Layered Decoding Research Articles

DeepRoute+: Modeling Couriers’ Spatial-temporal Behaviors and Decision Preferences for Package Pick-up Route Prediction

Over 10 billion packages are picked up every day in China. A fundamental task raised in the emerging intelligent logistics systems is the couriers’ package pick-up route prediction, which is beneficial for package dispatching, arrival-time estimation and overdue-risk evaluation, by leveraging the predicted routes to improve those downstream tasks. In the package pick-up scene, the decision-making of a courier is affected by strict spatial-temporal constraints (e.g., package location, promised pick-up time, current time, and courier’s current location). Furthermore, couriers have different decision preferences on various factors (e.g., time factor, distance factor, and balance of both), based on their own perception of the environments and work experience. In this article, we propose a novel model, named DeepRoute+, to predict couriers’ future package pick-up routes according to the couriers’ decision experience and preference learned from the historical behaviors. Specifically, DeepRoute+ consists of three layers: (1) The representation layer produces experience- and preference-aware representations for the unpicked-up packages, in which a decision preference module can dynamically adjust the importance of factors that affects the courier’s decision under the current situation. (2) The transformer encoder layer encodes the representations of packages while considering the spatial-temporal correlations among them. (3) The attention-based decoder layer uses the attention mechanism to generate the whole pick-up route recurrently. Experiments on a real-world logistics dataset demonstrate the state-of-the-art performance of our model.

Attention Guided Global Enhancement and Local Refinement Network for Semantic Segmentation

The encoder-decoder architecture is widely used as a lightweight semantic segmentation network. However, it struggles with a limited performance compared to a well-designed Dilated-FCN model for two major problems. First, commonly used upsampling methods in the decoder such as interpolation and deconvolution suffer from a local receptive field, unable to encode global contexts. Second, low-level features may bring noises to the network decoder through skip connections for the inadequacy of semantic concepts in early encoder layers. To tackle these challenges, a Global Enhancement Method is proposed to aggregate global information from high-level feature maps and adaptively distribute them to different decoder layers, alleviating the shortage of global contexts in the upsampling process. Besides, aLocal Refinement Module is developed by utilizing the decoder features as the semantic guidance to refine the noisy encoder features before the fusion of these two (the decoder features and the encoder features). Then, the two methods are integrated into a Context Fusion Block, and based on that, a novel Attention guided Global enhancement and Local refinement Network (AGLN) is elaborately designed. Extensive experiments on PASCAL Context, ADE20K, and PASCAL VOC 2012 datasets have demonstrated the effectiveness of the proposed approach. In particular, with a vanilla ResNet-101 backbone, AGLN achieves the state-of-the-art result (56.23% mean IOU) on the PASCAL Context dataset. The code is available at https://github.com/zhasen1996/AGLN.

A Hybrid LSTM-ResNet Deep Neural Network for Noise Reduction and Classification of V-Band Receiver Signals

Noise reduction is one of the most important process used for signal processing in communication systems. The signal-to-noise ratio (SNR) is a key parameter to consider for minimizing the bit error rate (BER). The inherent noise found in millimeter-wave systems is mainly a combination of white noise and phase noise. Increasing the SNR in wireless data transfer systems can lead to reliability and performance improvements. To address this issue, we propose to use a recurrent neural network (RNN) with a long short-term memory (LSTM) autoencoder architecture to achieve signal noise reduction. This design is based on a composite LSTM autoencoder with a single encoder layer and two decoder layers. A V-band receiver test bench is designed and fabricated to provide a high-speed wireless communication system. Constellation diagrams display the output signals measured for various random sequences of PSK and QAM modulated signals. The LSTM autoencoder is trained in real time using various noisy signals. The trained system is then used to reduce noise levels in the tested signals. The SNR of the designed receiver is of the order of 11.8dB, and it increases to 13.66dB using the three-level LSTM autoencoder. Consequently, the proposed algorithm reduces the bit error rate from 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−8</sup> to 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−11</sup> . The performance of the proposed algorithm is comparable to other noise reduction strategies. Augmented denoised signals are fed into a ResNet-152 deep convolutional network to perform the final classification. The demodulation types are classified with an accuracy of 99.93%. This is confirmed by experimental measurements.

Infrared Small UAV Target Detection Based on Depthwise Separable Residual Dense Network and Multiscale Feature Fusion

Unmanned aerial vehicles (UAVs) have been widely applied in military and civilian fields, but they also pose great threats to restricted areas, such as densely populated areas and airports. Thermal infrared imaging technology is capable of monitoring UAVs at a long range in both day and night conditions. Therefore, the anti-UAV technology based on thermal infrared imaging has attracted growing attention. However, the images acquired by infrared sensors often suffer from small and dim targets as well as heavy background clutter and noise. Conventional detection methods usually have a high false alarm rate and low detection accuracy. This paper proposes a detection method that formulates the UAV detection as predicting the residual image (i.e., background, clutter, and noise) by learning the nonlinear mapping from the input image to the residual image. The UAV target image is obtained by subtracting the residual image from the input infrared image. The constructed end-to-end U-shaped network exploits the depthwise separable residual dense blocks in the encoder stage to extract the abundant hierarchical features. Besides, the multi-scale feature fusion and representation block is introduced to fully aggregate multi-scale features from the encoder layers and intermediate connection layers at the same scale as well as the decoder layers at different scales to better reconstruct the residual image in the decoder stage. In addition, the global residual connection is adopted in the proposed network to provide long-distance information compensation and promote gradient backpropagation, which further improves the performance in reconstructing the image. The experimental results show that the proposed method achieves favorable detection performance in real-world infrared images and outperforms other state-of-the-art methods in terms of quantitative and qualitative evaluation metrics.

Lightweight Denoising Autoencoder Design for Noise Removal in Electrocardiography

This study proposes two denoising autoencoder (DAE) models with discrete cosine transform (DCT) and discrete wavelet transform (DWT), namely DCT–DAE and DWT–DAE, to remove electrode motion artifacts in noisy electrocardiography (ECG). Initially, the discrete cosine transform and discrete wavelet transform efficiently removed the high-frequency noise. The six encoder layers then retain important ECG features while the six decoder layers reconstruct the clean ECG. To improve the denoising performance, two network layers, the residual block and pixel adjustment, are added to the encoder and decoder layers to solve the vanishing gradient and improve subtle feature extraction. The proposed methods were applied to 66,000 real-recorded noisy ECG fragments. The experimental result indicates that DWT–DAE and DCT–DAE can improve the signal-to-noise ratio by 25.29 and 25.13 dB on average when the input signal-to-noise ratio is -6 dB.

Unsupervised 3-D Random Noise Attenuation Using Deep Skip Autoencoder

Effective random noise attenuation is critical for subsequent processing of seismic data, such as velocity analysis, migration, and inversion. Thus, the removal of seismic random noise with an uncertainty level is meaningful. Attenuating 3-D random noise in a supervised way based on deep learning (DL) is challenging because clean labels are difficult to obtain. Therefore, it is necessary to develop an adaptive unsupervised-based method for random noise attenuation. In this article, we propose a deep-denoising unsupervised learning (DDUL) network to attenuate random noise in 2-D/3-D seismic data. A patching technique is used to split 2-D/3-D seismic data into several patches to be fed into the network, which helps to expand the number of samples for training. We use the fully symmetrical structure of the autoencoder to construct the network. In each corresponding encoder and decoder layer, skip connections are added to enhance the learning of seismic data features. We construct three blocks to extract waveform features in seismic data, i.e., encoder, decoder, and skip blocks. Among them, the skip is connected between the encoder and decoder blocks of each hidden layer. The use of multiple blocks not only improves the network&#x2019;s ability to extract seismic data features but also solves the problem of excessive training parameters caused by hidden layer stacking. Five 2-D/3-D synthetic and field seismic datasets are used to test the denoising performance of our proposed method. The denoising results demonstrate that our proposed method has good signal-preserving and noise attenuation capabilities in real-world applications.

Convergence Analysis of Joint Channel Decoding and Physical Layer Network Coding

Convergence Analysis of Joint Channel Decoding and Physical Layer Network Coding

Cascaded Multiscale Structure With Self-Smoothing Atrous Convolution for Semantic Segmentation

Convolutional neural networks (CNNs) have attracted great attention in the semantic segmentation of very-high-resolution (VHR) images of urban areas. However, large-scale variation of objects in the urban areas often makes it difficult to achieve good segmentation accuracy. Atrous convolution and atrous spatial pyramid pooling composed of atrous convolution can alleviate this problem by exploring multiscale contextual information. Unfortunately, atrous convolution causes gridding artifacts, where actual receptive fields are separated unit sets and fail to cover all the receptive fields. To address this problem, in this article, we first propose a self-smoothing atrous convolution (SS-AConv) that intrinsically improves atrous convolution, unlike existing methods. SS-AConv enhances sampling rates with low computational costs by adding several key parameters in the spatial dimension of its filter. Then, in the backbone, the Xception network, all max pooling operations are replaced with SS-AConvs for a large and effective receptive field. Moreover, to extract diverse features at multiple scales, we propose the SS-AConv cascaded multiscale structure (SCMS) by integrating SS-AConvs with different rates and the residual correction scheme (RCS) into a cascaded spatial pyramid. Finally, to extract diverse features at dense multiple scales, the SS-AConv convolutional network (SS-ACNet) is constructed by integrating SCMS into both the encoder and decoder layers of the modified Xception network. Extensive experimental results show that SS-ACNet outperforms some state-of-the-art methods on four open challenge datasets: the ISPRS Vaihingen and Potsdam, Cityscapes, and PASCAL VOC 2012.

Dual supervised sampling networks for real-time segmentation of cervical cell nucleus

The morphology of the cervical cell nucleus is the most important consideration for pathological cell identification. And a precise segmentation of the cervical cell nucleus determines the performance of the final classification for most traditional algorithms and even some deep learning-based algorithms. Many deep learning-based methods can accurately segment cervical cell nuclei but will cost lots of time, especially when dealing with the whole-slide image (WSI) of tens of thousands of cells. To address this challenge, we propose a dual-supervised sampling network structure, in which a supervised-down sampling module uses compressed images instead of original images for cell nucleus segmentation, and a boundary detection network is introduced to supervise the up-sampling process of the decoding layer for accurate segmentation. This strategy dramatically reduces the convolution calculation in image feature extraction and ensures segmentation accuracy. Experimental results on various cervical cell datasets demonstrate that compared with UNet, the inference speed of the proposed network is increased by 5 times without losing segmentation accuracy. The codes and datasets are available athttps://github.com/ldrunning/DSSNet.

Deep Learning for Denoising: An Attempt to Recover the Effective Magnetic Resonance Sounding Signal in the Presence of High Level Noise

Magnetic resonance sounding (MRS) measurements commonly suffer from a notably low signal-to-noise ratio (SNR). In recent years, many denoising methods have been developed, which have demonstrated the useful capability to improve the SNR. However, when MRS measurements are implemented on the sites with high noise levels, i.e., the human living environment, the conventional methods are helpless to recover the effective MRS signal, which is submerged in extensive environmental noise. In addition, the conventional methods that depend on signal models and the corresponding prior assumptions commonly rely on manual experience, which brings obstacles to the automation and efficiency of signal processing. To solve the above problems, we attempt to apply an intelligent denoising framework with a novel neural network as the basic tool for deep learning, called Dn-ResUnet in this article. The network extracts the features of the MRS signal through the encoder and decoder layers of the Dn-ResUnet structure, in which residual learning is adopted to accelerate the training process and improve denoising performance. Once the training is completed, the deep learning realizes adaptive denoising with no need for 1) prior assumptions of the MRS signal and noise; 2) optimal filter parameter tuning; or 3) expensive time cost. The comparison experiments demonstrate that the Dn-ResUnet model provides superior noise cancellation performance, especially it can replace the conventional methods to recover the effective MRS signals in noise levels down to an SNR of −30 dB. In addition, the noise attenuation tests are performed on synthetic and real data. The results show that the framework of deep learning yields a convincing performance in MRS signal denoising.

Self-Attention Deep Image Prior Network for Unsupervised 3-D Seismic Data Enhancement

We develop a deep learning framework based on deep image prior (DIP) and attention networks for 3-D seismic data enhancement. First, the 3-D noisy data are divided into several overlapped patches. Second, the DIP network has a U-NET architecture, where the input patches are encoded to extract the significant latent features, while the decoder tries to reconstruct the input patches using these extracted features. Besides, the attention network is used to scale the extracted features from the encoder and the decoder. Third, the attention network output of the encoder is concatenated with that of the decoder to obtain high-order features and guide the network to extract the most significant information related to the seismic signals and discard the others. Finally, the 3-D seismic data are reconstructed using the output patches obtained by the DIP network. The proposed algorithm is an iterative and unsupervised approach, which does not require labeled data. We evaluate the proposed algorithm using several synthetic and field data examples. As a result, the proposed algorithm shows the ability to enhance the 3-D seismic data by attenuating the random noise and preserving the 3-D seismic signal with minimal signal leakage. Moreover, the proposed algorithm shows good denoising performance when tested using various types of events, e.g., linear, hyperbolic, low and high dominant frequencies, and weak amplitude. In addition, the proposed method outperforms the predictive filtering (PF) and damped rank-reduction (DRR) methods. To further understand the principle of the proposed method inside the DIP network, we analyze the weighting matrices in the encoder and decoder parts in detail. We attribute the denoising ability of the DIP network to the improvement of the extracted basis features from the encoder to the decoder layers through a deep network.

Gastrointestinal Disease Classification in Endoscopic Images Using Attention-Guided Convolutional Neural Networks

Gastrointestinal (GI) diseases constitute a leading problem in the human digestive system. Consequently, several studies have explored automatic classification of GI diseases as a means of minimizing the burden on clinicians and improving patient outcomes, for both diagnostic and treatment purposes. The challenge in using deep learning-based (DL) approaches, specifically a convolutional neural network (CNN), is that spatial information is not fully utilized due to the inherent mechanism of CNNs. This paper proposes the application of spatial factors in improving classification performance. Specifically, we propose a deep CNN-based spatial attention mechanism for the classification of GI diseases, implemented with encoder–decoder layers. To overcome the data imbalance problem, we adapt data-augmentation techniques. A total of 12,147 multi-sited, multi-diseased GI images, drawn from publicly available and private sources, were used to validate the proposed approach. Furthermore, a five-fold cross-validation approach was adopted to minimize inconsistencies in intra- and inter-class variability and to ensure that results were robustly assessed. Our results, compared with other state-of-the-art models in terms of mean accuracy (ResNet50 = 90.28, GoogLeNet = 91.38, DenseNets = 91.60, and baseline = 92.84), demonstrated better outcomes (Precision = 92.8, Recall = 92.7, F1-score = 92.8, and Accuracy = 93.19). We also implemented t-distributed stochastic neighbor embedding (t–SNE) and confusion matrix analysis techniques for better visualization and performance validation. Overall, the results showed that the attention mechanism improved the automatic classification of multi-sited GI disease images. We validated clinical tests based on the proposed method by overcoming previous limitations, with the goal of improving automatic classification accuracy in future work.

Improving the Performance of Vietnamese–Korean Neural Machine Translation with Contextual Embedding

With the recent evolution of deep learning, machine translation (MT) models and systems are being steadily improved. However, research on MT in low-resource languages such as Vietnamese and Korean is still very limited. In recent years, a state-of-the-art context-based embedding model introduced by Google, bidirectional encoder representations for transformers (BERT), has begun to appear in the neural MT (NMT) models in different ways to enhance the accuracy of MT systems. The BERT model for Vietnamese has been developed and significantly improved in natural language processing (NLP) tasks, such as part-of-speech (POS), named-entity recognition, dependency parsing, and natural language inference. Our research experimented with applying the Vietnamese BERT model to provide POS tagging and morphological analysis (MA) for Vietnamese sentences,, and applying word-sense disambiguation (WSD) for Korean sentences in our Vietnamese–Korean bilingual corpus. In the Vietnamese–Korean NMT system, with contextual embedding, the BERT model for Vietnamese is concurrently connected to both encoder layers and decoder layers in the NMT model. Experimental results assessed through BLEU, METEOR, and TER metrics show that contextual embedding significantly improves the quality of Vietnamese–Korean NMT.

Solving Traveling Salesman Problem with Time Windows Using Hybrid Pointer Networks with Time Features

This paper introduces a time efficient deep learning-based solution to the traveling salesman problem with time window (TSPTW). Our goal is to reduce the total tour length traveled by -*the agent without violating any time limitations. This will aid in decreasing the time required to supply any type of service, as well as lowering the emissions produced by automobiles, allowing our planet to recover from air pollution emissions. The proposed model is a variation of the pointer networks that has a better ability to encode the TSPTW problems. The model proposed in this paper is inspired from our previous work that introduces a hybrid context encoder and a multi attention decoder. The hybrid encoder primarily comprises the transformer encoder and the graph encoder; these encoders encode the feature vector before passing it to the attention decoder layer. The decoder consists of transformer context and graph context as well. The output attentions from the two decoders are aggregated and used to select the following step in the trip. To the best of our knowledge, our network is the first neural model that will be able to solve medium-size TSPTW problems. Moreover, we conducted sensitivity analysis to explore how the model performance changes as the time window width in the training and testing data changes. The experimental work shows that our proposed model outperforms the state-of-the-art model for TSPTW of sizes 20, 50 and 100 nodes/cities. We expect that our model will become state-of-the-art methodology for solving the TSPTW problems.

Multimodal medical image segmentation using multi-scale context-aware network

Multimodal medical image segmentation with different imaging devices is a key but challenging task in medical image visual analysis and reasoning. Recently, U-Net based networks achieved considerable success in semantic segmentation of medical image. However, U-Net utilizes a skip-connection to connect two symmetric encoder and decoder layers. Although the single granularity information of the encoder layer is preserved through skip connection, the rich multi-scale spatial information is ignored, which greatly affects its performance in the segmentation task. In this paper, a multi-scale context-aware network (CA-Net) for multimodal medical image segmentation is proposed, which captures rich context information with dense skip connection and assigns distinct weights to different channels. CA-Net consists of four key components, namely encoder module, multi-scale context fusion (MCF) module, decoder module, and dense skip connection module. The proposed MCF module extracts multi-scale spatial information through a spatial context fusion (SCF) block, and learn to balance channel-wise features through a Squeeze-and-Excitation (SE) block. Extensive experiments demonstrate that our model achieves state-of-the-art performance on three benchmark datasets of different modalities, including skin lesion segmentation in dermoscopy, lung segmentation in CT images, and blood vessel segmentation in retina images.

Neural Layered Decoding of 5G LDPC Codes

In this letter, we propose various low-complexity neural layered min-sum (NLMS) algorithms to improve the decoding performance of the fifth generation (5G) new radio (NR) low-density parity-check (LDPC) codes. Our proposals are based on the computationally-efficient normalized offset min-sum (NOMS) approach for the layered belief propagation (LBP). The main novelty of our proposals is a deep neural network (DNN) that implements the layered mode decoding and that additionally learns the normalization and the offset parameters of the NOMS scheme. The schemes that we propose use the DNN as well as adaptation and filtering for estimating the optimum values of these parameters. We show by simulation that our proposed decoders achieve a significant computational benefit compared to the standard (non-neural) LBP decoder without an appreciable performance penalty.

CADA: Multi-scale Collaborative Adversarial Domain Adaptation for unsupervised optic disc and cup segmentation

Recently, deep neural networks have demonstrated comparable and even better performance than board-certified ophthalmologists in well-annotated datasets. However, the diversity of retinal imaging devices poses a significant challenge: domain shift, which leads to performance degradation when applying the deep learning models trained on one domain to new testing domains. In this paper, we propose a domain adaptation framework comprising of multi-scale inputs along with multiple domain adaptors applied hierarchically in both feature and output spaces. The proposed training strategy and novel unsupervised domain adaptation framework, called Collaborative Adversarial Domain Adaptation (CADA), can effectively overcome the domain shift challenge. Multi-scale inputs can reduce the information loss due to the pooling layers used in the network for feature extraction, while our proposed CADA is an interactive paradigm that presents an exquisite collaborative adaptation through both adversarial learning and ensembling weights at different network layers. In particular, to produce a better prediction for the unlabeled target domain data, we simultaneously achieve domain invariance and model generalizability via adversarial learning at multi-scale outputs from different levels of network layers and maintaining an exponential moving average (EMA) of the historical weights during training. Without annotating any sample from the target domain, multiple adversarial losses in encoder and decoder layers guide the extraction of domain-invariant features to confuse the domain classifier. Meanwhile, the ensembling of weights via EMA reduces the uncertainty of adapting multiple discriminator learning. Comprehensive experimental results demonstrate that our CADA model incorporating multi-scale input training can overcome performance degradation and outperform state-of-the-art domain adaptation methods in segmenting retinal optic disc and cup from fundus images stemming from the REFUGE, Drishti-GS, and Rim-One-r3 datasets. The code is available at https://github.com/cswin/CADA.

A deep learning-based segmentation pipeline for profiling cellular morphodynamics using multiple types of live cell microscopy

SummaryTo accurately segment cell edges and quantify cellular morphodynamics from live-cell imaging data, we developed a deep learning-based pipeline termed MARS-Net (multiple-microscopy-type-based accurate and robust segmentation network). MARS-Net utilizes transfer learning and data from multiple types of microscopy to localize cell edges with high accuracy. For effective training on distinct types of live-cell microscopy, MARS-Net comprises a pretrained VGG19 encoder with U-Net decoder and dropout layers. We trained MARS-Net on movies from phase-contrast, spinning-disk confocal, and total internal reflection fluorescence microscopes. MARS-Net produced more accurate edge localization than the neural network models trained with single-microscopy-type datasets. We expect that MARS-Net can accelerate the studies of cellular morphodynamics by providing accurate pixel-level segmentation of complex live-cell datasets.

Reduced-complexity decoding implementation of QC-LDPC codes with modified shuffling

Layered decoding (LD) facilitates a partially parallel architecture for performing belief propagation (BP) algorithm for decoding low-density parity-check (LDPC) codes. Such a schedule for LDPC codes has, in general, reduced implementation complexity compared to a fully parallel architecture and higher convergence rate compared to both serial and parallel architectures, regardless of the codeword length or code-rate. In this paper, we introduce a modified shuffling method which shuffles the rows of the parity-check matrix (PCM) of a quasi-cyclic LDPC (QC-LDPC) code, yielding a PCM in which each layer can be produced by the circulation of its above layer one symbol to the right. The proposed shuffling scheme additionally guarantees the columns of a layer of the shuffled PCM to be either zero weight or single weight. This condition has a key role in further decreasing LD complexity. We show that due to these two properties, the number of occupied look-up tables (LUTs) on a field programmable gate array (FPGA) reduces by about 93% and consumed on-chip power by nearly 80%, while the bit error rate (BER) performance is maintained. The only drawback of the shuffling is the degradation of decoding throughput, which is negligible for low values of E_b/N_0 until the BER of 1e−6.

Fall Detection with CNN-Casual LSTM Network

Falls are one of the main causes of elderly injuries. If the faller can be found in time, further injury can be effectively avoided. In order to protect personal privacy and improve the accuracy of fall detection, this paper proposes a fall detection algorithm using the CNN-Casual LSTM network based on three-axis acceleration and three-axis rotation angular velocity sensors. The neural network in this system includes an encoding layer, a decoding layer, and a ResNet18 classifier. Furthermore, the encoding layer includes three layers of CNN and three layers of Casual LSTM. The decoding layer includes three layers of deconvolution and three layers of Casual LSTM. The decoding layer maps spatio-temporal information to a hidden variable output that is more conducive relative to the work of the classification network, which is classified by ResNet18. Moreover, we used the public data set SisFall to evaluate the performance of the algorithm. The results of the experiments show that the algorithm has high accuracy up to 99.79%.