Search

With the development of deep learning, the Super-Resolution (SR) reconstruction of microscopic images has improved significantly. However, the scarcity of microscopic images for training, the underutilization of hierarchical features in original Low-Resolution (LR) images, and the high-frequency noise unrelated with the image structure generated during the reconstruction process are still challenges in the Single Image Super-Resolution (SISR) field. Faced with these issues, we first collected sufficient microscopic images through Motic, a company engaged in the design and production of optical and digital microscopes, to establish a dataset. Secondly, we proposed a Residual Dense Attention Generative Adversarial Network (RDAGAN). The network comprises a generator, an image discriminator, and a feature discriminator. The generator includes a Residual Dense Block (RDB) and a Convolutional Block Attention Module (CBAM), focusing on extracting the hierarchical features of the original LR image. Simultaneously, the added feature discriminator enables the network to generate high-frequency features pertinent to the image's structure. Finally, we conducted experimental analysis and compared our model with six classic models. Compared with the best model, our model improved PSNR and SSIM by about 1.5 dB and 0.2, respectively. [ABSTRACT FROM AUTHOR]

The joint base matrix ${{{{B}}_{{\mathrm{J}}}}}$ of the joint source-channel coding (JSCC) system based on double protograph low-density parity-check (DP-LDPC) codes consists of four components, namely, the source code ${{{{B}}_{{\mathrm{s}}}}}$ , the channel code ${{{{B}}_{{\mathrm{c}}}}}$ , the type-1 connection edge ${{{{B}}_{{ \textit {L1}}}}}$ and the type-2 connection edge ${{{{B}}_{{ \textit {L2}}}}}$ , each having a non-negligible influence on the system performance. Different from the traditional component-specific design approach, we propose a joint design and optimization algorithm based on the idea of multi-objective differential evolution (MODE). Specifically, we consider the optimization of the DP-LDPC JSCC system through joint design of three components ${{{{B}}_{{\mathrm{s}}}}}$ , ${{{{B}}_{{\mathrm{c}}}}}$ , ${{{{B}}_{{ \textit {L1}}}}}$ and all four components ${{{{B}}_{{\mathrm{s}}}}}$ , ${{{{B}}_{{\mathrm{c}}}}}$ , ${{{\mathbf{B}}_{{ \textit {L1}}}}}$ , ${{{{B}}_{{ \textit {L2}}}}}$ , respectively. The proposed algorithm has low search complexity due to the reduction in size and element value of base matrices. The joint protograph extrinsic information transfer (JPEXIT) analyses and the simulation results demonstrate that the resulting JSCC system is free from a high error floor, requires fewer number of iterations for reaching the same bit error rate (BER) and achieves significant coding gains as compared to the state-of-the-art. Our DP-LDPC JSCC system is also shown to outperform its separation-based counterpart by a wide margin. [ABSTRACT FROM AUTHOR]

Sentiment recognition of online course reviews is valuable to understand emotions and feelings of learners. Nowadays, an increasing number of course reviews are being generated with the emergence of Massive Open Online Courses (MOOCs), which offers teachers a chance to analyze the opinions of learners and improve teaching strategies. However, the unstructured data contain large amounts of redundant features, which will significantly impact the performance of machine learning. To select effective emotional features, we adopt a multi-swarm particle swarm optimization (MSPSO) method, which generates multi diverse particle swarms on several cross training subsets. These swarms are utilized to find the best features by the F -Measure fitness function. The experimental results on the real-life dataset show that MSPSO can effectively reduce redundancy of text features and capture discriminative features. Compared with conventional feature selection methods, MSPSO can gain the better performance when selecting the same dimensions. Besides, the result of a user survey indicates that 72.19% of subjects approve of the usability of the recognition results and effectiveness of the feature selection. [ABSTRACT FROM AUTHOR]

With the rapid development of technology, more malicious traffic data brought negative influences on industrial areas. Modbus protocol plays a momentous role in the communications of Industrial Control Systems (ICS), but it's vulnerable to Denial of Service attacks(DoS). Traditional detection methods cannot perform well on fine-grained detection tasks which could contribute to locating targets of attacks and preventing the destruction. Considering the temporal locality and high dimension of malicious traffic, this paper proposed a Neural Network architecture named MODLSTM, which consists of three parts: input preprocessing, feature recoding, and traffic classification. By virtue of the design, MODLSTM can form continuous stream semantics based on fragmented packets, discover potential low-dimensional features and finally classify traffic at a fine-grained level. To test the model's performances, some experiments were conducted on industrial and public datasets, and the models achieved excellent performances in comparison with previous work(accuracy increased by 0.71% and 0.07% respectively). The results show that the proposed method has more satisfactory abilities to detect DoS attacks related to Modbus, compared with other works. It could help to build a reliable firewall to address a variety of malicious traffic in diverse situations, especially in industrial environments. [ABSTRACT FROM AUTHOR]

By using the residual source redundancy to achieve the shaping gain, a joint source-channel coded modulation (JSCCM) system has been proposed as a new solution for probabilistic amplitude shaping (PAS). However, the source and channel codes in the JSCCM system should be designed specifically for a given source probability to ensure optimal PAS performance, which is undesirable for systems with dynamically changing source probabilities. In this paper, we propose a new shaping scheme by optimizing the bit-labeling of the JSCCM system. Instead of the conventional fixed labeling, the proposed bit-labelings are adaptively designed according to the source probability and the source code. Since it is simple to switch between different labelings according to the source probability and the source code, the proposed design can be considered as a promising low complexity alternative to obtain the shaping gain for sources with different probabilities. Numerical results show that the proposed bit-labelings can significantly improve the bit-error rate (BER) performance of the JSCCM system. [ABSTRACT FROM AUTHOR]

In this paper, a joint group shuffled scheduling decoding (JGSSD) algorithm for a joint source-channel coding (JSCC) scheme based on double low-density parity-check (D-LDPC) codes is presented. The proposed algorithm considers the D-LDPC coding structure as a whole and applies shuffled scheduling to each group; the grouping relies on the types or the length of the variable nodes (VNs). By comparison, the conventional shuffled scheduling decoding algorithm can be regarded as a special case of this proposed algorithm. A novel joint extrinsic information transfer (JEXIT) algorithm for the D-LDPC codes system with the JGSSD algorithm is proposed, by which the source and channel decoding are calculated with different grouping strategies to analyze the effects of the grouping strategy. Simulation results and comparisons verify the superiority of the JGSSD algorithm, which can adaptively trade off the decoding performance, complexity and latency. [ABSTRACT FROM AUTHOR]

Estimating an accuracy spectrum from a linearly degraded and noisy data record is often desirable in infrared spectroscopy application. In this paper, we proposed a new framework which can estimate the desired spectrum and the instrument function simultaneously from the degraded data. Firstly, the instrument function is parametrically represented as a Lorentzian function for the first time. Then, we construct the cost functional with the DE regularization, and minimize the functional to obtain the desired spectrum and instrument function. It has been found that the program is very useful to determine accurate line widths and peak positions from degraded spectral spectrum. Comparative results including quantitative and qualitative analyses manifest that the proposed method outperforms the conventional methods on peak narrowing and noise suppression. [ABSTRACT FROM AUTHOR]

Taxonomy merging is an important work to provide a uniform schema for several heterogeneous taxonomies. Previous studies primarily focus on merging two taxonomies in a specific domain, while the merging of multiple taxonomies has been neglected. This article proposes a taxonomy merging approach to automatically merge multiple source taxonomies into a target taxonomy in an asymmetric manner. The approach adopts a strategy of breaking up the whole into parts to decrease the complexity of merging multiple taxonomies and employs a block-based method to reduce the scale of measuring semantic relations between concept pairs. In addition, for the problem of multiple inheritance, a method of topical coverage is proposed. Experiments conducted on synthetic and real-world scenarios indicate that the proposed merging approach is feasible and effective to merge multiple taxonomies. In particular, the proposed approach works well in the aspects of limiting the semantic redundancy and establishing high-quality hierarchical relations between concepts. [ABSTRACT FROM AUTHOR]

Abstract: The anonymous nature of online messages distribution causes a series of moral and legal issues. By analyzing identity cues people leave behind their texts, i.e., writeprint, potential authors can be identified individually. But writeprint identification is a difficult learning task, because of the high redundancy in stylistic feature set and high similarity of some authors’ writing-style. In this paper, we propose a novel method, called semi-random subspace (Semi-RS), to simultaneously address the two problems. Different from the conventional random subspace method (RSM) which samples features from the whole feature set in a completely random way, the proposed Semi-RS randomly samples features on each individual-author feature set (IAFS) partitioned from the whole feature set. More specifically, we first divide the whole feature set into several IAFSs in a deterministic way, then construct a set of base classifiers on different randomly sampled feature sets from each IAFS, and finally combine all base classifiers for the final decision. Experimental results on the benchmark dataset demonstrate the effectiveness of the proposed method which improves previously reported results. In addition, we analyze the diversity of algorithm, reveals that Semi-RS constructs more diverse base classifiers than conventional RSMs. [Copyright &y& Elsevier]

With the rapid development of MOOC platforms, the online learning resources are increasing. Because learners differ in terms of cognitive ability and knowledge structure, they cannot rapidly identify learning resources in which they are interested. Traditional collaborative filtering recommendation technologies perform poorly given sparse data and cold starts. Furthermore, the redundant recommended content and the high-dimensional and nonlinear data on online learning users cannot be effectively handled, leading to inefficient resource recommendations. To enhance learner efficiency and enthusiasm, this paper presents a highly accurate resource recommendation model (MOOCRC) based on deep belief networks (DBNs) in MOOC environments. This method deeply mines learner features and course content attribute features and incorporates learner behavior features to build user-course feature vectors as inputs to the deep model. Learner ratings of courses are processed as supervised labels with supervised learning. The MOOCRC model is trained by unsupervised pretraining and supervised feedback fine tuning; moreover, the model is obtained by adjusting the model parameters repeatedly. To verify the effectiveness of the MOOCRC, an experimental analysis is conducted using learner elective data obtained from the starC MOOC platform of Central China Normal University. Real course enrollment data are used to verify the classification accuracy of the MOOCRC. The experimental results show that the MOOCRC has greater recommendation accuracy and converges more quickly than traditional recommendation methods. [ABSTRACT FROM AUTHOR]

The emergence of massive open online courses not only changes the ecology of higher education, but also facilitates a blending learning paradigm, also known as small private online courses (SPOCs). In order to understand how college students interact with an SPOC platform, this study collects their online behaviors for a semester and adopts a lag sequential analysis approach to identify significant transitions between interactions with content, peers, and instructors. Regarding content, after entering courses, the students tend to access learning resources. Besides, the transitions between learning resources and personal performance are significantly interconnected to each other. Regarding peers, the interaction with classmates was mainly connected to the access of assignments and performance. Regarding instructors, the interaction with teachers was minor but connected to all other behaviors. In addition, the results also show that students’ online behavioral patterns in SPOCs may change over time. The implications of the findings for SPOCs research are discussed in this paper. [ABSTRACT FROM PUBLISHER]

Spectral signals often suffer from the common problems of band overlap and random Gaussian noise. To address these problems, we propose a sparsity regularization-based model that deconvolutes the degraded spectral signals. Sparsity regularization is achieved by fitting the probability density function of the gradient of the signal, and then, the iteratively reweighted least-squares method is used to solve the minimization problem. Results from experiments using real spectral signals showed that this algorithm separates the overlapping peaks and effectively suppresses the noise. The deconvoluted spectral signals will promote the practical application of infrared spectral analysis in the fields of target recognition, material identification, and chemometrics analysis. [ABSTRACT FROM AUTHOR]

• A novelty infrared imaging technology for facial express recognition with convolution neural network. • Discrete beamlet transform regularization is designed to retain infrared spectral structures. • The split Bregman iteration algorithm is introduced to optimize the proposed DBTR-BR model. • The proposed model is applied for infrared facial expression recognition problem in Oulu-CASIA dataset. Facial expression recognition is an important research topic in the field of human-machine interaction. Infrared imaging technology has illustrated good potential in the applications of facial expression recognition and computer vision. But the low-resolution spectral data has limited its applications, such as band overlap and random noises. To address the problems, a rapid blind restoration model with discrete beamlet transforms regularization is presented to reconstruct the infrared spectrum. To compare the sparsity between the observed infrared spectrum and ground-truth one in frequency domain, the discrete beamlet transforms is applied to analyze the different of their coefficients distributions. We propose an IR spectral deconvolution model with the sparsity coefficients regularization by L 0 -norm. We execute the proposed algorithm on the simulated and actual IR spectrum data, and the results demonstrate that can effectively suppress the Poisson noises and retain infrared spectral structure. The high-resolution IR spectrum can raise the recognition rate of facial expression classification task. [ABSTRACT FROM AUTHOR]