Branching Characteristics Research Articles

PhNAL1 Is Involved in Regulating Branch Development of Petunia

The branch number of plants is an important agronomic trait that directly influences the ornamental characters and production costs of ornamental plants. Shoot branching has always been a hot topic for Petunia hybrida. During our research, we isolated the homologous gene of narrow-leaf 1 (NAL1), denoted as PhNAL1. The expression level of PhNAL1 was higher in leaves and axils than in roots, stems, and flowers. Pertinent to shoot apex removal and 6-benzyladenine treatments, both interventions demonstrated a suppressive effect on the expression of PhNAL1. Through subcellular localization analysis, we found that PhNAL1 predominantly localized in the nucleus. By using RNA interference targeting PhNAL1, we induced a noticeable increase in branch number while concurrently reducing plant height of petunia. These findings demonstrate that PhNAL1 is involved in regulating branch development within petunia. This study provides genetic resources for the subsequent cultivation of new cultivars of petunia endowed with distinct branching characteristics.

Differences in photosynthetic characteristics of reproductive branches at different leaf positions of Rosa roxburghii on the Yunnan Guizhou Plateau

Differences in photosynthetic characteristics of reproductive branches at different leaf positions of Rosa roxburghii on the Yunnan Guizhou Plateau

Modelling branch surface area of Picea abies [L.] Karst

ABSTRACT Allometric models based on limited sub-samples are widely used for predicting forest-scale information. Here, we develop allometric models for the branch surface area of the widespread conifer species Picea abies [L.] Karst. Branch surface area is a proxy for the capacity of tree branches to intercept and store water and air pollutants. Based on “probability proportional to size” sampling, we measured the surface area for 285 branches and then calculated the branch surface area of 30 trees (and their 3298 branches). We developed allometric models to estimate the total surface area of branches, as well as their number and diameters, for trees across a range of diameters (DBH), heights, and crown ratios (CR). We show that DBH and CR play significant roles in branch characteristics. The branch surface area was linearly related to the stand basal area. Reducing stand density will proportionally reduce interception capacity. The approach outlined here may help stimulate further studies (more species, regions, and management practices) required to optimize stand density for ecosystem services related to crown characteristics, such as hydrology, forage quality, and quantity or capacity for air pollutants.

Multi-scale biomass estimation of an alpine desert shrub based on relative cover: A case study of Asterothamnus centraliasiaticus in the Qaidam Desert on the Qinghai–Tibet Plateau

Biomass estimates and the allocation characteristics of shrubs are essential indicators for studying the structure, function, and material and energy flows of alpine desert ecosystems. However, conventional destructive biomass estimation methods are unsuitable for large-scale application in alpine desert areas that are characterized by sparse vegetation and comprise extremely fragile ecosystems. In this study, we propose a multi-scale method for estimating the biomass of alpine desert shrubs based on relative cover and verified its accuracy, feasibility, and efficiency using data obtained from 100 individual plants and 27 quadrat plots of Asterothamnus centraliasiaticus, an important species growing in the Qaidam Desert on the Qinghai–Tibet Plateau. Having gained this information, we developed models to estimate the new branch (NBB), aboveground (AGB), and belowground (BGB) biomasses of A. centraliasiaticus at individual, quadrat, and regional scales. The modelling results revealed that an allometric relationship model (power law function) could be used to accurately represent the synergistic relationship between plant morphological variables and biomass increases at all assessed scales. Moreover, we established that at the individual scale, estimation models (NBB = 0.3015 × C1.0610 and AGB = 1.0046 × C1.4239) incorporating canopy area (C) as a parameter can better reflect the cumulative characteristics of the new branch and aboveground biomasses of A. centraliasiaticus, respectively, whereas canopy volume (CH, resembles a cylinder) is the most appropriate parameter for the model (BGB = 0.3262 × CH0.8822) used to estimate belowground biomass. In addition, at the sample scale, the total canopy cover (TC) of A. centraliasiaticus in the sample was established to be an appropriate parameter for estimating total biomass (TNBB = 0.0042 × TC1.4647, TAGB = 0.0033 × TC1.6308, and TBGB = 0.0110 × TC1.2554). Furthermore, at the regional scale, we found that incorporating relative canopy cover (RC) in models was effective for estimating the biomass density of A. centraliasiaticus (DNBB = 0.0012 × RC1.4647, DAGB = 0.0014 × RC1.6308, and DBGB = 0.0020 × RC1.2554) and can be combined with regional area estimates to determine the total biomass of A. centraliasiaticus at any spatial scale. On the basis of our findings, we established that the biomass estimation model for alpine desert shrubs, incorporating the parameter relative cover, can effectively solve the longstanding problem of scale conversion and facilitates the rapid and accurate estimation of the biomass components of alpine desert shrubs. In future studies, biomass could be efficiently estimated by combining models based on the relative cover data of shrub communities in relevant areas obtained using aerial photography and remote sensing. The results of this study have important implications for estimating the biomass of alpine desert shrubs, and will accordingly make a valuable contribution to the management of alpine desert ecosystems.

Bilateral network with rich semantic extractor for real-time semantic segmentation

Recently, owing to the requirements of inference speed, most real-time semantic segmentation networks often have shallow network depth, which limits the receptive field size of the model, leading to the limited acquisition of semantic information and resulting in intraclass inconsistency and ultimately a decrease in segmentation accuracy. Additionally, the shallow network depth also restricts the feature extraction capability of the network, reducing its robustness and ability to adapt to complex scenes. To address these issues, a bilateral network with a rich semantic extractor (RSE) for real-time semantic segmentation (BRSeNet) is presented to perform real-time semantic segmentation. First, to solve the problem of insufficient semantic feature information extraction, an RSE is proposed, which includes a multiscale global semantic extraction module (MGSEM) and a semantic fusion module (SFM). The MGSEM can extract rich global semantics and expand the effective receptive field. Simultaneously, the SFM efficiently integrates multiscale local semantics with multiscale global semantics, resulting in more comprehensive semantic information for the network. Finally, based on the characteristics of detail and semantic branches, a bilateral reconstruction aggregation module is designed to reconstruct the contextual information of detail features, model the interdependencies on semantic feature channels, and enhance feature representation. Comprehensive experiments on the challenging Cityscapes and ADE20K datasets are conducted. The experimental results show that the proposed BRSeNet achieves mean intersection over union of 74.9% and 35.7% at inference speeds of 74 and 65 frames per second, respectively, and ensures a favorable balance between segmentation accuracy and inference speed.

Mixed Convection Flow Analysis of Carreau Fluid Over a Vertical Stretching/Shrinking Sheet

The current investigation aspires to unravel the steady mixed convection flow of Carreau fluid over a permeable vertical stretching/shrinking sheet near a stagnation point. The system of governing equations is reduced into ODEs utilizing appropriate similarity transformations. The similarity transformations are obtained via the Lie scaling group of transformations. Dual similarity solutions are detected depending on the opposing flow parameter for stretching and shrinking cases. The effects of pertinent parameters on the skin friction coefficient, Nusselt number, velocity, and temperature fields are examined in detail. The influence of the suction parameter on the variations of skin friction coefficient for the stretching case shows various behavior than in the shrinking case. However, on the variations of the Nusselt number, a similar trend in both the stretching and shrinking cases is observed. The fluid velocity decreases, and the temperature rises with the increment of non-Newtonian parameter in the upper branch, whereas the lower branch depicts opposite trends. Due to the different characteristics of the lower branch than the upper branch, it is necessary to find a physically reliable solution branch. Thus, a linear temporal stability analysis is conducted based on the sign of the smallest eigenvalue. The smallest eigenvalues are determined numerically using the shooting technique, revealing that the upper branch is the only stable solution branch.

Experimental Study and 3-D Modeling of Positive Streamer Discharges Under the Effect of Humidity

The research on streamer discharges not only provides support for insulation design of high-voltage power transmission systems, but also lays a theoretical foundation for lightning discharges. The macroscopic parameters of streamer discharges have been investigated, but there is a lack of analysis on the development and branching characteristics of streamers under the effect of humidity. Besides, different types of models have been built to simulate streamer propagation. However, there are few 3D models which simulate branching process of multiple streamers at different humidity. In this work, a discharge experimental platform with controllable environmental parameters has been built. The effect of the humidity of 4.5~19.1 g/m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sup> on streamer propagation and branching has been experimentally studied. Then the 3D “discharge tree” model of collective streamers is improved, in which the streamer branching incorporates the photoionization in humid air. Furthermore, we compare the results of the 3D simulation model with the experimental data. Through the simulation model, the streamer parameters such as line charge density and electric field distribution are obtained. The results show that the electric field of streamer channels are larger when the humidity is higher, but the propagation velocities are reduced. The humidity has limited effect on the streamer branching ratio.

DETisSeg: A dual-encoder network for tissue semantic segmentation of histopathology image

Accurate segmentation of tissue regions in hematoxylin and eosin (H&E)-stained histopathology images is an essential step in computational pathology. It is still a challenging task due to the large-scale distribution, irregular morphological variations, and fuzzy boundaries between different tissues. Semantic segmentation methods with strong feature representative capability for extracting local fine-grained and global coarse-grained features are desired to tackle this challenge. In this paper, we propose a dual encoder network for tissue semantic segmentation of histopathology image, called DETisSeg, which aims to fully fuse the features of global context information by the Swin Transformer branch and local features by the convolutional neural network(CNN) branch. Particularly, we design an improved residual connection to recover the lost spatial information during the patch merging phase in the Swin Transformer branch. In addition, we employ a pyramid architecture decoder to generate a composite feature map with multiple scales. We perform experiments on two publicly available tissue semantic segmentation datasets, including the BCSS and LAUD-HistoSeg datasets, to evaluate the effectiveness of the proposed DETisSeg. Compared with eight state-of-the-art semantic segmentation methods, the results show that the proposed DETisSeg achieves improved performance. Particularly, it outperforms Swin transformer on the mean Dice and IoU with improvements of 1.36% and 1.79% on BCSS dataset, 0.96% and 4.48% on LUAD-HistoSeg dataset, respectively.

SOME OPTIONS OF FETAL TOPOGRAPHY OF THE LONG BRANCHES OF THE CROSS PLEXUS

In the conditions of wartime, the frequency of damage to the nerves of the lower extremity increased sharply. Open injuries make up the vast majority of peripheral nerve injuries. Understanding variations in the topography and the condition of the surrounding soft tissue structures allows for a detailed assessment of the condition of the peripheral nerves of the lower limb, to diagnose the exact site of damage, the type of neuropathy, and to choose treatment and rehabilitation tactics. The purpose of the study will be to establish variants of the fetal topography of the long branches of the sacral plexus. The material for the study of topographical and anatomical features of the long branches of the sacral plexus was 26 human fetuses of 5-7 months. The sciatic nerve is the largest nerve in humans, originating from the ventral division of spinal nerves L4 to S3 and containing fi bers from both the posterior and anterior lumbosacral plexuses. The sciatic nerve runs along the entire back surface of the lower limb up to the foot. The sciatic nerve innervates a signifi cant part of the skin and muscles of the thigh, lower leg, and foot. It was established that the long branches of the sacral plexus can be subjected to compression or damage in diff erent parts of its anatomical course. Among the variants of the anatomical variability of the long branches of the sacral plexus, the topography of the sciatic nerve is the most common. In the three studied fetuses of 180.0 mm, 215.0 mm, 255 mm TKD, a high branching variant of the sciatic nerve was detected, and the features of the long branches of the sacral plexus in the specifi ed fetuses were unilateral. In this case, when assessing clinical symptoms, practicing doctors should take into account options for bilateral asymmetry. Prospects for further research. Deepening the study of the variant anatomy of the long branches of the sacral plexus in human fetuses of diff erent ages.

Evaluation of facial nerve perineural spread from cutaneous squamous cell carcinoma using 3T MR neurography.

Perineural spread (PNS) is a rare but potentially fatal consequence of cutaneous squamous cell carcinoma (cSCC) of the head and neck. We aimed to evaluate the accuracy of 3T MR neurography in detecting and defining the extent of facial nerve (VII) PNS from cSCC, and highlight characteristic radiological features in peripheral branches to improve early diagnosis. Single-institution retrospective review of 38 patients with clinical, radiological, and/or histopathological findings consistent with VII PNS from cSCC who underwent pre-operative 3T MR neurography. Compared to histopathology (gold standard), 3T MR neurography had a sensitivity of 89% and positive predictive value of 97%. In true-positive cases (n = 33), zonal extent was correctly identified in 100%. Seventy-nine% had simultaneous trigeminal nerve (V) PNS, mostly involving the auriculotemporal branch of the mandibular nerve (64%). When the causative lesion was absent (n = 23), the extra-temporal VII demonstrated asymmetrical enhancement alone (n = 6), bulky expansion (n = 8), or extra-neural spread (n = 9). Peripheral VII branch involvement, particularly the buccal and zygomatic, was readily identified using known anatomical landmarks. 3T MR neurography is highly accurate in evaluating VII PNS from cSCC, and thus should be specifically requested by physicians if suspicious for disease. Coexistent V PNS was common, highlighting the need to examine V branches to allow complete treatment planning. The unique radiological patterns identified showcases disease progression. As early detection improves patient outcomes, the radiologist must look for peripheral VII involvement inspecific anatomical areas, which is within the capabilities of 3T MR neurography.

Theoretical Insights into Midchain Radicals and Branching Characteristics in Solution Copolymerization of Ethylene and Vinyl Acetate.

The kinetics of intermolecular chain transfer to polymer (CTP) and chain transfer to monomer (CTM), as well as backbiting in solution copolymerization of ethylene and vinyl acetate (VAc), are calculated by density functional theory. An intrinsic rate coefficient for each reaction type, whose Arrhenius parameters are expressed as functions of unreacted monomer composition, is extracted from calculated elementary reactions, and hence it applies to a wide range of fragment compositions of EVA. Backbiting controls the generation of midchain radicals (MCRs) on the backbone, and its rate maximizes at a medium fraction of VAc in the unreacted monomers. CTP plays an insignificant role in MCR generation even at high conversion rates due to its low pre-exponential factor. The concentration of MCRs is quantified and is close to that of ECRs at high conversions and high fractions of VAc in monomers. Additionally, branching characteristics are explored; concentrations of short-chain branching (SCB) and long-chain branching (LCB) are about 0.7-2.5 and 0.1-0.4%, respectively, and drop with VAc fraction rapidly. The role of the migration of MCRs is highlighted, which transforms about 10% of SCB points into LCB points. Disagreeing with insights from laboratory experiments, it is the migration, rather than CTP, which increases the LCB concentration at high conversions.

Segmentation of carotid artery vessel wall and diagnosis of carotid atherosclerosis on black blood magnetic resonance imaging with multi-task learning.

Atherosclerotic cardiovascular disease is the leading cause of death worldwide. Early detection of carotid atherosclerosis can prevent the progression of cardiovascular disease. Many (semi-) automatic methods have been designed for the segmentation of carotid vessel wall and the diagnosis of carotid atherosclerosis (i.e., the lumen segmentation, the outer wall segmentation, and the carotid atherosclerosis diagnosis) on black blood magnetic resonance imaging (BB-MRI). However, most of these methods ignore the intrinsic correlation among different tasks on BB-MRI, leading to limited performance. Thus, we model the intrinsic correlation among the lumen segmentation, the outer wall segmentation, and the carotid atherosclerosis diagnosis tasks on BB-MRI by using the multi-task learning technique and propose a gated multi-task network (GMT-Net) to perform three related tasks in a neural network (i.e., carotid artery lumen segmentation, outer wall segmentation, and carotid atherosclerosis diagnosis). In the proposed method, the GMT-Net is composed of three modules, including the sharing module, the segmentation module, and the diagnosis module, which interact with each other to achieve better learning performance. At the same time, two new adaptive layers, namely, the gated exchange layer and the gated fusion layer, are presented to exchange and merge branch features. The proposed method is applied to the CAREII dataset (i.e., 1057 scans) for the lumen segmentation, the outer wall segmentation, and the carotid atherosclerosis diagnosis. The proposed method can achieve promising segmentation performances (0.9677 Dice for the lumen and 0.9669 Dice for the outer wall) and better diagnosis accuracy of carotid atherosclerosis (0.9516 AUC and 0.9024 Accuracy) in the "CAREII test" dataset (i.e., 106 scans). The results show that the proposed method has statistically significant accuracy and efficiency. Even without the intervention of reviewers required for the previous works, the proposed method automatically segments the lumen and outer wall together and diagnoses carotid atherosclerosis with high performance. The proposed method can be used in clinical trials to help radiologists get rid of tedious reading tasks, such as screening review to separate normal carotid arteries from atherosclerotic arteries and to outline vessel wall contours.

Military Sexual Trauma and its Association with Mental Health Among Sexual Minority and Heterosexual Veterans in the United States.

Purpose: The purpose of this study was to examine the prevalence of military sexual trauma (MST) by sexual orientation, characteristics of sexual minority (SM) and heterosexual veterans who have experienced MST, and associations between sexual orientation and mental health symptoms among veterans who have experienced MST. Methods: Data were analyzed from a nationally representative web-based survey of 4069 U.S. veterans (4.9% SM), which assessed sociodemographic (e.g., age, sexual orientation) and military (e.g., branch) characteristics, and lifetime and current mental health symptoms. Bivariate analyses compared sociodemographic and military characteristics and mental health symptoms among veterans who have experienced MST by sexual orientation. Multivariable logistic regressions examined associations between sexual orientation and mental health symptoms among veterans who have experienced MST, while controlling for differences in sociodemographic characteristics and non-MST traumatic events. Results: The prevalence of MST (7.5% of weighted sample) was higher among SM veterans compared to heterosexual veterans, with bisexual/pansexual/queer veterans endorsing the highest prevalence (22.7%), followed by gay/lesbian (17.0%) and heterosexual (6.5%) veterans. SM veterans who experienced MST were more likely than heterosexual veterans to screen positive for lifetime and current posttraumatic stress disorder (odds ratio [OR] = 3.06 and 3.38, respectively), and current drug use disorder (OR = 3.53). Conclusions: This study adds to growing evidence that mental health symptoms associated with MST disproportionately impact SM veterans relative to heterosexual veterans. Approaches to reducing barriers and tailoring MST-related care to SM veterans, including through addressing cumulative effects of minority stress, are discussed.

Supplement and Suppression: Both Boundary and Nonboundary are Helpful for Salient Object Detection.

Current methods aggregate multilevel features from the backbone and introduce edge information to get more refined saliency maps. However, little attention is paid to how to suppress the regions with similar saliency appearances in the background. These regions usually exist in the vicinity of salient objects and have high contrast with the background, which is easy to be misclassified as foreground. To solve this problem, we propose a gated feature interaction network (GFINet) to integrate multiple saliency features, which can utilize nonboundary features with background information to suppress pseudosalient objects and simultaneously apply boundary features to supplement edge details. Different from previous methods that only consider the complementarity between saliency and boundary, the proposed network introduces nonboundary features into the decoder to filter the pseudosalient objects. Specifically, GFINet consists of global features aggregation branch (GFAB), boundary and nonboundary features' perception branch (B&NFPB), and gated feature interaction module (GFIM). According to the global features generated by GFAB, boundary and nonboundary features produced by B&NFPB and GFIM employ a gate structure to adaptively optimize the saliency information interchange between abovementioned features and, thus, predict the final saliency maps. Besides, due to the imbalanced distribution between the boundary pixels and nonboundary ones, the binary cross-entropy (BCE) loss is difficult to predict the pixels near the boundary. Therefore, we design a border region aware (BRA) loss to further boost the quality of boundary and nonboundary, which can guide the network to focus more on the hard pixels near the boundary by assigning different weights to different positions. Compared with 12 counterparts, experimental results on five benchmark datasets show that our method has better generalization and improves the state-of-the-art approach by 4.85% averagely in terms of the regional and boundary evaluation measures. In addition, our model is more efficient with an inference speed of 50.3 FPS when processing a 320 x 320 image. Code has been made available at https://github.com/lesonly/GFINet.

Multilevel Feature Fusion for End-to-End Blind Image Quality Assessment

In this paper, a framework based on two feature extraction networks and a multilevel feature fusion (MFF) network is proposed. Multilevel degradation features can be obtained through this method, and combined with the human visual perception system, the local and global feature information contained in these features can be captured, which is conducive to the prediction of distorted images. First, a restored image approximating a reference image is generated by a restorative generative adversarial network (GAN). Furthermore, the multilevel degradation features of distorted images and the restored image features are extracted by EfficientNet. Second, the features extracted by EfficientNet are input into the MFF network and are fully expressed by the top-down, bottom-up and third edge joining methods. Moreover, the features provide more low-level details and high-level semantic features for the prediction of image quality scores. In addition, after the MFF stage, the framework calculates the score of each branch feature and obtains the average quality score. Experimental results show that our method achieves greatly improved prediction accuracy and performance on five standard databases.

Visual camera relocalization using both hand-crafted and learned features

The localization of the camera is essential in AR, MR, and robotics. Diverse pipelines employ a hand-crafted or learning based way to predict the camera pose as per the task. In the localization process, both weaknesses and strengths are maintained. However, few current frameworks consider these two features simultaneously. In this study, a novel relocalization pipeline for RGB or RGB-D input is proposed, including a coarse stage with learned features, further refinement with hand-crafted features, and a stable process to measure the confidence of both stages for improving localization robustness. Instead of directly regressing the camera pose, the coarse procedure uses registration to the known source and predicted weighted target point cloud to obtain the initial result. Therefore, we design a deep network called PGNet to construct the weighted target point cloud with the image and previous poses as inputs. Moreover, in consideration of dynamic surroundings, we add a segmentation branch distinguishing each point as either fixed or dynamic with the purpose of promoting dynamic perception. Correspondingly, the segmentation-extended Chamfer Distance is added to optimize PGNet. During the pose refinement, the feature space is established via hand-crafted feature extraction and matching on the training set. Based on the coarse pose, we obtain the accurate pose by applying Kabsch or Perspective-n-Point (PnP) algorithm to point-to-point correspondences built through searching the space and matching Oriented Fast and Rotated Brief (ORB) features. Furthermore, an additional process is presented by defining coarse and refinement metrics to gain a more stable performance. Finally, experiments on both static and dynamic scenes are conducted. On the one side, the results demonstrate the state-of-the-art performance over other existing methods on 7 Scenes, INDOOR-6, Cambridge Landmarks and TUM RGB-D. On the other side, the positive effects of the pose learning part, dynamic branch, confidence regression and hand-crafted feature based refinement are also provided.

Feature-Decision Level Collaborative Fusion Network for Hyperspectral and LiDAR Classification

The fusion-based classification of hyperspectral (HS) and light detection and ranging (LiDAR) images has become a prominent research topic, as their complementary information can effectively improve classification performance. The current methods encompass pixel-, feature- and decision-level fusion. Among them, feature- and decision-level fusion have emerged as the mainstream approaches. Collaborative fusion of these two levels can enhance classification accuracy. Although various methods have been proposed, some shortcomings still exist. On one hand, current methods ignore the shared advanced features between HS and LiDAR images, impeding the integration of multimodal features and thereby limiting the classification performance. On the other hand, the existing methods face difficulties in achieving a balance between feature- and decision-level contributions, or they simply overlook the significance of one level and fail to utilize it effectively. In this paper, we propose a novel feature-decision level collaborative fusion network (FDCFNet) for hyperspectral and LiDAR classification to alleviate these problems. Specifically, a multilevel interactive fusion module is proposed to indirectly connect hyperspectral and LiDAR flows to refine the spectral-elevation information. Moreover, the fusion features of the intermediate branch can further enhance the shared-complementary information of hyperspectral and LiDAR to reduce the modality differences. In addition, a dynamic weight selection strategy is meticulously designed to adaptively assign weight to the output of three branches at the decision level. Experiments on three public benchmark datasets demonstrate the effectiveness of the proposed methods.

CoTrFuse: a novel framework by fusing CNN and transformer for medical image segmentation

Medical image segmentation is a crucial and intricate process in medical image processing and analysis. With the advancements in artificial intelligence, deep learning techniques have been widely used in recent years for medical image segmentation. One such technique is the U-Net framework based on the U-shaped convolutional neural networks (CNN) and its variants. However, these methods have limitations in simultaneously capturing both the global and the remote semantic information due to the restricted receptive domain caused by the convolution operation’s intrinsic features. Transformers are attention-based models with excellent global modeling capabilities, but their ability to acquire local information is limited. To address this, we propose a network that combines the strengths of both CNN and Transformer, called CoTrFuse. The proposed CoTrFuse network uses EfficientNet and Swin Transformer as dual encoders. The Swin Transformer and CNN Fusion module are combined to fuse the features of both branches before the skip connection structure. We evaluated the proposed network on two datasets: the ISIC-2017 challenge dataset and the COVID-QU-Ex dataset. Our experimental results demonstrate that the proposed CoTrFuse outperforms several state-of-the-art segmentation methods, indicating its superiority in medical image segmentation. The codes are available at https://github.com/BinYCn/CoTrFuse.

Multi-Scale Semantic and Detail Extraction Network for Lightweight Person Re-Identification

Exploring multi-level information to obtain fine-grained features is the key factor to improve the performance of person re-identification (Re-ID). However, existing models for person Re-ID only focus on learning the high-level semantic information while neglecting the low-level detail information. To alleviate this issue, we propose a lightweight person Re-ID method termed Multi-Scale Semantic and Detail Extraction Network (MSDENet) to obtain robustness and discriminative feature representation for the Re-ID task. Specifically, we design a Series Channel-Spatial Attention (SCSA) and embed it into the lightweight backbone network to focus on the key parts of pedestrian images. Meanwhile, we propose a Multi-Scale Semantic and Detail Extraction (MSDE) method to extract multi-scale features of semantic information and detail information, which can effectively capture the feature diversity of pedestrian images. Furthermore, we design a Feature Enhancement Fusion (FEF), which enhances and fuses the fine-grained features of semantic extraction and detail extraction branches to better obtain the discriminative feature representation. Extensive experiments conducted on popular datasets Market1501, MSMT17, and CUHK03 demonstrate that the proposed MSDENet has competitive performance compared with the state-of-the-art methods.

LBARNet: Lightweight bilateral asymmetric residual network for real-time semantic segmentation

Real-time semantic segmentation, as a key technique for scene understanding, has been an important research topic in the field of computer vision in recent years. However, existing models are unable to achieve good segmentation accuracy on mobile devices due to their huge computational overhead, which makes it difficult to meet actual industrial requirements. To address the problems faced by current semantic segmentation tasks, this paper proposes a lightweight bilateral asymmetric residual network (LBARNet) for real-time semantic segmentation. First, we propose the bilateral asymmetric residual (BAR) module. This module learns multi-scale feature representations with strong semantic information at different stages of the semantic information extraction branch, thus improving pixel classification performance. Secondly, the spatial information extraction (SIE) module is constructed in the spatial detail extraction branch to capture multi-level local features of the shallow network to compensate for the lost geometric information in the downsampling stage. At the same time, we design the attention mechanism perception (AMP) module in the jump connection part to enhance the contextual representation. Finally, we design the dual branch feature fusion (DBF) module to exploit the correspondence between higher-order features and lower-order features to fuse spatial and semantic information appropriately. The experimental results show that LBARNet, without any pre-training and pre-processing and using only 0.6M parameters, achieves 70.8% mloU and 67.2% mloU on the Cityscapes dataset and Camvid dataset, respectively. LBARNet maintain a high segmentation accuracy while using a smaller number of parameters compared to most existing state-of-the-art models.