Search

The existing change detection methods are difficult to balance local features and global features, and the boundary between change objects and backgrounds is blurred, so this paper proposed a remote sensing image building change detection method based on local-global feature coupling and boundary guidance. In the encoding stage, the method adopted parallel convolutional neural network and Transformer to extract the local features and global representation of remote sensing images, respectively. At different scales, the local-global feature coupling module fused local features and global feature representation to enhance the expression ability of image features. In addition, it introduced the boundary guidance branch to obtain the prior boundary information of the change objects, so that its guide change map can highlight the structural characteristics of the building and promote the accurate boundary location. This paper conducted experiments on the LEVIR-CD and WHU datasets, resulting in F1-score of 91.25% and 91.27%, IoU of 83.90% and 83.95%, respectively. The experimental results show that the method has a great improvement in the detection accuracy and good generalization ability. [ABSTRACT FROM AUTHOR]

Objectives: The quality of network structure is the key to obtain excellent classification performance. Most of existing deep models are designed by the trial-and-error method. When the same model processes different hyperspectral images (HSIs), different network hyperparameters need to be set artificially to adapt to the data characteristics of different images, greatly increasing the difficulty of network design. Aiming at the problem of deep learning model design in hyperspectral image classification, this paper proposes an automatic network structure search method for HSI classification. Methods: In the proposed method, the differentiable structure search technology is first used to search the network structure on the source HSI data sets, then the deep network model is constructed by stacking cells, and finally the classification performance is evaluated on the target HSIs. It should be pointed out that, the proposed method only conducts one search on the source HSI data sets, and the obtained deep model can be applied to other target HSI classification tasks, which can effectively improve the utilization rate of the model. To improve the generalization ability and classification accuracy of the model, multi-source and multi-resolution HSIs are employed to construct the precollected data set, and the partial channel connection operation is introduced to further improve the search efficiency. Results: Three HSIs inculding Houston, Chikusei and Pavia Center are selected as source data. Four HSIs including University of Pavia, Indian Pines, Salinas and Houston 2018 are selected as the target HSIs. The source data is used to determine the network model and the target HSIs are used to evaluate the classification performance. Experimental results show that the proposed method can automatically search out a universal deep network model suitable for HSI classification task, and this model can achieve better classification performance than the conventional deep learning models. The overall classification accuracies of University of Pavia, Indian Pines, Salinas and Houston 2018 are 98.15%, 98.74%, 97.30% and 74.47%, respectively. Conclusions: Compared with other deep models, the model in this paper is automatically determined by differentiable architecture search algorithm according to the data characteristics of HSIs, effectively improving the classification accuracy of target HSIs while avoiding complex network design. Extensive experiments verify the applicability and effectiveness of the proposed method in HSI classification. [ABSTRACT FROM AUTHOR]

Since the objects in remote sensing images have the characteristics of arbitrary direction, dense distribution and large scale differences, object detection in remote sensing images has become a challenging problem. Aiming at this problem, this paper systematically reviewed the work related to rotating object detection in deep learning remote sensing images in recent three years. Firstly, this paper introduced the representation method and characteristics of rotated bounding box. Then, it analyzed the existing methods of rotating object detection in remote sensing images from four aspects: the feature extraction network, the generation of rotating anchor and candidate bounding box, the label allocation and sampling strategy, and the loss function. Next, it introduced the commonly used remote sensing image datasets for rotated object detection, and compared the performance of different algorithms. Finally, this paper prospected rotating object detection in remote sensing images. [ABSTRACT FROM AUTHOR]

To improve the problem that existing stereo matching algorithms have difficulty in maintaining a good balance between higher parallax estimation accuracy and faster model inference speed, this paper proposed an efficient and accurate multi-layer cascaded recurrent stereo-matching network. Firstly, this paper designed a multilayer network to introduce position encoding and self-attention mechanisms on the higher resolution feature maps, then updated the disparity values using the disparity refinement strategies of hierarchical recurrent refinement, cascaded refinement, and recurrent refinement to recover the detailed information of stereo images better. In addition, this paper improved the method of disparity iterative update by using a light- weight group correlation layer at a low scale and an adaptive group correlation layer at a large scale to update the disparity values, which reduced the computational effort of parallax iterative update and improved the inference speed of the model. The experimental results show that the proposed algorithm has a faster model inference speed than others while achieving a competitive parallax estimation accuracy. [ABSTRACT FROM AUTHOR]

Aiming at the high computational cost of traditional compressed sensing algorithms, this paper proposed a sparse imaging algorithm based on depth expansion model from the perspective of sparse signal recovery. Firstly, it constructed a complex sparse reconstruction network VAMP-Net. In VAMP-Net, it divided complex regressive echo signal into real part and imaginary part as input. Secondly, it substituted the input into the iterative block based on VAMP algorithm. Finally, it carried out the optimal nonlinear sparse transformation by convolutional neural module to obtain the recovered real part and imaginary part signals, and then merged them to obtain the restored target image. As for the proposed algorithm, this paper used artificial data sets to conduct simulation experiments under different target density, iteration times and noise environment, and compared with the traditional iterative shrinkage threshold algorithm and deep learning reconstruction algorithm. Then it used the measured data with different sparsity for field measurement verification. Experimental results show that the image reconstructed by this algorithm has better performance in NMSE, TBR, reconstruction speed and memory usage. [ABSTRACT FROM AUTHOR]

Underwater image quality degradation has seriously hindered the development of marine engineering technology. Aiming at the problems of underwater image color distortion, low contrast and uneven brightness, in order to solve the special underwater imaging environment cause the above degradation problems, and facilitate the subsequent research work based on underwater images, this paper proposed an underwater image enhancement network model based on self-calibrated illumination framework. The network model integrated self-calibrated illumination module on the basis of convolutional neural network, which could brighten images quickly and flexibly. The method established a cascaded lighting learning process with weight sharing in the self-calibrated illumination module to process underwater images. Considering the computational burden of cascaded pattern, the method also constructed a self-calibrated module to realize the convergence of the results of each stage. This self-calibrated module reduced the computational cost greatly and realized the automatic adjustment of underwater image brightness. Finally, this paper analyzed the experimental results qualitatively and quantitatively, and compared eight advanced enhancement algorithms. The results show that the proposed method is obviously superior to other compared algorithms. The results of the proposed method improve the contrast, correct the color, improve the uneven brightness, and enhance the visual effect. The qualitative and quantitative analysis results show the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Recognizing the phase transition of matter is an important problem in physics research. Convolutional neural network algorithm of confusion label scheme is used to identify quantum phase transition point of two- component Bose-Einstein condensates (BEC) in this paper. By calculating the output accuracy of neural net- work, the W-shape performance curve is obtained. The maximum value in the middle of W-shape performance curve corresponds to the critical point of quantum phase transition. The research results show that the critical point obtained by deep learning is consistent with the analytic calculation results. The deep learning method of confusion label scheme can be applied to the phase transition system of existing two phases. [ABSTRACT FROM AUTHOR]

Aiming at the problem of low interlayer resolution in medical tomographic images, this paper proposed an interlayer super-resolution algorithm based on deformation estimation and motion compensation. The method aimed to enhance interlayer resolution by generating interslice images. Firstly, the algorithm employed U-Net for multi-scale feature extraction and fusion of two adjacent images. To handle the complex deformation of interlayer images, it estimated slice bidirectional deformation by utilizing a warping module based on adaptive collaboration of flows. Multi-scale information fusion module performed feature aggregation on pyramid feature layers, to compensate for the motion of the generated map. Finally, it employed a post-processing network to reduce pixel outliers. The results on two CT datasets show that the PSNR of the proposed algorithm reaches 35.59 dB and 30.76 dB, the output effectively restores image details. Comparative experiments with existing methods demonstrate the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Convolution operations lead to a significant increase in the network size, which makes CNN models difficult to deploy to the embedded hardware platform, and different granularity data is not coordinated with the underlying hardware structure, which leads to low computing efficiency. Based on the reconfigurable array with the computing units supporting multiple bit widths, through software hardware cooperation and reconfigurable computing methods, this paper defined the quantization threshold using KL divergence and random integer method, proposed a strategy for finding the best basis point, designed an instruction set and a parallel mapping scheme supporting multiple bit widths to realize three distinct bit widths in data quantization. The results show the quantization scheme with 8 bit weight and feature map can compress model parameter quantity to about 50% with 2% accuracy loss. The acceleration ratios of quantifying the test images to three different bit widths reach 1.012, 1.273, and 1.556, respectively, which can shorten the execution time by up to 35.7% and reduce memory access times by 56.2%, while only bringing less than 1% relative error. This indicates that this method can achieve efficient neural network computation under three quantization bit widths, thereby implementing hardware acceleration and model compression. [ABSTRACT FROM AUTHOR]

To address the low utilization of raw data and insufficient feature extraction capability of multi-dimensional data in existing remaining useful life prediction methods, this paper proposed a convolutional neural network model based on feature enhancement and spatio-temporal information embedding. Firstly, it adopted a feature enhancement module to extract additional operating condition features and manual features from raw data as auxiliary features. Then, it introduced the spatio-temporal embedding module to encode the spatio-temporal information, embedding the time series information and spatial feature information into the original data. Finally, it concatenated the aforementioned features, and it employed a regression prediction module to capture the inherent relationships in the data and obtain regression prediction results. It evaluated the predictive effectiveness of the proposed model on the commonly used commercial modular aero-propulsion system simulation(C-MAPSS) dataset. The experimental results show that the root mean square error of the proposed model decreases by 8.8% on average over the four subsets compared with other mainstream deep learning methods, and it also outperforms existing state-of-the-art algorithms in prediction accuracy under multiple operating conditions and fault types. The experiments fully verify the effectiveness and accuracy of the proposed model in predicting the remaining useful life of turbofan engines. [ABSTRACT FROM AUTHOR]

In recent years, convolutional neural networks have been widely used in various fields of computer vision and achieved remarkable results. Regularization method is an important part of convolutional neural network, which helps to avoid the overfitting phenomenon of convolutional neural network in the process of model training. There are fewer reviews on regularization methods for convolutional neural networks, and most of them lack a summary of the newly proposed regularization methods. Firstly, this paper conducted a detailed summary of the literature on regularization methods in convolutional neural networks, and classified the regularization methods into parameter regularization, data regularization, label regularization and combinatorial regularization. After that, on the public datasets such as ImageNet, it compared and analyzed the advantages and disadvantages of different regularization methods based on evaluation indexes such as top-1 accuracy and top-5 accuracy. Finally, it discussed the future research trends and work directions of regularization methods for convolutional neural network. [ABSTRACT FROM AUTHOR]

Congenital heart disease is one of the main causes of infant mortality. Echocardiography is the preferred inspection method, but its complexity and repeatability often affect the quality of screening and preoperative diagnosis. Based on deep learning The artificial intelligence method has recently become a research hotspot in the diagnosis of congenital heart disease by echocardiography. This paper reviews the application progress of deep learning in the diagnosis of pediatric congenital heart disease by echocardiography [ABSTRACT FROM AUTHOR]

Most of the Encryption-then-Compression (ETC) methods can only obtain several limited fixed compression ratios. However, arbitrary encryption compression ratios instead of limited fixed compression ratios are more suitable for practical requirements. To this end, this paper proposed a lossy compression algorithm for encrypted color images with arbitrary compression ratios. It combines the uniform and random downsampling to compress the encrypted images, obtaining arbitrary compression ratios of an encrypted image. The receiver receives the compressed sequence of the encrypted image and obtains the decrypted image by decompression and decryption. The proposed scheme then characterizes the lossy reconstruction of the original image from the decrypted image as an optimization problem with the downsampling compression-based constraint. This scheme designs a convolutional neural network-based image reconstruction model for lossy ETC to resolve this problem, which is denoted the ETC-oriented Reconstruction Network (ETRN). ETRN consists of shallow feature extraction (SFE), residual in residual (RIR), residual content supplementation (RCS), and down-sampling constraint (DC). The experimental simulation results show that the encrypted color image lossy compression algorithm proposed in this paper can obtain excellent compression and reconstruction performance, which fully demonstrates the feasibility and effectiveness of this method. [ABSTRACT FROM AUTHOR]

Although high-quality and high-resolution depth maps can significantly improve the performance of computer vision tasks in various natural scenes, the limitations of depth camera hardware make the depth maps captured by consumer-level depth cameras have problems such as low resolution, poor quality, invalid holes, etc. Depth super-resolution reconstruction (DSR) is a kind of technology that can effectively improve the resolution and quality of depth maps, and DSR has become a research hot-spot in the field of computer vision. Firstly, this paper introduced the definition of DSR and the research progress of DSR algorithm at home and abroad in recent years, and then mainly stated and analyzed deep learning DSR reconstruction algorithms. Next, it introduced the depth image quality evaluation criteria. Finally, this paper prospected the application fields of DSR and the challenges as well as the opportunities in the future. [ABSTRACT FROM AUTHOR]

Objective To understand the research hotspots and research trends about convolutional neural networks in the field of oncology imaging diagnosis by analyzing the characteristics of published literature at home and abroad over the past decade. Methods The SCI-E database was used as the data source to retrieve literature about convolutional neural networks in the field of oncology imaging diagnosis published from 2012 to 2022. The distribution characteristics of countries, institutions, journals, co-cited authors, and keywords of the studies were analyzed by CiteSpace software. Results A total of 1088 papers were eventually included, and they were mostly from China, the United States, and India. A total of 39 papers were published by Sun Yat-sen University, the research institution with the highest number of publications. Radiology Nuclear Medicine Medical Imaging was the journal with the highest number of publications. A total of 25 high-frequency keywords and 15 burst keywords were obtained. The formation of 12 author co-citation clusters such as image segmentation and lung nodule, as well as 11 keyword clusters such as automatic segmentation and breast cancer, was observed. Conclusion Current research on convolutional neural networks for oncology imaging diagnosis focuses on oncology segmentation, lung-nodule recognition, assisted diagnosis of breast cancer, and other high-frequency oncology. [ABSTRACT FROM AUTHOR]

Deploying convolutional neural networks (CNN) on embedded devices with limited computing and storage resources poses challenges such as slow execution speed, low computational efficiency, and high power consumption. This paper proposes a novel CNN acceleration architecture based on a heterogeneous platform, and designs and implements a lightweight CNN acceleration system based on MobileNet. Firstly, to reduce hardware resource consumption and data transmission costs, a design method combining dynamic fixed-point quantization and batch normalization fusion is employed to optimize the network model and reduce the hardware design complexity of the acceleration system. Secondly, by implementing convolutional block partitioning, parallel convolutional computation, and data flow optimization, the efficiency of convolutional operations and system throughput are effectively improved. Experimental results on the PYNQ-Z2 platform demonstrate that the MobileNet network inference acceleration scheme implemented by this acceleration system achieves a recognition time of 0.18 seconds per image and a system power consumption of 2.62 watts, representing a 128-fold improvement in acceleration performance compared to an ARM single-core processor. [ABSTRACT FROM AUTHOR]

[Purpose/Significance] The development of the Internet has led to the rapid development of social networks, providing users with a convenient channel for the release, dissemination and acceptance of information. However, its low-threshold characteristics have also given rise to a group of the "Internet water army"・・online social spammers, who are paid to post online comments with particular content and spread false information on purpose. They have become a major problem for the Internet ecology. It is of great significance to detect the Internet water army, prevent their malicious attacks, and combat and eliminate their negative effects on the security of the online public opinion. [Method/Process] First, we analyzed the development process and characteristics of the online social spammers, summarized the algorithms used in previous studies and the characteristics mentioned, and sorted out three research starting points: text features, interaction features and graph structure features. Then, an online social spammer detection method based on deep learning was proposed. Combined with the three aspects of user basic information, historical remarks and interaction behavior, six types of features were extracted from the basic information, recent remarks, social intimacy, interaction behavior, microblog number and membership level. Through feature depth extraction and vector splicing and fusion, the user feature vectors were formed with the same length. Finally, a convolutional neural network was used as a classifier to build an automatic, high-precision and high-efficiency spammer detection model. Two Chinese online spammer datasets collected from the Sina Weibo platform were selected for the experiment. The features of the datasets were spliced and aligned to form the Weibo Spammer 2023 dataset as the model training dataset, which prevented the data features of a single dataset from being too discrete and reducing modle generalization. Considering the overfitting problem in the model training process, we solved the problem by adding abandoned layers. [Results/Conclusions] The online spammer detection model constructed in this experiment has significantly improved in terms of metrics such as precision and accuracy. At the same time, the ablation experiment shows that the six features extracted in this e迎eriment have a positive effect on the detection process. Through empirical analysis, the model constructed in this paper has a high detection accuracy and detection efficiency, which can provide certain technical support and theoretical guidance for online spammer identification. By using machine learning methods to actively identify online social spammer accounts, real-time monitoring and prevention of key spammer accounts can prevent the occurrence of malicious network events more timely and effectively and reduce the risk of illegal forces damaging the public opinion ecology. [ABSTRACT FROM AUTHOR]

Aiming at the problem that the traditional deep learning fish recognition method has a low accuracy rate and the parameters cannot be determined adaptively during the model training process. This paper proposes an improved ShuffleNet fish identification method based on Particle Swarm Optimization( PSO) and Transfer Learning(TL). The research takes 20 species of fish as the object, uses the particle swarm algorithm to take the model's loss function as the fitness function, optimizes the two hyperparameters of batch size and learning rate, and uses the transfer learning method for training to construct TL-PSO-ShuffleNet model. The research shows that compared with the models of AlexNet, MobileNet, and ShuffleNet, the recognition accuracy rate is increased by 57. 89%, 30. 43% and 23. 28%, respectively. The fish identification method proposed in this paper has the characteristics of high accuracy and self-adaptive parameter setting, which provides a reference for the research on automatic fish identification. [ABSTRACT FROM AUTHOR]

Aiming at the problem that the existing single image super segmentation reconstruction methods are difficult to capture complete and effective information in the image, this paper proposed a joint image-frequency supervised image super-resolution reconstruction algorithm, which aimed to utilize the complementary information between multiple domains, thereby obtaining a more complete image feature representation. Firstly, by analyzing the characteristics of image pixels after converting to frequency domain space, this paper proposed a new frequency domain distance supervision loss according to its complex representation, which effectively applied spectral information to the optimization process of convolutional neural networks. Based on the characterization characteristics of different frequency bands in the domain, this paper constructed the spectral weighting loss on the basis of the distance loss in the frequency domain, and applied it to the two frequency bands of low frequency and high frequency respectively. Finally, combined with the supervision of the image domain, it formed a joint optimization of multiple domains and got good performance. The results on Setl4, BlOO and Kodak show that the PSNR and SSIM of the proposed algorithm reach 33.47 dB and 0.985 9, respectively, and the proposed algorithm achieves the best performance compared with several image super-resolution methods. [ABSTRACT FROM AUTHOR]

s an important carrier of network public opinion, social media platforms can provide timely and effective information for emergency management of major disaster events such as earthquakes. However, there is still a lack of research on extracting and analyzing the disaster situation with fine-grained topic information contained in social media data. This paper extracts fine-grained seismic topic information contained in social media data based on deep learning algorithms. Take the “Dali Yangbi Ms6.4 earthquake” event on May 21, 2021 as an example, build a framework of topic information extraction process, by optimizing the training of the model, the overall accuracy reaches more than 80%. The newly acquired Qinghai seismic data set was input into the model for verification, and the expected effect was also achieved. The verification results of two earthquakes show that the model is feasible in the extraction of disaster theme information. Finally, combined with the auxiliary public opinion analysis methods such as keyword mining technology, sentiment analysis and kernel density estimation, the earthquake theme was visualized in space and time. It provides new ideas for disaster research and judgment and emergency management. [ABSTRACT FROM AUTHOR]

[Objectives] Under the high-density breeding mode, broilers in cages are prone to respiratory diseases, which are difficult to prevent. In this paper, a cough detection method for whole-life cycle cough of white-feathered broiler based on vocalization characteristics and deep learning was designed to automatically monitor the cough vocalization of broilers and provide early warning information in time. [Methods] Digital filtering, spectral subtraction denoising and endpoint detection were performed on the vocalization signals of broilers at three different growth stages: 10-19 days old, 20-29 days old and after 30 days of age, and the filter bank (FBank), Mel-frequency cepstral coefficients (MFCC) features and their respective first-order and second-order differences were used as the input of the VGG16 and ResNet18 neural networks to complete the three-category model training of cough, tweet and other vocalizations. [Results] The recognition models constructed by using different vocalization features and neural networks in each age group could accurately realize the vocalization classification, and the FBank-VGG16 model was better at 10-19 days old and 20-29 days old, and the accuracy rates were respectively 94.29% and 97.65%. The accuracy rate of the MFCC-ResNet18 model after 30 days was higher than other models at 98.66%. As the age increased, the overall recognition accuracy of each model increased, ranging from 3% to 7%. [Conclusions] This experimental method could quickly and accurately identify the cough sounds of broilers at different life stages in the actual production environment, and provided technical support for the early detection of respiratory diseases in caged broilers. [ABSTRACT FROM AUTHOR]

To solve the problem of combustion state recognition in solid oxide fuel cell (SOFC) combustor, this paper proposed a combustion state recognition method based on attention mechanism and image feature pyramid. The method adopted the adaptive gamma correction with weighting distribution (AGCWD) to standardize the input images. It combined two single full connections with 1×1 convolution to replace the squeeze and excitation structure, and proposed a hybrid attention structure combined with spatial attention structure to enhance the ability of feature extraction. To provide multi-scale information communication capability, it constructed the multi-scale bidirectional fusion pyramid by means of bidirectional computation and multi-scale fusion. The experimental results show that the proposed method reaches 99.22% accuracy under the premise of 3.98 M parameters and 397 M floating point operations (FLOPs), and effectively identifies the combustion state in SOFC combustor. [ABSTRACT FROM AUTHOR]

In order to solve the problem of strong popularity bias in recommendation lists generated by Bayesian personalized ranking algorithm under the condition of unbalanced data, this paper designed a hybrid recommendation algorithm based on feature embedding to remove popularity bias. Firstly, this paper used convolutional neural network to extract user and item features, determine user preferences, and fill original unbalanced data according to user preferences. Secondly, this paper embedded the user preference features extracted from convolutional neural network into the Bayesian personalized ranking algorithm for hybrid recommendation. Finally, it trained the mixed recommendation model with score filled data, and obtained the personalized ranking list without popularity bias. In order to verify the performance of the algorithm, it conducted the analysis and comparison experiments on Movielens-100 K and Movielens-1 M. And the experimental results show that the popularity bias is reduced by about 50%～60% and the accuracy is improved by about twice. [ABSTRACT FROM AUTHOR]

Action recognition (AR) is a hot research area in computer vision field, and has an extensive application prospect for security monitoring, autopilot, production safety etc. Firstly, this paper analysed the connotation and denotation of AR and put forward the technical challenges; Secondly, the paper analysed and compared the working principles of AR from three aspects: time feature extraction, efficient optimization and long-term feature capture; Thirdly, in order to select suitable AR models for different application scenarios, this paper compared the performance characterization of 43 benchmark AR methods in recent ten years based on UCF101, HMDB51, Something-Something and Kinetics400 data sets. Finally, this paper pointed out the future development direction of AR field, and the research results can provide theoretical reference and technical support for video feature extraction and visual content understanding. [ABSTRACT FROM AUTHOR]

In order to solve the problems of high computational complexity, large computational cost and large number of parameters, this paper proposed the parallel group convolution algorithm based on the domestic SW26010P multi-core processor. The core idea was to use the unique data layout, through the multi-core mapping processing, parallel computing. Experimental results show that compared with single-core serial algorithm, the proposed parallel group convolution algorithm can achieve the highest speed-up ratio of 79.5 and the maximum effective computing power of 186.7MFLOPS. After data parallel optimization of the parallel group convolution algorithm by SIMD instruction, the algorithm obtains the highest speed-up ratio of 10.2 compared with the parallel group convolution algorithm before optimization. [ABSTRACT FROM AUTHOR]

As an important research direction in the field of machine learning, deep learning technology has been infiltrated into various fields of scientific research due to its powerful feature of learning ability and excellent portability. In particular, the achievements in the frontier fields of deep learning have played a positive role in the recognition, classification, segmentation and quantification of medical images. Optical coherence tomography (OCT) technology has the characteristics of non-invasive, high resolution, non-invasive acquisition of micron-level high resolution 3D sectional image of biological tissue, and has been widely used in ophthalmology, vascular, dermatology and other biomedical fields. This paper first introduces the basic principles of OCT and deep learning and the advantages of their combination in the field of skin, and then elaborates the application of deep learning in the field of skin OCT medical images, including the analysis and classification of skin cancer, beauty, fingertip image segmentation, and the qualitative and quantitative evaluation of damaged skin. [ABSTRACT FROM AUTHOR]

Expert recommendation is one of the research hotspots in the online Q&A community. However, the majority of the currently used algorithms concentrate on matching user's static interest and question information, ignoring the efficient capture of user' s dynamic interest representation information, which results in a lack of recommendation accuracy. In order to address the aforementioned issues, this paper proposed an expert recommendation algorithm for the Q&A community that incorporated multi-head self-attention. Firstly, this algorithm built out question encoders of convolutional neural networks and attention mechanisms to process target question and user' s historically answered questions, in order to extract the corresponding question representations. Secondly, it treated the sequence of user's historically answered questions as a time series and used the multi-head self-attention mechanism to learn the dynamic interest representations hidden in the sequence, and combined with the user' s static interest representations to obtain the user' s comprehensive interest representations. Finally, it calculated the similarity between the target question representation and the comprehensive user representation to generate recommendation results. Based on the real data from the Zhihu Q&A community, the experimental results show that the performance of this algorithm is significantly better than the current popular expert recommendation algorithm. [ABSTRACT FROM AUTHOR]

Gas chromatography data of oil-immersed transformers is a kind of multivariate time series, but such data is often missing due to equipment or network failures. Imputation is usually required to form a complete dataset for further business analysis and research. However, the existing imputation models cannot deal with multivariate time series data conveniently to guarantee the efficiency and effect from the inherent characteristics of temporal irregularity and temporal bidirectionality. In this paper, a model Conv-WGAIN is proposed based on the Generative Adversarial Imputation Nets (GAIN). Through the constructed imputation feature map, 2D convolution can be used to learn temporal bidirectional features and simultaneously deal with irregular time intervals. The Wasserstein distance is introduced in discriminator for judgement to improve the stability of the model. Experiments on gas chromatography datasets from a real project and 3 public datasets show that our work is universal for data imputation on multivariate time series missing, and Conv-WGAIN outperforms other baselines with 20.75% to 73.37% in metric RMSE. [ABSTRACT FROM AUTHOR]

Accurate information on the spatial distribution of water is of great significance for monitoring water resources and applications, urban planning, and social and economic development. Remote sensing image segmentation technology based on convolutional neural networks has become an important approach for extracting the spatial distribution of water from remote sensing images. When only convolutional neural networks are used to extract spatial distributions of water from remote sensing images, there are often large differences between the features of edge and internal pixels of water objects, resulting in high noise, fuzzy boundaries, and large differences in the accuracy of extraction of internal and edge pixels. Improving the precision of edge pixel segmentation is the key to improving the precision of the whole extraction result. In this paper, the edge extraction algorithm is used to generate edge images from original images, and remote sensing images and edge images are taken as inputs to establish a water extraction model of high resolution based on semantic feature and edge feature fusion. A semantic and edge feature fusion network, SEF-NET (Semantic Feature and Edge Feature Fusion Network), is used to extract water objects from high-resolution remote sensing images. SEF-NET consists of an encoder, a multi-parallel cavity convolution module, a decoder, and a classifier. The encoder contains a group of semantic feature extraction units and a group of edge feature extraction units, and each group of feature extraction units can extract 4-level features. The multi-parallel cavity convolution module is composed of four extended convolution layers of different cavity sizes in series, which can obtain feature maps at four scales and add them together with the initial input feature maps to obtain multi-scale semantic feature maps. A 4-level decoding unit is set up for the decoder, which splices semantic feature images and edge feature images in series, and then performs feature fusion and upsampling. This strategy can reduce the feature difference between the edge pixel and the inner pixel of the object to obtain high inter-class discrimination and intra-class consistency. SoftMax was used as a classifier to complete pixel classification and generate the final segmentation results. In this paper, the Gaofen Image Dataset, the high-resolution visible light image water object dataset of the 2020 "Xingtucup" High-resolution Remote Sensing Image Interpretation Software Competition, and eight Gaofen-2 images from 2020 were selected for comparative experiments to extract water. SegNet, DeepLabV3, Refinenet and HED-H CNN were the comparison models. The recall rates (91.97%, 92.07%, 93.97%), accuracy rates (91.12%, 98.37%, 97.88%), precision rates (89.56%, 95.07%, 94.06%) and F1 scores (91.54%, 95.12%, 95.88%) were better than those in the comparison models, indicating that the SEF-NET model had greater accuracy and generalization ability in extracting water from high-resolution remote sensing images. Thus, the SEF-NET model served government decision-making and monitoring water pollution better than the comparison models did. [ABSTRACT FROM AUTHOR]

Mild cognitive decline is the early stage of Alzheimer's disease, and the feature extraction and classification of mild cognitive decline using EEG signals is an important method for diagnosing mild cognitive decline. In the automatic detection technology of mild cognitive decline based on electroencephalogram artificial intelligence, the existing research only extracted a certain feature in the electroencephalogram signal or simply concatenates multiple features, which caused these methods to fail to well consider the correlation between different features and it would cause the problem of dimensional disaster. This paper proposed a convolutional neural network based automatic detection algorithm for resting state electroencephalogram data of mild cognitive decline. By extracting the power spectrum and brain network features of the electroencephalogram, it fused the two features by matrix operation, and designed a convolutional neural network to classify the fused features. This method achieves a high accuracy rate on the data set collected by a hospital in Shanghai. In addition, by inputting different subsets of the feature set, this method found the few groups of features that contribute the most to mild cognitive decline, thereby it also had a certain interpretability. Experiments on the dataset of this paper, it p roves the advantages of the power brain network for the automatic diagnosis of mild cognitive decline. [ABSTRACT FROM AUTHOR]

This paper proposed a natural language sentence matching method that combined high-level and low-level semantic information to solve the problems about current natural language sentence matching method fail to integrate common semantic information and it is difficult to capture deep-semantic information. First of all, the method used pre-trained word vector GloVe and character-level word vector to obtained the word embedding representation of sentence P and sentence Q. Secondly, this paper encode red P and Q with bidirectional LSTM, then it contained low-level semantic information through preliminary fusion of P and Q. Thirdly, this paper calculated bidirectional attention between P and Q, then spliced them together to get semantic representation, afterwards it calculated its self-attention to obtained high-level semantic information. Finally, this paper used a heuristic fusion function to fuse the low-level semantic information with the high-level semantic information to obtain the final semantic representation, and it used a convolutional neural network to prediction answers. This paper evaluated the proposed model on two tasks, such as recognition textual entailment, paraphrase recognition. This paper conducted experiments on the SNLI dataset and the Quora dataset. The results show that the accuracy of the proposed algorithm on the SNLI test set is 87. 1%, and the accuracy of the Quora test set is 86. 8%, which verifies the effectiveness of the algorithm in the task of natural language sentence matching. [ABSTRACT FROM AUTHOR]

Most of the existing steganalysis methods have weak generalization ability and cannot effectively detect unknown steganalysis algorithms, which makes the accuracy of their classification greatly reduced in the practical application process. To solve this problem, this paper proposed an image steganalysis method based on group convolution and snapshot ensembling（Snapshot Ensembling steganalysis network, SENet）. Firstly, residual convolution block and group convolution block are used to extract the features of the image. Secondly, the model with the best performance is obtained as the snapshot model in each training period. Finally, the selected snapshot models are integrated to classify the images. This method uses the techniques of group convolution and snapshot ensembling to avoid the high training cost of traditional integration methods and the limited generalization ability of a single classifier. Experimental results show that this method can improve the accuracy of steganographic analysis model, and can effectively classify when the training set and test set are mismatched, and has high model generalization ability. [ABSTRACT FROM AUTHOR]

In the field of spectrum monitoring, it is difficult to identify wideband spectrum signals because of wide frequency band and limited receiver sampling steps. This paper proposed a spectrum signal identification method combining frequency-domain superposition preprocessing and object detection. The domain superposition method accumulated multiple frames of spectrum data to highlight weak signals in the spectrum, then sent the spectrum images to the improved object detection network for signal type identification. Experiments show that the proposed method can effectively identify 7 kinds of spectrum signals. The proposed frequency-domain superposition preprocessing can improve the accuracy of the object detection algorithm and improve the identification ability of weak signals in the spectrum. When the signal-to-noise ratio is 6 dB,the algorithm can achieve an average recognition rate of 89. 7%. [ABSTRACT FROM AUTHOR]

Traditional manually designed feature descriptors are prone to cause a large number of mismatches in the face of multi temporal remote sensing images with large geometric and radiation differences and obvious changes in ground features. This paper developed a convolutional neural network matching algorithm with double screening and proposed double constraints to extract the deep semantic feature map of the image through CNN, it selected the key points that satisfied both the priority maximum principle and the precise extreme value principle from the feature map, and extracted 512-dimension descriptors from the corresponding positions. In the feature matching stage, the strategy of combining the rough to fine reverse matching constraints and RANSAC constraints ensured that the false matching points were eliminated effectively, and retained a certain number of correct matching points, which improved the matching accuracy. The experimental results show that the number of matches is moderate, the correct matching rate is improved compared with other algorithms selected in the experiment, and the image registration error is reduced by about 2 pixels. The running time is obviously improved, matching effect is good, with certain adaptability and robustness. [ABSTRACT FROM AUTHOR]

Malicious face manipulation has a negative impact on social security and stability, and it is a very important issue to accurately detect video images after face tampering. In order to solve the problem of poor real-time performance of video manipulation detection model, this paper proposes a face manipulation video detection model based on ensemble learning dual-stream recurrent neural network, and introduces the voting mechanism in ensemble learning. The model first receives a small number of consecutive sequence frames, extracts spatial features through a convolutional neural network, and introduces central differential convolution to enhance tampering artifacts in the spatial domain. The model then differentiates consecutive sequence frames to enhance tampering artifacts in the temporal domain, while temporal feature extraction is performed through a convolutional neural network. Then, the model splices the dual-stream feature vectors in the spatial domain and the time domain, and performs feature extraction through a recurrent neural network. During the feature extraction process of the recurrent neural network, the frame-by-frame feature information is retained as the input of the subsequent auxiliary frame-level classifier, while the final output of the recurrent neural network is used as the input of the video-level discriminator. Finally, the model introduces the voting mechanism of the integrated model to integrate the outputs of multiple auxiliary frame-level discriminators and video-level discriminators, and introduces a weight hyperparameter γ to balance the importance of the auxiliary framelevel discriminator and video-level discriminator, helping the model to improve detection accuracy. On the FaceForensics++ dataset, the experimental results show that the proposed model improves the average accuracy by 0.4% and 1.0% compared with mainstream detection model. At the same time, the proposed model can only use fewer consecutive frames for manipulation detection, which improves the real-time performance of the model. [ABSTRACT FROM AUTHOR]

To solve the problem of insufficient real-time performance of existing optical flow estimation methods, this paper proposed a lightweight improved PSC-NET model with deep separable convolution ( depth pyramid, warping and cost volume, DS-PWC). One of the improvements was decoupling the conventional two-dimensional convolutional network layer into a deep separable convolutional layer. The other one was adding a weight coefficient based on the number of layers in the pyramid layer by DS-PWC, which greatly reduced the number of model parameters in the network structure without loss of precision. In addition, in the training process, the paper applied data enhancement technologies such as I+ ORE to further improve the generalization ability of estimation and prediction results. The experimental results show that the DS-PWC model was tested in the dataset and the operating efficiency reaches about 58 fps while maintaining the quality. To verify the effectiveness of the algorithm, this paper carried out the ablation experiments of model structure and data enhancement. The results verify the effectiveness of the DS-PWC model. [ABSTRACT FROM AUTHOR]

The image rain removal algorithm analyzes and processes rain images to remove rain streaks and restore clean background scenes, which helps to improve the recognition accuracy of computer vision task, so that it becomes a hot spot in current research. In order to systematically understand the current research status and development trends in this field. Firstly, this paper introduced some typical synthetic rain models. Secondly, it analyzed typical image rain removal algorithm models and methods from two aspects, such as model-driven and data-driven. Then it compared the algorithm evaluation metrics and public rain datasets respectively. Finally, this paper prospected the future development trend of single image rain removal algorithms. [ABSTRACT FROM AUTHOR]

This paper proposed a self-encoding salient region extraction algorithm based on multi-layer convolutional to solve the problem that the existing convolutional neural network wasn’t sufficient in extracting image features, which led to the poor effect of salient extraction. When convolutional networks were used to extract image features, the shallow convolution features generally extracted the detailed features of the image, such as color, texture and location features, and the deep convolution features were generally referred to the semantic features of the image. On the coding layer, this algorithm downsampled the shallow convolution features and fused them into the deep convolution features, and upsampled the deep convolution features to the shallow convolution feature. Experiments show that this method can greatly improve the coding quality. In the decoding, it also fused the convolutional features by encoding to obtain the lost information in the decoding and also obtained a better decoded image. In addition, this paper designed a layer-by-layer supervision method to guide the training of the decoding layer, which used the standard region extraction map by down-sampling as the standard image for supervision training. The experimental results show that this algorithm can increase the F-measure by 0.071 on the basis of PAGRN, and reduce the MEA by 0.031 on average. [ABSTRACT FROM AUTHOR]

The currently widely used lightweight neural networks still have the problems such as large number of calculations and parameters, which make cheap mobile devices with low computation resources cannot quickly complete the image classification tasks. Aiming at this problem, this paper proposed a lightweight neural network that was more suitable for using on cheap mobile devices with low computation resources. This paper introduced cheap linear operations and concatenations of feature maps to reduce the number of calculations and parameters of the neural network. Also, this paper introduced improved residual structure, attention mechanism and label smoothing to improve the accuracy of result. Experiments based on PD-38 dataset show that the proposed neural network can achieve higher classification accuracy by using a smaller number of calculations and parameters than traditional lightweight neural networks. Experiments on the public dataset CIFAR-10 further show that the neural network is universal. [ABSTRACT FROM AUTHOR]

For ESC, this paper proposed a convolutional neural network method based on multi-resolution features and time-frequency attention module. Firstly, compared with the single-resolution spectrogram, multi-channel and multi-resolution features could enrich feature information, realize information complementarity among different feature resolutions, and enhance the expression ability of features. Secondly, for sound signals, this paper proposed a time-frequency attention module. The module firstly used different sizes of one-dimensional convolution to focus on the effective information in the time domain and frequency domain, and then used two-dimensional convolution to fuse the two-domain information to suppress the background noise in the environment sound and eliminate redundant information interference caused by multi-channel and multi-resolution features. The experimental results show that the classification accuracy rates on the two benchmark data sets of ESC-10 and ESC-50 have reached 98.50% and 88.46%, which are 2.70% and 0.76% higher than the latest methods. [ABSTRACT FROM AUTHOR]

In order to recognize complex interactive actions, this paper proposed a feature learning method based on depth information, and used a two-layer classification strategy to solve similar action recognition problems. The method analyzed the depth image action sequence from the frequency domain, extracted the frequency domain features, used VAE for spatial feature compression representation, established HMM to simulate time series changes and performed the first layer action recognition. In order to solve the problem of similar actions, this paper introduced three-dimensional joint point features for the second layer action recognition. Experimental results show that these two types of features can effectively represent the meaning of gestures on the SBU-Kinect action data set, the strategy is simple and effective and can get high recognition accuracy. [ABSTRACT FROM AUTHOR]

In order to solve the problem that the Transformer model cannot apply to mobile devices with limited computational resources due to its huge number of parameters and computational complexity, this paper proposed a mobile-friendly lightweight image recognition network called FuseNet. FuseNet utilized convolutional neural network to extract local feature information and self-attention to excel in modeling global information, and it integrated the features of both local and global representations into a single feature extraction module, which efficiently combined the advantages of the two different structures to achieve a high accuracy with a small model size. Experiments demonstrate that FuseNet with different model sizes can achieve good performance without using pre-training and it can well apply to mobile devices. FuseNet-B achieves 80. 5% accuracy with 14. 8 M parameters on ImageNet-1 K dataset, the performance exceeds the same volume of Transformer models and convolutional neural networks. [ABSTRACT FROM AUTHOR]

To accurately recognize and evaluate tennis actions, by combining computer vision with tennis related knowledge, this paper proposes a tennis action recognition and evaluation method based on PoseC3D. Firstly, a pose estimation model based on resnet-50 is used to detect human targets in tennis video and extract bone key points. Secondly, the PoseC3D model is trained through the video data set collected in the professional tennis court, so that it can classify the sub actions of tennis. Thirdly, the dynamic time warping algorithm is used to evaluate the classified actions. Finally, based on the collected video data set, a large number of experiments are carried out. The results show that the Top1 accuracy of the proposed tennis action recognition method based on PoseC3D can reach 90.8%. Compared with the methods based on graph convolution network, such as AGCN and ST-GCN, it has stronger generalization ability. Moreover, the proposed scoring algorithm based on dynamic time warping can give real-time and accurate evaluation scores for corresponding actions after action classification, reducing the work intensity of tennis teachers and effectively improving the quality of tennis teaching. [ABSTRACT FROM AUTHOR]

At present deep learning has played an important role in various human activity recognition (HAR) tasks. However, the activity data has the particularity of time series and includes body movements. The existing convolutional neural network (CNN) will cause the data to be highly correlated when performing convolutional operations. As the network affects the next layer, the accuracy of network recognition is limited. In order to solve this phenomenon, this paper proposes an improved convolutional neural network with covariance matrix for HAR scenario. It builds a de-correlated network structurethrough matrixtransformationtoeliminatecorrelationproblems. When the network performance is poor, the network can replace the existing BN layer to normalize data. The verification experimentsarefinishedonfourHARpublicdatasets. The proposedneural network is compared withtraditionalCNN modelandBNlayermodel. The results show that the improved neural network is improved by 1% to 2% compared with the previous deep learning networks, which proves that the improved neural network is effective. Furthermore, the application is transplanted to the mobile terminal for real-time activity recognition. [ABSTRACT FROM AUTHOR]

The function of recommendation system is to help users actively finding personalized items that meet theirs preferences and recommend them to users. Collaborative filtering algorithm is a classic algorithm in recommender system, but it is limited by cold start of data and sparsity and has disadvantages such as poor interpretability and poor model generalization ability . This paper studied its shortcomings. By taking the original score matrix in the form of user-project bipartite graph as input, designing the figure convolution neural network as a variant of graph autoencoder, which was obtained by the latent vector of user and item by iteratively aggregating neighbor node information, and combining CNN, this paper proposed a recommendation algorithm based on convolution matrix decomposition to improve the interpretability of the model and generalization ability. And also solved the auxiliary information fusion the data sparseness, and made recommendation performance improved by 1.4% and 1.7% . It provides a new idea for recommendation direction based on graph neural network in the future. [ABSTRACT FROM AUTHOR]

The onset of atrial fibrillation is sudden and often accompanied with serious complications (such as stroke and myocardial infarction) which greatly threats health. With the development of deep learning technology, deep neural network has been widely used in atrial fibrillation classification algorithm due to its advantages including automatic feature extraction. This paper proposes a deep learning framework based on convolutional neural network (CNN) and bi-directional long short-term memory (Bi-LSTM) network. It could predict atrial fibrillation according to ECG. By using CNN, we extracted the morphological characteristics of ECG signals and made sequence reconstruction. The reconstructed sequence was input into Bi-LSTM network, and the rhythm changes of positive and negative timing were analyzed. It could effectively predict the data within 30 minutes before the occurrence of atrial fibrillation, normal sinus rhythm data and the data at the onset of atrial fibrillation. The proposed algorithm was trained and validated with the data of long-term atrial fibrillation database, MIT-BIH atrial fibrillation database and MIT-BIH normal sinus rhythm database; the accuracy reached 93. 3%. [ABSTRACT FROM AUTHOR]

China is the major producer and consumer of pork all around the world. In China, pork production has accounted for more than 60% of meat production for a long time. Nowadays, the pig industry has gradually shifted from decentralized and extensive breeding to intensive large-scale breeding, promoting production efficiency. At present, as people's demand for the quantity and quality of pork is increasing, China's pig industry faces the problem that the output and quality of pork can not meet people's daily needs. With the rise of artificial intelligence technology in recent years, deep learning technology has been developed rapidly and widely used in image and audio recognition, natural language processing, robotics, bioinformatics, chemistry, finance, and other fields. It is also an essential tool in developing precision livestock farming. The body condition, behavior and health status of pigs could directly affect the income level of pig farms, so through using deep learning technology, we can quickly and accurately acquire the relevant information about pigs, and carry out precise management to improve the feeding efficiency and welfare levels of pigs. This paper expounds on the research progress and application status of deep learning technology used in target pig detection, pig image segmentation, pig body condition and abnormal monitoring, as well as pig behavior recognition. Then we put forward the improvement strategy of deep learning technology used in the process of pig feeding, which make it easier for researchers to understand. At the same time, we summarize and analyze the data sources and datasets, application scope, and model optimization of previous works using deep learning technology in pig breeding. In the field of data sources and datasets, mobile computer vision systems are currently more suitable than systems with many fixed cameras when they are applied in pig houses, therefore, further research could focus on how to use deep learning technology to mobile computer vision systems. Deep learning technology requires a large amount of data to learn data features, which is a massive disadvantage for pig applications. To get pre-training weights suitable for agricultural data sets, a large number of public datasets and a unified dataset standard suited to the pig field should be established. In terms of application scope, due to the short application time of deep learning in the process of raising pigs, many critical occasions are not involved or seldom involved in existing research, so the application scope of deep learning should further expand. In model optimization, optimizing the model to meet the needs of practical application scenarios is the direction of future research. Optimizing the model to achieve a better balance between size and model performance requires further study. Optimizing the model to locate the start and end time of the behavior in unclipped video and recognize the behavior of target pigs is the challenge in identifying the behavior of pigs in future. China's research on deep learning is at the top level, so the application prospect of deep learning in pig farming is highly expected. Combined with the actual production scene of pigs, deep learning technology will significantly contribute to the development of precision livestock farming. [ABSTRACT FROM AUTHOR]

The research in 3D human pose estimation from 2D images has achieved great success in recent years. However, the performance of existing 3D human pose estimation methods may degrade significantly in complicated scenarios. To improve the accuracy and robustness of 3D human pose estimation in unconstrained scenarios, this paper proposed ARGP-Pose, a monocular 3D pose estimation framework by exploring the relationship between the joint points of a 3D human pose. To be more specific, the proposed method included a new joint point preprocessing method and a 3D pose estimation network. The preprocessing method enhanced structural features and extracted the relationship among joint points, which were used as input of the following pose estimation network. Also, the proposed network fused local information of each joint point and the global information of the overall pose for rich feature extraction. Additionally, the proposed method extracted the temporal information by using a self-attention module, which achieved further performance boost. Last, for a complex human pose, this method decomposed the prediction of the whole pose into the prediction of each point, which again improved the estimation accuracy for human bodies with complex pose variations. The experimental results obtained on several well-known benchmarking datasets, such as Human3.6M and HumanEva-I, demonstrate the merits and superiority of the proposed method. [ABSTRACT FROM AUTHOR]

Due to the low randomness of character distribution and high randomness of word combination in word-based DGA domain names, the traditional detection methods are difficult to identify them. The method based on deep learning such as LSTM can capture domain's sequence features, but it lacks the word combination features and has low detection accuracy. In order to solve above problems, this paper proposed a new DGA domain name detection method based on fusion embedding layer. In domain embedding stage, it carried out a fusion vector representation based on characters and word segmentation technology . Then, it constructed a hybrid deep learning model based on CNN and BiGRU. The experimental results show that the proposed model has increased by 3 . 1 % and 4. 3% respectively in accuracy compared with the models based on CNN or LSTM alone. When dealing with the matsnu, suppobox and ngioweb DGAs, the proposed model is more effective. [ABSTRACT FROM AUTHOR]

The image dehazing method based on deep learning has achieved better results than the traditional methods, and has been used widely. In this paper, image prior information is integrated into convolutional neural network, and a multi-scale image dehazing network based on non-local and prior constrain ts is proposed. In this network, a non-local and multi-scale reconstruction module is designed to capture the image s elf-similarity and reconstruction information at different scales. A loss function which contains dark channel prior, mean square error and content difference is proposed. At the same time, a phased optimization method is proposed to effectively improve the training efficiency. Compared with other dehazing methods, the proposed method can achieve better dehazing images. [ABSTRACT FROM AUTHOR]