Query：
学者姓名：李建微
Refining：
Year
Type
Indexed by
Source
Complex
Former Name
Co-
Language
Clean All
Abstract ：
Accurate and reliable segmentation of the pancreas and its lesions on computed tomography (CT) images is crucial in medical imaging for preoperative diagnosis, surgical planning, and postoperative monitoring. However, there are limited studies that address simultaneous segmentation of the pancreas and pancreatic tumors. Moreover, existing studies have not fully utilized the feature potential of the original images and have neglected the exploration of semantic information with strong representation. To overcome these limitations, we propose the Strongly Representative Semantic-guided Segmentation Network (SRSNet). Specifically, we employ intermediate semantic information to generate strongly representative high-resolution pre-segmented images, effectively reducing channel redundancy across different resolutions. We utilize various mechanisms to extract distinct representative features, and with the guidance of these features, SRSNet effectively supplements high-resolution detailed information for features of different resolutions, provides auxiliary features for the pixel decision phase of the network, and detects large-scale changes in the pancreas and pancreatic tumors. Additionally, we design a loss function that enhances SRSNet's sensitivity to boundary pixels and attenuates the effect of class imbalance. Our method is evaluated on Task07 Pancreas and NIH Pancreas datasets. In the experiment of combined pancreas and tumor segmentation in the MSD dataset, we achieved Dice, Recall, Precision, and MIoU scores of 78.60%, 79.64%, 81.72%, and 71.47%, respectively. Extensive experiments demonstrate that our algorithm not only outperforms state-of-the-art algorithms for pancreas segmentation but also exhibits excellent performance for pancreas and pancreatic tumor segmentation.
Keyword ：
High resolution High resolution Lightweight Lightweight Pancreas Pancreas Pancreatic cyst Pancreatic cyst Priori probability Priori probability
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Cao, Luyang , Li, Jianwei . Strongly representative semantic-guided segmentation network for pancreatic and pancreatic tumors [J]. | BIOMEDICAL SIGNAL PROCESSING AND CONTROL , 2024 , 87 . |
| MLA | Cao, Luyang 等. "Strongly representative semantic-guided segmentation network for pancreatic and pancreatic tumors" . | BIOMEDICAL SIGNAL PROCESSING AND CONTROL 87 (2024) . |
| APA | Cao, Luyang , Li, Jianwei . Strongly representative semantic-guided segmentation network for pancreatic and pancreatic tumors . | BIOMEDICAL SIGNAL PROCESSING AND CONTROL , 2024 , 87 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Background Extreme wildfires pose a serious threat to forest vegetation and human life because they spread more rapidly and are more intense than conventional wildfires. Detecting extreme wildfires is challenging due to their visual similarities to traditional fires, and existing models primarily detect the presence or absence of fires without focusing on distinguishing extreme wildfires and providing warnings.Aims To test a system for real time detection of four extreme wildfires.Methods We proposed a novel lightweight model, called LEF-YOLO, based on the YOLOv5 framework. To make the model lightweight, we introduce the bottleneck structure of MobileNetv3 and use depthwise separable convolution instead of conventional convolution. To improve the model's detection accuracy, we apply a multiscale feature fusion strategy and use a Coordinate Attention and Spatial Pyramid Pooling-Fast block to enhance feature extraction.Key results The LEF-YOLO model outperformed the comparison model on the extreme wildfire dataset we constructed, with our model having excellent performance of 2.7 GFLOPs, 61 FPS and 87.9% mAP.Conclusions The detection speed and accuracy of LEF-YOLO can be utilised for the real-time detection of four extreme wildfires in forest fire scenes.Implications The system can facilitate fire control decision-making and foster the intersection between fire science and computer science. We tested a lightweight architecture called LEF-YOLO for detecting four extreme wildfires. We found improved detection accuracy through multi-scale fusion and attention mechanism, and constructed four extreme wildfire datasets and compared these with multiple object detection models and lightweight feature extraction networks. This method is beneficial for the development of extreme wildfire field robots.
Keyword ：
convolutional neural networks convolutional neural networks deep learning deep learning extreme wildfire extreme wildfire fire safety fire safety lightweight lightweight multiscale feature fusion multiscale feature fusion object detection object detection YOLO (LEF-YOLO) YOLO (LEF-YOLO)
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Li, Jianwei , Tang, Huan , Li, Xingdong et al. LEF-YOLO: a lightweight method for intelligent detection of four extreme wildfires based on the YOLO framework [J]. | INTERNATIONAL JOURNAL OF WILDLAND FIRE , 2024 , 33 (1) . |
| MLA | Li, Jianwei et al. "LEF-YOLO: a lightweight method for intelligent detection of four extreme wildfires based on the YOLO framework" . | INTERNATIONAL JOURNAL OF WILDLAND FIRE 33 . 1 (2024) . |
| APA | Li, Jianwei , Tang, Huan , Li, Xingdong , Dou, Hongqiang , Li, Ru . LEF-YOLO: a lightweight method for intelligent detection of four extreme wildfires based on the YOLO framework . | INTERNATIONAL JOURNAL OF WILDLAND FIRE , 2024 , 33 (1) . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
To recognize and locate fire accurately, we developed a wildfire-locating method based on superpixel features. Previous research detected fire at the pixel level of the image using deep learning, whereas we preprocessed and classified superpixels for fire recognition and locating. Firstly, a simple linear iterative clustering algorithm was created to segment the image into superpixel blocks. Then, color, texture, and shape features were extracted from each superpixel, and a convolutional neural network (CNN) was employed to classify the superpixels into two categories: fire superpixels and background superpixels. Finally, superpixels were refined based on the superpixel adjacency relationship. Experimental results demonstrated that the combined approach of superpixel features and CNNs performed satisfactory segmentation performance with an accuracy of 96.58%, which was effective in wildfire-locating. © 2024 IEEE.
Keyword ：
Clustering algorithms Clustering algorithms Convolution Convolution Convolutional neural networks Convolutional neural networks Deep learning Deep learning Feature extraction Feature extraction Fires Fires Image segmentation Image segmentation Iterative methods Iterative methods Location Location Superpixels Superpixels Textures Textures
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Tang, Huan , Zheng, Xiaogan , Guo, Wei et al. Fire-locating Based on Superpixel Features by Convolutional Neural Network [C] . 2024 : 200-205 . |
| MLA | Tang, Huan et al. "Fire-locating Based on Superpixel Features by Convolutional Neural Network" . (2024) : 200-205 . |
| APA | Tang, Huan , Zheng, Xiaogan , Guo, Wei , Wan, Jiali , Li, Jianwei . Fire-locating Based on Superpixel Features by Convolutional Neural Network . (2024) : 200-205 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
深度估计在虚拟现实、场景重建、自动驾驶和目标检测等领域发挥着重要作用。全景图像包含全向视野信息，逐渐成为深度估计领域的研究热点。但是，全景图像存在图像畸变的问题，而且深度数据采集、标注较为困难。对此，提出采用自监督方式，利用自监督深度学习算法，引入通道优化多空间融合注意力机制，增强远距离特征提取，以获取全局和局部信息。同时，引入全景感受野块，扩充感受野以获取多尺度信息。
Keyword ：
全景图像 全景图像 深度估计 深度估计 深度学习 深度学习 自监督 自监督
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 陈思喜 , 张延吉 , 李建微 . 基于自监督深度学习的全景图像深度估计研究 [J]. | 电视技术 , 2024 , 48 (03) : 34-38,43 . |
| MLA | 陈思喜 et al. "基于自监督深度学习的全景图像深度估计研究" . | 电视技术 48 . 03 (2024) : 34-38,43 . |
| APA | 陈思喜 , 张延吉 , 李建微 . 基于自监督深度学习的全景图像深度估计研究 . | 电视技术 , 2024 , 48 (03) , 34-38,43 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
人群分布不均、遮挡和背景干扰等问题使得人群计数成为了一项复杂且具有挑战性的任务.针对这些问题,提出了一种多尺度特征融合的位置关注网络(Position-Aware Network based on Multi-Scale Feature Fusion,MSF-PANet).首先,设计了一种多尺度特征融合模块,以在不同感受野下提取并融合人群密度图的多尺度特征,同时提取出前景信息,来应对人群计数中的遮挡和背景干扰问题;然后,通过位置注意力分配网络提高模型对人群区域的关注度,有效地应对人群分布不均的问题;最后,为了辅助模型训练,减小背景噪声带来的干扰,引入了一种结构交叉损失用于强化模型对人群结构的学习.实验结果表明:MSF-PANet在Shanghai Tech Part A、Shanghai Tech Part B、UCF-QNRF和UCF_CC_50 上平均绝对误差分别为 59.5、7.8、103、182.7,均方误差分别为 96.7、13.6、177、237.7,验证了所提模块在提高人群计数准确率上的有效性.
Keyword ：
人群密度估计 人群密度估计 人群计数 人群计数 多尺度特征 多尺度特征 注意力机制 注意力机制 背景分割 背景分割
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 谢劭卓 , 李建微 . 基于多尺度特征融合与位置关注网络的人群计数研究 [J]. | 微电子学与计算机 , 2024 , 41 (8) : 22-30 . |
| MLA | 谢劭卓 et al. "基于多尺度特征融合与位置关注网络的人群计数研究" . | 微电子学与计算机 41 . 8 (2024) : 22-30 . |
| APA | 谢劭卓 , 李建微 . 基于多尺度特征融合与位置关注网络的人群计数研究 . | 微电子学与计算机 , 2024 , 41 (8) , 22-30 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Wildfires, a persistent environmental menace, are a significant source of harmful gases and particulate emissions. This study leverages the fire radiative power (FRP) method to delineate a comprehensive wildfire emission inventory for Southwest China from 2001 to 2020. Daily fire radiative power data derived from 1 km MODIS Thermal Anomalies/Fire products (MOD14/MYD14) were used to calculate the FRE and combusted biomass. Available emission factors were assigned to three biomass burn types: forest, grass, and shrub fires. Over the span of two decades, we have compiled data and estimated the annual emissions of carbon dioxide (CO2), carbon monoxide (CO), methane (CH4), sulfur dioxide (SO2), ammonia (NH3), nitrogen oxides (NOx), total particulate matter (TPM), black carbon (BC), organic carbon (OC), and non-methane volatile organic compounds (NMVOCs) to be 9809.13, 566.82, 25.79, 5.37, 12.25, 16.67, 133.53, 4.16, 41.81, and 97.23 Gg per year (Gg yr(-1)), respectively. In terms of fire type, forest fires accounted for the largest portion of total CO2 emissions (59.23%), with grass fires and shrub fires coming in second and third, accounting for 20.41% and 20.36%, respectively. Geographically, Yunnan Province were identified as the major contributor in Southwest China, accounting for 69.67% of the total emissions. Temporally, the maximum emission occurred in 2010 (24263.33 Gg), and the minimum emission occurred in 2017 (2917.66 Gg). And the emissions were mainly concentrated in February (23.33%), March (25.52%), and April (22.61%), which accounted for nearly three-fourths of the total emissions. The results of this study are much higher than those obtained by the burned area method, almost three times as high. In contrast, the results of this study are close to the fire emission data from the GFED4s and GFASv1.2 and QFEDv2.5r1 databases.
Keyword ：
Emission inventory Emission inventory Fire radiative power Fire radiative power Forest fire Forest fire Southwest China Southwest China Wildfire Wildfire
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Ning, Xincen , Li, Jianwei , Zhuang, Pengkun et al. Wildfire combustion emission inventory in Southwest China (2001-2020) based on MODIS fire radiative energy data [J]. | ATMOSPHERIC POLLUTION RESEARCH , 2024 , 15 (11) . |
| MLA | Ning, Xincen et al. "Wildfire combustion emission inventory in Southwest China (2001-2020) based on MODIS fire radiative energy data" . | ATMOSPHERIC POLLUTION RESEARCH 15 . 11 (2024) . |
| APA | Ning, Xincen , Li, Jianwei , Zhuang, Pengkun , Lai, Shifu , Zheng, Xiaogan . Wildfire combustion emission inventory in Southwest China (2001-2020) based on MODIS fire radiative energy data . | ATMOSPHERIC POLLUTION RESEARCH , 2024 , 15 (11) . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Existing neural network segmentation schemes perform well in the task of segmenting images of organs with large areas and clear morphology, such as the liver and lungs. However, it is difficult to segment organs with variable morphology and small target area, such as pancreas and tumors. In order to achieve accurate seg-mentation of pancreas and its cysts, MDAG-Net (Multi-dimensional Attention Gate Network) is proposed in this paper. Combining three attention mechanisms: spatial, channel and multi-dimensional feature map input, MDAG (Multi-dimensional Attention Gate) obtains the global distribution of semantic information in spatial and channel dimensions, filters redundant information in shallow feature maps, realizes feature response, and recalibrates convolution kernel parameters. In addition, the WML(Weighted cross entropy and MIoU loss function) loss can adaptively assign the weight of category loss and count the classification error of global pixels, which can in-crease the error attention of the target area and improving the segmentation accuracy of the network. The al-gorithm is experimented on the Task07_Pancreas dataset, compared with U-Net under the same conditions, the Dice coefficient, Precision, Recall rate and MIoU (Mean Intersection over Union) of MDAG-Net are improved by 5.3%, 1.5%, 12.7% and 7.6% respectively. The results show that MDAG-Net can accurately segment the region of pancreas and its cyst in CT(Computed Tomography) images, which proves that MDAG has better segmentation efficiency for such small target regions.
Keyword ：
Attention mechanism Attention mechanism Cross entropy loss Cross entropy loss Multi-object segmentation Multi-object segmentation Pancreas Pancreas Pancreatic tumor Pancreatic tumor Small target detection Small target detection
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Cao, Luyang , Li, Jianwei , Chen, Shu . Multi-target segmentation of pancreas and pancreatic tumor based on fusion of attention mechanism [J]. | BIOMEDICAL SIGNAL PROCESSING AND CONTROL , 2023 , 79 . |
| MLA | Cao, Luyang et al. "Multi-target segmentation of pancreas and pancreatic tumor based on fusion of attention mechanism" . | BIOMEDICAL SIGNAL PROCESSING AND CONTROL 79 (2023) . |
| APA | Cao, Luyang , Li, Jianwei , Chen, Shu . Multi-target segmentation of pancreas and pancreatic tumor based on fusion of attention mechanism . | BIOMEDICAL SIGNAL PROCESSING AND CONTROL , 2023 , 79 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
When a firefighting incident occurs in a wild complex mountain with no obvious roads or sparse roads, it is crucial to plan a safe and fast route through the complex mountain environment. Aiming at the problem that Ant Colony Optimization (ACO) is easy to fall into local optimum and the search time is long for complex mountain path planning, our study proposes an ACO algorithm for hiking emergency rescue path planning, which is suitable for fine-grained wild mountain environments. Firstly, our study analyzed the relationship between surface information and human movement speed based on existing literature and designed the objective function and heuristic function of the optimization algorithm considering two factors: surface shrub cover and terrain slope. Then, we used a combination of plane and field of view ant search combined with heuristic function and pheromone concentration to determine the next grid to be selected in the optimization process of the improved algorithm. Finally, the improved algorithm used a Laplace distribution to adjust the initial pheromone to improve the quality of the algorithm's initial solution. For the deadlock problem, the improved algorithm added isolated pheromones to prevent the next ant from falling into a deadlock dilemma. The improved algorithm used a genetic operator with grouping to update the global regular pheromone to avoid the ant colony from falling into a local optimum dilemma. In our study, we applied four ACO to the wild mountain environment of 400×400 grids, 1000 grids×1000 grids, 5000 grids×5000 grids, and 10 000 grids×10 000 grids for comparison, and set different starting and ending points for each environment. The experimental results show that each ACO using a combined planar and visual field search approach can obtain feasible paths in all four experiments, which verified the feasibility of the method. The quality of the paths using the improved algorithms was better than the other three algorithms, with improvements of 0.52%~4.95%, 4.71%~5.39%, 2.26%~13.11%, and 3.84%~9.16% in the four experiments, respectively, and the improved algorithm had shorter search time and convergence time. In addition, the combined planar and visual field search approach reduced the search space and improved the computational efficiency of the algorithm in the field 3D mountain environment. This search method was faster than the 8-connected method and reduced the average time consumption by more than 90%. Our algorithm is suitable for hiking path planning research in large 3D mountain scenes, with reduced planning time and improved path quality, providing technical support for the work of finding the best 3D mountain hiking paths without road networks. © 2023 Journal of Geo-Information Science. All rights reserved.
Keyword ：
Ant colony optimization Ant colony optimization Computational efficiency Computational efficiency Genetic algorithms Genetic algorithms Heuristic algorithms Heuristic algorithms Landforms Landforms Motion planning Motion planning
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Wu, Yuefei , Li, Jianwei , Bi, Sheng et al. Research on Improved Ant Colony Algorithm for Mountain Hiking Emergency Rescue Path Planning [J]. | Journal of Geo-Information Science , 2023 , 25 (1) : 90-101 . |
| MLA | Wu, Yuefei et al. "Research on Improved Ant Colony Algorithm for Mountain Hiking Emergency Rescue Path Planning" . | Journal of Geo-Information Science 25 . 1 (2023) : 90-101 . |
| APA | Wu, Yuefei , Li, Jianwei , Bi, Sheng , Zhu, Xin , Wang, Qianfeng . Research on Improved Ant Colony Algorithm for Mountain Hiking Emergency Rescue Path Planning . | Journal of Geo-Information Science , 2023 , 25 (1) , 90-101 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Detecting and tracking dynamic objects in a scene using point cloud data collected by LiDAR and estimating the motion state of objects with high accuracy are challenges for autonomous driving technology. In this study, a motion detection method based on point cloud registration is investigated to detect motion through the overlapping relationship between source and target point clouds after registration and extract moving objects using clustering and scale analysis by combining the object information of interest acquired by deep learning networks. Next, object association is achieved by object motion information and geometric and texture features. Then, a point cloud registration method flow is designed to estimate the motion state of the object with high accuracy by point cloud registration. The detection, tracking and estimation of the accurate motion state of moving objects are achieved.
Keyword ：
Autonomous vehicles Autonomous vehicles environmental perception environmental perception Feature extraction Feature extraction Motion detection Motion detection motion estimation motion estimation object tracking object tracking point cloud point cloud Point cloud compression Point cloud compression Radar tracking Radar tracking registration registration Sensors Sensors Target tracking Target tracking Tracking Tracking
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Li, Jianwei , Huang, Xin , Zhan, Jiawang . High-Precision Motion Detection and Tracking Based on Point Cloud Registration and Radius Search [J]. | IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS , 2023 , 24 (6) : 6322-6335 . |
| MLA | Li, Jianwei et al. "High-Precision Motion Detection and Tracking Based on Point Cloud Registration and Radius Search" . | IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 24 . 6 (2023) : 6322-6335 . |
| APA | Li, Jianwei , Huang, Xin , Zhan, Jiawang . High-Precision Motion Detection and Tracking Based on Point Cloud Registration and Radius Search . | IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS , 2023 , 24 (6) , 6322-6335 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
头发分割是图像分割领域的一大挑战,头发的自动分割对辅助性别分类、身份识别、医疗影像分析以及头部重构、AR染发等都有着重要的意义.基于机器学习方法对头发进行自动化分割是该领域的常用方法,具有效率高性能好的优点.文章梳理了基于早期机器学习的传统头发自动分割方法与基于深度学习的头发自动分割方法的发展历程,重点分析了贝叶斯网络图模型、区域生长算法、聚类算法、图割算法等传统分割方法以及全连接神经网络、全卷积神经网络、U-Net、MobileNet等基于深度学习的分割方法,并归纳对比各方法的分割效果、优缺点和发展方向.基于深度学习的头发分割方法需要使用大体量的数据集对网络进行训练,文章整理了头发分割常用公开数据集的各项属性,并对各方法使用不同数据集的各项分割性能进行对比.在此基础上,对基于机器学习的头发自动分割所面临的困难和挑战进行梳理和分析,针对存在的问题提出解决思路,对该领域的发展前景加以展望.
Keyword ：
图像分割 图像分割 头发分割 头发分割 机器学习 机器学习 深度学习 深度学习 神经网络 神经网络
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 林霞 , 李建微 , 陈溶漾 . 基于机器学习的头发自动分割研究进展 [J]. | 微电子学与计算机 , 2023 , (04) : 18-29 . |
| MLA | 林霞 et al. "基于机器学习的头发自动分割研究进展" . | 微电子学与计算机 04 (2023) : 18-29 . |
| APA | 林霞 , 李建微 , 陈溶漾 . 基于机器学习的头发自动分割研究进展 . | 微电子学与计算机 , 2023 , (04) , 18-29 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Export
| Results： |
Selected to |
| Format： |
