Search

The East Asia (or Physospermopsis) clade was recognized in previous molecular phylogenetic investigations into the higher-level relationships of Apiaceae subfamily Apioideae. The composition of this clade, the phylogenetic relationships among its constituent taxa, and the placement of species previously determined to be problematic have yet to be resolved. Herein, nrDNA ITS sequences were obtained for 150 accessions of Apioideae, representing species whose distributions are in East Asia or genera having one or more species included within the East Asia clade. These data, along with published ITS sequences from other Apioideae (for 3678 accessions altogether), were subjected to maximum likelihood and Bayesian inference analyses. The results show that the East Asia clade contains representatives of 11 currently recognized genera: Hansenia, Hymenolaena, Keraymonia, Sinolimprichtia, Acronema, Hymenidium, Physospermopsis, Pimpinella, Sinocarum, Tongoloa, and Trachydium. However, the latter seven genera have members falling outside of the East Asia clade, including the generic types of all except Tongoloa. Within the clade, the species comprising these seven genera are widely intermingled, greatly increasing confusion among relationships than previously realized. The problematic species Physospermopsis cuneata is confirmed as falling within the East Asia clade, whereas P. rubrinervis allies with the generic type in tribe Pleurospermeae. Physospermopsis kingdon-wardii is confirmed as a member of the genus Physospermopsis, whereas the generic attributions of P. cuneata and Tongoloa stewardii remain unclear. Two species of Sinocarum (S. filicinum and S. wolffianum) are transferred into the genus Meeboldia. This is the most comprehensive molecular phylogenetic investigation of the East Asia clade to date, and while the results increase systematic understanding of the clade, they also highlight the need for further studies of one of the most taxonomically intractable groups in Apioideae.

Abstract Background Sium L. (Apiaceae) is a small genus distributed primarily in Eurasia, with one species also occurring in North America. Recently, its circumscription has been revised to include 10 species, however, the phylogenetic relationships within its two inclusive clades were poorly supported or collapsed in previous studies based on nuclear ribosomal DNA ITS or cpDNA sequences. To identify molecular markers suitable for future intraspecific phylogeographic and population genetic studies, and to evaluate the efficacy of plastome in resolving the phylogenetic relationships of the genus, the complete chloroplast (cp) genomes of six Sium species were sequenced. Results The Sium plastomes exhibited typical quadripartite structures of Apiaceae and most other higher plant plastid DNAs, and were relatively conserved in their size (153,029–155,006 bp), gene arrangement and content (with 114 unique genes). A total of 61–67 SSRs, along with 12 highly divergent regions (trnQ, trnG-atpA, trnE-trnT, rps4-trnT, accD-psbI, rpl16, ycf1-ndhF, ndhF-rpl32, rpl32-trnL, ndhE-ndhG, ycf1a and ycf1b) were discovered in the plastomes. No significant IR length variation was detected showing that plastome evolution was conserved within this genus. Phylogenomic analysis based on whole chloroplast genome sequences produced a highly resolved phylogenetic tree, in which the monophyly of Sium, as well as the sister relationship of its two inclusive clades were strongly supported. Conclusions The plastome sequences could greatly improve phylogenetic resolution, and will provide genomic resources and potential markers useful for future studies of the genus.

Abstract Extracting robust feature representation is one of the key challenges for person re‐identification (ReID) task. Although convolution neural network (CNN)‐based methods have achieved great success, they still cannot handle the part occlusion and misalignment caused by limited receptive field. Recently, pure transformer models have shown its power in the person ReID task. However, current transformer models adopt patches of equal‐scale as input, and cannot solve the problem of cross‐scale interaction properly. To overcome this problem, an adaptive cross‐scale transformer from a perspective of the graph signal, named ACSFormer, is proposed. Specifically, the self‐attention module is first treated as an undirected fully connected graph. And then, “node variation” is introduced as an indicator to adaptively merge neighbourhood tokens. To the best of the authors’ knowledge, their ACSFormer is the first work to attempt to combine pure transformers and graph signal processing in the field of person ReID. Extensive evaluations are conducted on three person ReID datasets to validate the performance of ACSFormer. Experiments demonstrate that this ACSFormer performs on par with state‐of‐the‐art CNN‐based methods and consistently improves transformer‐based baseline, for example, surpassing ViT‐baseline by 2.5%, 2.7% and 4.8% mAP on Market1501, DukeMTMC‐reID and MSMT17, respectively.

Ligusticopsis acaulis, belonging to the family Apiaceae (Umbelliferae), is endemic to China. The complete chloroplast genome sequence of L. acaulis was assembled and annotated for the first time in this study. The results showed that the plastome was 148,509 bp in length and consisted of a pair of inverted repeat regions (IRs: 19,468 bp), a large single-copy region (LSC: 91,902 bp), and a small single-copy region (SSC: 17,671 bp). A total of 114 unique genes were annotated, including 80 protein-coding, 30 tRNA, and four rRNA genes. According to the phylogenetic analysis, L. acaulis belongs to the tribe Selineae, with a close relationship to Ligusticum hispidum (Franch.) Wolff.

The taxonomic placement of Pleurospermum lecomteanum, previously synonymized with Pleurospermum wilsonii, was carefully examined using herbarium specimens and molecular evidence. The results showed that Pleurospermum lecomteanum is distinguished from P. wilsonii by several morphological characters. Its phylogenetic position is separate from P. wilsonii in the ML tree. Therefore, Pleurospermum lecomteanum should be restored as a distinct species.

Abstract The fusion of multiple monitoring sensors is crucial to improve the accuracy and robustness of machinery fault diagnosis. However, existing fault diagnosis methods may underestimate the interference of noise in the multi‐sensor fusion process, leading to unsatisfied performance. To handle this problem, this paper proposes a deep model based on the frequency adaptive wavelet pyramid. First, an adaptive frequency selection strategy is designed to prune the seriously polluted frequencies and only retain some key frequencies. Then, the self‐attention mechanism is used to perform information fusion on the selected frequency bands of different sensors. Finally, a wavelet fusion pyramid is adopted by repeating the fusion process at multiple wavelet decomposition levels. In this way, different sensors can be fused in a more fine‐grained manner. The experimental results on two multi‐sensor‐based fault diagnosis datasets demonstrate the anti‐noise capability of the proposed method.

Background and aims: Renal damage in severe coronavirus disease 2019 (COVID-19) is highly associated with mortality. Finding relevant therapeutic candidates that can alleviate it is crucial. Angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin-receptor blockers (ARBs) have been shown to be harmless to COVID-19 patients, but it remains elusive whether ACEIs/ARBs have protective benefits to them. We wished to determine if ACEIs/ARBs had a protective effect on the renal damage associated with COVID-19, and to investigate the mechanism.Methods: We used the envelope (E) protein of severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) to induce COVID-19-like multiple organ damage and observed renal fibrosis. We induced the epithelial–mesenchymal transformation of HK-2 cells with E protein, and found that olmesartan could alleviate it significantly. The protective effects of olmesartan on E protein-induced renal fibrosis were evaluated by renal-function assessment, pathologic alterations, inflammation, and the TGF-β1/Smad2/3 signaling pathway. The distribution of high-mobility group box (HMGB)1 was examined after stimulation with E protein and olmesartan administration.Results: E protein stimulated HMGB1 release, which triggered the immune response and promoted activation of TGF-β1/Smad2/3 signaling: both could lead to renal fibrosis. Olmesartan regulated the distribution of HMGB1 under E protein stimulation. Olmesartan inhibited the release of HMGB1, and reduced the inflammatory response and activation of TGF-β1/Smad2/3 signaling. Olmesartan increased the cytoplasmic level of HMGB1 to promote the autophagic degradation of TGF-β1, thereby alleviating fibrosis further.Conclusion: Olmesartan alleviates E protein-induced renal fibrosis by regulating the release of HMGB1 and its mediated autophagic degradation of TGF-β1.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused a devastating global pandemic, which has seriously affected human health worldwide. The discovery of therapeutic agents is extremely urgent, and the viral structural proteins are particularly important as potential drug targets. SARS-CoV-2 envelope (E) protein is one of the main structural proteins of the virus, which is involved in multiple processes of the virus life cycle and is directly related to pathogenesis process. In this review, we present the amino acid sequence of the E protein and compare it with other two human coronaviruses. We then explored the role of E protein in the viral life cycle and discussed the pathogenic mechanisms that E protein may be involved in. Next, we summarize the potential drugs against E protein discovered in the current studies. Finally, we described the possible effects of E protein mutation on virus and host. This established a knowledge system of E protein to date, aiming to provide theoretical insights for mitigating the current COVID-19 pandemic and potential future coronavirus outbreaks.

Abstract This study addresses the problem of the unsupervised pre‐training of video representation learning. The authors' focus is on two common approaches: knowledge distillation and self‐supervised learning. The insight provided is that knowledge distillation and the rapidly advancing self‐supervised technique can be mutually beneficial. Combining these two approaches, a unified framework of self‐supervised learning and image‐based distillation (SSID) for unsupervised video pre‐training is proposed. The effectiveness of SSID in comparison to both image‐based distillation methods and the existing self‐supervised pre‐training baseline is demonstrated. In particular, the authors' model leverages three signals from the unlabelled data. First, the authors distil from the classifier of a 2D pre‐trained model as a soft label. To regularize the training process, the authors then build a novel positive pair of contrastive learning on the representation of the 2D/3D model. Finally, a self‐supervised pretext task is introduced to enhance the authors' model to become aware of the temporal evolution. The authors' experiment results showed that the learnt features achieved the best performance when transferred to action recognition tasks on UCF101 and HMDB51, reaching increases of 2.4% and 1.9% compared to the existing unsupervised pre‐training model, respectively.

Infrared small target superresolution (SR) aims to recover a reliable and detailed high-resolution image with high-contrast targets from its low-resolution counterparts. Since the infrared small target lacks color and fine structure information, it is significant to exploit the supplementary information among sequence images to enhance the target. In this article, we propose the first infrared small target SR method named local motion and contrast prior driven deep network (MoCoPnet) to integrate the domain knowledge of the infrared small target into deep network, which can mitigate the intrinsic feature scarcity of infrared small targets. Specifically, motivated by the local motion prior in the spatio-temporal dimension, we propose a local spatio-temporal attention module to perform implicit frame alignment and incorporate the local spatio-temporal information to enhance the local features (especially for small targets). Motivated by the local contrast prior in the spatial dimension, we propose a central difference residual group to incorporate the central difference convolution into the feature extraction backbone, which can achieve center-oriented gradient-aware feature extraction to further improve the target contrast. Extensive experiments have demonstrated that our method can recover accurate spatial dependence and improve the target contrast. Comparative results show that the MoCoPnet can outperform the state-of-the-art video SR and single image SR methods in terms of both SR performance and target enhancement. Based on the SR results, we further investigate the influence of SR on infrared small target detection and the experimental results demonstrate that the MoCoPnet promotes the detection performance.

Abstract Background and aims Renal fibrosis is the common outcome in all progressive forms of chronic kidney disease. Unfortunately, the pathogenesis of renal fibrosis remains largely unexplored, among which metabolic reprogramming plays an extremely crucial role in the evolution of renal fibrosis. Ceria nanoparticles (CeNP-PEG) with strong ROS scavenging and anti-inflammatory activities have been applied for mitochondrial oxidative stress and inflammatory diseases. The present study aims to determine whether CeNP-PEG has therapeutic value for renal fibrosis. Methods The unilateral ureteral obstructive fibrosis model was used to assess the therapeutic effects in vivo. Transforming growth factor beta1-induced epithelial-to-mesenchymal transition in HK-2 cells was used as the in vitro cell model. The seahorse bioscience X96 extracellular flux analyzer was used to measure the oxygen consumption rate and extracellular acidification rate. Results In the present study, CeNP-PEG treatment significantly ameliorated renal fibrosis by increased E-cadherin protein expression, and decreased α-SMA, Vimentin and Fibronectin expression both in vitro and in vivo. Additionally, CeNP-PEG significantly reduced the ROS formation and improved the levels of mitochondrial ATP. The seahorse analyzer assay demonstrated that the extracellular acidification rate markedly decreased, whereas the oxygen consumption rate markedly increased, in the presence of CeNP-PEG. Furthermore, the mitochondrial membrane potential markedly enhanced, hexokinase 1 and hexokinase 2 expression significantly decreased after treatment with CeNP-PEG. Conclusions CeNP-PEG can block the dysregulated metabolic status and exert protective function on renal fibrosis. This may provide another therapeutic option for renal fibrosis. Graphical Abstract

Abstract Light field (LF) cameras capture scenes from multiple views and provide additional angular information for image super‐resolution (SR). Existing CNN‐based LF image SR methods commonly develop specific models for different angular resolutions. However, since the angular resolution can vary significantly with different LF devices, these methods have limited flexibility for real‐world applications. Here, an angular‐flexible network to use a single model to super‐resolve LF images of arbitrary angular resolution is proposed. In this method, spatial and angular feature extractors are designed to achieve angular‐flexible feature extraction, and develop a decouple‐and‐fuse module for SR reconstruction. Moreover, a mixed‐angular‐resolution training strategy is proposed to further enhance the angular flexibility. Experimental results on five public datasets demonstrate the state‐of‐the‐art performance of the method. Source codes are available at https://github.com/ZhengyuLiang24/LF‐AFnet. Here, an angular‐flexible network for light field image super‐resolution is proposed. The network can handle LFs captured by different kinds of devices with arbitrary angular resolutions. Experimental results demonstrate the effectiveness and superior performance of the method.

Thin-film composite (TFC) polyamide membranes have a wide range of applications in forward osmosis, but tuning the water flux remains a significant challenge due to concentration polarization. The generation of nano-sized voids within the polyamide rejection layer can change the roughness of the membrane. In this experiment, the micro-nano structure of the PA rejection layer was adjusted by adding sodium bicarbonate to the aqueous phase to generate nano-bubbles, and the changes of its roughness with the addition of sodium bicarbonate were systematically demonstrated. With the enhanced nano-bubbles, more and more blade-like and band-like features appeared on the PA layer, which could effectively reduce the reverse solute flux of the PA layer and improve the salt rejection of the FO membrane. The increase in roughness raised the area of the membrane surface, which led to a larger area for concentration polarization and reduced the water flux. This experiment demonstrated the variation of roughness and water flux, providing an effective idea for the preparation of high-performance FO membranes.

Early time series classification (ETSC) is crucial for real-world time-sensitive applications. This task aims to classify time series data with least timestamps at the desired accuracy. Early methods used fixed-length time series to train the deep models, and then quit the classification process by setting specific exiting rules. However, these methods may not adapt to the length variation of flow data in ETSC. Recent advances have proposed end-to-end frameworks, which leveraged the Recurrent Neural Networks to handle the varied-length problems, and the exiting subnets for early quitting. Unfortunately, the conflict between the classification and early exiting objectives is not fully considered. To handle these problems, we decouple the ETSC task into the varied-length TSC task and the early exiting task. First, to enhance the adaptive capacity of classification subnets to the data length variation, a feature augmentation module based on random length truncation is proposed. Then, to handle the conflict between classification and early exiting, the gradients of these two tasks are projected into a unified direction. Experimental results on 12 public datasets demonstrate the promising performance of our proposed method.

The negation of probability distribution is a new perspective from which to obtain information. Dempster–Shafer (D–S) evidence theory, as an extension of possibility theory, is widely used in decision-making-level fusion. However, how to reasonably construct the negation of basic probability assignment (BPA) in D–S evidence theory is an open issue. This paper proposes a new negation of BPA, logarithmic negation. It solves the shortcoming of Yin’s negation that maximal entropy cannot be obtained when there are only two focal elements in the BPA. At the same time, the logarithmic negation of BPA inherits the good properties of the negation of probability, such as order reversal, involution, convergence, degeneration, and maximal entropy. Logarithmic negation degenerates into Gao’s negation when the values of the elements all approach 0. In addition, the data fusion method based on logarithmic negation has a higher belief value of the correct target in target recognition application.

The realistic resistance zone of water and salt molecules to transport across a TFC-RO membrane is the topmost polyamide nanofilm. The existence of hollow voids in the fully aromatic polyamide (PA) film gives its surface ridge-and-valley morphologies, which confuses the comprehensions of the definition of the PA thickness. The hollow voids, however, neither participate in salt–water separation nor hinder water penetrating. In this paper, the influence of intrinsic thickness (single wall thickness) of the PA layer on water permeability was studied by adjusting the concentration of reacting monomers. It confirms that the true permeation resistance of water molecules originates from the intrinsic thickness portion of the membrane. The experimental results show that the water permeability constant decreases from 3.15 ± 0.02 to 2.74 ± 0.10 L·m−2·h−1·bar−1 when the intrinsic thickness of the membrane increases by 9 nm. The defects on the film surface generate when the higher concentration of MPD is matched with the relatively low concentration of TMC. In addition, the role of MPD and TMC in the micro-structure of the PA membrane was discussed, which may provide a new way for the preparation of high permeability and high selectivity composite reverse osmosis membranes.

Wavelet transform is a well-known multi-resolution tool to analyze the time series in the time-frequency domain. Wavelet basis is diverse but predefined by manual without taking the data into the consideration. Hence, it is a great challenge to select an appropriate wavelet basis to separate the low and high frequency components for the task on the hand. Inspired by the lifting scheme in the second-generation wavelet, the updater and predictor are learned directly from the time series to separate the low and high frequency components of the time series. An adaptive multi-scale wavelet neural network (AMSW-NN) is proposed for time series classification in this paper. First, candidate frequency decompositions are obtained by a multi-scale convolutional neural network in conjunction with a depthwise convolutional neural network. Then, a selector is used to choose the optimal frequency decomposition from the candidates. At last, the optimal frequency decomposition is fed to a classification network to predict the label. A comprehensive experiment is performed on the UCR archive. The results demonstrate that, compared with the classical wavelet transform, AMSW-NN could improve the performance based on different classification networks.

Recently, some researchers adopted the convolutional neural network (CNN) for time series classification (TSC) and have achieved better performance than most hand-crafted methods in the University of California, Riverside (UCR) archive. The secret to the success of the CNN is weight sharing, which is robust to the global translation of the time series. However, global translation invariance is not the only case considered for TSC. Temporal distortion is another common phenomenon besides global translation in time series. The scale and phase changes due to temporal distortion bring significant challenges to TSC, which is out of the scope of conventional CNNs. In this paper, a CNN architecture with an elastic matching mechanism, which is named Elastic Matching CNN (short for EM-CNN), is proposed to address this challenge. Compared with the conventional CNN, EM-CNN allows local time shifting between the time series and convolutional kernels, and a matching matrix is exploited to learn the nonlinear alignment between time series and convolutional kernels of the CNN. Several EM-CNN models are proposed in this paper based on diverse CNN models. The results for 85 UCR datasets demonstrate that the elastic matching mechanism effectively improves CNN performance.

Visual place recognition (VPR) in changing environments is an urgent challenge for long-term autonomous navigation. One recent ConvNet landmark-based approach exploits region landmarks coupled with ConvNet features to match images, and the approach has shown promising results under significant environmental and viewpoint changes. In this paper, we propose a robust ConvNet landmark-based system for VPR in changing outdoor roadway environments by extension of this approach from the following two aspects. First, our method utilizes more discriminative landmarks obtained by a novel refinement method called SemLandmarks, which leverages roadway scene semantic information to screen landmarks directly detected by an existing object proposal method. Second, our method improves the accuracy of image matching by introducing consistent spatial constraints based on the use of geometry-preserving landmark pairs. Experimental results demonstrate that our method significantly improves the state of the art in VPR in terms of recognition accuracy on three challenging benchmark data sets with various environmental and viewpoint changes.

Conflicting evidence affects the final target recognition results. Thus, managing conflicting evidence efficiently can help to improve the belief degree of the true target. In current research, the existing approaches based on belief entropy use belief entropy itself to measure evidence conflict. However, it is not convincing to characterize the evidence conflict only through belief entropy itself. To solve this problem, we comprehensively consider the influences of the belief entropy itself and mutual belief entropy on conflict measurement, and propose a novel approach based on an improved belief entropy and entropy distance. The improved belief entropy based on pignistic probability transformation function is named pignistic probability transformation (PPT) entropy that measures the conflict between evidences from the perspective of self-belief entropy. Compared with the state-of-the-art belief entropy, it can measure the uncertainty of evidence more accurately, and make full use of the intersection information of evidence to estimate the degree of evidence conflict more reasonably. Entropy distance is a new distance measurement method and is used to measure the conflict between evidences from the perspective of mutual belief entropy. Two measures are mutually complementary in a sense. The results of numerical examples and target recognition applications demonstrate that our proposed approach has a faster convergence speed, and a higher belief degree of the true target compared with the existing methods.

Recent advances in time series classification (TSC) have exploited deep neural networks (DNN) to improve the performance. One promising approach encodes time series as recurrence plot (RP) images for the sake of leveraging the state-of-the-art DNN to achieve accuracy. Such an approach has been shown to achieve impressive results, raising the interest of the community in it. However, it remains unsolved how to handle not only the variability in the distinctive region scale and the length of sequences but also the tendency confusion problem. In this paper, we tackle the problem using Multi-scale Signed Recurrence Plots (MS-RP), an improvement of RP, and propose a novel method based on MS-RP images and Fully Convolutional Networks (FCN) for TSC. This method first introduces phase space dimension and time delay embedding of RP to produce multi-scale RP images; then, with the use of asymmetrical structure, constructed RP images can represent very long sequences (>700 points). Next, MS-RP images are obtained by multiplying designed sign masks in order to remove the tendency confusion. Finally, FCN is trained with MS-RP images to perform classification. Experimental results on 45 benchmark datasets demonstrate that our method improves the state-of-the-art in terms of classification accuracy and visualization evaluation.

We propose a constant false alarm rate (CFAR) algorithm for moving target detection in synthetic aperture radar (SAR) images based on the spreading characteristics of interferograms on the magnitude-phase (M-P) plane. This method is based on the observation that, in practice, both moving and stationary targets along with clutter are located at different regions in the M-P plane, and hence reasonable partitions of the M-P plane can help in detecting moving targets. To ensure efficient CFAR detection and to resolve the effect of factors that influence detection results, the proposed algorithm is divided into three distinct stages: coarse detection, fine detection, and post-processing. First, to accurately describe the statistical behavior of clutter, a global censoring strategy, called coarse detection, is introduced to adaptively eliminate the influences of the moving and stationary target points from the given data. Then, to acquire fine detection results, a novel CFAR detector is developed on the basis of the fits of a known theoretical M-P joint probability density function (PDF) against the two-dimensional (2-D) histogram of the censored clutter. The joint PDF’s projected contour line that satisfies the desirable probability of false alarm (PFA) corresponds to the required threshold of detection in the M-P plane. Finally, two filters, the magnitude and phase filters, are applied to reduce the false alarms generated from the previous procedures. The effectiveness of the proposed algorithm is validated through experimental results obtained from a two-channel SAR complex image.

Recent impressive studies on using ConvNet landmarks for visual place recognition take an approach that involves three steps: (a) detection of landmarks, (b) description of the landmarks by ConvNet features using a convolutional neural network, and (c) matching of the landmarks in the current view with those in the database views. Such an approach has been shown to achieve the state-of-the-art accuracy even under significant viewpoint and environmental changes. However, the computational burden in step (c) significantly prevents this approach from being applied in practice, due to the complexity of linear search in high-dimensional space of the ConvNet features. In this article, we propose two simple and efficient search methods to tackle this issue. Both methods are built upon tree-based indexing. Given a set of ConvNet features of a query image, the first method directly searches the features’ approximate nearest neighbors in a tree structure that is constructed from ConvNet features of database images. The database images are voted on by features in the query image, according to a lookup table which maps each ConvNet feature to its corresponding database image. The database image with the highest vote is considered the solution. Our second method uses a coarse-to-fine procedure: the coarse step uses the first method to coarsely find the top- N database images, and the fine step performs a linear search in Hamming space of the hash codes of the ConvNet features to determine the best match. Experimental results demonstrate that our methods achieve real-time search performance on five data sets with different sizes and various conditions. Most notably, by achieving an average search time of 0.035 seconds/query, our second method improves the matching efficiency by the three orders of magnitude over a linear search baseline on a database with 20,688 images, with negligible loss in place recognition accuracy.

Currently, huge quantities of remote sensing images (RSIs) are becoming available. Nevertheless, the scarcity of labeled samples hinders the semantic understanding of RSIs. Fortunately, many ground-level image datasets with detailed semantic annotations have been collected in the vision community. In this paper, we attempt to exploit the abundant labeled ground-level images to build discriminative models for overhead-view RSI classification. However, images from the ground-level and overhead view are represented by heterogeneous features with different distributions; how to effectively combine multiple features and reduce the mismatch of distributions are two key problems in this scene-model transfer task. Specifically, a semi-supervised manifold-regularized multiple-kernel-learning (SMRMKL) algorithm is proposed for solving these problems. We employ multiple kernels over several features to learn an optimal combined model automatically. Multi-kernel Maximum Mean Discrepancy (MK-MMD) is utilized to measure the data mismatch. To make use of unlabeled target samples, a manifold regularized semi-supervised learning process is incorporated into our framework. Extensive experimental results on both cross-view and aerial-to-satellite scene datasets demonstrate that: (1) SMRMKL has an appealing extension ability to effectively fuse different types of visual features; and (2) manifold regularization can improve the adaptation performance by utilizing unlabeled target samples.

Vehicle detection with orientation estimation in aerial images has received widespread interest as it is important for intelligent traffic management. This is a challenging task, not only because of the complex background and relatively small size of the target, but also the various orientations of vehicles in aerial images captured from the top view. The existing methods for oriented vehicle detection need several post-processing steps to generate final detection results with orientation, which are not efficient enough. Moreover, they can only get discrete orientation information for each target. In this paper, we present an end-to-end single convolutional neural network to generate arbitrarily-oriented detection results directly. Our approach, named Oriented_SSD (Single Shot MultiBox Detector, SSD), uses a set of default boxes with various scales on each feature map location to produce detection bounding boxes. Meanwhile, offsets are predicted for each default box to better match the object shape, which contain the angle parameter for oriented bounding boxes’ generation. Evaluation results on the public DLR Vehicle Aerial dataset and Vehicle Detection in Aerial Imagery (VEDAI) dataset demonstrate that our method can detect both the location and orientation of the vehicle with high accuracy and fast speed. For test images in the DLR Vehicle Aerial dataset with a size of 5616 × 3744 , our method achieves 76.1% average precision (AP) and 78.7% correct direction classification at 5.17 s on an NVIDIA GTX-1060.

Synthetic aperture radar (SAR) is one of the most important techniques for ocean monitoring. Azimuth ambiguities are a real problem in SAR images today, which can cause performance degradation in SAR ocean applications. In particular, littoral zones can be strongly affected by land-based sources, whereas they are usually regions of interest (ROI). Given the presence of complexity and diversity in littoral zones, azimuth ambiguities removal is a tough problem. As SAR sensors can have a repeat cycle, multi-temporal SAR images provide new insight into this problem. A method for azimuth ambiguities removal in littoral zones based on multi-temporal SAR images is proposed in this paper. The proposed processing chain includes co-registration, local correlation, binarization, masking, and restoration steps. It is designed to remove azimuth ambiguities caused by fixed land-based sources. The idea underlying the proposed method is that sea surface is dynamic, whereas azimuth ambiguities caused by land-based sources are constant. Thus, the temporal consistence of azimuth ambiguities is higher than sea clutter. It opens up the possibilities to use multi-temporal SAR data to remove azimuth ambiguities. The design of the method and the experimental procedure are based on images from the Sentinel data hub of Europe Space Agency (ESA). Both Interferometric Wide Swath (IW) and Stripmap (SM) mode images are taken into account to validate the proposed method. This paper also presents two RGB composition methods for better azimuth ambiguities visualization. Experimental results show that the proposed method can remove azimuth ambiguities in littoral zones effectively.

Detecting vehicles in aerial imagery plays an important role in a wide range of applications. The current vehicle detection methods are mostly based on sliding-window search and handcrafted or shallow-learning-based features, having limited description capability and heavy computational costs. Recently, due to the powerful feature representations, region convolutional neural networks (CNN) based detection methods have achieved state-of-the-art performance in computer vision, especially Faster R-CNN. However, directly using it for vehicle detection in aerial images has many limitations: (1) region proposal network (RPN) in Faster R-CNN has poor performance for accurately locating small-sized vehicles, due to the relatively coarse feature maps; and (2) the classifier after RPN cannot distinguish vehicles and complex backgrounds well. In this study, an improved detection method based on Faster R-CNN is proposed in order to accomplish the two challenges mentioned above. Firstly, to improve the recall, we employ a hyper region proposal network (HRPN) to extract vehicle-like targets with a combination of hierarchical feature maps. Then, we replace the classifier after RPN by a cascade of boosted classifiers to verify the candidate regions, aiming at reducing false detection by negative example mining. We evaluate our method on the Munich vehicle dataset and the collected vehicle dataset, with improvements in accuracy and robustness compared to existing methods.

If equipped with several radar emitters, a target will produce more than one measurement per time step and is denoted as an extended target. However, due to the requirement of all possible measurement set partitions, the exact probability hypothesis density filter for extended target tracking is computationally intractable. To reduce the computational burden, a fast partitioning algorithm based on hierarchy clustering is proposed in this paper. It combines the two most similar cells to obtain new partitions step by step. The pseudo-likelihoods in the Gaussian-mixture probability hypothesis density filter can then be computed iteratively. Furthermore, considering the additional measurement information from the emitter target, the signal feature is also used in partitioning the measurement set to improve the tracking performance. The simulation results show that the proposed method can perform better with lower computational complexity in scenarios with different clutter densities.

Detection and tracking surrounding moving obstacles such as vehicles and pedestrians are crucial for the safety of mobile robotics and autonomous vehicles. This is especially the case in urban driving scenarios. This paper presents a novel framework for surrounding moving obstacles detection using binocular stereo vision. The contributions of our work are threefold. Firstly, a multiview feature matching scheme is presented for simultaneous stereo correspondence and motion correspondence searching. Secondly, the multiview geometry constraint derived from the relative camera positions in pairs of consecutive stereo views is exploited for surrounding moving obstacles detection. Thirdly, an adaptive particle filter is proposed for tracking of multiple moving obstacles in surrounding areas. Experimental results from real-world driving sequences demonstrate the effectiveness and robustness of the proposed framework.

The performances on the applications of synthetic aperture radar (SAR) data strongly depend on the statistical characteristics of the pixel amplitudes or intensities. In this paper, a new empirical model, called simply ℋ𝒢o, has been proposed to characterize the statistical properties of SAR clutter data over the wide range of homogeneous, heterogeneous, and extremely heterogeneous returns of terrain classes. A particular case of the ℋ𝒢odistribution is the well-known 𝒢o distributions. We also derived analytically the estimators of the presented ℋ𝒢o model by applying the “method of log cumulants” (MoLCs). The performance of the proposed model is verified by using some measured SAR images.

This paper proposes a joint distribution of magnitude and phase for multilook SAR interferogram in extremely heterogeneous clutter. The presented theoretical distribution, called simply here distribution and derived from the multiplicative model and the known joint distribution of the homogeneous clutter, is shown to be able to model the extremely heterogeneous clutter areas. Moreover, we estimate distribution parameters by means of the well-known method-of-log-cumulants (MoLC). The experimental results applied on actual dual-channel SAR images prove the good performance of the proposed distribution.

A constant false alarm rate (CFAR) detecting method for ships in high-resolution dual-polarization synthetic aperture radar (SAR) amplitude images has been proposed in this paper. First, by the production of amplitude images from two polarimetric channels, a novel detector simply called the PMA detector has been constructed. We testified that the PMA detector could improve the signal-to-clutter ratio (SCR) and make the discrimination of a ship from clutter more easily. Second, the PMA detector’s statistical model has been described by the well-known distribution when facing complex sea background. The experiments performed on measured dual-polarization TerraSAR-X images demonstrate the good performance of the proposed CFAR detecting method.

The superpixel segmentation algorithm, as a preprocessing technique, should show good performance in fast segmentation speed, accurate boundary adherence and homogeneous regularity. A fast superpixel segmentation algorithm by iterative edge refinement (IER) works well on optical images. However, it may generate poor superpixels for Polarimetric synthetic aperture radar (PolSAR) images due to the influence of strong speckle noise and many small-sized or slim regions. To solve these problems, we utilized a fast revised Wishart distance instead of Euclidean distance in the local relabeling of unstable pixels, and initialized unstable pixels as all the pixels substituted for the initial grid edge pixels in the initialization step. Then, postprocessing with the dissimilarity measure is employed to remove the generated small isolated regions as well as to preserve strong point targets. Finally, the superiority of the proposed algorithm is validated with extensive experiments on four simulated and two real-world PolSAR images from Experimental Synthetic Aperture Radar (ESAR) and Airborne Synthetic Aperture Radar (AirSAR) data sets, which demonstrate that the proposed method shows better performance with respect to several commonly used evaluation measures, even with about nine times higher computational efficiency, as well as fine boundary adherence and strong point targets preservation, compared with three state-of-the-art methods.

With the rapid development of spaceborne synthetic aperture radar (SAR) and the increasing need of ship detection, research on adaptive ship detection in spaceborne SAR imagery is of great importance. Focusing on practical problems of ship detection, this paper presents a highly adaptive ship detection scheme for spaceborne SAR imagery. It is able to process a wide range of sensors, imaging modes and resolutions. Two main stages are identified in this paper, namely: ship candidate detection and ship discrimination. Firstly, this paper proposes an adaptive land masking method using ship size and pixel size. Secondly, taking into account the imaging mode, incidence angle, and polarization channel of SAR imagery, it implements adaptive ship candidate detection in spaceborne SAR imagery by applying different strategies to different resolution SAR images. Finally, aiming at different types of typical false alarms, this paper proposes a comprehensive ship discrimination method in spaceborne SAR imagery based on confidence level and complexity analysis. Experimental results based on RADARSAT-1, RADARSAT-2, TerraSAR-X, RS-1, and RS-3 images demonstrate that the adaptive scheme proposed in this paper is able to detect ship targets in a fast, efficient and robust way.

The simple linear iterative clustering (SLIC) method is a recently proposed popular superpixel algorithm. However, this method may generate bad superpixels for synthetic aperture radar (SAR) images due to effects of speckle and the large dynamic range of pixel intensity. In this paper, an improved SLIC algorithm for SAR images is proposed. This algorithm exploits the likelihood information of SAR image pixel clusters. Specifically, a local clustering scheme combining intensity similarity with spatial proximity is proposed. Additionally, for post-processing, a local edge-evolving scheme that combines spatial context and likelihood information is introduced as an alternative to the connected components algorithm. To estimate the likelihood information of SAR image clusters, we incorporated a generalized gamma distribution (GГD). Finally, the superiority of the proposed algorithm was validated using both simulated and real-world SAR images.