检索结果-内蒙古大学图书馆

An Optimized LSTM Model for Diagnosis Prediction of Lower Respiratory Tract Infections Using a Minimalistic data Set

引用

IEEE ACCESS 2025年 13卷 60341-60354页

作者： Bukenya, Lukman Eyobu, Odongo Steven Oyana, Tonny J. Makerere Univ Coll Comp & Informat Sci Dept Networks Kampala Uganda Makerere Univ Coll Comp & Informat Sci Geospatial Data & Computat Intelligence Lab Kampala Uganda Makerere Univ Coll Comp & Informat Sci Internet Things IoT Res & Training Lab Kampala Uganda

data scarcity presents a significant challenge in developing effective data-driven diagnostic systems, particularly in healthcare settings where comprehensive feature capture is often limited. In medical contexts, vital features may be absent or unmeasurable due to constraints in laboratory capabilities and limited data collection infrastructure. This study addresses these limitations by deriving a minimalistic feature set to predict lower respiratory tract infections (LRTI's), including diseases like pneumonia, bronchitis, and tuberculosis, using essential but widely available features. We employ age, gender, and initial diagnosis at admission-key indicators extracted from the MIMIC-III dataset-to train a Long Short-Term Memory (LSTM) model, demonstrating that minimal but relevant features can still yield accurate predictive outcomes. Compared to conventional models like Bi-LSTM, Random Forest, and XGBoost, our LSTM approach achieves superior training and validation performance, with an F1 score of 90% even without extensive optimization. The study validates our model across both LRTI and non-LRTI cases, underscoring its potential as an efficient, reliable diagnostic tool for resource-limited healthcare environments.

关键词： Long short term memory data models Predictive models Medical services Logic gates Diseases Training Reliability Neural networks Mortality lower respiratory tract infections prediction datasets

来源：评论

学校读者我要写书评

暂无评论

Mother: a maternal online technology for health care dataset

引用

BMC RESEARCH NOTES 2025年第1期18卷 1-3页

作者： Eyobu, Odongo Steven Nyanga, Brian Angoda Bukenya, Lukman Ongom, Daniel Oyana, Tonny J. Makerere Univ Coll Comp & Informat Sci Geospatial Data & Computat Intelligence Lab POB 7062 Kampala Uganda Makerere Univ Sch Comp & Informat Technol Dept Networks POB 7062 Kampala Uganda

ObjectivesThese data enable the development of both textual and speech based conversational machine learning models that can be used by expectant mothers to provide answers to challenges they face during the different trimesters of their pregnancy. Such models are key to the improvement of the lives of pregnant mothers, specifically in low resourced settings where doctors advise is limited by access to hospitals and language barrier. These data were used to develop a conversational chatbot model tailored for mothers in their first, second and third trimesters of *** description503 question and answer pairs on maternal health were collected through a survey of challenges facing pregnant mothers in a rural and semi-urban area of Uganda. The answers to the questions were provided and validated by professional medical personnel. The participants were purposively sampled, focusing on women in their 1st, 2nd and 3rd trimesters, with a 94% response rate. The dataset addresses common health concerns, symptoms, and conditions associated with pregnancy, particularly for women without immediate access to medical personnel. It targets maternal health outcomes such as pregnancy, morbidity, and mortality, specifically among women of reproductive age.

关键词： Maternal health Pregnancy Question and answer knowledge base Electronic health (E-health) Conversational chatbots

来源：评论

学校读者我要写书评

暂无评论

ACGRIME: adaptive chaotic Gaussian RIME optimizer for global optimization and feature selection

引用

CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS 2025年第1期28卷 1-39页

作者： Batis, Mohammed Chen, Yi Wang, Mingjing Liu, Lei Heidari, Ali Asghar Chen, Huiling Wenzhou Univ Key Lab Intelligent Informat Safety & Emergency Zh Wenzhou 325035 Peoples R China Wenzhou Univ Technol Sch Data Sci & Artificial Intelligence Wenzhou 325000 Peoples R China Sichuan Univ Coll Comp Sci Chengdu 610065 Sichuan Peoples R China Univ Tehran Coll Engn Sch Surveying & Geospatial Engn Tehran Iran

Feature selection (FS) is a crucial data preprocessing technique that selects important features to enhance learning efficiency, yet it encounters challenges due to the high-dimensional search space. This paper introduces an improved version of the RIME optimization algorithm called ACGRIME. Inspired by rime-ice formation, RIME initially showed good convergence and search capabilities but struggled with diversity and local optima in FS. ACGRIME integrates chaos theory, adaptive weighting, and Gaussian mutation to address these limitations to enhance exploration, balance exploration-exploitation, and improve solution quality. Extensive experiments were conducted using 29 CEC 2017 benchmark functions, comparing ACGRIME against the original RIME, two recently enhanced RIME variants, and 11 advanced metaheuristic algorithms. Further, balance and diversity analyses of ACGRIME, RIME, and each strategy were conducted to demonstrate its exploration and exploitation capabilities. Following this, ACGRIME is then adapted into its binary form, bACGRIME, to evaluate its performance in FS on 22 datasets, testing it with different machine learning classifiers and benchmarking it against six well-known algorithms. The experimental results on the benchmark functions, balance, and diversity analyses demonstrate that ACGRIME achieves better convergence speed and balanced exploration-exploitation. Moreover, bACGRIME's performance for FS shows superior classification accuracy with fewer features. The source code of ACGRIME and bACGRIME is publicly accessible at https://***/batis1/ACGRIME.

关键词： RIME Chaos theory Adaptive weighting Gaussian mutation Optimization Feature selection

来源：评论

学校读者我要写书评

暂无评论

DALI-SLAM: Degeneracy-aware LiDAR-inertial SLAM with novel distortion correction and accurate multi-constraint pose graph optimization

引用

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING 2025年 221卷 92-108页

作者： Wu, Weitong Chen, Chi Yang, Bisheng Zou, Xianghong Liang, Fuxun Xu, Yuhang He, Xiufeng Hohai Univ Sch Earth Sci & Engn Nanjing 211100 Peoples R China Wuhan Univ State Key Lab Informat Engn Surveying Mapping & Re Wuhan 430079 Peoples R China Wuhan Univ Minist Educ China Engn Res Ctr Spatio temporal Data Smart Acquisit & Wuhan 430079 Peoples R China Wuhan Univ Inst Geospatial Intelligence Wuhan 430079 Peoples R China

LiDAR-Inertial simultaneous localization and mapping (LI-SLAM) plays a crucial role in various applications such as robot localization and low-cost 3D mapping. However, factors including inaccurate motion distortion estimation and pose graph constraints, and frequent LiDAR feature degeneracy present significant challenges for existing LI-SLAM methods. To address these issues, we propose DALI-SLAM, an accurate and robust LI-SLAM that consists of degeneracy-aware LiDAR-inertial odometry (DA-LIO) with a dual spline-based motion distortion correction (DS-MDC) module, and multi-constraint pose graph optimization (MC-PGO). Considering the cumulative errors of micro-electromechanical systems (MEMS) inertial measurement unit (IMU) integration, two continuous-time trajectories in the sliding window are fitted to update the discrete IMU poses for accurate motion distortion correction. In the LiDAR-inertial fusion stage, LiDAR feature degeneracy is detected by analyzing the Jacobian matrix and a remapping strategy is introduced into the updating of error state Kalman Filter (ESKF) to mitigate the influence of degeneracy. Furthermore, in the back-end optimization stage, three types of submap constraints are accurately built with dedicated strategy through a robust variant of the iterative closest point (ICP) method. The proposed method is comprehensively validated using data collected from a helmet-based laser scanning system (HLS) in representative indoor and outdoor environments. Experiment results demonstrate that the proposed method outperforms the SOTA methods on the test data. Specifically, the proposed DS-MDC module reduces trajectory root mean square errors (RMSEs) by 7.9 %, 5.8 %, and 3.1 %, while the degeneracy-aware update strategy achieves additional reductions of 43.3 %, 17.7%, and 4.9 %, respectively, across three typical sequences compared to existing methods, thereby effectively improving trajectory accuracy. Furthermore, the results of DA-LIO demonstrate a maxi

关键词： LiDAR degeneracy LiDAR-inertial SLAM Motion distortion Degeneracy-aware Pose graph optimization

来源：评论

学校读者我要写书评

暂无评论

Adaptive density-based clustering for many objective similarity or redundancy evolutionary optimization

引用

EXPERT SYSTEMS WITH APPLICATIONS 2025年 266卷

作者： Wang, Mingjing Heidari, Ali Asghar Chen, Long Wang, Ruili Liu, Mingzhe Shao, Lizhi Chen, Huiling Wenzhou Univ Technol Sch Data Sci & Artificial Intelligence Wenzhou 325000 Peoples R China Univ Tehran Coll Engn Sch Surveying & Geospatial Engn Tehran Iran Southeast Univ Sch Comp Sci & Engn Nanjing 211189 Peoples R China Southeast Univ Key Lab Comp Network & Informat Integrat Minist Educ Nanjing 211189 Peoples R China Chinese Acad Sci Key Lab Mol Imaging Inst Automat Beijing 100190 Peoples R China Wenzhou Univ Coll Comp Sci & Artificial Intelligence Wenzhou 325035 Zhejiang Peoples R China

With the increase in the number of objectives, the curse of dimensionality will eventually occur in some practical multi-objective optimization problems. This situation will become even worse when the multi- objective changes into many-objective optimization problems (MaOPs) with more than 15 objectives, which makes it more difficult for evolutionary computing. The persistent focus within the field has been on how to reduce the scale of problem-solving and alleviate the complexity of problems by analyzing the linear or nonlinear relationships between various objectives. Traditionally, the objectives relationships are typically presumed to be in conflict with each other, and yet similarity or redundancy may exist in some MaOPS. In this study, an adaptive evolutionary many-objective optimization algorithm with similarity or redundancy reduction based on adaptive density clustering denoted as AENec is proposed for MaOPs. In the AENec, the similarity or redundancy between objectives is analyzed for MaOPs for the first time and the redundance analyses method for objectives(RAMO) is designed with the limited Pareto front structure information during classic MaOPs evolutionary optimization(CMEVO). An adaptive density-based clustering algorithm (AEDBSCAN) is designed to configure the number of adaptive density clusters and automatically determine the extent of objective reduction. In addition, to avoid inconsistencies between the Pareto solutions post-reduction and the original MaOPs, the efficient non-redundant objective aggregation function (NROAF) is also devised. The parameters and algorithmic components of AENec are successfully analyzed on random test cases. The statistical results demonstrate that the proposed method outperforms other competitive algorithms on most test benchmarks and also especially verifies its effectiveness on MaOPs with similar or redundant objective relationships.

关键词： Multi-objective optimization Many-objective optimization Similarity or redundancy Objectives relationships Redundance analyses Objective reduction Objective aggregation

来源：评论

学校读者我要写书评

暂无评论

PylonModeler: A hybrid-driven 3D reconstruction method for power transmission pylons from LiDAR point clouds

引用

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING 2025年 220卷 100-124页

作者： Wu, Shaolong Chen, Chi Yang, Bisheng Yan, Zhengfei Wang, Zhiye Sun, Shangzhe Zou, Qin Fu, Jing Wuhan Univ State Key Lab Informat Engn Surveying Mapping & Re Wuhan Peoples R China Minist Educ PRC Engn Res Ctr Space Time Data Capturing & Smart App Wuhan Peoples R China Wuhan Univ Inst Geospatial Intelligence Wuhan Peoples R China Wuhan Univ Sch Comp Sci Wuhan Peoples R China China Elect Power Res Inst Wuhan Branch Wuhan Peoples R China

As the power grid is an indispensable foundation of modern society, creating a digital twin of the grid is of great importance. Pylons serve as components in the transmission corridor, and their precise 3D reconstruction is essential for the safe operation of power grids. However, 3D pylon reconstruction from LiDAR point clouds presents numerous challenges due to data quality and the diversity and complexity of pylon structures. To address these challenges, we introduce PylonModeler: a hybrid-driven method for 3D pylon reconstruction using airborne LiDAR point clouds, thereby enabling accurate, robust, and efficient real-time pylon reconstruction. Different strategies are employed to achieve independent reconstructions and assemblies for various structures. We propose Pylon Former, a lightweight transformer network for real-time pylon recognition and decomposition. Subsequently, we apply a data-driven approach for the pylon body reconstruction. Considering structural characteristics, fitting and clustering algorithms are used to reconstruct both external and internal structures. The pylon head is reconstructed using a hybrid approach. A pre-built pylon head parameter model library defines different pylons by a series of parameters. The coherent point drift (CPD) algorithm is adopted to establish the topological relationships between pylon head structures and set initial model parameters, which are refined through optimization for accurate pylon head reconstruction. Finally, the pylon body and head models are combined to complete the reconstruction. We collected an airborne LiDAR dataset, which includes a total of 3398 pylon data across eight types. The dataset consists of transmission lines of various voltage levels, such as 110 kV, 220 kV, and 500 kV. PylonModeler is validated on this dataset. The average reconstruction time of a pylon is 1.10 s, with an average reconstruction accuracy of 0.216 m. In addition, we evaluate the performance of PylonModeler on public a

关键词： Power Grid Automatic 3D Reconstruction LiDAR Point Cloud Pylon

来源：评论

学校读者我要写书评

暂无评论

Cross-view geolocalization and disaster mapping with street-view and VHR satellite imagery: A case study of Hurricane IAN

引用

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING 2025年 220卷 841-854页

作者： Li, Hao Deuser, Fabian Yin, Wenping Luo, Xuanshu Walther, Paul Mai, Gengchen Huang, Wei Werner, Martin Tech Univ Munich Sch Engn & Design Professorship Big Geospatial Data Management D-85521 Munich Bavaria Germany Univ Bundeswehr Munich Inst Distributed Intelligent Syst D-85579 Neubiberg Bavaria Germany China Univ Min & Technol Sch Environm & Spatial Informat Xuzhou 221116 Peoples R China Univ Texas Austin Dept Geog & Environm Spatially Explicit Artificial Intelligence Lab Austin TX 78712 USA Tongji Univ Coll Surveying & Geoinformat Shanghai 200092 Peoples R China

Nature disasters playa key role in shaping human-urban infrastructure interactions. Effective and efficient response to natural disasters is essential for building resilience and sustainable urban environment. Two types of information are usually the most necessary and difficult to gather in disaster response. The first information is about the disaster damage perception, which shows how badly people think that urban infrastructure has been damaged. The second information is geolocation awareness, which means how people's whereabouts are made available. In this paper, we proposed a novel disaster mapping framework, namely CVDisaster, aiming at simultaneously addressing geolocalization and damage perception estimation using cross-view Street-View Imagery (SVI) and Very High-Resolution satellite imagery. CVDisaster consists of two cross-view models, where CVDisaster-Geoloc refers to a cross-view geolocalization model based on a contrastive learning objective with a Siamese ConvNeXt image encoder and CVDisaster-Est is a cross-view classification model based on a Coupled Global Context Vision Transformer (CGCViT). Taking Hurricane IAN as a case study, we evaluate the CVDisaster framework by creating a novel cross-view dataset (CVIAN) and conducting extensive experiments. As a result, we show that CVDisaster can achieve highly competitive performance (over 80% for geolocalization and 75% for damage perception estimation) with even limited fine-tuning efforts, which largely motivates future cross- view models and applications within a broader GeoAI research community. The data and code are publicly available at: https://***/tum-bgd/CVDisaster.

关键词： Cross-view Disaster response GeoAI Human-urban interaction Street-view imagery Geolocalization Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

DeCoTa: a lightweight decentralized cross-chain transfer scheme based on hash-locking

引用

SCIENTIFIC REPORTS 2025年第1期15卷 1-15页

作者： Zou, Huiying Sun, Hua Chen, Sheng Ren, Wei Zheng, Xianghan China Univ Geosci Sch Comp Sci Wuhan 430074 Peoples R China Soochow Univ Adv Comp Prov Univ Key Lab Data Intelligence Suzhou Peoples R China Sichuan Univ Key Lab Data Protect & Intelligent Management Minist Educ Chengdu Peoples R China Third Surveying & Mapping Inst Hunan Prov Changsha Hunan Peoples R China Hunan Engn Res Ctr Geog Informat Secur & Applicat Changsha Hunan Peoples R China Hubei Univ Econ Hubei Key Lab Digital Finance Innovat Wuhan Peoples R China CASM State Key Lab Geoinformat Engn Key Lab Surveying & Mapping Sci & Geospatial Infor Beijing Peoples R China Fuzhou Univ Coll Comp & Big Data Fuzhou Peoples R China Sanya Coll Sch Informat & Intelligent Engn Sanya Peoples R China

With the wide application of blockchain comes the challenge of cross-chain interaction. For example, the isolation between the information stored in different blockchains can result in the "isolated islands of value" effect in blockchains. In addition, conventional cross-chain methods have a number of drawbacks such as centralization, or have a very restricted application scope. (e.g., hash-locking can only be utilized for asset exchanges, but asset transfer is a more common and demanding requirement). Other solutions like sidechain may be an overkill (e.g., too complex) for simple applications. To address these challenges, we propose an cross-chain transaction scheme that integrates hash-locking with notary mechanisms to enable decentralized asset transfers. By embedding notary scheme within the hash-lock structure, our approach ensures support for asset transfers while mitigating centralization risks. Additionally, we enhance the traditional hash-locking mechanism, making it more flexible and suitable for diverse cross-chain scenarios. This combination of hash-locking and notary scheme allows our scheme to overcome the original shortcomings of them both, while maintaining high efficiency and security. In addition to designing the transaction method, we also constructed a transaction model, conducted experiments on transaction gas consumption, performed a detailed security analysis, and presented an in-depth discussion of our scheme while proposing directions for future work. Experiment and analysis demonstrate the practicability of our scheme in terms of performance and security.

关键词： Cross-chain Blockchain Hash-locking Notary scheme Decentralization Smart contract

来源：评论

学校读者我要写书评

暂无评论

PatchOut: A novel patch-free approach based on a transformer-CNN hybrid framework for fine-grained land-cover classification on large-scale airborne hyperspectral images

引用

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 2025年 138卷

作者： Ji, Renjie Tan, Kun Wang, Xue Tang, Shuwei Sun, Jin Niu, Chao Pan, Chen East China Normal Univ Key Lab Spatial Temporal Big Data Anal & Applicat Minist Nat Resources Shanghai 200241 Peoples R China East China Normal Univ Key Lab Geog Informat Sci Minist Educ Shanghai 200241 Peoples R China East China Normal Univ Sch Geog Sci Shanghai 200241 Peoples R China East China Normal Univ Sch Geospatial Artificial Intelligence Shanghai 200241 Peoples R China Shanghai Municipal Inst Surveying & Mapping Shanghai 200063 Peoples R China

Airborne hyperspectral systems can provide high-resolution hyperspectral images (HSIs) covering large scenes, enabling fine-grained land-cover classification. However, the most popular patch-based methods are limited by low computational efficiency and broken classification results, which hinders the full utilization of this powerful technology in Earth observation applications. Therefore, in this paper, considering the efficiency requirements for large-scale land-cover classification, a novel patch-free approach based on a Transformer-CNN hybrid (PatchOut) framework is proposed. The proposed PatchOut framework adopts an encoder-decoder architecture, enabling rapid semantic segmentation for HSI classification. For the encoder module, we introduce a computationally efficient reduced Transformer module integrated with convolutional neural network (CNN), to leverage their complementary strengths for long-range and local feature extraction, respectively. A multi-scale spatial-spectral feature fusion (MSSSFF) module is also proposed to amalgamate the characteristics of different levels from the encoder, which enhances the overall feature representation. Then, to address the loss of semantic detail and resolution inherent in multi-level feature extraction, a novel feature reconstruction module (FRM) is applied to recover high-quality semantic features. Finally, a large-scale benchmark dataset, Qingpu-HSI, is presented, comprising airborne HSIs covering 33.91 km2 with 20 land-cover classes. Experiments on the Qingpu-HSI and another public dataset demonstrate the superior accuracy and efficiency of our proposed PatchOut framework, outperforming several well-known patch-free and patch-based methods. The Qingpu HSI dataset, along with the PatchOut framework code will be released at https://***/busbyjrj/PatchOut.

关键词： Hyperspectral image classification Patch-free Semantic segmentation Multi-scale feature fusion

来源：评论

学校读者我要写书评

暂无评论

Broad Area Search and Detection of Surface-to-Air Missile Sites Using Spatial Fusion of Component Object Detections From Deep Neural Networks

引用

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING 2020年 13卷 4728-4737页

作者： Cannaday, Alan B., II Davis, Curt H. Scott, Grant J. Ruprecht, Blake Anderson, Derek T. Univ Missouri Ctr Geospatial Intelligence Columbia MO 65211 USA Univ Missouri Mizzou Informat & Data Fus Lab Columbia MO 65211 USA

Here, we demonstrate how deep neural network (DNN) detections of multiple constitutive or component objects that are part of a larger, more complex, and encompassing feature can be spatially fused to improve the search, detection, and retrieval (ranking) of the larger complex feature. First, scores computed from a spatial clustering algorithm are normalized to a reference space so that they are independent of image resolution and DNN input chip size. Then, multiscale DNN detections from various component objects are fused to improve the detection and retrieval of DNN detections of a larger complex feature. We demonstrate the utility of this approach for broad area search and detection of surface-to-air missile (SAM) sites that have a very low occurrence rate (only 16 sites) over a similar to 90 000 km(2) study area in SE China. The results demonstrate that spatial fusion of multiscale componentobject DNN detections can reduce the detection error rate of SAM Sites by >85% while still maintaining a 100% recall. The novel spatial fusion approach demonstrated here can be easily extended to a wide variety of other challenging object search and detection problems in large-scale remote sensing image datasets.

关键词： Broad area search data fusion deep neural networks (DNN) information retrieval object detection spatial clustering

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：