feature maps by 1×1 convolution and replaces the 7×7 convolution structure with the asymmetric structure of 1×7 convolution,7×1 convolution,and 7×7 convolution in parallel.The spatial-temporal graph convolutional network divides the multi-layer convolution into dense blocks,and the convolutional layers in each dense block are connected,thus improving the feature transitivity,enhancing the network’s ability to extract features,thus improving the detection accuracy.Two representative datasets,Multiple Cameras Fall dataset(MCF),and Nanyang Technological University Red Green Blue+Depth Action Recognition dataset(NTU RGB+D),are selected for our experiments,among which NTU RGB+D has two evaluation benchmarks.The results show that the proposed model is superior to the current fall detection models.The accuracy of this network on the MCF dataset is 96.3%,and the accuracies on the two evaluation benchmarks of the NTU RGB+D dataset are 85.6%and 93.5%,respectively.

Speech Enhancement via Mask-Mapping Based Residual Dense Network

Masking-based and spectrum mapping-based methods are the two main algorithms of speech enhancement with deep neural network(DNN).But the mapping-based methods only utilizes the phase of noisy speech,which limits the upper bound of speech enhancement performance.Maskingbased methods need to accurately estimate the masking which is still the key problem.Combining the advantages of above two types of methods,this paper proposes the speech enhancement algorithm MM-RDN(maskingmapping residual dense network)based on masking-mapping(MM)and residual dense network(RDN).Using the logarithmic power spectrogram(LPS)of consecutive frames,MM estimates the ideal ratio masking(IRM)matrix of consecutive frames.RDN can make full use of feature maps of all layers.Meanwhile,using the global residual learning to combine the shallow features and deep features,RDN obtains the global dense features from the LPS,thereby improves estimated accuracy of the IRM matrix.Simulations show that the proposed method achieves attractive speech enhancement performance in various acoustic environments.Specifically,in the untrained acoustic test with limited priors,e.g.,unmatched signal-to-noise ratio(SNR)and unmatched noise category,MM-RDN can still outperform the existing convolutional recurrent network(CRN)method in themeasures of perceptual evaluation of speech quality(PESQ)and other evaluation indexes.It indicates that the proposed algorithm is more generalized in untrained conditions.

Speech Enhancement via Residual Dense Generative Adversarial Network

Generative adversarial networks(GANs)are paid more attention to dealing with the end-to-end speech enhancement in recent years.Various GANbased enhancement methods are presented to improve the quality of reconstructed speech.However,the performance of these GAN-based methods is worse than those of masking-based methods.To tackle this problem,we propose speech enhancement method with a residual dense generative adversarial network(RDGAN)contributing to map the log-power spectrum(LPS)of degraded speech to the clean one.In detail,a residual dense block(RDB)architecture is designed to better estimate the LPS of clean speech,which can extract rich local features of LPS through densely connected convolution layers.Meanwhile,sequential RDB connections are incorporated on various scales of LPS.It significantly increases the feature learning flexibility and robustness in the time-frequency domain.Simulations show that the proposed method achieves attractive speech enhancement performance in various acoustic environments.Specifically,in the untrained acoustic test with limited priors,e.g.,unmatched signal-to-noise ratio(SNR)and unmatched noise category,RDGAN can still outperform the existing GAN-based methods and masking-based method in the measures of PESQ and other evaluation indexes.It indicates that our method is more generalized in untrained conditions.

基于YOLOv3-Tiny-D算法的偏光片缺陷检测

随着偏光片的应用日益广泛,对于其生产质量的要求也愈加严苛。采用深度学习的目标检测算法对偏光片的三类瑕疵缺陷进行检测,以解决传统方法检测精度低、硬件成本高的问题,从而优化生产工艺。基于YOLOv3-Tiny算法,采用Dense Block模块与SPP-Net模块对其特征提取网络进行优化,并与待检测目标的实际情况相结合调整优化网络的检测模块,提出一种改进后的算法YOLOv3-Tiny-D。实验表明,所提方法在偏光片数据集上测试时,单张图片在保证检测速度的同时(18ms/张),脏污、划痕、标记3类缺陷的检测正确率为90.74%、98.90%、97.52%,平均正确率95.72%,较原算法提高7%。

United block sequence mapping-based visible light positioning for dense small cell deployment

Visible light positioning （VLP） is an emerging candidate for indoor positioning, which carl simultaneously meet the requirements for accuracy, cost, coverage area, and security. However, intercell interference causcd by light intensity superposition linfits the application of VLP. In this Letter, we propose a united block sequence mapping （UBSM）-based VLP that utilizes superposition to integrate the multidimensional information from dense small cells into 2D information. The experimental result shows that UBSM-based VLP can achieve an accuracy of 1.5 cm with a 0.4 m row spacing and 0.35 m column spacing of LED lights.

3D brain glioma segmentation in MRI through integrating multiple densely connected 2D convolutional neural networks

To overcome the computational burden of processing three-dimensional(3 D)medical scans and the lack of spatial information in two-dimensional(2 D)medical scans,a novel segmentation method was proposed that integrates the segmentation results of three densely connected 2 D convolutional neural networks(2 D-CNNs).In order to combine the lowlevel features and high-level features,we added densely connected blocks in the network structure design so that the low-level features will not be missed as the network layer increases during the learning process.Further,in order to resolve the problems of the blurred boundary of the glioma edema area,we superimposed and fused the T2-weighted fluid-attenuated inversion recovery(FLAIR)modal image and the T2-weighted(T2)modal image to enhance the edema section.For the loss function of network training,we improved the cross-entropy loss function to effectively avoid network over-fitting.On the Multimodal Brain Tumor Image Segmentation Challenge(BraTS)datasets,our method achieves dice similarity coefficient values of 0.84,0.82,and 0.83 on the BraTS2018 training;0.82,0.85,and 0.83 on the BraTS2018 validation;and 0.81,0.78,and 0.83 on the BraTS2013 testing in terms of whole tumors,tumor cores,and enhancing cores,respectively.Experimental results showed that the proposed method achieved promising accuracy and fast processing,demonstrating good potential for clinical medicine.

基于改进深度残差收缩网络的风电机组滚动轴承故障诊断方法

滚动轴承是风电机组关键部件,其运行工况复杂,故障类型难以准确识别。针对传统深度神经网络在强噪声环境下特征学习能力不足的问题,提出一种基于稠密连接模块的改进深度残差收缩网络(Deep residual shrinkage network based on dense block,DB-DRSN),实现强噪声、不同负载工况下滚动轴承故障的高效诊断。首先,将添加不同等级噪声的振动信号间隔采样并矩阵化,构建二维灰度图作为输入样本。然后,基于Dense block构造稠密连接残差收缩模块层(Residual shrinkage block unit based on dense block,DB-RSBU),利用Bottleneck层替代残差收缩模块中的卷积隐层,并加入Concat连接,达到对浅层和深层特征的充分利用。在每次稠密连接后通过1×1卷积进行降维,利用注意力模块和软阈值对逐通道特征赋不同阈值并降噪。最后,输入样本经过卷积池化层和DB-RSBU层堆叠的网络得到分类结果。试验表明,DB-DRSN模型在CWRU与PU滚动轴承数据集上不同噪声等级下的平均诊断准确率分别达到99.80%和96.44%,相比其他模型有更高的准确率、更快的收敛速度和更强的抗干扰能力。引入稠密连接核心思想对网络结构的改进可为基于数据驱动的风电机组滚动轴承故障诊断方法提供新思路。

Adaptive Densely Residual Network for Image Super-Resolution

Many networks are designed to stack a large number of residual blocks,deepen the network and improve network performance through short residual connec-tion,long residual connection,and dense connection.However,without consider-ing different contributions of different depth features to the network,these de-signs have the problem of evaluating the importance of different depth features.To solve this problem,this paper proposes an adaptive densely residual net-work(ADRNet)for the single image super resolution.ADRN realizes the evalua-tion of distributions of different depth features and learns more representative features.An adaptive densely residual block(ADRB)was designed,combining 3 residual blocks(RB)and dense connection was added.It learned the attention score of each dense connection through adaptive dense connections,and the at-tention score reflected the importance of the features of each RB.To further en-hance the performance of ADRB,a multi-direction attention block(MDAB)was introduced to obtain multidirectional context information.Through comparative experiments,it is proved that theproposed ADRNet is superior to the existing methods.Through ablation experiments,it is proved that evaluating features of different depths helps to improve network performance.

Lightweight and Efficient Attention-Based Superresolution Generative Adversarial Networks

To address the problems of lack of high-frequency information and texture details and unstable training in superresolution generative adversarial net-works,this paper optimizes the generator and discriminator based on the SRGAN model.First,the residual dense block is used as the basic structural unit of the gen-erator to improve the network’s feature extraction capability.Second,enhanced lightweight coordinate attention is incorporated to help the network more precisely concentrate on high-frequency location information,thereby allowing the gener-ator to produce more realistic image reconstruction results.Then,we propose a symmetric and efficient pyramidal segmentation attention discriminator network in which the attention mechanism is capable of derivingfiner-grained multiscale spatial information and creating long-term dependencies between multiscale chan-nel attentions,thus enhancing the discriminative ability of the network.Finally,a Charbonnier loss function and a gradient variance loss function with improved robustness are used to better realize the image’s texture structure and enhance the model’s stability.Thefindings from the experiments reveal that the reconstructed image quality enhances the average peak signal-to-noise ratio(PSNR)by 1.59 dB and the structural similarity index(SSIM)by 0.045 when compared to SRGAN on the three test sets.Compared with the state-of-the-art methods,the reconstructed images have a clearer texture structure,richer high-frequency details,and better visual effects.

Advanced Generative Adversarial Network for Image Superresolution

The superresolution(SR)method based on generative adversarial networks(GANs)cannot adequately capture enough diversity from training data,resulting in misalignment between input low resolution(LR)images and output high resolution(HR)images.GAN training has difficulty converging.Based on this,an advanced GAN-based image SR reconstructionmethod is presented.First,the dense connection residual block and attention mechanism are integrated into the GAN generator to improve high-frequency feature extraction.Meanwhile,an added discriminator is added into the GAN discriminant network,which forms a dual discriminator to ensure that the process of training is stable.Second,the more robust Charbonnier loss is used instead of the mean square error(MSE)loss to compare similarities between the obtained image and actual image,and the total variation(TV)loss is employed to smooth the training results.Finally,the experimental results indicate that global structures can be better reconstructed using the method of this paper and texture details of images compared with other SOTA methods.The peak signal-to-noise ratio(PSNR)values by the method of this paper are improved by an average of 2.24 dB,and the structural similarity index measure(SSIM)values are improved by an average of 0.07.