Search

Aiming at the problem that feature extraction is inaccurate and the reconstruction image is not natural enough in blind super-resolution reconstruction, this paper proposed an image blind super-resolution reconstruction based on degradation aware and sequence residuals. This paper proposed a mini-residual group combined degeneration aware and sequence residuals as the backbone network. Then the method constructed a symmetrical enhanced multi-scale residual block. In the image reconstruction part, this paper used the bottleneck attention module and the sub-pixel convolutional module to emphasize the multi-dimensional elements of the image. Finally, the method made a global residual connection. Compared with the current representative algorithm DASR, experiments show that the PSNR and SSIM of the proposed algorithm are improved 0.145dB and 0.0014 on Set14×2, and the PSNR of the proposed algorithm is improved 1.898dB and 0.252dB on Set14×3/4, respectively. The proposed algorithm achieves better performance than several current image super-resolution algorithms on five standard test sets. [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]

Aiming at the problem of insufficient feature extraction ability of deep clustering algorithm on multivariate time series data( MTS), This paper proposes a new deep clustering structure model (MDTC. In order to extract the key features of MTS and realize dimensionality reduction, one-dimensional convolution learns the attribute and temporal dimension feature representation of MTS and the autoencoder structure composed of network layers such as recurrent neural network; To improve the model’s ability to represent temporal features, this paper proposes MCBAM temporal attention module, which is used to enhance the representation features of different time periods in the MTS sequence. This paper conducts experiments on nine publicly available UEA multivariate time series datasets, compared with other autoencoders, the autoencoder structure of the model improves by 2%-9% on 7 data sets; Compared with other attention modules, the MCBAM module of the model improves by 0.3%-2% on the six datasets. Experiments show the effectiveness of the MDTC model structure and MCBAM module, and the model has excellent performance compared with other clustering algorithms. [ABSTRACT FROM AUTHOR]

To address the problem of unbalanced computational resource consumption and accuracy of semantic segmentation models using attention mechanism, this paper proposed a lightweight attention enhancement algorithm for semantic segmentation. Firstly, it designed a striped dimensional attention mechanism based on the shape characteristics of objects in driving scenes, used striped pooling instead of traditional square convolution, and combined dimensionality reduction operations to extract long-range semantic associations in each dimension to cut down the model computation. Then it fused the attention on channel domain and spatial domain to form a lightweight multidimensional attention fusion module that could be superimposed and disassembled to extract feature information in all directions and further improve the model accuracy. Finally, it inserted the module into the ResNet-101 backbone based encoding-decoding network to guide the semantic fusion of high and low layers, correct the feature map edge information, and supplement the prediction details. The experiments show that the module has strong robustness and generalization ability, cutting about 90% of the number of parameters and 80% of the computation compared with the same type of attention mechanism, and the segmentation accuracy still achieves a stable improvement. [ABSTRACT FROM AUTHOR]

Aiming at the problems of low detection accuracy in massive unbalanced data and poor model generalization ability caused by data heterogeneity in traditional anomaly detection methods in IoT scenarios, this paper proposed an adversarial dual-coding anomaly detection network based on federated learning(GB-AEnet-FL). Firstly, this paper proposed an active feature distribution learning algorithm based on abnormal data, which actively learns the potential feature distribution of data, expands abnormal data through data reconstruction, and balances the proportion of positive and negative samples. Secondly, this paper introduced an adversarial training mechanism in the latent feature layer adding consistency enhancement constraints and shrinkage constraints to improve the accuracy of feature extraction. Finally, this paper designed a federated learning algorithm based on dynamic model selection which compares the confidence scores of the local model and the global model, dynamically selects part of the federated bodies to participate, accelerates the aggregation of models, and protects user privacy to a certain extent. Validated on four different datasets, the results show that the proposed algorithm is better than the traditional algorithm in detection accuracy, and the generalization ability is improved accordingly. [ABSTRACT FROM AUTHOR]

In order to solve the problems of large number of parameters and high GPU resource cost in binocular vision stereo matching method based on neural network in robot and unmanned driving fields, this paper proposed a multi-scale aggregation stereo matching method. Firstly, this paper proposed a multi-scale feature extraction network to obtain richer features without changing the resolution by using dilated convolution, and introduced the attention mechanism was. Then, features at different resolutions were cross-fused to improve the feature information. Secondly, the acquisition method of cost volume was changed, the cost volume was obtained by aggregation at low scale, and continuously combined the high-scale similar information to update iteratively, and cross fused multiple cost volumes to obtain the final cost volume. Finally, combined with the refinement module of attention mechanism, the outliers and discontinuous regions in the initial disparity map are corrected to obtain the final disparity map. Experimental results show that the algorithm can run under low parameter number and low cost GPU resources, and obtain good matching accuracy. [ABSTRACT FROM AUTHOR]

In dynamic gesture recognition methods based on deep learning, aiming at the problems of large network scale, simple dimension of extracted spatiotemporal features, and insufficient extraction ability of effective features, this paper proposed a deep network framework. Firstly, this paper applied a novel spatiotemporal convolution module based on multiscale information fusion in the time domain to improve the 3D residual network structure, greatly reduced the size of the network and obtained rich spatiotemporal receptive field characteristics. Then it introduced a spatiotemporal feature channel attention mechanism with global information synchronization, and used a few parameters to construct the global dependency between feature maps, and the module could obtain the key features of dynamic gestures more efficiently. The experimental results on the self-built gesture dataset SHC and the public gesture dataset SKIG show that the proposed gesture recognition method has fewer parameters and more powerful multiscale spatiotemporal feature extraction ability, and gains a higher recognition rate than the current mainstream algorithms. [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]

Due to low line extraction and matching accuracy, existing point-line fusion visual SLAM（simultaneous localization and mapping） methods have low localization accuracy. This paper improved existing point-line fusion stereo visual SLAM method from multiple dimensions. Firstly, this paper improved the LSD（line segment detection） extraction quality through internal parameters adjustment and length threshold filtering. Secondly, on the basis of geometric constraints, it abstracted the line feature matching as a sparse minimization problem to find best match of line segment. Thirdly, on the basis of image texture score and number of successfully matched features, this paper improved back-end point-line feature fusion strategy by assigning point-line feature weights. Compared with LSD algorithm and LBD（line binary descriptor） matching method, the number of short line segments decreased by 67% on average and matching accuracy increased by 18% on average. Compared with PL-SLAM, localization error dropped by about 15% and 45% respectively on the EuRoC and KITTI datasets. Experiments demonstrate that the proposed method enhances localization accuracy of SLAM system effectively by improving various modules performance of point-line SLAM method in multiple dimensions. [ABSTRACT FROM AUTHOR]

Aiming at the feature extraction problem of feature-based visual odometry, this paper proposed a feature selection algorithm based on mutual information and semantic segmentation uncertainty. Based on semantic and geometric information of features, the algorithm points extracted from potential static objects. It corrected the semantic uncertainty of features based on the semantic context, and selected the features based on the information entropy changes of features. Finally, this paper used the public KITTI visual odometry dataset to evaluate the algorithm and compared the experiment results with other algorithms. The results prove that the proposed algorithm has better accuracy of pose estimation, which verifies the effectiveness and feasibility of the algorithm. [ABSTRACT FROM AUTHOR]

Aiming at the problems of the current semantic segmentation network's spatial and channel feature mismatch, as well as the pixel loss of small target objects, this paper designed a dual-path semantic segmentation algorithm based on spatial feature extraction and attention mechanism. The spatial information path used four times downsampling to retain high-resolution features, and introduced a spatial feature extraction module to fuse multi-scale spatial information, thereby strengthening the network's ability to recognize small target objects. In addition, it used a semantic context path combined with two-stage channel attention to extract discriminative features, so that deep features could guide shallow features to capture more accurate semantic information, thereby reducing accuracy loss. This paper verified the algorithm on the CamVid dataset and Aeroscapes dataset, the mean intersection over union can reach 70. 5% and 51. 8% respectively. Compared with the current mainstream dual-path semantic segmentation model, the results verify the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Single-view 3D reconstruction based on deep learning is a research hot spot at present. In order to discover more high-frequency details, SDF-SRN algorithm introduces positional encoding, but neural network is easy to overfit without accurate supervision, and reconstructs uneven surface. To solve the problem, this paper proposed the network model based on sparse feature. The model enabled the network that preferred to overfitting to predict high-frequency residual by residual learning. The feature extraction network extracted sparse features and the global features. Then one hypernetwork took the sparse features as input and generated prediction shallow head. This shallow head predicted low-frequency part of signed distance function. Another hypernetwork took global features as input and generated another shallow head. This shallow head predicted high-frequency residual. It fused two predictions of shallow heads into final signed distance function. Spectrum analysis shows that the design purpose of network is achieved. Compared with other smooth surface reconstruction schemes, the network can achieve smoother surface reconstruction with enough details. It overcomes the overfitting of SDF-SRN. The qualitative and quantitative comparison with other advanced single-view reconstruction approaches show the superiority of the proposed approach. [ABSTRACT FROM AUTHOR]

In recommender systems, extracting higher order information about the graph using methods such as graph convolutional network alleviates the cold start problem. In order to fuse the deep feature extraction capability of neural network collaborative filtering on this basis, this paper proposed a two-channel collaborative filtering recommendation algorithm (GCNCF-2C) based on graph convolution. Firstly, it divided the recommendation problem into an upstream task and a downstream task. Secondly, to form a feature representation of nodes, in the upstream task, a pre-trained encoder extracted node features and higher-order features of the graph using a one-dimensional convolutional layer containing residuals and multiple graph convolutional layers in two independent channels. Finally, the decoder performed end-to-end training by node features for rating prediction. Experiments on the datasets MovieLens-100k and MovieLens-1M show that compared to the baseline models, the algorithm improves RMSE metrics by 1.72% and 1.76% on average and MAE metrics by 2.7% and 1.98% on average on the two datasets. Meanwhile, it improves RMSE metrics by 5.9% on average in the cold-start experiments based on users and projects, which shows overall performance. [ABSTRACT FROM AUTHOR]

The current ADMET classification methods have shortcomings in classifying the ADMET of compounds data with multiple characteristics and feature correlation. Moreover, the classification results of ADMET are not explainable. To solve these problems, this paper proposed a classification model based on capsule network（CapsMC). CapsMC model first proposed the Feature-to-Image algorithm. It used this algorithm to consider the correlation and dependence between features into the classification basis, and realized the multi-level extraction of features. Second, it explored the advanced application of capsule network, and designed a cognitive reasoning mechanism. It used this mechanism to carry out cognitive reasoning on features, and realized explainable classification of ADMET. Experimental results on five ADMET datasets show that CapsMC can achieve the explainable classification of ADMET well. [ABSTRACT FROM AUTHOR]

In the large data environment, there are many problems in the parallel deep convolution neural network（DCNN） algorithm, such as excessive data redundancy, slow convolution layer operation and poor convergence of loss function. This paper proposed a parallel deep convolution neural network optimization algorithm based on the Im2 col method. First, the algorithm proposed a parallel feature extraction strategy based on Marr-Hildreth operator to extract target features from data as input of convolution neural network, which can effectively avoid the problem of excessive data redundancy. Secondly, the algorithm designed a parallel model training strategy based on the Im2 col method. The redundant convolution kernel is removed by designing the Mahalanobis distance center value, and the convolution layer operation speed is improved by combining the MapReduce and Im2 col methods. Finally, the algorithm proposed an improved small-batch gradient descent strategy, which eliminates the effect of abnormal data on the batch gradient and solves the problem of poor convergence of the loss function. The experimental results show that IA-PDCNNOA algorithm performs well in deep convolution neural network calculation under large data environment and is suitable for parallel DCNN model training of large datasets. [ABSTRACT FROM AUTHOR]

In view of the large amount of vehicle detection model parameters and missing detection of small targets and occluded targets. This paper presents a vehicle detection algorithm based on lightweight MobileVit. Firstly, in the data preprocessing stage, the method used GridMask image enhancement to improve the performance of occluded target detection; Secondly, the method used MobileVit network as the backbone feature extraction network of the model to fully extract the feature information and make the model lightweight; Finally, in the prediction layer network, the method used multiscale vehicle detection and recognition based on PANet network to improve the detection performance of the model for small targets. The experimental results show that the average detection accuracy of this algorithm is 98.24%, the detection speed is 58 ms per picture, and the model size is 136 MB, compared with the comparison algorithm, the comprehensive performance is better. [ABSTRACT FROM AUTHOR]

The purpose of aspect-level sentiment analysis is to determine the sentiment of specific aspect words in a sentence . In recent years, many methods have adopted syntactic dependency tree combined with graph convolutional network modeling. But the use of syntactic dependency structures is too direct and ignores the noise effect that accompanies the spanning tree, which limits the use of syntactic relations. This paper proposed an emotional classification model ( PCB-GCN) with progressive enhancement combined with a bidirectional graph convolution module. Firstly, it designed a progressive enhancement algorithm to obtain richer syntactic relations, used Bi-LSTM to extract semantics, and used bidirectional graph convolution module for feature extraction for syntactic graph structures in different directions. Finally, it combined the syntactic features and context semantics through a collaborative network, combined them for the final classification. The model has been tested on multiple public data sets, and all have achieved better results than the current baseline model. [ABSTRACT FROM AUTHOR]

Aiming at the problem of missing detection and false detection caused by inaccurate motion feature extraction of existing video key frame extraction algorithms, this paper proposed a reinforcement learning key frame extraction algorithm combining multi-channel feature and attention mechanism. The algorithm extracted the human skeleton joint points from the video sequence through the human posture recognition algorithm firstly. Then it used the S-GCN and ResNet50 network to extract the motion features and static features in the video sequence respectively, and performed a weighted fusion of the two. Finally, it applied the attention mechanism to calculate the importance of the video frame of the feature sequence, and used reinforcement learning to extract and optimize key frames. The experimental results show that the algorithm can solve the problem of missing and false detection in the key frame extraction of motion video. It performs well in the detection of video frames containing key actions, with high accuracy and strong stability. [ABSTRACT FROM AUTHOR]

Using spikes to represent and convey information, SNN is more biologically plausible than traditional artificial neural networks. However, a classical SNN has limited feature extraction ability due to the shallow network structure, leading to inferior classification performance to CNN especially on multi-class classification tasks such as object categorization. Inspired by the powerful convolutional structure of CNN, this paper proposed a SCSNN. By exploiting convolutional structures and dynamic impulse triggered property of biological neurons, the proposed SCSNN organized integrate-and-fire models in a convolutional fashion, and trained by a new surrogate gradient back-propagation algorithm directly. It validated the proposed SCSNN on the MNIST and Fashion-MNIST dataset respectively, obtaining superior performance to state-or-the-art SNN on both datasets. The classification accuracy reaches 99. 62% on the MNIST dataset, and 93. 52% on the Fashion-MNIST dataset, which verifies the effectiveness of the proposed model. [ABSTRACT FROM AUTHOR]

Aiming at the problem of inefficient operation of the point-line visual odometry (PL-VO) in the simultaneous location and mapping technology when the environmental texture changes, this paper designed feature extraction adaptive optimizer based on environmental information entropy to improve the efficiency and robustness of the original PL-VO algorithm. The optimizer used image information entropy as the main influencing factor to determine the optimal extraction features of the odometry, and generated a strategy map with included feature extraction options. It could also perform predictive calculations on the unexplored area, and quickly matched it with the strategy map to obtain the best feature extraction strategy for the area. It tested the average processing time and mapping effect of the PL-VO with an optimizer (APL-VO) in the TUM dataset. Experimental results show that APL-VO has stronger robustness and mapping efficiency in a hybrid environment than the original algorithm. [ABSTRACT FROM AUTHOR]

To solve the problems of difficult initialization and low positioning accuracy of SLAM algorithm caused by sparse feature points and fast camera movement in practical application scenes, this paper proposed a visual-inertial SLAM algorithm. It used EDLines feature extraction and matching algorithm to introduce rich line features to improve the robustness of feature tracking. Using the complementary advantages of the inertial sensor and the visual sensor, it carried out the optimization in three steps in the initial stage through the visual-inertial initialization by steps method, so as to improve the accuracy and speed of the initialization. Experiments show that the line feature extraction and matching algorithm used in this algorithm provides faster matching speed and lower mismatching rate than the traditional LSD algorithm. Moreover, the proposed algorithm has higher accuracy and stronger robustness in the estimation of camera pose in the visual-inertial data set. [ABSTRACT FROM AUTHOR]

This paper discussed existing gesture-based emotion recognition researches from three aspects, such as gesture-related emotion expression features, emotion recognition algorithm and adopted emotional gesture datasets. It analyzed the difficulties and challenges for gesture-based emotion recognition research, and named two key tasks of gesture-based emotion recognition: extraction of emotion-expressive features from body posture and the build-up of emotion body gesture dataset. It discussed the potential directions of gesture-based emotion recognition research at the end. [ABSTRACT FROM AUTHOR]

Aiming at the problem of different sizes and different shapes of skin lesions in skin disease segmentation, this paper proposed a double U-shaped skin disease segmentation algorithm based on multi-scale feature fusion. The algorithm consisted of two parts:coarse U-shaped network and subdivided U-shaped network. Firstly, the coarse U-shaped network encoder used the pre-trained VGG-19 model to extract multi-scale features for related features. In the decoder stage, the improved attention residual block effectively mapped and merged the low-level and high-level information to obtain a preliminary Mask. Then it aggregated the preliminary Mask with the original image, and input it into the multi-path feature extraction encoder for secondary feature distillation. The subdivided U-shaped network decoder simultaneously integrated with the feature mapping of the course U shaped network encoder and the subdivided U-shaped network encoder to ensure that the network could aggregate more effective features. Finally, the Focal Tversky loss function further improved the segmentation effect. Experimental results on the ISBI 2016 dataset show that the accuracy, sensitivity, specificity, Dice coefficient and Jaccard coefficient of the proposed method are 96.11%.93.59% 97. 10%.93.14% and 87. 17%, respectively, which can effectively segment the skin disease area. [ABSTRACT FROM AUTHOR]

The identification of encryption algorithms is the primary task of cryptanalysis. The current identification schemes cover the identification of block ciphers, the identification of stream ciphers, and the identification of hybrid algorithms of sym-metric ciphers and public key ciphers. With the popularization of Chinese national cryptographic algorithms, their security has also attracted much attention. For the national secret algorithm, this paper proposed the SM4 block cipher system identification based on random characteristics, and identified the national secret SM4 algorithm with the main international standard block cipher algorithm. It used the randomness test method to perform feature extraction on the ciphertext to obtain feature vectors. These feature vectors formed the feature space. Finally, it used machine learning algorithms to train and test the feature space. Experimental results show that the recognition rate of SM4 and other block cipher algorithms are over 90%. This work provides new possibilities for cryptanalysis and at the same time increases the diversity of password identification schemes. [ABSTRACT FROM AUTHOR]

In order to fuse heterogeneous data with high precision, then express the buildings roof and facade information fully, this paper proposed a registration method of terrain laser scanning point cloud and image matching point cloud based on building footprint features. It extracted image matching point cloud building roof edge by boundary estimation, got terrain laser scanning point cloud building eaves boundary by a-shape algorithm, completed two footprint point cloud' s coarse registration by principal component analysis algorithm, centroid constraint and Rodrigues' s formula, finished fine registration according to the ICP algorithm. Experimental results show that the method can combine the advantages of cross-modal data, and improve the calculation efficiency and registration accuracy of image and point cloud registration effectively. [ABSTRACT FROM AUTHOR]

Aiming at the problem of high complexity and long coding time of fractal image coding algorithm, the paper proposed orthogonal sparse coding and texture feature extraction to represent image blocks. Firstly, the orthogonal sparse transformation of gray level improved the image reconstruction quality and decoding time. Secondly, the correlation coefficient matrix measured the coefficient of variation characteristics between the range block and the domain block to reduce redundancy and coding time. The simulation results show that the proposed method has better image reconstruction quality and faster coding speed than the traditional fractal image coding algorithm. [ABSTRACT FROM AUTHOR]

At present, the protection of finished embroidery patterns plays an important role in the study of digital p reservation of embroidery heritage. However, the essence of embroidery culture is not only in the pattern, but also in the organic combination of stitching and pattern, which can better transmit the cultural characteristics of ethnic minorities . Therefore, finding a method to extract and match embroidery stitches and pattern features is an urgent problem for digital embroidery preservation. To address this problem, this paper proposed a feature extraction model by studying the commonly used embroidery needlework and feature extraction algorithms, with the guide filter as the scale transformation function. Traditional scale invariant feature transform (SIFT) was a basic model in this algorithm, which replaced the Gaussian scale transform function with guided filter function to enhance the algorithm response on edge pixels, so as to accurately extract the key features of the stitch pattern. Experiments show that the key features of the pattern is accurate at the edge of the pattern, and the matching accuracy can reach more than 88 % on average, which is 7 % - 25 % higher than the SIFT algorithm. [ABSTRACT FROM AUTHOR]