Zhenxing Dong

Bio: Zhenxing Dong is an academic researcher. The author has contributed to research in topics: Computer science & Iterative reconstruction. The author has an hindex of 1, co-authored 1 publications receiving 31 citations.

Deep Sentiment Representation Based on CNN and LSTM

Abstract: Traditional machine learning techniques, including support vector machine (SVM), random walk, and so on, have been applied in various tasks of text sentiment analysis, which makes poor generalization ability in terms of complex classification problem. In recent years, deep learning has made a breakthrough in the research of Natural Language Processing. Convolutional neural network (CNN) and recurrent neural networks (RNNs) are two mainstream methods of deep learning in document and sentence modeling. In this paper, a model of capturing deep sentiment representation based on CNN and long short-term memory recurrent neural network (LSTM) is proposed. The model uses the pre-trained word vectors as input and employs CNN to gain significant local features of the text, then features are fed to two-layer LSTMs, which can extract context-dependent features and generate sentence representation for sentiment classification. We evaluate the proposed model by conducting a series of experiments on dataset. The experimental results show that the model we designed outperforms existing CNN, LSTM, CNN-LSTM (our implement of one-layer LSTM directly stacked on one-layer CNN) and SVM (support vector machine).

Fourier-inspired neural module for real-time and high-fidelity computer-generated holography.

Abstract: Learning-based computer-generated holography (CGH) algorithms appear as novel alternatives to generate phase-only holograms. However, most existing learning-based approaches underperform their iterative peers regarding display quality. Here, we recognize that current convolutional neural networks have difficulty learning cross-domain tasks due to the limited receptive field. In order to overcome this limitation, we propose a Fourier-inspired neural module, which can be easily integrated into various CGH frameworks and significantly enhance the quality of reconstructed images. By explicitly leveraging Fourier transforms within the neural network architecture, the mesoscopic information within the phase-only hologram can be more handily extracted. Both simulation and experiment were performed to showcase its capability. By incorporating it into U-Net and HoloNet, the peak signal-to-noise ratio of reconstructed images is measured at 29.16 dB and 33.50 dB during the simulation, which is 4.97 dB and 1.52 dB higher than those by the baseline U-Net and HoloNet, respectively. Similar trends are observed in the experimental results. We also experimentally demonstrated that U-Net and HoloNet with the proposed module can generate a monochromatic 1080p hologram in 0.015 s and 0.020 s, respectively.

Vision transformer-based, high-fidelity, computer-generated holography

Abstract: Current learning-based Computer-Generated Holography (CGH) algorithms often utilize Convolutional Neural Networks (CNN)-based architectures. However, the CNN-based non-iterative methods mostly underperform the State-Of-The-Art (SOTA) iterative algorithms such as Stochastic Gradient Descent (SGD) in terms of display quality. Inspired by the global attention mechanism of Vision Transformer (ViT), we propose a novel unsupervised autoencoder-based ViT for generating phase-only holograms. Specifically, for the encoding part, we use Uformer to generate the holograms. For the decoding part, we use the Angular Spectrum Method (ASM) instead of a learnable network to reconstruct the target images. To validate the effectiveness of the proposed method, numerical simulations and optical reconstructions are performed to compare our proposal against both iterative algorithms and CNN-based techniques. In the numerical simulations, the PSNR and SSIM of the proposed method are 26.78 dB and 0.832, which are 4.02 dB and 0.09 higher than that of the CNN-based method, respectively. Moreover, the proposed method contains less speckles and features a higher display quality than other CGH methods in experiments. We suggest the improvement might be ascribed to the ViT’s global attention mechanism, which is more suitable for learning the cross-domain mapping from image (spatial) domain to hologram (Fourier) domain. We believe the proposed ViT-based CGH algorithm could be a promising candidate for future real-time high-fidelity holographic displays.

Deep learning empowered highly compressive SS-OCT via learnable spectral-spatial sub-sampling.

Abstract: With the rapid advances of light source technology, the A-line imaging rate of swept-source optical coherence tomography (SS-OCT) has experienced a great increase in the past three decades. The bandwidths of data acquisition, data transfer, and data storage, which can easily reach several hundred megabytes per second, have now been considered major bottlenecks for modern SS-OCT system design. To address these issues, various compression schemes have been previously proposed. However, most of the current methods focus on enhancing the capability of the reconstruction algorithm and can only provide a data compression ratio (DCR) up to 4 without impairing the image quality. In this Letter, we proposed a novel design paradigm, in which the sub-sampling pattern for interferogram acquisition is jointly optimized with the reconstruction algorithm in an end-to-end manner. To validate the idea, we retrospectively apply the proposed method on an ex vivo human coronary optical coherence tomography (OCT) dataset. The proposed method could reach a maximum DCR of ∼62.5 with peak signal-to-noise ratio (PSNR) of 24.2 dB, while a DCR of ∼27.78 could yield a visually pleasant image with a PSNR of ∼24.6 dB. We believe the proposed system could be a viable remedy for the ever-growing data issue in SS-OCT.

I and i

Abstract: There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality. Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

Sentiment analysis using deep learning architectures: a review

Abstract: Social media is a powerful source of communication among people to share their sentiments in the form of opinions and views about any topic or article, which results in an enormous amount of unstructured information. Business organizations need to process and study these sentiments to investigate data and to gain business insights. Hence, to analyze these sentiments, various machine learning, and natural language processing-based approaches have been used in the past. However, deep learning-based methods are becoming very popular due to their high performance in recent times. This paper provides a detailed survey of popular deep learning models that are increasingly applied in sentiment analysis. We present a taxonomy of sentiment analysis and discuss the implications of popular deep learning architectures. The key contributions of various researchers are highlighted with the prime focus on deep learning approaches. The crucial sentiment analysis tasks are presented, and multiple languages are identified on which sentiment analysis is done. The survey also summarizes the popular datasets, key features of the datasets, deep learning model applied on them, accuracy obtained from them, and the comparison of various deep learning models. The primary purpose of this survey is to highlight the power of deep learning architectures for solving sentiment analysis problems.

Deep learning CNN–LSTM framework for Arabic sentiment analysis using textual information shared in social networks

Abstract: Recently, the world has witnessed an exponential growth of social networks which have opened a venue for online users to express and share their opinions in different life aspects. Sentiment analysis has become a hot-trend research topic in the field of natural language processing due to its significant roles in analyzing the public’s opinion and deriving effective opinion-based decisions. Arabic is one of the widely used languages across social networks. However, its morphological complexities and varieties of dialects make it a challenging language for sentiment analysis. Therefore, inspired by the success of deep learning algorithms, in this paper, we propose a novel deep learning model for Arabic language sentiment analysis based on one layer CNN architecture for local feature extraction, and two layers LSTM to maintain long-term dependencies. The feature maps learned by CNN and LSTM are passed to SVM classifier to generate the final classification. This model is supported by FastText words embedding model. Extensive experiments carried out on a multi-domain corpus demonstrate the outstanding classification performance of this model with an accuracy of 90.75%. Furthermore, the proposed model is validated using different embedding models and classifiers. The results show that FastText skip-gram model and SVM classifier are more valuable alternatives for the Arabic sentiment analysis. The proposed model outperforms several well-established state-of-the-art approaches on relevant corpora with up to $$+\,20.71\%$$ accuracy improvement.

Network text sentiment analysis method combining LDA text representation and GRU-CNN

Abstract: In order to improve the performance of internet public sentiment analysis, a text sentiment analysis method combining Latent Dirichlet Allocation (LDA) text representation and convolutional neural network (CNN) is proposed. First, the review texts are collected from the network for preprocessing. Then, using the LDA topic model to train the latent semantic space representation (topic distribution) of the short text, and the short text feature vector representation based on the topic distribution is constructed. Finally, the CNN with gated recurrent unit (GRU) is used as a classifier. According to the input feature matrix, the GRU-CNN strengthens the relationship between words and words, text and text, so as to achieve high accurate text classification. The simulation results show that this method can effectively improve the accuracy of text sentiment classification.

A Hybrid CNN-LSTM Model for Aircraft 4D Trajectory Prediction

Abstract: The 4D trajectory is a multi-dimensional time series with plentiful spatial-temporal features and has a high degree of complexity and uncertainty. Aiming at these features of aircraft flight trajectory and the problem that it is difficult for existing trajectory prediction methods to extract spatial-temporal features from the trajectory data at the same time, we propose a novel 4D trajectory prediction hybrid architecture based on deep learning, which combined Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM). An 1D convolution is used to extract the spatial dimension feature of the trajectory, and LSTM is used to mine the temporal dimension feature of the trajectory. Hence the high-precision prediction of the 4D trajectory is realized based on the sufficient fusion of the above features. We use real Automatic Dependent Surveillance -Broadcast (ADS-B) historical trajectory data for experiments and compare the proposed method with a single LSTM model and BP model on the same data set. The experimental results show that the trajectory prediction accuracy of the CNN-LSTM hybrid model is superior to a single model. The prediction error is reduced by an average of 21.62% compared to the LSTM model and by an average of 52.45% compared to the BP model. It provides a certain reference for the trajectory prediction research and Air Traffic Management decision-making.