Search

Childhood demyelinative diseases classification (DDC) with brain magnetic resonance imaging (MRI) is crucial to clinical diagnosis. But few attentions have been paid to DDC in the past. How to accurately differentiate pediatric-onset neuromyelitis optica spectrum disorder (NMOSD) from acute disseminated encephalomyelitis (ADEM) based on MRI is challenging in DDC. In this paper, a novel architecture M-DDC based on joint U-Net segmentation network and deep convolutional network is developed. The U-Net segmentation can provide pixel-level structure information, that helps the lesion areas location and size estimation. The classification branch in DDC can detect the regions of interest inside MRIs, including the white matter regions where lesions appear. The performance of the proposed method is evaluated on MRIs of 201 subjects recorded from the Children's Hospital of Zhejiang University School of Medicine. The comparisons show that the proposed DDC achieves the highest accuracy of 99.19% and dice of 71.1% for ADEM and NMOSD classification and segmentation, respectively. [ABSTRACT FROM AUTHOR]

Chemists have been pursuing general mathematical laws to explain and predict molecular properties for a long time. However, most of the traditional quantitative structure‐activity relationship (QSAR) models have limited application domains; for example, they tend to have poor generalization performance when applied to molecules with parent structures different from those of the trained molecules. This paper attempts to develop a new QSAR method that is theoretically possible to predict various properties of molecules with diverse structures. The proposed deep electron cloud‐activity relationships (DECAR) and deep field‐activity relationships (DFAR) methods consist of three essentials: (1) a large number of molecule entities with activity data as training objects and responses; (2) three‐dimensional electron cloud density (ECD) or related field data by the accurate density functional theory methods as input descriptors; and (3) a deep learning model that is sufficiently flexible and powerful to learn the large data described above. DECAR and DFAR are used to distinguish 977 sweet and 1965 non‐sweet molecules (with 6‐fold data augmentation), and the classification performance is demonstrated to be significantly better than the traditional least squares support vector machine (LS‐SVM) models using traditional descriptors. DECAR and DFAR would provide a possible way to establish a widely applicable, cumulative, and shareable artificial intelligence‐driven QSAR system. They are likely to promote the development of an interactive platform to collect and share the accurate ECD and field data of millions of molecules with annotated activities. With enough input data, we envision the appearance of several deep networks trained for various molecular activities. Finally, we could anticipate a single DECAR or DFAR network to learn and infer various properties of interest for chemical molecules, which will become an open and shared learning and inference tool for chemists. Deep electron cloud‐activity relationships (DECAR) and deep field‐activity relationships (DFAR) methods are quantitative structure‐activity relationship (QSAR) models that involve deep learning of thousands (or hundreds of thousands) of molecule entities using augmented and high‐quality 3D electron cloud or related field data as molecular structure descriptors. [ABSTRACT FROM AUTHOR]

[Display omitted] • Tackle the under-studied semi-supervised image regression problem to relief the heavy workload of manually annotation for regression tasks. • Applying manifold smoothness to image regression tasks that transfers the SSL problem to kernel learning tasks. • A novel and efficient algorithm, ManiDKL, learns the regression parameters in a Bayesian manner. • Outperforms all existing semi-supervised methods for image regression tasks. Semi-supervised learning (SSL) aims at utilizing the vast unlabeled data to help the supervised training. While existing SSL methods have shown promising results on image classification tasks, most of them rely on the cluster assumption that does not apply to image regression tasks. In this paper, we address the under-studied semi-supervised image regression problem, of which the outputs are continuous values instead of categorical distributions. To tackle this challenging task, we propose an algorithm, called ManiDKL, with the idea that the prediction function should be smooth with respect to the intrinsic manifold of data distribution and behave similarly on both labeled and unlabeled data. In particular, we propose a framework that implements the Tikhonov regularization with generative manifold learning to ensure manifold smoothness of regression function and also reduces the problem to kernel learning. Then a semi-supervised non-parametric Bayesian based deep kernel learning algorithm is proposed, in which unlabeled data are incorporated through posterior regularization. We show the effectiveness of ManiDKL with extensive experiments. It shows that ManiDKL performs comparatively with state-of-the-art SSL image classification methods. Most importantly, we show the superiority of ManiDKL over all existing SSL regression methods on public image datasets. [ABSTRACT FROM AUTHOR]

In order to more effectively monitor and interfere with enemy signals, it is particularly important to accurately and efficiently identify the intercepted signals and estimate their parameters in the increasingly complex electromagnetic environment. Therefore, in non-cooperative situations, it is of great practical significance to study how to accurately detect direct sequence spread spectrum (DSSS) signals in real time and estimate their parameters. The traditional time-delay correlation algorithm encounters the challenges such as peak energy leakage and false peak interference. As an alternative, this paper introduces a Pseudo-Noise (PN) code period estimation method utilizing a one-dimensional (1D) convolutional neural network based on the residual network (CNN-ResNet). This method transforms the problem of spread spectrum code period estimation into a multi-classification problem of spread spectrum code length estimation. Firstly, the In-phase/Quadrature(I/Q) two-way of the received DSSS signals is directly input into the CNN-ResNet model, which will automatically learn the characteristics of the DSSS signal with different PN code lengths and then estimate the PN code length. Simulation experiments are conducted using a data set with DSSS signals ranging from −20 to 10 dB in terms of signal-to-noise ratios (SNRs). Upon training and verifying the model using BPSK modulation, it is then put to the test with QPSK-modulated signals, and the estimation performance was analyzed through metrics such as loss function, accuracy rate, recall rate, and confusion matrix. The results demonstrate that the 1D CNN-ResNet proposed in this paper is capable of effectively estimating the PN code period of the non-cooperative DSSS signal, exhibiting robust generalization abilities. [ABSTRACT FROM AUTHOR]

With the wide application of direct sequence spread spectrum (DSSS) signals, the comprehensive performance of DSSS communication systems has been continuously improved, making the electronic reconnaissance link in communication countermeasures more difficult. Electronic reconnaissance technology, as the fundamental means of modern electronic warfare, mainly includes signal detection, recognition, and parameter estimation. At present, research on DSSS detection algorithms is mostly based on the correlation characteristics of DSSS signals, and autocorrelation algorithm is the most mature and widely used method in practical engineering. With the continuous development of deep learning, deep-learning-based methods have gradually been introduced to replace traditional algorithms in the field of signal processing. This paper proposes a spread spectrum signal detection method based on convolutional neural network (CNN). Through experimental analysis, the detection performance of the CNN model proposed in this paper on DSSS signals in various situations has been compared and analyzed with traditional autocorrelation detection methods for different signal-to-noise ratios. The experiments verified the estimation performance of the model in this paper under different signal-to-noise ratios, different spreading code lengths, different spreading code types, and different modulation methods and compared it with the autocorrelation detection algorithm. It was found that the detection performance of the model in this paper was higher than that of the autocorrelation detection method, and the overall performance was improved by 4 dB. [ABSTRACT FROM AUTHOR]

Reliable and timely rice distribution information is of great value for real-time, quantitative, and localized control of rice production information. Synthetic aperture radar (SAR) has all-weather and all-day observation capability to monitor rice distribution in tropical and subtropical areas. To improve the physical interpretability and spatial interpretability of the deep learning model for SAR rice field extraction, a new SHapley Additive exPlanation (SHAP) value-guided explanation model (SGEM) for polarimetric SAR (PolSAR) data was proposed. First, a rice sample set was produced based on field survey and optical data, and the physical characteristics were extracted using decomposition of polarimetric scattering. Then a SHAP-based Physical Feature Interpretable Module (SPFIM) combing the long short-term memory (LSTM) model and SHAP values was designed to analyze the importance of physical characteristics, a credible physical interpretation associated with rice phenology was provided, and the weight of physical interpretation was combined with the weight of original PolSAR data. Moreover, a SHAP-guided spatial interpretation network (SSEN) was constructed to internalize the spatial interpretation values into the network layer to optimize the spatial refinement of the extraction results. Shanwei City, Guangdong Province, China, was chosen as the study area. The experimental results showed that the physical explanation provided by the proposed method had a high correlation with the rice phenology, and spatial self-interpretation for finer extraction results. The overall accuracy of the rice mapping results was 95.73%, and the kappa coefficient reached 0.9143. The proposed method has a high interpretability and practical value compared with other methods. [ABSTRACT FROM AUTHOR]

• A stratified framework is proposed to extract information from large area of image. • The theoretical concept of geographical thematic scene is firstly presented. • A deep learning network is proposed to fast divide the geographical thematic scene. • A well-performed boundary delineation accuracy assessment method is proposed. Extraction of cultivated land information from high spatial resolution remote sensing images is increasingly becoming an important approach to digitization and informatization in modern agriculture. The continuous development of deep learning technology has made it possible to extract information of cultivated land parcels by an intelligent way. Aiming at fine extraction of cultivated land parcels within large areas, this article builds a framework of geographical thematic scene division according to the rule of territorial differentiation in geography. A deep learning semantic segmentation network, improved U-net with depthwise separable convolution (DSCUnet), is proposed to achieve the division of the whole image. Then, an extended multichannel richer convolutional features (RCF) network is involved to delineate the boundaries of cultivated land parcels from agricultural functional scenes obtained by the former step. In order to testify the feasibility and effectiveness of the proposed methods, this article implemented experiments using Gaofen-2 images with different spatial resolution. The results show an outstanding performance using methods proposed in this article in both dividing agricultural functional scenes and delineating cultivated land parcels compared with other commonly used methods. Meanwhile, the extraction results have the highest accuracy in both the traditional evaluation indices (like Precision, Recall, F 1 , and IoU) and geometric boundary precision of cultivated land parcels. The methods in this article can provide a feasible solution to the problem of finely extracting cultivated land parcels information within large areas and complex landscape conditions in practical applications. [ABSTRACT FROM AUTHOR]

Anatomical Therapeutic Chemical (ATC) classification for compounds/drugs plays an important role in drug development and basic research. However, previous methods depend on interactions extracted from STITCH dataset which may make it depend on lab experiments. We present a pilot study to explore the possibility of conducting the ATC prediction solely based on the molecular structures. The motivation is to eliminate the reliance on the costly lab experiments so that the characteristics of a drug can be pre-assessed for better decision-making and effort-saving before the actual development. To this end, we construct a new benchmark consisting of 4545 compounds which is with larger scale than the one used in previous study. A light-weight prediction model is proposed. The model is with better explainability in the sense that it is consists of a straightforward tokenization that extracts and embeds statistically and physicochemically meaningful tokens, and a deep network backed by a set of pyramid kernels to capture multi-resolution chemical structural characteristics. Its efficacy has been validated in the experiments where it outperforms the state-of-the-art methods by 15.53% in accuracy and by 69.66% in terms of efficiency. We make the benchmark dataset, source code and web server open to ease the reproduction of this study. [ABSTRACT FROM AUTHOR]

The elimination of hunger is the top concern for developing countries and is the key to maintain national stability and security. Paddy rice occupies an essential status in food supply, whose accurate monitoring is of great importance for human sustainable development. As one of the most important paddy rice production countries in the world, Thailand has a favorable hot and humid climate for paddy rice growing, but the growth patterns of paddy rice are too complicated to construct promising growth models for paddy rice discrimination. To solve this problem, this study proposes a large-scale paddy rice mapping scheme, which uses time-series Sentinel-1 data to generate a convincing annual paddy rice map of Thailand. The proposed method extracts temporal statistical features of the time-series SAR images to overcome the intra-class variability due to different management practices and modifies the U-Net model with the fully connected Conditional Random Field (CRF) to maintain the edge of the fields. In this study, 758 Sentinel-1 images that covered the whole country from the end of 2018 to 2019 were acquired to generate the annual paddy rice map. The accuracy, precision, and recall of the resultant paddy rice map reached 91%, 87%, and 95%, respectively. Compared to SVM classifier and the U-Net model based on feature selection strategy (FS-U-Net), the proposed scheme achieved the best overall performance, which demonstrated the capability of overcoming the complex cultivation conditions and accurately identifying the fragmented paddy rice fields in Thailand. This study provides a promising tool for large-scale paddy rice monitoring in tropical production regions and has great potential in the global sustainable development of food and environment management. [ABSTRACT FROM AUTHOR]

Immunofixation Electrophoresis (IFE) analysis is of great importance to the diagnosis of Multiple Myeloma, which is among the top-9 cancer killers in the United States, but has rarely been studied in the context of deep learning. Two possible reasons are: 1) the recognition of IFE patterns is dependent on the co-location of bands that forms a binary relation, different from the unary relation (visual features to label) that deep learning is good at modeling; 2) deep classification models may perform with high accuracy for IFE recognition but is not able to provide firm evidence (where the co-location patterns are) for its predictions, rendering difficulty for technicians to validate the results. We propose to address these issues with collocative learning, in which a collocative tensor has been constructed to transform the binary relations into unary relations that are compatible with conventional deep networks, and a location-label-free method that utilizes the Grad-CAM saliency map for evidence backtracking has been proposed for accurate localization. In addition, we have proposed Coached Attention Gates that can regulate the inference of the learning to be more consistent with human logic and thus support the evidence backtracking. The experimental results show that the proposed method has obtained a performance gain over its base model ResNet18 by 741.30% in IoU and also outperformed popular deep networks of DenseNet, CBAM, and Inception-v3. [ABSTRACT FROM AUTHOR]

In recent years, synthetic aperture radar (SAR) has been a widely used data source in the remote sensing field due to its ability to work all day and in all weather conditions. Among SAR satellites, Sentinel-1 is frequently used to monitor large-scale ground objects. The Mekong Delta is a major agricultural region in Southeast Asia, so monitoring its cropland is of great importance. However, it is a challenge to distinguish cropland from other ground objects, such as aquaculture and wetland, in this region. To address this problem, the study proposes a statistical feature combination from the Sentinel-1 dual-polarimetric (dual-pol) data time series based on the m/χ decomposition method. Then the feature combination is put into the proposed Omni-dimensional Dynamic Convolution Residual Segmentation Model (ODCRS Model) of high fitting speed and classification accuracy to realize the cropland extraction of the Mekong Delta region. Experiments show that the ODCRS model achieves an overall accuracy of 93.85%, a MIoU of 88.04%, and a MPA of 93.70%. The extraction results show that our method can effectively distinguish cropland from aquaculture areas and wetlands. [ABSTRACT FROM AUTHOR]

Change detection (CD) is a challenging task considering land surface cover changes. Recently, deep learning has been introduced into CD. However, these methods suffer from shortcomings such as insufficient utilization of shallow features and inadequate aggregation of semantic context, which result in pseudo changes and poor integrity of change objects. To address this, an enhanced shallow feature difference and semantic context network (ESDSCNet) is proposed in this study for remote sensing CD. It uses HRNet to extract multi-scale features from bi-temporal images to obtain shallow and deep features. To fully exploit the shallow features, they are input into the difference statistical texture learning (DSTL) module to extract more discriminative features. Subsequently, the features enhanced by DSTL are fed into the change object contextual representation (COCR) module along with the deep difference features extracted by HRNet to characterize the contextual information of the change object. To verify the performance of ESDSCNet in different scenes, this paper takes building change detection as a case and conducts experiments based on three datasets. The experimental results reveal that ESDSCNet outperforms six other advanced methods, regardless of the intersection over union or the harmonic mean of precision rate and recall rate, which further confirms the effectiveness of the proposed network. In addition, another speciality of this paper is that ESDSCNet can identify not only 'Where has changed' but also 'What has changed what', so it has potential application in semantic change detection besides building change detection. [ABSTRACT FROM AUTHOR]

Classifications based on deep learning have been widely applied in the estimation of the direction of arrival (DOA) of signal. Due to the limited number of classes, the classification of DOA cannot satisfy the required prediction accuracy of signals from random azimuth in real applications. This paper presents a Centroid Optimization of deep neural network classification (CO-DNNC) to improve the estimation accuracy of DOA. CO-DNNC includes signal preprocessing, classification network, and Centroid Optimization. The DNN classification network adopts a convolutional neural network, including convolutional layers and fully connected layers. The Centroid Optimization takes the classified labels as the coordinates and calculates the azimuth of received signal according to the probabilities of the Softmax output. The experimental results show that CO-DNNC is capable of acquiring precise and accurate estimation of DOA, especially in the cases of low SNRs. In addition, CO-DNNC requires lower numbers of classes under the same condition of prediction accuracy and SNR, which reduces the complexity of the DNN network and saves training and processing time. [ABSTRACT FROM AUTHOR]

Built-up area (BA) extraction using synthetic aperture radar (SAR) data has emerged as a potential method in urban research. Currently, typical deep-learning-based BA extractors show high false-alarm rates in the layover areas and subsurface bedrock, which ignore the surrounding information and cannot be directly applied to large-scale BA mapping. To solve the above problems, a novel transformer-based BA extraction framework for SAR images is proposed. Inspired by SegFormer, we designed a BA extractor with multi-level dual-attention transformer encoders. First, the hybrid dilated convolution (HDC) patch-embedding module keeps the surrounding information of the input patches. Second, the channel self-attention module is designed for dual-attention transformer encoders and global modeling. The multi-level structure is employed to produce the coarse-to-fine semantic feature map of BAs. About 1100 scenes of Gaofen-3 (GF-3) data and 200 scenes of Sentinel-1 data were used in the experiment. Compared to UNet, PSPNet, and SegFormer, our model achieved an 85.35% mean intersection over union (mIoU) and 94.75% mean average precision (mAP) on the test set. The proposed framework achieved the best results in both mountainous and plain terrains. The experiments using Sentinel-1 shows that the proposed method has a good generalization ability with different SAR data sources. Finally, the BA map of China for 2020 was obtained with an overall accuracy of about 86%, which shows high consistency with the global urban footprint. The above experiments proved the effectiveness and robustness of the proposed framework in large-scale BA mapping. [ABSTRACT FROM AUTHOR]

Underwater ghost imaging based on deep learning can effectively reduce the influence of forward scattering and back scattering of water. With the help of data-driven methods, high-quality results can be reconstructed. However, the training of the underwater ghost imaging requires enormous paired underwater datasets, which are difficult to obtain directly. Although the Cycle-GAN method solves the problem to some extent, the blurring degree of the fuzzy class of the paired underwater datasets generated by Cycle-GAN is relatively unitary. To solve this problem, a few-shot underwater image generative network method is proposed. Utilizing the proposed few-shot learning image generative method, the generated paired underwater datasets are better than those obtained by the Cycle-GAN method, especially under the condition of few real underwater datasets. In addition, to reconstruct high-quality results, an underwater deblurring ghost imaging method is proposed. The reconstruction method consists of two parts: reconstruction and deblurring. The experimental and simulation results show that the proposed reconstruction method has better performance in deblurring at a low sampling rate, compared with existing underwater ghost imaging methods based on deep learning. The proposed reconstruction method can effectively increase the clarity degree of the underwater reconstruction target at a low sampling rate and promotes the further applications of underwater ghost imaging. [ABSTRACT FROM AUTHOR]

Timely and accurate information on rice cultivation makes important contributions to the profound reform of the global food and agricultural system, and promotes the development of global sustainable agriculture. With all-day and all-weather observing ability, synthetic aperture radar (SAR) can monitor the distribution of rice in tropical and subtropical areas. To solve the problem of misclassification of rice with no marked signal during the flooding period in subtropical hilly areas, this paper proposes a new feature combination and dual branch bi-directional long short-term memory (DB-BiLSTM) model to achieve high-precision rice mapping using Sentinel-1 time series data. Based on field investigation data, the backscatter time series curves of the rice area were analyzed, and a characteristic index (VV − VH)/(VV + VH) (VV: vertical emission and vertical receipt of polarization, VH: vertical emission and horizontal receipt of polarization) for small areas of hilly land was proposed to effectively distinguish rice and non-rice crops with no marked flooding period. The DB-BiLSTM model was designed, ensuring the independent learning of multiple features and effectively combining the time series information of both (VV − VH)/(VV + VH) and VH features. The city of Shanwei, Guangdong Province, China, was selected as the study area. Experimental results showed that the overall accuracy of the rice mapping results was 97.29%, and the kappa coefficient reached 0.9424. Compared to other methods, the rice mapping results obtained by the proposed method maintained good integrity and had less misclassification, which demonstrated the proposed method's practical value in accurate and effective rice mapping tasks. [ABSTRACT FROM AUTHOR]

[Display omitted] • A novel machine vision method for real-time monitoring of deposition height in noise conditions is proposed. • A novel self-attention temporal convolutional network (SA-TCN) is proposed for predicting future deposition height. • Quantifying complex physical phenomena with specific data has significantly contributed to enhancing the prediction accuracy of a deep learning model. Laser Directed Energy Deposition (L-DED) has garnered significant attention due to its high flexibility and rapid processing capabilities. However, complex physical flow fields and drastic temperature variations are present during L-DED processing, leading to variations in deposition height at different layers and positions under the same processing parameters. Therefore, real-time monitoring of deposition height and timely knowledge of future deposition height are crucial for controlling geometries and arranging processing time effectively. To address this issue, a machine vision method for real-time monitoring of deposition height in noisy environments is proposed, demonstrating a remarkable similarity of 99.22% compared to values measured by a laser scanner. Addressing the complex physical phenomena during processing, specific data quantification was performed. A novel self-attention temporal convolutional network (SA-TCN) was then introduced as a data-driven model to replace physical models for predicting future deposition height, achieving an impressive accuracy of 99.71%. Experiments show that quantifying different physical phenomena with specific data to some extent improves the model prediction accuracy, providing significant support for future deposition height prediction and processing time control of parts in actual production. [ABSTRACT FROM AUTHOR]

• An efficient sampling strategy optimization method with high-quality FSPI reconstruction is implemented. • Numerical Simulations and experiments proved the effectiveness, which has advantages over deterministic sampling strategies. • The geometric topologies of the optimal sampling strategy for different categories of datasets are analyzed. • The geometric topologies of the optimal sampling strategy for different categories of datasets are summarized. • It provides subsequent for sampling strategy choice of Fourier single pixel imaging. It is necessary to artificially set the sampling strategy of Fourier single-pixel imaging (FSPI) to obtain more useful information of the image under the assumption that the image information is mainly concentrated in the low-frequency part of the spatial frequency domain. However, due to the empiricism sampling strategy, it is easy to sample the spatial frequencies with low contribution to image reconstruction, resulting in a waste of sampling resources. To solve this problem, a one-stage FSPI reconstruction network with sampling strategy optimization is designed to obtain the efficient sampling strategy and improve the reconstruction quality. The proposed one-stage FSPI reconstruction network contains the sampling strategy optimization module and reconstruction module, which are jointly trained to obtain high-quality reconstructed results and efficient sampling strategy simultaneously. Experiments and numerical simulations demonstrate the effectiveness of the proposed method. Moreover, the geometric topology of the optimal sampling strategy for FSPI reconstruction trained on various categories of datasets is summarized. It is found that the dataset with similar categories have the similar geometric topology of the optimal sampling strategy, which provides support for sampling strategy selection in subsequent Fourier single-pixel imaging studies. [ABSTRACT FROM AUTHOR]