检索结果-内蒙古大学图书馆

Explainable AI for Medical Data: Current Methods, Limitations, and Future Directions

ACM COMPUTING SURVEYS 2025年第6期57卷 1-46页

作者： Hossain, Md Imran Zamzmi, Ghada Mouton, Peter r. Salekin, Md Sirajus Sun, Yu Goldgof, Dmitry Univ S Florida Dept Comp Sci & Engn ENB 4202 E Fowler Ave Tampa FL 33620 USA SRC Biosci 1810 W Kennedy Blvd Tampa FL 33606 USA

With the power of parallel processing, large datasets, and fast computational resources, deep neural networks (DNNs) have outperformed highly trained and experienced human experts in medical applications. However, the large global community of healthcare professionals, many of whom routinely face potentially life-or-death outcomes with complex medicolegal consequences, have yet to embrace this powerful technology. The major problem is that most current AI solutions function as a metaphorical black-box positioned between input data and output decisions without a rigorous explanation for their internal processes. With the goal of enhancing trust and improving acceptance of artificial intelligence- (AI) based technology in clinical medicine, there is a large and growing effort to address this challenge using eXplainable AI (XAI), a set of techniques, strategies, and algorithms with an explicit focus on explaining the "hows and whys" of DNNs. Here, we provide a comprehensive review of the state-of-the-art XAI techniques concerning healthcare applications and discuss current challenges and future directions. We emphasize the strengths and limitations of each category, including image, tabular, and textual explanations, and explore a range of evaluation metrics for assessing the effectiveness of XAI solutions. Finally, we highlight promising opportunities for XAI research to enhance the acceptance of DNNs by the healthcare community.

关键词： Explainability medical data responsible AI deep neural networks pretable AI

来源：评论

学校读者我要写书评

暂无评论

Remote sensing scene classification with relation-aware dynamic graph neural networks

引用

ENGINEERING applications OF artificial INTELLIGENCE 2025年 150卷

作者： Huang, Qionghao Jiang, Fan Huang, Changqin Zhejiang Normal Univ Zhejiang Key Lab Intelligent Educ Technol & Applic 688 Yingbin Ave Jinhua 321004 Peoples R China Guangdong Polytech Normal Univ Sch Educ Sci 155 155 Xibei Rd Guangzhou 510665 Peoples R China Zhejiang Normal Univ Sch Comp Sci 688 Yingbin Ave Jinhua 321004 Peoples R China Zhejiang Univ 866 Yuhangtang Rd Hangzhou 310063 Peoples R China

Remote sensing scene classification (RSSC) is challenging due to the complexity and diversity of scenes. Existing methods need help to capture long-range and structural relationships among image regions, limiting their performance. This paper proposes a novel graph-based model that learns relation-aware dynamic graph representations for remote sensing scene classification tasks. The proposed model consists of three main components: Multi-Scale Feature Extraction (MSFE), Relation-Aware Graph processing (RAGP), and Scene Classification with Weighted Pooling (SCWP). MSFE uses a multi-scale feature extraction strategy to generate low-level feature nodes from remote sensing images. RAGP applies several cascaded graph processing blocks to dynamically learn the relations between nodes in high-level semantic spaces using relation-aware graph convolutional and node feature update operators. SCWP performs weighted pooling on the learned node features from RAGP to obtain global representations of remote sensing images and makes scene decisions using a fully feed-forward network-based classifier. We evaluate our model on three benchmark datasets and compare it with state-of-the-art RSSC methods. Our experimental results show that our model outperforms existing methods on all three datasets, demonstrating the effectiveness of a graph-based model with the proposed techniques for RSSC tasks.

关键词： Scene classification Graph neural networks Remote sensing Multi-scale feature extraction Graph convolution neural networks

来源：评论

学校读者我要写书评

暂无评论

Patch- and Class-Wise Hyperspectral Knowledge Learning: A Composite Consistency-Constrained Self-Ensemble Framework for Change Detection

引用

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 2025年 63卷

作者： Zhao, Xiaoyang Li, Siyao Liu, Xinyue Song, Chuanming Wang, Xianghai Liaoning Normal Univ Sch Geog Sci Dalian 116029 Peoples R China Liaoning Normal Univ Sch Comp & Artificial Intelligence Dalian 116029 Peoples R China Dalian Univ Sch Informat Engn Dalian 116622 Peoples R China

Obtaining fine land surface change information from multitemporal hyperspectral images (HSIs) is a key goal pursued in remote sensing image processing. Recently, HSI change detection (HSI-CD) methods based on convolutional neural networks (CNNs) have achieved surprising detection results. One of the reasons is the support of large-scale labeled samples for network learning. However, the existence of mixed pixels greatly increases the difficulty of HSI interpretation, resulting in accurate pixel-level labeling work with a heavy burden and unable to meet the needs of time-sensitive applications. For this reason, achieving stable and high-precision CD with fewer samples is a difficult issue in this field. To address the above problems, a composite consistency-constrained self-ensemble framework (C3SelF) for HSI-CD is proposed, to alleviate the problems of low detection accuracy and instability caused by small samples. The framework mainly comprises two lightweight networks with the same structure aiming at accelerating the model inference process and thus improving the processing timeliness. The composite learning mode implements patch-wise classification loss, class-wise consistency loss on labeled samples, and patch-wise consistency loss on unlabeled samples under a multilevel noise perturbation strategy, which improves the classification results and reduces the labeling cost. Moreover, to exploit the multidimensional features contained in HSIs, a lightweight selective spatial-spectral feature joint network (S3Net) is designed to overcome over-fitting, and to deeply mine the discriminative information in unlabeled samples, a new sample screening strategy is designed to ensure the stability of the network during training unlabeled samples. Extensive experiments prove that the proposed C3SelF outperforms the state-of-the-art (SOTA) methods at a sampling rate of 0.1%, reaching 93.44% Kappa and 97.26% overall accuracy (OA) on the Farmland dataset. The source code of the

关键词： Accuracy Training Feature extraction Hyperspectral imaging Transformers Data mining Vectors Semisupervised learning Convolutional neural networks Unsupervised learning Change detection (CD) convolutional neural networks (CNNs) deep learning hyperspectral image (HSI) processing semi-supervised learning

来源：评论

学校读者我要写书评

暂无评论

SoN: Selective Optimal Network for smartphone-based indoor localization in real-time

引用

EXPERT SYSTEMS WITH applications 2025年 272卷

作者： Lee, Kyungsu Lee, Haeyun Hwang, Jae Youn Jeonbuk Natl Univ Dept Comp Sci & Artificial Intelligence Jeonju 54896 South Korea Korea Univ Technol & Educ Sch Comp Sci & Engn Cheonan 31253 South Korea Daegu Gyeongbuk Inst Sci & Technol DGIST Dept Elect Engn & Comp Sci Interdisciplinary Studies Artificial Intelligence Daegu 42988 South Korea

Deep learning-based scene recognition algorithms have been developed for real-time application in indoor localization systems. However, owing to the slow calculation time resulting from the deep structure of convolutional neural networks, deep learning-based algorithms have limitations in the usage of real-time applications, despite their high accuracy in classification tasks. To significantly reduce the computation time of these algorithms and slightly improve their accuracy, we thus propose a path-selective deep learning network, denoted as Selective Optimal Network (SoN). The SoN selectively uses the depth-variable networks depending on anew indicator, denoted as the classification-complexity of a source image. The SoN reduces the prediction time by selecting optimal depth for the baseline networks corresponding to the input samples. The network was evaluated using two public datasets and two custom datasets for indoor localization and scene classification, respectively. The experimental results indicated that, compared to other deep learning models, the SoN exhibited improved accuracy and enhanced the processing speed by up to 78.59%. Additionally, the SoN was applied to a smartphone-based indoor positioning system in real-time. The results indicated that the SoN shows excellent performance for rapid and accurate classification in real-time applications of indoor localization systems.

关键词： Deep learning Semi-supervised learning Efficient learning Scene recognition Indoor localization

来源：评论

学校读者我要写书评

暂无评论

applications of artificial neural networks in image processing viiI (Proceedings of SPIE)

Applications of Artificial Neural Networks in Image Processi...

引用

2003年

作者： Nasser M Nasrabadi Aggelos Konstantinos Katsaggelos

来源：评论

学校读者我要写书评

暂无评论

image Classification with Regularized Likelihood Estimation and Large Interval Discriminant Dimensionality Reduction

引用

JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 2025年第2期41卷 385-400页

作者： Shen, Ying Zhu, Wei-hua Fu, Si-yong Wu, Guang-sheng Xinyu Univ Zhongxing Commun & Informat Coll Xinyu 338004 Peoples R China Xinyu Univ Sch Math & Comp Sci Xinyu 338004 Peoples R China

With the expansion of image application scale and the development of artificial intelligence industry, the demand for high precision and high robustness of image feature expression is gradually increasing. Convolutional neural networks (CNNS) have attracted much attention in the field of computer vision in recent years because of their excellent image feature learning ability. In order to further improve the accuracy and robustness of image classification, a second-order statistical convolutional neural network with multilayer fusion and cross-convolutional layer pooling are proposed in this study to obtain feature information of different dimensions in different convolutional layers. Due to the increase of computing and storage costs, an image classification method based on robust covariance of deep convolution features is proposed, and two regularized maximum likelihood estimators are introduced. Finally, the proposed method is tested by fine-grained experiments, image retrieval experiments and image classification experiments. The results show that MFBP (4,5) structure achieves 92.4%, 74.2% and 72.9% accuracy in finegrained image classification. Compared with MFBP (3,4,5), despite a slight decrease in accuracy, FPS performance is improved by more than 40%, demonstrating the advantages of this method in real-time applications. In addition, RMIE-VND and RMIE-IM methods improved by 4.1%, 8.4%, 7.8% and 7.1%, 12%, 9.8%, respectively, compared with LW methods on different data sets. The experimental results of image classification show that MFBP+RLME (VND+IM) has better performance and applicability than the current mainstream classification methods. This research is of great value for improving the efficiency and accuracy of image processing in big data environment, and provides a new research direction and practical tool for future real-time image application and computer vision system development.

关键词： image classification image dimensionality reduction convolutional neural net- work bilinear pooling regularized likelihood estimation

来源：评论

学校读者我要写书评

暂无评论

An integrated approach to apple leaf disease detection: leveraging convolutional neural networks for accurate diagnosis

引用

Multimedia Tools and applications 2025年 1-36页

作者： Rohith, Dasari Saurabh, Praneet Bisen, Dhananjay Manipal University Jaipur Rajasthan Jaipur India Madhav Institute of Technology & Science Deemed University M.P Gwalior India

Apple, a globally revered fruit with diverse cultivars, often falls victim to various leaf diseases that compromise crop yield and quality. Precise and timely disease identification is crucial for effective orchard management and timely intervention. Recent years has witnessed the exponential growth of artificial Intelligence, Machine Learning and its subsequent applications to various diversified applications including image processing. This paper intends to utilise the potential of Deep Learning to transform apple leaf disease detection and classification. In this work, the best Convolutional neural Network pretrained model for identification of apple leaf disease is found using rich visual dataset of 3,200 apple leaf images, meticulously curated to encompass four distinct categories: healthy leaves, those afflicted by apple scab, black rot, and cedar rust. Experimental results revealed that Residual neural Network achieved highest validation accuracy of 98.9%, significantly exceeding the performance of benchmark models like Visual Geometry Group 19 (96%), InceptionV3 (97%), EfficientNet (97.5%), Visual Geometry Group 16 (98.1%), MobileNet (98.2%) and the traditional baseline Convolutional neural Network (95%). The extensive experimentation also revealed that an 80:20 training-to-validation split yields the highest classification accuracy. The exceptional accuracy achieved by Residual neural Network powered model, coupled with the robustness of the image dataset, opens doors for the development of advanced orchard management tools with real-time identification and combating harmful leaf diseases. This can potentially revolutionize agricultural practices, minimizing crop losses and ensuring sustainable apple production. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.

关键词： Fruits

来源：评论

学校读者我要写书评

暂无评论

A Power-Efficient 0.5668 TOPS/W Digital Logic Accelerator Implemented Using 40-nm CMOS Process for Underwater Object Recognition Usage

引用

IEEE ACCESS 2025年 13卷 28123-28134页

作者： Wang, Chua-Chin Luo, Shih-Heng Wu, Hsin-Che Sangalang, Ralph Gerard B. Jou, Chewn-Pu Hsia, Harry Cho, Lan-Chou Natl Sun Yat sen Univ Dept Elect Engn Kaohsiung 80424 Taiwan Natl Sun Yat sen Univ Inst Undersea Technol Kaohsiung 80424 Taiwan Vel Tech Chennai 600062 Tamil Nadu India Batangas State Univ Natl Engn Univ Dept Elect Engn Batangas 4200 Philippines Batangas State Univ Natl Engn Univ Elect Syst Res Ctr Batangas 4200 Philippines Taiwan Semicond Mfg Co TSMC Hsinchu 30078 Taiwan

DNN (deep neural network) and CNN (convolution neural network) have been widely used in real-time artificial intelligent (AI) applications, particularly image or video recognitions, because they have been proved physically in many occasions. However, most prior AI hardware works either suffered from high on-silicon area cost or low usage thereof. This investigation presents a power efficient and high performance implementation of a digital logic accelerator (DLA) for the real-time underwater object recognition. The proposed DLA is also featured with 2-dimensional PE (processing element) array to increase the processing throughput by the enhancement of parallelism. The DLA design was realized and fabricated using TSMC 40-nm CMOS process. Not only the post-layout simulation results are shown, the on-silicon measurement outcome as well as the system validation in water are also demonstrated to prove the function correctness and the performance. The area efficiency (GOPS/mm(2)) is 4.562, and the power efficiency (TOPS/W) is 0.5668 on silicon, which both are the best to date.

关键词： Kernel Hardware Logic Convolution Throughput Random access memory Object recognition artificial intelligence Optical computing Optical buffering Deep neural networks (DNN) digital logic accelerator (DLA) high degree of parallelism underwater object recognition low power

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Deep Learning Framework for Automatic Detection of Brain Tumours Using Different Modalities

引用

IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE 2025年第2期9卷 1216-1225页

作者： Sahu, Adyasha Das, Pradeep Kumar Paul, Indraneel Meher, Sukadev VIT Bhopal Univ Sch Comp Sci Engn & Artificial Intelligence Bhopal 466114 India Vellore Inst Technol Sch Elect Engn Vellore 632014 India Natl Inst Technol Rourkela Dept Elect & Commun Engn Rourkela 769008 India

Nowadays, deep convolutional neural networks (DCNNs) are the focus of substantial research for classification and detection applications in medical image processing. However, the limited availability and unequal data distribution of publicly available datasets impede the broad use of DCNNs for medical image processing. This work proposes a novel deep learning-based framework for efficient detection of brain tumors across different openly accessible datasets of different sizes and modalities of images. The introduction of a novel end-to-end Cumulative Learning Strategy (CLS) and Multi-Weighted New Loss (MWNL) function reduces the impact of unevenly distributed datasets. In the suggested framework, the DCNN model is incorporated with regularization, such as DropOut and DropBlock, to mitigate the problem of over-fitting. Furthermore, the suggested augmentation approach, Modified RandAugment, successfully deals with the issue of limited availability of data. Finally, the employment of K-nearest neighbor (KNN) improves the classification performance since it retains the benefits of both deep learning and machine learning. Moreover, the effectiveness of the proposed framework is also validated over small and imbalanced datasets. The proposed framework outperforms others with an accuracy of up to $ 99.70\%$.

关键词： Tumors Databases Accuracy Magnetic resonance imaging Training Brain modeling Nearest neighbor methods Brain tumors convolutional neural network data augmentation deep learning machine learning MRI

来源：评论

学校读者我要写书评

暂无评论

Acoustic-to-hyper-spectral: real-time perimeter intrusion detection system monitoring through learnable filters and hyper-spectral image generation from distributed acoustic sensing systems

引用

OPTICS EXPRESS 2025年第3期33卷 4109-4126页

作者： Pierau, Ruth-emely Katsifolis, Jim Meehan, Alaster Rezatofighi, Hamid Stuckey, Peter J. Monash Univ Clayton Vic 3800 Australia Univ Melbourne ARC Training Ctr Optimisat Technol Integrated Meth Parkville Vic 3052 Australia AVA Risk Grp Co Future Fibre Technol 10 Hartnett CL Mulgrave Vic 3170 Australia

This paper presents an integrated distributed acoustic sensing (DAS) system with artificial intelligence to provide real-time system monitoring for fence perimeter and buried system applications. The DAS system is a Rayleigh backscatter based fibre optic sensing system that has been deployed in two real-world, commercial applications to detect acoustic wave propagation and scattering along perimeter lines, and classify intrusions accurately. What we believe to be three novel signal processing methods are proposed to train filters for automatically selecting frequency bands from the power spectrum and generating hyper-spectral images from the data gathered by the DAS system without expert knowledge. The hyper-spectral images are analyzed by a neural network based object detection model. The system achieves 81.8% accuracy on a fence perimeter installation and 60.4% accuracy on a buried system application in detecting and classifying various intrusion events. The evaluation interval of the integrated DAS system framework between event sensing and detection does not exceed 5 s. (c) 2025 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

关键词： Deep learning Hyperspectral imaging Machine learning neural networks Signal processing Wave propagation

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：