Subjects: Computer Science >> Computer Application Technology submitted time 2020-09-30
Abstract： "
Peer Review Status:
Awaiting Review
Subjects: Psychology >> Cognitive Psychology submitted time 2020-09-29
Abstract： 研究通过分析视觉搜索任务的首次注视点和行为反应时，探讨无关工作记忆表征的负性情绪信息对视觉注意选择的影响。实验1发现在反映早期注意选择的首次注视点百分率指标上，不管工作记忆表征的情绪效价如何，均出现了显著的注意捕获效应；实验2发现当采用中性情绪靶子刺激时，首次注视点百分率指标上仍表现出了稳健的注意捕获效应；在首次注视点持续时间指标上，实验1和实验2均发现记忆匹配条件的干扰刺激显著小于控制条件的干扰刺激，表现出注意的快速脱离；而在行为反应时指标上，早期的注意捕获效应消失（实验1），甚至被反转为注意抑制效应（实验2）。这些结果表明在早期注意选择阶段，记忆驱动的注意捕获效应不受工作记忆表征情绪效价的影响，但认知控制会在早期注意捕获之后促使注意快速脱离记忆匹配的干扰刺激，其作用效果受靶子刺激情绪效价的调节。
Peer Review Status:Awaiting Review
Subjects: Psychology >> Social Psychology submitted time 2020-09-29
Abstract： Increasing divorce rates in China have led to greater numbers of children growing up in single-parent homes. Previous studies have indicated that such single-parent children reported greater senses of relative deprivation and more psychological adjustment problems than their counterparts in undivided families. However, few studies have yet examined associations between relative deprivation and psychological adjustment and their directions. We thus explored characteristics of relative deprivation, psychological adjustment, and associations among them over 1.5 years beginning March, 2017. A sample of 273 single-parent children (50.5% boys) was recruited from two primary schools and two junior middle schools in Hubei, China. Attrition was relatively minor, namely 93.4% of participants completed all surveys during three assessment waves. Participants provided self-report data on individual and group cognitive and individual and group affective relative deprivation, and depression, loneliness, social anxiety, and self-esteem, as well as demographic variables (i.e., gender, academic period, and family economic status). All the measures had good reliability and validity. Results indicated that relative deprivation of single-parent children was not obvious, and psychological adjustment was generally good. Boys reported higher levels of depression and loneliness than girls. Moreover, single-parent children with poor family economic status reported higher levels of relative deprivation, depression, and loneliness, as well as lower levels of self-esteem than their counterparts. To explore the possible reciprocal relations between relative deprivation and psychological adjustment, as well as to separate between-person effects from within-person effects, we analyzed data by using the random intercepts cross-lagged panel model (RI-CLPM). Results showed that there were reciprocal relations between relative deprivation and psychological adjustment in within-person level when controlling for between-person effects and key demographic variables. Specifically, initial relative deprivation significantly negatively predicted psychological adjustment at Time 2, which in turn negatively predicted relative deprivation at Time 3. Moreover, relative deprivation at Time 2 also negatively predicted psychological adjustment at Time 3. These reciprocal relations between relative deprivation and psychological adjustment did not differ by gender and academic period (i.e., primary or secondary school). However, the association between psychological adjustment and relative deprivation was stronger for single-parent children with poor family economic status than for those with good family economic status. These observations expand understanding of the complex relations between relative deprivation and psychological adjustment among single-parent children in China. Additionally, they have important implications for intervention and improvement of mental health for vulnerable groups, especially single parent children. For instance, programs that aim to improve the mental health of single-parent children and to reduce the levels of relative deprivation among this vulnerable group may be helpful in breaking the detrimental cycle between relative deprivation and psychological adjustment.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： In complex environment, the object tracking algorithm of fully-convolutional siamese network is prone to track drift or even track failure. In order to solve the problem, this paper proposed a siamese network tracking algorithm based on hard sample mining. On the basis of SiamFC, this method first used a feature fusion module for feature fusion to enhance the robustness of feature representation, and then proposed a novel loss function to strengthen the learning ability of network to hard samples and alleviate the problem of imbalance between positive and negative samples. To verify the validity, this method was tested on OTB2015 benchmark and GOT10k dataset. The results of OTB2015 show that this method increases the success rate by 2.6% and the accuracy by 2% compared with SiamFC. On the GOT10k dataset, the mAO of this method is 0.429, which is 3.7% higher than the SiamFC. It illustrates that this method has a better performance in the case of illumination variation, object deformation, and similar background interference.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Based on convolution neural network, this paper proposed a new asymmetric two branch underwater biological classification model to solve the problems of low visibility, poor illumination conditions and no obvious differences among species in the underwater environment. In the model, the interactive branch used different convolution neural network to extracted local features and interacted with local features through the interactive module to enhanced the classification model. Convolutional neural network branch could effectively learned the global characteristics of the target and made up for the global information ignored in the interactive branch. Finally, this model obtains 98.9%, 98.3% and 97.9% of the accuracy on the three data sets of fish4 knowledge (f4k) , Eilat and RAMAS, which are significantly improved compared with the previous methods. visual interpretation also verifies that the model can effectively capture local features and eliminates the background influence. Finally, it shows that the model has good classification performance in underwater environment.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Painterly image compositing aims to harmonize a foreground image inserted into a background painting, which is done by local style transfer. The chief drawback of the existing methods is the high computational cost, which makes real-time operation difficult. To overcome this drawback, this paper proposed a feed-forward model based on generative adversarial network (GAN) , called PainterGAN. PainterGAN introducesd a self-attention network and a U-net to control the semantic content in the generated image. Meanwhile, adversarial learning guaranteed a faithful transfer of style. PainterGAN also introduced a pre-trained network within the generator to extract features. This allowed PainterGAN to dramatically reduce training-time and storage. Experiments show that, compared to state-of-art methods, PainterGAN generated images hundreds of times faster with comparable or superior quality. Therefore, it is effective and efficient for local style transfer.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Focused on the partly issue that in the process of single-frame image super-resolution reconstruction, such as insufficient utilization of feature information during image super-resolution reconstruction, the interdependence between the channels of the feature map is difficult to determine, and reconstruction errors existing at high-resolution image reconstructed, this paper proposed an single image super resolution methods based on depth residual backprojection attention network. It used the residual learning to ease the training difficulty and fully discover the feature information of the image, and used the back-projection method to learn the interdependence between the high- and low-resolution images. In addition, it introduced the attention mechanism to assign each feature map with different attention to discover more high-frequency information, and learnt the interdependence between the channels of the feature map. The experimental results show that compared with most single-frame image super-resolution methods, the proposed method not only has a significant improvement in objective indicators, but also the reconstructed predicted image has richer texture information.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： This paper presented a novel discriminative K-SVD network (DKSVDN) for face recognition. It embedded discriminative information into traditional K-SVD algorithm by special design of dictionary as well as sparse representation coefficients on the dictionary. The dictionary consisted of label specific atoms and descriptive atoms, while sparse codes contained one-hot label vectors and descriptive codes. In addition, as sparse representation algorithms were time-consuming, DKSVDN attached a co-trained feed-forward neural network to discriminative dictionary learning model to predict sparse codes. Moreover, with generative module in DKSVDN, this work also designed a new dreaming training phase to improve the robustness of DKSVDN for unknown pattern in known class. The experiment results on public face image datasets verified effectiveness of this method.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Recently, deep neural networks need to be deployed with low memory and computing resources, so it is necessary to design an efficient and compact network structure. This paper proposed a model compression method (KE) based on improved attention transfer for the design of compact neural networks, which mainly used a wide residual teacher network (WRN) to guide a compact student network (KENet) by extracting both spatial and channel-wise attention to improve the performance, and applied this method to real-time object detection. The image classification experiment on CIFAR verified that the knowledge distillation method with improved attention transfer can improve the performance of the compact model. The object detection experiment on VOC verified that the model KEDet has good accuracy (72.7mAP) and time performance (86FPS) . The experimental results show that the object detection model based on improved attention transfer has good accuracy and real-time performance.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Person re-identification is to judge whether the pedestrian across different cameras belongs to the same person or not. While it is challenging task due to the large variations in person pose, occlusion, background clutter, etc. And several deep learning based person re-identification have been proposed and achieved remarkable performance. However, these methods are only considered separately from the local or global features of the pedestrian, ignoring the relationship between the features. So this paper proposed the enhanced feature convergent network (EFCN) . In the global branch, the paper used to employ the new attention to pay close attention to the global feature of pedestrians. In the local branch, it proposed the gated recurrent unit change network(GRU-CN) to obtain more robust local features, and then this paper used feature fusion to connect the extracted global and local features. Extensive comparative experiments show that EFCN can achieve better experimental results on three standard person Re-ID datasets. The proposed enhanced feature convergent network can extract highly discriminative pedestrian features. This model can be applied to the problem of Re-ID under non-overlapping multi-cameras in large scenes. It has high recognition ability and accuracy. The method can extract robust features for pedestrian images with changing background.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Aiming at the problems that the action recognition methods based on two-stream ignore the inter-relationship between feature channels, and have large amount of redundant spatio-temporal information, this paper proposed an end-to-end action recognition model based on two-stream network with spatio-temporal attention mechanism (T-STAM) , which realized the full utilization of the key spatio-temporal information in the video. Firstly, this paper introduced the channel attention mechanism to the two-stream basic network, and calibrated the channel information by modeling the dependencies between feature channels to improve the ability of future expression. Secondly, this paper proposed a CNN-based temporal attention model to learn the attention score of each frame with fewer parameters, which can focuses on the frames with significant amplitude of motion. At the same time, it proposed a multi-spatial attention model, which calculated the attention score of each position in frame from different angles to extract motion saliency areas. Then, temporal and spatial features were fused to further enhance the feature representation of video. Finally, the fused features were input into the classification network, and the results of each stream are fused according to different weights to obtain the recognition results. The experimental results on the datasets HMDB51 and UCF101 show that T-STAM can effectively recognize actions in video.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： In order to solve the problems of uneven population distribution and different target scales affecting the crowd counting in public places, this paper proposed a novel crowd counting model based on image field division. Firstly, it divided the image scene into two parts: the near and far field of vision area. For the near field of vision area, it used the YOLO based network for pedestrian detection and added scene constraints to avoid repeated counting in the near and far field of vision. For the far field of vision area, it used the improved MobileNets to extract the population density distribution characteristics, and introduced the super-resolution reconstruction module to improve the quality of the population density map. Finally, it obtained the population in the whole image by calculating the sum of the two. This paper tested the proposed model on Shanghai Tech and Mall datasets, and the results show that the model has a significant improvement in accuracy and robustness. Experiments show that the model is feasible.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： In order to solve the problem that in the process of image reconstruction destroys the edge structure and losses the texture detail , this paper proposed an image interpolation algorithm for preserving the texture detail and the edge structure. Firstly, image divides area by using eight direction edge detection based on adaptive threshold. Secondly, construct a bivariate rational function model, which can convert into rational model and polynomial model. Finally, based on the local asymmetry edge data and gradient feature, this paper proposed a method to adjust the distance of points to be interpolated. By this method, adjusting the coordinates of the points to be interpolated and using rational model to realize interpolation, as for the non edge part, adopt polynomial model to interpolation. Experiments show that, the peak signal to noise ratio of the algorithm increases 0.48db-2.17db on average, and the structural similarity index increases 0.004-0.028 on average. This algorithm obtains high objective evaluation data. In this algorithm, the original spatial distance invariant interpolation is modified to the spatial distance varying interpolation, which effectively keeps the edge structure and texture details of the image, and makes the reconstruction result have better visual effect.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Aiming at the problem of low face recognition accuracy caused by poor image quality and loss of detailed information of face pictures obtained in security surveillance scene, this paper proposed a low-resolution face recognition algorithm based on super-resolution reconstruction. The algorithm included two sub-networks: super-resolution reconstruction and face recognition, which could respectively realize super-resolution reconstruction of low-resolution face image and extraction of face features. Firstly, the algorithm increased the number of feature maps before the activation function of super-resolution reconstruction sub-network to achieve wide activation and ensure effective transfer of information flow, so as to reconstruct high-resolution images containing more effective detailed information. Then, the algorithm combined image content loss and identity loss during training to retain more identity information while reconstructing image, which could make extracted face features more discriminative. Experimental results show that the algorithm improves accuracy of low-resolution face recognition and has better performance than traditional algorithms on surveillance face dataset QMUL-SurFace.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： In order to solve the problem that the residual network can not extract image information well and the segmentation effect is poor in semantic segmentation, this paper proposed a joint feature pyramid model (JFP) to integrate the output features of the residual network, and then further extract the features in combination with the atrous spatial pyramid pooling module (ASPP) . In the decoding part, this paper applied a simple decoding structure to recover the image size to complete the semantic segmentation This paper also used attention module as the auxiliary semantic segmentation network to assist the training of the neural network. This method trains the network in the Pascal VOC 2012 data set and the enhanced Pascal VOC 2012 data set respectively, and tests it on the verification set of Pascal VOC 2012. The average ratio of intersection and Union (Miou) is 78.55% and 80.14% respectively, which shows that proposed method has good semantic segmentation performance.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Splatting is a classical direct volume rendering method based on object order, which volume data exists in layers and each layer of data has similar lines. The amount of calculation data restricts the speed of image rendering. In order to further improve the rendering speed, this paper used a method based on the combination of similarity between adjacent layers and empty voxel jump to speed up the algorithm. It filtered the 3D texture data of the image in the process of reading the data, and then used the footstep table in the 3D texture data after filtering was projected in two dimensions. It calculated the gray value of each point by using the similarity between adjacent layers, and classified the data according to the gray value of each point to calculate the empty voxel that had no effect on the imaging, skiped the rendering process and speeded up the algorithm. The experimental results show that the optimized algorithm can solve and improve the spatial correlation and operation efficiency of splatting algorithm to a certain extent on the basis of ensuring the quality of the drawn image.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Poor communication conditions and frequent terminal movements in search, rescue, military and other environments make it difficult for traditional security authentication systems to achieve secure access to terminals. To solve this problem, this paper proposed a scheme about security connection with terminals in the mobile tactical environment. This scheme used a certificateless key management mechanism to analyze the security authentication process after the terminals were moved and the secure processing method after terminals and gateways were damaged or invaded. Simulation results show that the scheme improves the security of authorization and authentication between the gateway and the terminal, can well resist some known attacks, solves the problem of lack of mutual authentication and key escrow in the tactical environment, and the certificateess key algorithm used in this sheme has better security and less computational overhead than other algorithms, and can balance the security of accessing a gateway with energy consumption during a terminal movement.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： Zero-RTT (Zero Round-Trip Time) key exchange protocol allows clients to send encrypted protected payloads and the first key exchange protocol message at zero round-trip time, which has the advantages of non-interactive and off-line. In order to reduce the round-trip time of key exchange, this paper proposes a zero-RTT key exchange protocol on lattice based on the idea of penetrating encryption. Firstly, utilizing the one-time signatures algorithm and the hierarchical identity-based key encapsulation mechanism to construct penetrable forward secret key encapsulation scheme, and then using the penetrable forward secret key encapsulation scheme to design a 0-RTT key exchange protocol. The protocol only requires the client side to authenticate the server one-way, and can effectively resist the quantum attack and replay attack. Compared with similar protocols, the proposed protocol has penetrable full forward secrecy, reduces the number of communication rounds and improves the communication efficiency.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： This paper proposed a multiscale feature fusion approach for malicious HTTP request detection. Firstly, it models the HTTP request in both word-level and character-level. Secondly, it extracts the high level sematic information in HTTP request by using a specially designed convolutional neural network (CNN) . Thirdly, it jointly learns the multiscale representation for HTTP request with the help of multimodal learning techniques. Finally, a linear classifier is adopted for classification. Extensive experiments conducted on public HTTP CSIC 2010 dataset and WAF dataset show large improvement on the performance against existing state-of-the-art methods.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2020-09-28 Cooperative journals: 《计算机应用研究》
Abstract： With the widespread application of embedded cryptographic equipment, Side Channel Analysis (SCA) has become one of its security threats. The key information is recovered by analyzing the leaked information during the physical implementation of the cryptographic algorithm. Furthermore, the security of the cryptographic algorithm can be evaluated. Multi-layer perceptron (MLP) is an artificial neural network structure. In order to streamline the MLP network structure for energy analysis and reduce the training parameters and training time of the model. We studied the models based on Hamming weight (HW) and bit-based neural networks, and the output categories were reduced from 256 to 9 and 2 respectively. The power trace during the operation of the AES cryptographic algorithm was collected through experiments. We train and test the proposed MLP neural network. The results show that the model can reduce the training parameters of the MLP neural network by 84% and the training time by 28%, and reduce the number of power traces required during the key recovery phase, while ensuring the prediction accuracy. At least only one power trace is needed to complete the recovery of the AES algorithm's complete key. The validity of the model is verified by experiments, and the security of the block cipher algorithm can be analyzed and evaluated by using the model.
Hits
Hits
Downloads
Comment