检索结果-内蒙古大学图书馆

Robust Stochastically-Descending Unrolled networks

IEEE TRANSACTIONS ON SIGNAL processing 2024年 72卷 5484-5499页

作者： Hadou, Samar Naderializadeh, Navid Ribeiro, Alejandro Univ Penn Dept Elect & Syst Engn Philadelphia PA 19104 USA Duke Univ Dept Biostat & Bioinformat Durham NC 27705 USA

Deep unrolling, or unfolding, is an emerging learning-to-optimize method that unrolls a truncated iterative algorithm in the layers of a trainable neural network. However, the convergence guarantees and generalizability of the unrolled networks are still open theoretical problems. To tackle these problems, we provide deep unrolled architectures with a stochastic descent nature by imposing descending constraints during training. The descending constraints are forced layer by layer to ensure that each unrolled layer takes, on average, a descent step toward the optimum during training. We theoretically prove that the sequence constructed by the outputs of the unrolled layers is then guaranteed to converge for in-distribution problems. We then analyze the generalizability to certain out-of-distribution (OOD) shifts in the optimization problems being solved. Our analysis shows that the descending nature imposed by the proposed constraints is transferable under these distribution shifts, subject to a generalization error, thereby providing the unrolled networks with OOD robustness. We numerically assess unrolled architectures trained with the proposed constraints in two different applications, including the sparse coding using learnable iterative shrinkage and thresholding algorithm (LISTA) and image inpainting using proximal generative flow (GLOW-Prox), and demonstrate the performance and robustness advantages of the proposed method.

关键词： Convergence Standards Trajectory Training Perturbation methods Optimization Signal processing algorithms Iterative algorithms Robustness artificial neural networks Algorithm unrolling deep unfolding learning to optimize constrained learning

来源：评论

学校读者我要写书评

暂无评论

Systematic literature review of AI algorithms applied to unmanned aerial vehicle images

引用

INTERNATIONAL JOURNAL OF image AND DATA FUSION 2025年第1期16卷

作者： El Khadir, Kenza Ait Fadil, Abdelhamid El Brirchi, El Hassan Hassania Sch Publ Works LaGes Lab Km 7 Jadida RdBP 8108 Casablanca Morocco

artificial Intelligence (AI) combined with image processing has shown significant improvements through new techniques such as Machine Learning (ML) models. This paper introduces the key methods and algorithms used for Drone image processing. We discuss the benefits and limitations of using ML models instead of classical techniques. Our goal is to classify, categorize and describe the methods that are used in realistic settings of diverse domains of applications. We conducted a systematic literature review where systems presented in the papers were analysed based on their domain, task, technology, and efficiency. By extensively reviewing the existing literature, we successfully identified key themes and trends that emerged across the various research questions. The overall findings of the research emphasise the potential of AI and drone imagery in numerous fields. However, the review also uncovered several challenges that necessitate attention, such as issues related to data quality and the requirement for more advanced AI algorithms. The paper outlines significant innovations in the field and offers recommendations for future research directions. By highlighting cross-disciplinary insights, it delves into methodological approaches, exploring commonalities in AI algorithms and UAVs technologies.

关键词： artificial intelligence ML image processing computer vision UAVs (unmanned aerial vehicles) drones image recognition fault detection object detection inspection agriculture environmental pollution CNNs (convolutional neural networks) infrastructure

来源：评论

学校读者我要写书评

暂无评论

The State of Retinal image Analysis: Deep Learning Advances and applications

引用

IEEE ACCESS 2025年 13卷 65066-65093页

作者： Soomro, Toufique Ahmed Bin Ubaid, Faisal Zardari, Shehnila Shah, Ahmed Ali Alansari, Marwah M. Nagro, Shimaa A. Charles Univ Prague Artificial Intelligence & Cyber Futures Inst Bathurst NSW 2795 Australia Univ Larkano Dept Elect Engn Larkana 77280 Pakistan Sukkur IBA Univ Comp Sci Dept Sukkur 65200 Pakistan NED Univ Engn & Technol Dept Software Engn Karachi 75270 Pakistan Sukkur IBA Univ Elect Engn Dept Sukkur 65200 Pakistan Saudi Elect Univ Coll Comp & Informat Riyadh 11673 Saudi Arabia

This study explores deep learning techniques that can help diagnose eye conditions, focusing on diabetic retinopathy (DR). By using cutting-edge technology in image processing and artificial intelligence, it aims to better understand and classify various eye issues like microaneurysms, hemorrhages, and others. Despite progress, challenges like poor image quality and noise persist. The paper stresses the need for improved methods to detect small blood vessels while reducing interference. It also addresses the complexity of uneven lighting in extracting blood vessel images from the retina. The main contribution of this study is its thorough examination of existing methods and its detailed investigation into using deep learning for detecting retinal blood vessels. It highlights the effectiveness of different deep learning models and their potential in accurately classifying DR grades. This research underscores the importance of grading methods in analyzing medical images, especially retinal images, to enhance the accuracy of diagnosing eye diseases.

关键词： Retina Biomedical imaging Blood vessels Feature extraction Accuracy image segmentation Hemorrhaging Deep learning Classification algorithms Retinal vessels Diabetic retinopathy retinal blood vessels color fundus images deep learning microaneurysms hemorrhages exudates image processing techniques convolutional neural networks medical image analysis

来源：评论

学校读者我要写书评

暂无评论

networks of random lasers: current perspective and future challenges [Invited]

引用

OPTICAL MATERIALS EXPRESS 2023年第4期13卷 1060-1076页

作者： Consoli, Antonio Caselli, Niccolo Lopez, Cefe CSIC Inst Ciencia Mat Madrid ICMM Sor Juana Ines Cruz 3 Madrid 28049 Spain Univ Rey Juan Carlos ETSI Telecomunicac Camino Molino 5 Madrid 28942 Spain Univ Complutense Madrid Dept Quim Fis Ciudad Univ S-N Madrid Spain

artificial neural networks are widely used in many different applications because of their ability to deal with a range of complex problems generally involving massive data sets. These networks are made up of nodes, connections, and nonlinear response connections, which are typically implemented as software code running on ordinary electronic computers. In such systems, electrons, with their advantages and drawbacks, are in charge of storing, processing, and transmitting information. Signal processing in the optical domain can provide ultrafast, parallel operation, nonlinear dynamics, and high energy efficiency, making photonics a suitable technology for the realization of neuroinspired computing platforms. This advantage stimulated the development of photonics neural networks based on single and multiple lasers with classical optical cavities. Recently, networks made of random lasers emerged as a novel concept that uses randomly placed scattering elements to create nonlinearity and complexity in photonics neural networks. In this review paper, we present the general framework for networks of coupled lasers, discuss recent advances in networks of random lasers, and outline future directions in this area. We also examine the challenges and limitations of using random lasers in photonic networks, as well as potential solutions. By harnessing the properties of random lasers, such as their unique spectral characteristics in pulsed emission mode and their robustness against noise, networks of interacting random lasers can explore new and exciting possibilities for photonics technology that could find applications in a variety of fields, including image recognition and encryption.(c) 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

关键词： Diode laser arrays Fiber lasers image encryption neural networks Optical feedback Random lasers

来源：评论

学校读者我要写书评

暂无评论

Pixel-level block classification and crack detection from 3D reconstruction models of masonry structures using convolutional neural networks

引用

ENGINEERING STRUCTURES 2024年 310卷

作者： Loverdos, Dimitrios Sarhosis, Vasilis Univ Leeds Sch Civil Engn Woodhouse Ln Leeds LS2 9DY England

Inspection and documentation of masonry structures is a time-consuming and expensive process that heavily relies on an engineer's expertise. This paper introduces a computer vision-based method that automates the creation of classified point clouds from 3D reconstruction models obtained through photogrammetry and/or laser scanning. By leveraging Convolutional neural networks (CNN) on 3D renders, the proposed methodology can be used to classify structural features (i.e., blocks, mortar, cracks) with greater accuracy and less effort than conventional methods. Moreover, this approach is flexible and can include further classifications by incorporating additional CNN models, allowing broader applications across various materials (i.e., concrete, steel, timber) and defects. Additionally, a precise methodology for accurate crack quantification featuring a manual annotation tool was introduced which was validated using outputs from a full-scale masonry arch bridge test carried out in the laboratory. The results emphasize the robustness of the classification approach and highlight the useful geometric information that can be gained from full-scale masonry structures. The proposed approach has the potential to use image data from UAVs/static cameras and revolutionize the way we document and inspect structures in the future.

关键词： Masonry image processing Documentation Inspection artificial intelligence Convolutional neural networks Reality mesh 3D mesh

来源：评论

学校读者我要写书评

暂无评论

Dual-function switchable terahertz metamaterial device assisted by a deep neural network

引用

JOURNAL OF THE OPTICAL SOCIETY OF AMERICA A-OPTICS image SCIENCE AND VISION 2025年第4期42卷 529-541页

作者： Ren, Jiali Zhang, Ting Li, Jing Shanxi Vocat Univ Engn Sci & Technol Sch Comp Engn Jinzhong 030619 Peoples R China Shanxi Vocat Univ Engn Sci & Technol Sch Informat Engn Jinzhong 030619 Peoples R China North Univ China Sch Instrument & Elect Taiyuan 030051 Peoples R China

A dual-functional switchable terahertz metamaterial (THz MM) device with broadband absorption and polarization conversion capabilities is proposed, leveraging the insulator-to-metal phase transition of vanadium dioxide (VO2). In its metallic state, the device functions as an absorber, achieving a bandwidth of 3.93 THz within the frequency range of 2.53-6.46 THz and maintaining an absorption rate of >= 90%. In its insulating state, the device operates as a polarization converter, providing two conversion modes: linear-to-cross-polarization (LTX) over the range of 1.17-3.63 THz, and linear-to-circular polarization (LTC) within the ranges of 1.04-1.14 THz and 3.71-3.87 THz. This study presents a significant advancement over conventional THz devices by overcoming the limitations of single-functionality through innovative design strategies and offering what we believe is a novel approach to multifunctional integration. Leveraging the switchable properties of VO2, the device achieves outstanding performance in both broadband absorption and polarization conversion. Furthermore, deep neural networks (DNNs) significantly reduce the computational time required by traditional optimization methods by efficiently navigating multi-dimensional parameter spaces. DNNs can also capture complex nonlinear relationships between parameters that are often difficult to model with conventional techniques. This highlights the potential of artificial intelligence in the design of functional materials. With these capabilities, the device shows promising applications in fields such as THz stealth technology, energy harvesting, THz communication, polarization imaging, and optical information processing. (c) 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), artificial Intelligence (AI) training, and similar technologies, are reserved.

关键词： Anderson localization Information processing Metamaterial devices Mode conversion neural networks Polarimetric imaging

来源：评论

学校读者我要写书评

暂无评论

A Quality-Aware Voltage Overscaling Framework to Improve the Energy Efficiency and Lifetime of TPUs Based on Statistical Error Modeling

引用

IEEE ACCESS 2024年 12卷 92181-92197页

作者： Senobari, Alireza Vafaei, Jafar Akbari, Omid Hochberger, Christian Shafique, Muhammad Tarbiat Modares Univ Dept Elect & Comp Engn Tehran 14115111 Iran Univ Tehran Sch Elect & Comp Engn Tehran 14395515 Iran Tech Univ Darmstadt Dept Elect Engn D-64289 Darmstadt Germany NYU Abu Dhabi NYU AD Div Engn Abu Dhabi U Arab Emirates

Deep neural networks (DNNs) are a type of artificial intelligence models that are inspired by the structure and function of the human brain, designed to process and learn from large amounts of data, making them particularly well-suited for tasks such as image and speech recognition. However, applications of DNNs are experiencing emerging growth due to the deployment of specialized accelerators such as the Google's Tensor processing Units (TPUs). In large-scale deployments, the energy efficiency of such accelerators may become a critical concern. In the voltage overscaling (VOS) technique, the operating voltage of the system is scaled down beyond the nominal operating voltage, which increases the energy efficiency and lifetime of digital circuits. The VOS technique is usually performed without changing the frequency resulting in timing errors. However, some applications such as multimedia processing, including DNNs, have intrinsic resilience against errors and noise. In this paper, we exploit the inherent resilience of DNNs to propose a quality-aware voltage overscaling framework for TPUs, named X-TPU, which offers higher energy efficiency and lifetime compared to conventional TPUs. The X-TPU framework is composed of two main parts, a modified TPU architecture that supports a runtime voltage overscaling, and a statistical error modeling-based algorithm to determine the voltage of neurons such that the output quality is retained above a given user-defined quality threshold. We synthesized a single-neuron architecture using a 15-nm FinFET technology under various operating voltage levels. Then, we extracted different statistical error models for a neuron corresponding to those voltage levels. Using these models and the proposed algorithm, we determined the appropriate voltage of each neuron (the voltage level of each column of the X-TPU). Results show that running a DNN on X-TPU can achieve 32% energy saving for only 0.6% accuracy loss.

关键词： Voltage control Accuracy artificial neural networks Energy efficiency Approximate computing Power demand Computer architecture Statistical analysis Error analysis Voltage overscaling deep neural networks approximate computing accuracy TPU energy efficiency statistical error analysis

来源：评论

学校读者我要写书评

暂无评论

Tchebichef Transform Domain-Based Deep Learning Architecture for image Super-Resolution

引用

IEEE TRANSACTIONS ON neural networks AND LEARNING SYSTEMS 2024年第2期35卷 2182-2193页

作者： Kumar, Ahlad Singh, Harsh Vardhan Khare, Vijeta Dhirubhai Ambani Inst Informat & Commun Technol D Dept Informat & Commun Technol Gandhinagar 382007 Gujarat India Dhirubhai Ambani Inst Informat & Commun Technol D Gandhinagar 382007 Gujarat India Adani Inst Infrastruct Engn Adalaj 382421 Gujarat India

Recent advances in the area of artificial intelligence and deep learning have motivated researchers to apply this knowledge to solve multipurpose applications in the area of computer vision and image processing. Super-resolution (SR), in the past few years, has produced remarkable results using deep learning methods. The ability of deep learning methods to learn the nonlinear mapping from low-resolution (LR) images to their corresponding high-resolution (HR) images leads to compelling results for SR in diverse areas of research. In this article, we propose a deep learning-based image SR architecture in the Tchebichef transform domain. This is achieved by integrating a transform layer into the proposed architecture through a customized Tchebichef convolutional layer (TCL). The role of TCL is to convert the LR image from the spatial domain to the orthogonal transform domain using Tchebichef basis functions. The inversion of the transform mentioned earlier is achieved using another layer known as the inverse TCL (ITCL), which converts back the LR images from the transform domain to the spatial domain. It has been observed that using the Tchebichef transform domain for the task of SR takes the advantage of high and low-frequency representation of images that makes the task of SR simplified. Furthermore, a transfer learning-based approach is adopted to enhance the quality of images by considering Covid19 medical images as an additional experiment. It is shown that our architecture enhances the quality of X-ray and CT images of COVID-19, providing a better image quality that may help in clinical diagnosis. Experimental results obtained using the proposed Tchebichef transform domain SR (TTDSR) architecture provides competitive results when compared with most of the deep learning methods employed using a fewer number of trainable parameters.

关键词： Transforms Deep learning Computer architecture Convolutional neural networks COVID-19 Training Kernel Convolutional neural network (CNN) COVID-19 deep learning image super-resolution (SR) Tchebichef moments

来源：评论

学校读者我要写书评

暂无评论

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

引用

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

Deep spiking neural networks based on model fusion technology for remote sensing image classification

引用

ENGINEERING applications OF artificial INTELLIGENCE 2025年 142卷

作者： Niu, Li-Ye Wei, Ying Zhao, Liping Hu, Keli Zhejiang Chinese Med Univ Zhejiang Prov Hosp Chinese Med Dept Radiol Affiliated Hosp 1 Hangzhou 310006 Peoples R China Northeastern Univ Coll Informat Sci & Engn Shenyang 110819 Peoples R China Shaoxing Univ Dept Comp Sci & Engn Shaoxing 312000 Peoples R China

The spiking neural network (SNN) based on brain inspiration, as the third-generation neural network, has attracted great research interest due to its ultra-low power event-driven data processing method. How to obtain high-accuracy deep networks has always been a challenge in the field of SNN. At present, there are two main methods for training deep spiking neural networks (SNNs). The first is the spike-based spatiotemporal backpropagation (STBP), and the other indirect is to convert the trained artificial neural network (ANN) into SNN (ANN-SNN). Algorithms that directly train SNNs are usually inefficient, while the ANN-SNN-based models require along inference time and also suffer from performance losses. The model fusion technology (MFT) proposed in this paper, combined with ANN-SNN and STBP, provides a new training paradigm for obtaining deep SNNs. We propose a bilateral multi-strength integrate-and-fire (BM-IF) spiking neuron for ANN-SNN to simplify the conversion operation. Under the same network architecture and encoding time window, the ResNet34 based on our algorithm achieved State-of-the-art (SOTA) in imageNet. At the same time, we combined transfer learning methods to achieve SOTA results on three remote sensing scene classification datasets. Our results also indicate that the MFT-based SNN outperforms the optimal inference accuracy provided by ANNSNN in terms of network architecture and dataset inference time steps reduced by 512-1250 times. It can be seen that this algorithm provides a low-latency and high-precision training scheme for the development of deep SNN.

关键词： Spiking neural network Model fusion technology Low-latency High-precision Transfer learning

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：