检索结果-内蒙古大学图书馆

ieee Symposium on Nuclear Science (NSS/MIC)

作者： M. K. Schütz S. J. Useche G. A. Roque J. Jakubek S. Procz M. Fiederle Freiburg Materials Research Centre (FMF) Albert-Ludwigs-University Freiburg Freiburg im Breisgau Baden-Württemberg Germany ADVACAM s.r.o Prague Czech Republic

This work introduces a new approach to data processing and recording with photon-counting Timepix3 detectors, addressing challenges related to high count rates and large amounts of data when measuring in data-driven readout mode. The 12x1 Timepix3 detector array combined with a 1mm CdTe sensor allows an event-based, data-driven readout mode. The energy information, arrival time, and position of each registered photon are stored in the ASIC. This data-driven readout causes a challenging amount of data which requires new techniques to handle long-lasting scans performed with a fully spectroscopic working detection mode. Our strategy limits the number of collected photons per pixel based on a selectable signal-to-noise ratio. It applies a simple but effective clustering algorithm that pre-processes data before storing it on the hard drive. This approach significantly reduces the amount of data in two different steps, allowing the performance of fully spectroscopic CT scans using standard laboratory computer setups. To measure the absorption of an underlying object considering the active time of each pixel from the start of the recording until it reaches the total amount of desired counts, was introduced. The results reveal the potential of the Timepix3 detector array with CdTe sensors in performing spectroscopic CT scans and provide insights into future applications in materials science and other relevant fields.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Writer Identification: Deep Learning with ResNet50 for Offline Urdu Handwritten Documents

Writer Identification: Deep Learning with ResNet50 for Offli...

引用

international conference on image Information processing

作者： Syed Tufael Nabi Paramjeet Singh Munish Kumar Department of Computer Science and Engineering GZS Campus College of Engineering and Technology Maharaja Ranjit Singh Punjab Technical University Bathinda Punjab INDIA Department of Computational Sciences Maharaja Ranjit Singh Punjab Technical University Bathinda Punjab INDIA

ISBN: (数字)9798350371406

ISBN: (纸本)9798350371413

The field of writer identification, a crucial aspect of document analysis, has undergone a significant evolution through the incorporation of deep learning models. This paper presents an innovative writer identification system, meticulously designed around the ResNet50 deep learning architecture, with a specific emphasis on its adaptability for processing offline Urdu handwritten documents. The system underwent rigorous training and thorough testing using an Urdu handwritten dataset that encompasses contributions from 418 Urdu writers. Our ResNet50-based model demonstrates an exceptional level of accuracy, achieving an impressive identification rate of 99.26%. Within this paper, we delve into a comprehensive examination of the model’s architecture, intricate feature extraction methods, and the subtle intricacies of its training methodologies, all of which collectively contribute to this outstanding performance. Beyond pushing the boundaries of current writer identification capabilities, the application of deep learning, particularly the ResNet 50 model, to Urdu handwritten documents holds significant promise across various domains, including historical document preservation, forensic analysis, and robust authentication. This paper not only introduces a state-of-the-art writer identification system but also underscores the broader impact of deep learning in the field of document analysis. The remarkable accuracy demonstrated by our ResNet50-based model highlights the transformative potential of deep learning, enhancing the precision and effectiveness of writer identification tasks. Furthermore, it illuminates new avenues for research and practical applications, spanning diverse linguistic and cultural contexts.

关键词： Deep learning Training Analytical models Adaptation models Text analysis Information processing Linguistics

来源：评论

学校读者我要写书评

暂无评论

Recognition of Hybrid Digital Watermarking Techniques using Digital image processing using Deep Learning Techniques

Recognition of Hybrid Digital Watermarking Techniques using ...

引用

Augmented Intelligence and Sustainable Systems (ICAISS), international conference on

作者： Chilumala Indira RM Noorullah Praveen Kumar Balguri Sowjanya Krishna Pinisetty Department of Computer Science and Engineering Institute of Aeronautical Engineerring Dundigal Hyderabad India Department of Aeronautical Engineering Institute of Aeronautical Engineerring Dundigal Hyderabad India Institute of Fundamental and Frontier Science University of Electronic Science and Technology of China Chengdu China

Digital watermarking is the procedure to identify the source and creator of the data. This is a form of digital media marker such as video, audio or image format. It helps to protect the data from various external unauthorized access. This is done through embedded digital information in the carrier signal to identify ownership and copyright protection. Security and privacy concerns need to be improved due to the rapid increase in the transmission and receiving of digital data. Various sectors such as the navy, military, intelligence, and police are sending and receiving numerous data through online mode. The existing system utilizes Arnold map with QWT-DCT techniques which are unable to obtain invisibility in attack detection. To overcome the limitations of the existing system the proposed system is implemented. They are highly confidential and need to be secured and protected thus hybrid digital watermarking techniques such as nested forms of watermarking are employed. If the third parties tend to make any modification, then it is termed as an attack. The removal of digital watermarking by third parties through certain modifications are denoted through copyright protection application. For security applications and to avoid various types of attacks, Discrete wavelet Transform (DWT) and Non-sampled Contourlet Transform (NSCT) are employed. Security and privacy is obtained through the Quaternion Fourier Transform (QFT) based algorithm with human speech recognition. The pixel displacement with pixel encryption is enhanced to save the privacy of images. Digital watermarking requires security, robustness, watermarking keys and computational cost parameter. Thus to protect the data by neglecting various conventional attacks, digital watermarking techniques are implemented through deep convolutional neural networks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Towards a standard for PET raw data

Towards a standard for PET raw data

引用

ieee Symposium on Nuclear Science (NSS/MIC)

作者： K. Thielemans A. Kesner E. Asma J. Cabello M. Cook M. Hansen J. Jones N. A. Karakatsanis E. K. S. Leung P. Markiewicz S. Prevrhal A. Rahmim B. Saboury J. Stairs S. Stute H. Tashima G. Wells Institute for Nuclear Medicine University College London London Centre for Medical Image Computing University College London London United Kingdom Memorial Sloan Kettering Cancer Center New York Washington United States of America Canon Medical Research USA Vernon Hills Washington United States of America Siemens Medical Solutions USA Inc. Knoxville Tennessee United States of America GE HealthCare Waukesha Wisconsin United States of America Microsoft Research Redmond Washington United States of America Weill Cornell Medical College Cornell University New York New York United States of America UIH America Inc. Houston Texas United States of America Dept Medical Physics and Bioengineering University College London London United Kingdom Philips Research Hamburg Germany University of British Columbia Vancouver British Columbia Canada National Institutes of Health (NIH) Clinical Center Bethesda Maryland United States of America CRCI2NA Nantes University Hospital Nantes France National Institutes for Quantum Science and Technology Chiba Japan University of Ottawa Heart Institute Ottawa Ontario Canada

As the capabilities of computer hardware and software progresses, the importance of standardized data is becoming more relevant. While DICOM is widely used as a standard for PET images, no agreed format exists for PET raw data. In addition, all current vendor formats differ, to accommodate different architectures and processing *** international working group comprised of PET imaging experts from academia and industry was formed in 2022 to propose a new standard format. Efforts include establishing the key informational elements to include in the standard, data container formats, and integration initiatives. An important feature of the proposal is the use of a meta-language (called Yardl) for defining data structure and protocols for accessing, transferring, and storing data, allowing logical separation of data elements and container formats. All portions of the format, tools for accessing the data in the standardized format, and example data will be made open source and publicly *** this work, we will provide a status report of the progress of the working group. The effort is ongoing, and we welcome interest, feedback, and support from the greater community. We hope that this standard will facilitate a new paradigm for PET innovation, including new opportunities for inter-scanner and inter-vendor harmonization and AI applications, and aide in the advancement of novel PET applications and analysis tools.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Crop-Guided Neural Network Segmentation of High-Resolution Skin Lesion images 2nd

Crop-Guided Neural Network Segmentation of High-Resolution ...

引用

2nd international conference on Digital Transformation in Education and Artificial Intelligence applications, MoStart 2024

作者： Benčević, Marin Habijan, Marija Galić, Irena Faculty of Electrical Engineering Computer Science and Information Technology Osijek Josip Juraj Strossmayer University of Osijek Osijek Croatia Department of Telecommunications and Information Processing TELIN-GAIM Ghent University Ghent Belgium

ISBN: (纸本)9783031620577

Medical images are often exceedingly large in width and height, limiting the maximum batch size when training convolutional neural networks and requiring models with a large number of parameters. Typically, images are uniformly downsampled, leading to losing fine-detailed information. Instead of uniformly downsampling images, we introduce a two-stage end-to-end segmentation network utilizing image crops to reduce network input size. Initially, a uniformly downscaled image is first segmented with a rough segmentation module, and the rough segmentation is used as a saliency map to crop the original high-resolution image to a region of interest. This crop is then re-segmented with a fine segmentation module. Our method’s effectiveness is demonstrated in segmenting lesion boundaries in clinical images across two datasets. We establish that this technique maintains comparable segmentation quality to a baseline model while reducing the network input size. Furthermore, our approach enhances the robustness of segmentation outcomes with smaller input sizes, outperforming uniformly downscaled images and baseline models. This improvement is consistent in both in-sample and out-of-sample evaluations. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Crops

来源：评论

学校读者我要写书评

暂无评论

WMDRS: Workload-Aware Performance Model Based Multi-Task Dynamic-Quota Real-Time Scheduling for Neural processing Units

WMDRS: Workload-Aware Performance Model Based Multi-Task Dyn...

引用

international conference on Parallel and Distributed Systems (ICPADS)

作者： Chong Liu Yuan Yao Yi Dang Gang Yang Wei Jia Xinyu Tian Xingshe Zhou School of Computer Science Northwestern Polytechnical University Xi’an China Department of Aviation Electronics Xian Aisheng Technology Group Co.Ltd Xi’an China

To further improve the capacity of airborne embedded system for dealing with deep learning (DL) applications and reduce overall power consumption, it is necessary to equip Neural processing Units (NPUs). Comparing with the cloud system, the airborne embedded system usually has a fixed application set, but strict real-time constraints. Unfortunately, the inherent NPU scheduler does not consider the application priority, which cannot provide the sufficient real-time capability for the airborne embedded system. At present, there are few researches on multi-task real-time scheduling for NPUs. Therefore, we propose WMDRS, a workload-aware performance model multi-task dynamic-quota real-time scheduling for Neural processing Units. The NPU performance model based on workload-awareness can accurately predict the remaining execution time of a task, which is running concurrently with other tasks on NPU. The multi-task dynamic-quota real-time scheduling algorithm can provide the approximate preemption by dynamically adjusting NPU computing resources for active applications. In addition, we implement a prototype NPU scheduler without any hardware extension. Furthermore, the proposed NPU performance model and real-time scheduling algorithm are evaluated in realistic application sets. Experimental results demonstrate that WMDRS can achieve low prediction error and high scheduling success ratio.

关键词： Embedded systems Scheduling algorithms Computational modeling Atmospheric modeling Prototypes Predictive models Multitasking

来源：评论

学校读者我要写书评

暂无评论

Robot Cleaner Localization Using RFID and Vision System

Robot Cleaner Localization Using RFID and Vision System

引用

AFRICON conference

作者： Farida Helaly Sarah Bassiouny Yasmine Zaghloul Department of Mechanical Engineering German International University Egypt

Localization is a fundamental element required in various applications across fields such as vehicle navigation, smart factories, automation systems, and product shipping services. This paper discusses the fusion of two popular systems- RFID and vision systems to maximize their individual advantages and minimize their respective drawbacks. Using the RFID system, the Received Signal Strength Indicator (RSSI) and tag ID are collected using the fingerprinting technique by utilizing two different types of antennas the far-field antenna (transmits and receives in range of meters) and the near-field antenna (transmits and receives in range of centimetres) and deploying passive tags. After comparing the two in terms of location error accuracy, it was found that the near-field antenna achieved significantly better results, with an accuracy level 55 % better than that of the far-field antenna. In the vision system, a fixed camera captures images for image processing algorithms to detect and localize the cleaner using the appearance-based approach. The accuracy level of this system alone was found to be 81.8%. However, upon fusing both systems, the accuracy levels improved to 90.9 % as each system contributed to the other to reduce each system's drawbacks eliminating the time incompatibility present in the vision system due to computational cost and the sensitivity of the RFID system from the materials used in the environment also, to get the best of RFID system two types of antennas where compared to choose the most accurate, the fusion resulted in an improved and a decreased average error localization system.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Production of MusicXML from Locally Inclined Sheet-music Photo image by Using Measure-based Multi-modal Deep-learning-driven Assembly Method

引用

Transactions of the Japanese Society for Artificial Intelligence 2023年第3期38卷

作者： Shishido, Tomoyuki Fati, Fehmiju Tokushige, Daisuke Ono, Yasuhiro Kumazawa, Itsuo Biomedical AI Research Unit Tokyo Institute of Technology Shishido&Associates Japan SK Intellectual Property Law Firm Japan Enspirea LLC Japan Laboratory for Future Interdisciplinary Research of Science and Technology Tokyo Institute of Technology Japan

Deep learning has been applied to optical music sheet recognition (OMR). However, OMR processing from various sheet-music images still lacks precision to be widely applicable. We propose a measure-based multimodal deep-learning-driven assembly (MMdA) method enabling end-to-end OMR processing from various images including inclined photo images. Using this method, measures are extracted using a deep-learning model, aligned, and resized to be used for inference of given musical-symbol components by using multiple deep-learning models in sequence or in parallel. The use of each standardized measure enables efficient training of the deep-learning models and accurate adjustment of five staff lines in each measure, which enables locally inclined sheet-music images to be precisely positioned. Thus, a score can be reproduced from the inclined image with the proposed MMdA method while current OMR applications cannot. Multiple musical-symbol-component deep-learning feature-category models with a small number of feature types can represent a diverse set of notes and other musical symbols including chords. The proposed MMdA method provides a solution to end-to-end OMR processing and enhances the utility of OMR of mobile phone-based sheet-music photo images. © 2023, Japanese Society for Artificial Intelligence. All rights reserved.

关键词： Music

来源：评论

学校读者我要写书评

暂无评论

Computational Efficient Brain PET image Denoising by Diffusion Probabilistic Models

Computational Efficient Brain PET Image Denoising by Diffusi...

引用

ieee Symposium on Nuclear Science (NSS/MIC)

作者： A. Sanaat H.R. Mazandarani Z. Mansouri M. Amini Y. Salimi I. Shiri H. Zaidi Division of Nuclear Medicine and Molecular Imaging Geneva University Hospital Geneva Genève Switzerland Department of Computer Engineering Amirkabir University of Technology Tehran Iran Islamic Republic of

Given the accomplishments of Denoising Diffusion Probabilistic Models (DDPMs) in a diverse range of generative artificial intelligence applications, including those related to medical imaging, their limited training and inference efficiency remain critical shortcomings. To address this issue, we propose an effective approach for generating high-quality, full-dose (FD) Positron Emission Tomography (PET) images from low-resolution, noisy low-dose (LD) PET images. Instead of initiating the denoising process by sampling from a Gaussian distribution, as in conventional DDPMs, our method commences with the output of a trained Variational Autoencoder (VAE) model, thereby enhancing computational efficiency. To assess the efficacy of our approach, we employed a dataset consisting of 75 18 F-Flutemetamol PET raw data from patients diagnosed with neurodegenerative diseases. We reconstructed FD images and corresponding LD images, which contained only 5% of the total events. Subsequently, we conducted a quantitative analysis using well-established metrics, including peak signal-to-noise ratio (PSNR), structural similarity index metric (SSIM), and root mean square error (RMSE). Our results revealed that the proposed DDPM-based denoising model can enhance SSIM and PSNR values up to 26% and 31%, respectively. This approach facilitates a reduction in the injection dose administered to patients by up to 95% while simultaneously improving quantitative and qualitative factors.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Skin Cancer Detection Using Deep Learning Technique

Skin Cancer Detection Using Deep Learning Technique

引用

Innovation in Technology (INOCON), ieee international conference for

作者： Dilli Hemalatha Kaile Nishi Latha Pandala.Madhavi Latha Department of Information Technology Velagapudi Ramakrishna Siddhartha Engineering College Vijayawada AP India

One of the most deadly forms of cancer is skin cancer. Due to its relatively high difficult and expensive diagnosis, as well as human subjectivity and challenges, dermatological diseases are one of the most pressing medical issues of the 21st century. When it comes to lethal diseases like melanoma, early detection is crucial for assessing the likelihood of recovery. Skin cancer is brought on by DNA breaks in skin cells that are not repaired, which results in genetic faults or mutations. Since skin cancer is more treatable in its early stages and has a tendency to gradually spread to other body parts, it is best detected at an early stage. Early detection of skin cancer symptoms is necessary due to the rising incidence of cases, high death rate, and expensive medical treatments. Given the severity of these difficulties, researchers have created a number of early. The automatic classification of melanoma images utilizing machine learning and computing vision methods has been the subject of numerous investigations. The segmentation of the skin lesion and the features identified for the classification approach have a significant impact on the classification performance of traditional machine learning and computer vision detection techniques for skin cancer, despite the fact that these studies yield encouraging results. Lesion characteristics including symmetry, color, size, shape, and others are used to distinguish between benign and malignant skin cancer. Experts believe the use of automated approaches will aid in early diagnosis, particularly when a batch of photos has a variety of diagnosis. In-depth investigation of deep learning techniques for skin cancer early detection is provided in this study. The three stages of our model creation are data collection and augmentation, model construction, and prediction. To improve the structure and increase accuracy to 84%, we used image processing tools in conjunction with the Inception V3 method.

关键词： image segmentation Melanoma Predictive models Feature extraction Skin Real-time systems Lesions

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：