Double Branch Research Articles

LIANet: Layer Interactive Attention Network for RGB-D Salient Object Detection

RGB-D salient object detection (SOD) usually describes two modes’ classification or regression problem, namely RGB and depth. The existing RGB-D SOD methods use depth hints to increase the detection performance, meanwhile they focus on the quality of little depth maps. In practical application, the interference of various problems in the acquisition process affects the depth map quality, which dramatically reduces the detection effect. In this paper, to minimize interference in depth mapping and emphasize prominent objects in RGB images, we put forward a layered interactive attention network (LIANet). The whole network model adopts the idea of double branch structure to integrate RGB information and Depth information. The network has three parts: feature coding, layered fusion mechanism, and feature decoding. In the feature coding stage, a simple attention module (SAM) is added, which defines the energy function considering the weights of channel and space dimensions. This module enables the network to learn more discriminating neurons without adding parameters. By refining these neurons, the high-level semantic features of images can be fully mined. The layered interactive fusion module (LIFM) is the most critical part of this paper. This module effectively enhances the cross-modal interaction between RGB features and depth features. The RGB-depth-RGB modulation feedback mechanism successfully eliminates interference in the depth map and accurately highlights the features of salient objects. In addition, we also used mixed losses to optimize further and train our model. Finally, a mass of experiments on six standard datasets demonstrated the importance of the method, and a timely detection speed reaches 30 fps on every dataset.

CDANet: Contextual Detail-Aware Network for High-Spatial-Resolution Remote-Sensing Imagery Shadow Detection

Shadow detection automatically marks shadow pixels in high-spatial-resolution (HSR) imagery with specific categories based on meaningful colorific features. Accurate shadow mapping is crucial in interpreting images and recovering radiometric information. Recent studies have demonstrated the superiority of deep learning in very-high-resolution satellite imagery shadow detection. Previous methods usually overlap convolutional layers but cause the loss of spatial information. In addition, the scale and shape of shadows vary, and the small and irregular shadows are challenging to detect. In addition, the unbalanced distribution of the foreground and the background causes the common binary cross-entropy loss function to be biased, which seriously affects model training. A contextual detail-aware network (CDANet), a novel framework for extracting accurate and complete shadows, is proposed for shadow detection to remedy these issues. In CDANet, a double branch module is embedded in the encoder–decoder structure to effectively alleviate low-level local information loss during convolution. The contextual semantic fusion connection with the residual dilation module is proposed to provide multiscale contextual information of diverse shadows. A hybrid loss function is designed to retain the detailed information of the tiny shadows, which per-pixel calculates the distribution of shadows and improves the robustness of the model. The performance of the proposed method is validated on two distinct shadow detection datasets, and the proposed CDANet reveals higher portability and robustness than other methods.

Digital Implementation of Single Current Sensor-Based IPMSM Drive With Sampling Error Delay Compensation

Single current sensor-based three-phase drives are gaining huge interest from industries for sensitivity mismatch issues and cost reduction. However, the current sampling error caused by multiple sampling points during current reconstruction from various single current sensor (SCS) topologies led to increase in speed fluctuation and torque ripple. In this paper, a computationally efficient and accurate model-based sampling error compensation scheme based on a new SCS with double branch current sampling technique is proposed for an interior permanent magnet synchronous motor (IPMSM) drive. In order to improve of current reconstruction precision, the compensation strategy includes rotor angle compensation of IPMSM and dead-time effect mitigation for reduction of non-ideal behavior of inverter, which does not require major computation. Also, dc-link voltage utilization is improved significantly without any external dc-link control. The efficacy of the proposed compensation strategy is verified by an experimental setup consisting of 1.5 kW IPMSM drive, which shows the effectiveness of proposed current reconstruction technique from SCS by reducing current sampling error with minimum computational burden.

Hyperspectral Image Classification Based on Fast Double Branch Dense Connection Network and Double Attention Mechanism

Hyperspectral Image Classification Based on Fast Double Branch Dense Connection Network and Double Attention Mechanism

DBMF: A Novel Method for Tree Species Fusion Classification Based on Multi-Source Images

Multi-source data remote sensing provides innovative technical support for tree species recognition. Tree species recognition is relatively poor despite noteworthy advancements in image fusion methods because the features from multi-source data for each pixel in the same region cannot be deeply exploited. In the present paper, a novel deep learning approach for hyperspectral imagery is proposed to improve accuracy for the classification of tree species. The proposed method, named the double branch multi-source fusion (DBMF) method, could more deeply determine the relationship between multi-source data and provide more effective information. The DBMF method does this by fusing spectral features extracted from a hyperspectral image (HSI) captured by the HJ-1A satellite and spatial features extracted from a multispectral image (MSI) captured by the Sentinel-2 satellite. The network has two branches in the spatial branch to avoid the risk of information loss, of which, sandglass blocks are embedded into a convolutional neural network (CNN) to extract the corresponding spatial neighborhood features from the MSI. Simultaneously, to make the useful spectral feature transfer more effective in the spectral branch, we employed bidirectional long short-term memory (Bi-LSTM) with a triple attention mechanism to extract the spectral features of each pixel in the HSI with low resolution. The feature information is fused to classify the tree species after the addition of a fusion activation function, which could allow the network to obtain more interactive information. Finally, the fusion strategy allows for the prediction of the full classification map of three study areas. Experimental results on a multi-source dataset show that DBMF has a significant advantage over other state-of-the-art frameworks.

Interfacial properties of nonionic hybrid surfactant with double branches of fluorocarbon and hydrocarbon

To obtain superior surface-active properties while minimizing environmental risks, new fluorinated surfactants (F2HX-nEO) were designed to include a hybrid multi-chain with a short fluorocarbon (–CF2CF3) and nonionic ethylene oxide. They were simply synthesized by a straightforward reaction involving the addition of ethylene oxide to a hybrid fluorine alcohol, and their surface-active properties were optimized by controlling the hydrocarbon length and number of ethylene oxide units. Therefore, the desirable values of critical micelle concentration (0.3 mmol/L) and surface tension (24.3 mN/m) were obtained in spite of containing the short fluorocarbon chain. Particularly, F2HX–nEO showed excellent wetting properties on hydrophobic surfaces and very low toxicity level, which can be used in various fields as a promising alternative to long fluoroalkyl chain surfactants.

Aspergillus fumigatus Fumagillin Contributes to Host Cell Damage

The activity of fumagillin, a mycotoxin produced by Aspergillus fumigatus, has not been studied in depth. In this study, we used a commercial fumagillin on cultures of two cell types (A549 pneumocytes and RAW 264.7 macrophages). This toxin joins its target, MetAP2 protein, inside cells and, as a result, significantly reduces the electron chain activity, the migration, and the proliferation ability on the A549 cells, or affects the viability and proliferation ability of the RAW 264.7 macrophages. However, the toxin stimulates the germination and double branch hypha production of fungal cultures, pointing out an intrinsic resistant mechanism to fumagillin of fungal strains. In this study, we also used a fumagillin non-producer A. fumigatus strain (∆fmaA) as well as its complemented strain (∆fmaA::fmaA) and we tested the fumagillin secretion of the fungal strains using an Ultra High-Performance Liquid Chromatography (UHPLC) method. Furthermore, fumagillin seems to protect the fungus against phagocytosis in vitro, and during in vivo studies using infection of immunosuppressed mice, a lower fungal burden in the lungs of mice infected with the ∆fmaA mutant was demonstrated.

Variations in the Number and Drainage Levels of Retroaortic Left Renal Veins

Abstract Aim: Circumaortic left renal vein (CLRV) and retroaortic left renal vein (RLRV) are the most common variations of the left renal vein and have a retroaortic course. We aimed to detect the anatomical variations of retroaortic left renal veins in terms of their number and entrance levels into the inferior vena cava (IVC). Materials and Methods: The computed tomographic (CT) images of 2804 patients who had undergone abdominal CT examinations in our hospital between May 1, 2017, and June 30, 2017, were evaluated. Results: CLRV variations were detected in 139 (4.96%) patients, while in another 132 (4.71%) RLRV variations were detected. Double branches were present in 15 (10.8%) patients in the RLRV group, and 3 patients (2.3%) in the CLRV group. The most common drainage levels into the IVC for these variations were the L3 vertebra level and the middle 1/3 infrarenal IVC segment. Conclusion: CLRV and RLRV may possess single or double retroaortic branches. The entrance levels into the IVC are various in these variations. Knowing these variations is important for surgical and interventional procedures.

Left Anterior Fascicular Block After Transcatheter Closure of Ventricular Septal Defect in Children.

Background: Arrhythmia is the most common complication after transcatheter closure of a ventricular septal defect (VSD). However, the effects of postprocedural left anterior fascicular block are not clear. This study presents the clinical characteristics, prognosis, and related risk factors of left anterior fascicular block after transcatheter closure of a VSD in children.Methods: The clinical and follow-up data of the patients in the Heart Center of Children's Hospital of Chongqing Medical University from June 2009 to October 2018 were reviewed. And 30 cases were eligible out of all 1,371 cases.Results: An electrocardiogram showed a left anterior fascicular block within 3 days, and most patients gradually returned to normal within 1–2 years, showing a dynamic change. Left ventricular end-diastolic dimension Z-score ranged from −2 to 2 in all children, and no decrease of left ventricular ejection fraction was found in all children. The high ratio between VSD size and body surface area [p < 0.05, odds ratio (OR) 2.6, 95% CI: 1.136–6.113] and large diameter difference between the occluder size and VSD size (p < 0.05, OR 2.1, 95% CI: 1.036–4.609) were independent risk factors for postprocedural left anterior fascicular block.Conclusions: The incidence of postprocedural left anterior fascicular block is not that low, and the overall prognosis is quite good at the current follow-up stage. No progressive severity has been found, such as complete left bundle branch block, double (triple) bundle branch block, and atrioventricular block, to have an influence on cardiac systolic and diastolic function.

Bilateral Anterolateral Thigh Myocutaneous Flaps for Giant Complex Chest Wall Reconstruction.

Extensive reconstruction of complex full-thickness chest wall oncological defects is challenging. Bilateral free anterolateral thigh (ALT) myocutaneous flap transfer for the complex reconstruction of a large area of the chest wall is discussed. We reported a single unit's experience in 1-staged multilayered reconstruction of large full-thickness chest wall defects in 22 patients (16 primary chest wall tumor cases, 5 locally advanced breast cancer cases, and 1 osteoradionecrosis case) treated between 2011 and 2018. Bilateral ALT myocutaneous flaps together with traditional cement implant or unmovable/movable joint conformable titanium struts were used for chest wall reconstruction. The anatomical characteristics of pedicle origin and pattern of the venae comitantes of the ALT myocutaneous flap, recipient vessels, and anastomosis patterns were described. Bilateral ALT myocutaneous flaps were used for soft tissue reconstruction in 22 cases. Different methods of flap harvesting and vascular anastomosis were selected as needed. No vein grafts or arteriovenous loops were required. We observed 3 vascular patterns of the flap pedicle, including 1 oblique branch and descending branch (59.1%, n = 26), 2 single descending branch (9.1%, n = 4), and 3 double branches of the descending branch (31.8%, n = 14). The flap was harvested pedicled with solely the oblique branch in 7 (15.9%) cases, solely the descending branch in 28 (63.6%) cases to minimize the donor site morbidity, and pedicled with the oblique and descending branch in 9 (20.5%) cases to achieve multiple vascular anastomosis choices. Stable skeletal reconstructions were achieved using traditional cement implant (13.6%, n = 3) or conformable titanium struts (86.4%, n = 19), with good fixation strength. Complication risk was low. An algorithmic approach to management is presented and recommended. Various forms of bilateral ALT myocutaneous flap transfer with different skeletal reconstruction presents as a reliable treatment for patients with large full-thickness chest wall defects. Anatomical variations in the pedicle and pattern of venae comitantes of the ALT myocutaneous flap are reported. In some challenging cases, finding the vessels in the recipient area is difficult. The clinical significance of each vascular pattern is delineated, and surgical technical considerations are discussed on the basis of the recipient area requirements and types of a flap's vascular anatomy.

Complete sub-wavelength flexural wave band gaps in plates with periodic acoustic black holes

Acoustic Black Hole (ABH) effect shows promise for vibration control, but mainly limited to a relatively high frequency range. Though achievable in 1D periodic ABH structures, complete sub-wavelength band gaps (BGs) have not yet been realized in 2D configuration. Capitalizing on the unique wave propagation characteristics of the ABH, we propose a new type of plates containing periodically arranged double-layer ABH cells which offer complete and omnidirectional BGs. The phenomena originate from the combined effects of the ABH-specific local resonances and Bragg scattering, which are made possible through a dual process: a proper channeling of the wave propagation path and an impaired coupling between the ABH-induced local resonances and the global vibration of the unit cells. The former is warranted by a proper structural tailoring of the unit cells and the latter by the dynamics of the double-layer ABH design. It is shown that the BGs can be tuned through adjusting ABH parameters. Meanwhile, attaching the centers of the double ABH branches with a connecting cylinder can further broaden and lower the frequencies of the BGs as a result of the enhanced Bragg scattering. It is also demonstrated numerically and experimentally that remarkable vibration attenuation and energy insulation can be achieved in a plate with only a small number of ABH cells, thus pointing at the possibility of achieving sub-wavelength vibration control in structures with reasonable dimensions.

Compact Global-Local Convolutional Network With Multifeature Fusion and Learning for Scene Classification in Synthetic Aperture Radar Imagery

Feature learning of convolutional neural networks (CNNs) has gained considerable attention and achieved good performance on synthetic aperture radar (SAR) image scene classification. However, the performance of the existing convolutional feature learning methods is limited for generating the distinguishable feature representations because such techniques inherently suffer from shortcomings, i.e., they do not consider the local feature distribution of deep orderless feature statistics and deep orderless multifeature learning style. To alleviate these drawbacks, we propose a compact global-local convolutional network with multifeature fusion and learning (CGML) for SAR image scene classification, which contains double branches of convolutional feature learning net (C-net) and local feature distribution learning net (L-net). L-net employs the localized and parameterized affine subspace coding layer for local feature distribution learning and captures the feature statistics of each cluster center via detailed local feature division. The standard convolutional feature map is utilized for the convolutional feature learning in C-net. Subsequently, the compact multifeature fusion and learning strategy captures the compact global second-order orderless feature representation and allows the double branches to interact with each other via the tensor sketch algorithm. Especially, the feature learning strategy of L-net is defined in affine subspace which fully characterizes the feature distribution inside each cluster space. Finally, we concatenate the outputs of the multifeature fusion and learning network, then pool and feed them into softmax loss. Based on extensive evaluations on TerraSAR-X1 and TerraSAR-X2 image scene classification datasets, CGML can yield superior performances when compared with those of several state-of-the-art networks.

Vertical U-shaped heat exchangers for construction engineering: New promises and future challenges

The ground source heat pump system is a promising technology for building heating and cooling by using underground energy, however, many technical problems, e. g. low energy utilization rate and inefficiency operation, have greatly hindered its wide applications. This paper studies the key technologies of vertical U-shaped heat exchangers in construction engineering to solve the bottleneck problem and optimizes the system. Three commonly used backfill methods are analyzed, and a double U-shaped branch pipe is recommended to improve construction efficiency. The paper sheds a bright light on green energy utilization with high-efficiency, energy-saving, and environmental protection.

LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification

Deep learning (DL) algorithms have been demonstrated to have great potential in hyperspectral image (HSI) classification. However, most DL methods mainly focus on improving classification performance, neglecting the computational cost. In order to broaden the application scenarios of DL-based HSI classification methods, it is necessary to develop lightweight and fast models to fit the deployment needs of computationalresource-limited platforms. Taking into account the efficency and accuracy, this paper designs a lightweight network architecture based on spectral and channel-wise attention modules, namely LiteSCANet, for HSI classification. It contains a residual double branch structure, which makes the model effectively extract spectral-spatial features and achieve good performance with fast inference speed and low computational consumption (i.e., floating-point operations, number of parameters and graphics processing unit memory usage). The experiment results on four benchmark data sets show that our proposed model achieves an excellent trade-off between efficiency and accuracy compared with the other six existing networks.

HResNetAM: Hierarchical Residual Network With Attention Mechanism for Hyperspectral Image Classification

This article proposes a novel hierarchical residual network with attention mechanism (HResNetAM) for hyperspectral image (HSI) spectral-spatial classification to improve the performance of conventional deep learning networks. The straightforward convolutional neural network-based models have limitations in exploiting the multiscale spatial and spectral features, and this is the key factor in dealing with the high-dimensional nonlinear characteristics present in HSIs. The proposed hierarchical residual network can extract multiscale spatial and spectral features at a granular level, so the receptive fields range of this network will be increased, which can enhance the feature representation ability of the model. Besides, we utilize the attention mechanism to set adaptive weights for spatial and spectral features of different scales, and this can further improve the discriminative ability of extracted features. Furthermore, the double branch structure is also exploited to extract spectral and spatial features with corresponding convolution kernels in parallel, and the extracted spatial and spectral features of multiple scales are fused for hyperspectral image classification. Four benchmark hyperspectral datasets collected by different sensors and at different acquisition time are employed for classification experiments, and comparative results reveal that the proposed method has competitive advantages in terms of classification performance when compared with other state-of-the-art deep learning models.

双分支残差去马赛克网络

双分支残差去马赛克网络

A cadaveric study of coronary arteriesin Rayalaseema population of Andhra Pradesh

Introduction: Majority of the sudden cardiac deaths were due to anomalous origin and distribution of the coronary arteries. Aim: To study the arterial vascular pattern of the human heart by using dissection method in Rayalaseema population. Materials and Methods: A total of 80 human fresh heart specimens were obtained from the Department of Forensic Medicine, S.V. Medical College, Tirupati for the present study. Results: We have observed both the coronary arteries arising from single coronary ostia in 76 specimens whereas 4 heart specimens noted with coronary arteries originated from two ostia. Right and left dominance circulation was noted more in males than females. We have also observed double posterior interventricular branches arising from the right and left coronary arteries showing the balanced type of coronary circulation in 2 (Male-1; Female-1) hearts. Conclusion: Arterial pattern and its variations are important to prevent false interpretation of the arterial angiograms in management of coronary artery diseases. Keywords: Arteries, Branches, Coronary, Dominance.

Building Double Branch Dyes with Acetylenic Bonds: An Effective Strategy to Enhance the Efficiency of Dye‐Sensitized Solar Cells

Double branch (DB) dyes could reduce intermolecular interactions and realize good photovoltaic performances. Herein, symmetric DB dyes (FSDD‐1 and FSDD‐2) featuring with one and two acetylenic bonds as the isolation groups are reported. The corresponding single branch (SB) dye FSSD‐EH is synthesized for comparison. For the adsorption spectra, bathochromic shifts and higher ε are observed for DB dyes FSDD‐1 and FSDD‐2 compared with SB dye FSSD‐EH, which suggests better light‐harvesting capabilities for the DB dyes. Dye‐sensitized solar cells (DSSCs) sensitized by DB dyes FSDD‐1 and FSDD‐2 exhibit significant increases for the Jsc and Voc compared with SB dye FSSD‐EH, so much better η are achieved by the DB dyes (5.34, 7.14, and 2.46% for FSDD‐1, FSDD‐2, and FSSD‐EH, respectively). FSDD‐2 with a diyne unit exhibits higher η compared with FSDD‐1 with an yne unit as the isolation groups. The research shows that a much better photovoltaic performance of DSSCs can be achieved after connecting two dye molecules to be a DB dye by adopting acetylenic bonds as the isolation group. A longer distance between the two branches realized by a diyne unit is beneficial and effective to achieve better photovoltaic performance.

Novel glucoamylase-resistant gluco-oligosaccharides with adjacent α-1, 6 branches at the non-reducing end discovered in Japanese rice wine, sake

Sake, a traditional Japanese rice wine, contains various oligosaccharides (Sake oligosaccharides; SAOs) derived from rice starch. We previously found that SAOs reach a high degree of polymerization (DP). In this study, we developed a hydrophilic interaction liquid chromatography-time-of-flight/mass spectrometry (HILIC-TOF/MS) based analytical method to separate isomeric SAOs. Isomers of SAOs with DP = 6, 7, and 8, which were named DP6-1, DP7-1, DP8-1 and DP8-2, respectively, were purified from sake and their structures were determined by two-dimensional NMR spectroscopy. These were novel oligosaccharides containing two α-1, 6 bonded branches on an α-1, 4-linked glucose main chain. Interestingly, adjacent double α-1, 6 branches that have not been identified in starch, were found in DP6-1, DP7-1, and DP8-1, suggesting the presence of the branching pattern in starch. DP6-1 was poorly digested by fungal glucoamylase, and this may be attributed to its adjacent double branches at the non-reducing end.

Ensemble model with cascade attention mechanism for high-resolution remote sensing image scene classification.

Scene classification of high-resolution remote sensing images is a fundamental task of earth observation. And numerous methods have been proposed to achieve this. However, these models are inadequate as the number of labelled training data limits them. Most of the existing methods entirely rely on global information, while regions with class-specific ground objects determine the categories of high-resolution remote sensing images. An ensemble model with a cascade attention mechanism, which consists of two kinds of the convolutional neural network, is proposed to address these issues. To improve the generality of the feature extractor, each branch is trained on different large datasets to enrich the prior knowledge. Moreover, to force the model to focus on the most class-specific region in each high-resolution remote sensing image, a cascade attention mechanism is proposed to combine the branches and capture the most discriminative information. By experiments on four benchmark datasets, OPTIMAL-31, UC Merced Land-Use Dataset, Aerial Image Dataset and NWPU-RESISC45, the proposed end-to-end model cascade attention-based double branches model in this paper achieves state-of-the-art performance on each benchmark dataset.