检索结果-内蒙古大学图书馆

DMAeEDNet: Dense Multiplicative Attention Enhanced Encoder Decoder Network for Ultrasound-Based Automated Breast Lesion Segmentation

引用

IEEE ACCESS 2024年 12卷 60541-60555页

作者： Saini, Manali Afrin, Humayra Sotoudehnia, Setayesh Fatemi, Mostafa Alizad, Azra Mayo Clin Coll Med & Sci Dept Radiol Rochester MN 55905 USA Mayo Clin Coll Med & Sci Dept Physiol & Biomed Engn Rochester MN 55905 USA

Automated and precise segmentation of breast lesions can facilitate early diagnosis of breast cancer. Recent research studies employ deep learning for automatic segmentation of breast lesions using ultrasound imaging. Numerous studies introduce somewhat complex modifications to the well adapted segmentation network, U-Net for improved segmentation, however, at the expense of increased computational time. Towards this aspect, this study presents a low complex deep learning network, i.e., dense multiplicative attention enhanced encoder decoder network, for effective breast lesion segmentation in the ultrasound images. For the first time in this context, two dense multiplicative attention components are utilized in the encoding layer and the output layer of an encoder-decoder network with depthwise separable convolutions, to selectively enhance the relevant features. A rigorous performance evaluation using two public datasets demonstrates that the proposed network achieves dice coefficients of 0.83 and 0.86 respectively with an average segmentation latency of $19 ms$ . Further, a noise robustness study using an in-clinic recorded dataset without pre-processing indicates that the proposed network achieves dice coefficient of 0.72. Exhaustive comparison with some commonly used networks indicate its adeptness with low time and computational complexity demonstrating feasibility in real time.

关键词： Lesions Breast image segmentation Ultrasonic imaging Decoding Feature extraction real-time systems Ultrasound breast lesion segmentation deep learning U-Net convolution neural network

来源：评论

学校读者我要写书评

暂无评论

Underwater image enhancement using lightweight vision transformer

引用

MULtimeDIA TOOLS AND APPLICATIONS 2024年第31期83卷 75603-75625页

作者： Daud, Muneeba Afzal, Hammad Mahmood, Khawir Natl Univ Sci & Technol Islamabad Pakistan

deep learning-based models have recently shown a strong potential in Underwater image Enhancement (UIE) that are satisfying and have the right colors and details, but these methods significantly increase the parameters and complexity of the image processing models and therefore cannot be deployed directly to the edge devices. Vision Transformers (ViT) based architectures have recently produced amazing results in many vision tasks such as image classification, super-resolution, and image restoration. In this study, we introduced a lightweight Context-Aware Vision Transformer (CAViT), based on the Mean Head tokenization strategy and uses a self-attention mechanism in a single branch module that is effective at simulating long-distance dependencies and global features. To further improve the image quality we proposed an efficient variant of our model which derived results by applying White Balancing and Gamma Correction methods. We evaluated our model on two standard datasets, i.e., Large-Scale Underwater image (LSUI) and Underwater image Enhancement Benchmark Dataset (UIEB), which subsequently contributed towards more generalized results. Overall findings indicate that our real-time UIE model outperforms other deep learning based models by reducing the model complexity and improving the image quality (i.e., 0.6 dB PSNR improvement while using only 0.3% parameters and 0.4% float operations).

关键词： Tokenization Feature extraction image enhancement Vision transformers

来源：评论

学校读者我要写书评

暂无评论

real-time deep-learning inversion of seismic full waveform data for CO2 saturation and uncertainty in geological carbon storage monitoring

引用

GEOPHYSICAL PROSPECTING 2024年第1期72卷 199-212页

作者： Um, Evan Schankee Alumbaugh, David Lin, Youzuo Feng, Shihang Lawrence Berkeley Natl Lab Earth & Environm Sci Berkeley CA 94720 USA Los Alamos Natl Lab Earth & Environm Div Los Alamos NM 87545 USA Los Alamos Natl Lab Theoret Div Los Alamos NM 87545 USA

deep-learning inversion has recently drawn attention in geological carbon storage research due to its potential of imaging and monitoring carbon storage in real time, significantly improving efficiency and safety of carbon storage operations. We present a deep-learning full waveform inversion method that after the neural network has been trained can image CO2 saturation and its uncertainty in real time. Our deep-learning inversion method is based on the U-Net architecture with the neural network trained on pairs of synthetic seismic data and CO2 saturation models. Accordingly, our training establishes a mapping relationship between seismic data and CO2 saturation models and once fully trained directly estimates CO2 saturation as a function of subsurface location. We further quantify uncertainties of CO2 saturation estimates using the Monte Carlo dropout method and a bootstrap aggregating method. For this proof-of-concept study, the CO2 training models and data are derived from the Kimberlina 1.2 model, a hypothetical 3D geological carbon storage model that is constructed based on various geological and hydrological data from the Southern San Joaquin Basin, California. We perform deep-learning inversion experiments using noise-free and noisy training and test data sets and compare the results. Our modelling experiments show that (1) the deep-learning inversion can estimate 2D distributions of CO2 fairly well even in the presence of Gaussian random noise and (2) both CO2 saturation imaging and uncertainty quantification can be done in real time. Our results suggest that the deep-learning inversion method can serve as a robust real-time monitoring tool for geological carbon storage and/or other time-varying reservoir/aquifer properties that result from injection, extraction, and/or other subsurface transport phenomena.

关键词： full waveform inversion monitoring

来源：评论

学校读者我要写书评

暂无评论

learning Optimum Binary Color Filter Arrays for Demosaicing with Neural Networks

Learning Optimum Binary Color Filter Arrays for Demosaicing ...

引用

Conference on real-time image processing and deep learning

作者： Ayna, Cemre Omer Gurbuz, Ali Cafer Mississippi State Univ Mississippi State MS 39762 USA

ISBN: (纸本)9781510673878;9781510673861

Modern digital color cameras depend on Color Filter Arrays (CFA) for capturing color information. The majority of the commercial CFAs are designed by hand with different physical and application-specific considerations. The available machine learning (ML)-based CFA learning architectures dismiss the considerations of a physical camera device. This study aims to develop an alternative approach to jointly learn binary Color Filter Arrays (CFA) in a deep learning-based filtering-demosaicing pipeline. The proposed approach provides higher reconstruction performance than the compared hand-designed filters while learning physically applicable CFAs. This paper includes the learned binary CFAs for various color configurations and training data size, their analysis with common reconstruction metrics, and a short discussion on future works.

关键词： color filter array deep learning demosaicing hard thresholding straight through estimator gradient estimation

来源：评论

学校读者我要写书评

暂无评论

Authentix Trustify: Pioneering deep learning for Text and image processing and Integrity

Authentix Trustify: Pioneering Deep Learning for Text and Im...

引用

Emerging Trends in Industry 4.0 (ETI 4.0)

作者： P Kumar Rhikshitha K Samitha S Sanjay A Department of CSE Rajalakshmi Engineering College Chennai India

ISBN: (数字)9798350373783

ISBN: (纸本)9798350373790

In a world of electronic data and communication, there is an urgent need for safe, fast, and automatic review on documents. With this in mind, this project is meant to develop an all-encompassing system that can accurately detect text information in submitted documents as well as verify signatures. The system incorporates a set of external records supplied by consumers like the document under consideration and reference signatures. The system has two main objectives: the anchor model is used in the first case while authenticating signatures in the second one. The objective is to establish if the signature is genuine or contains defects pointing at falsification or illegal tampering. Generally, these processes give rise to comprehensive reports. These reports give credence and authenticity of the submitted documents while providing clarification whether a submitted document is valid or not through determination of match between document’s signature and reference signatures, as well as any inconsistencies noted throughout the evaluation process. This paper will help ensure safe and faster authentication of documents at a time when the world depends on accuracy and speed of document verification.

关键词： deep learning Technological innovation Adaptation models Accuracy Reviews Authentication Streaming media Security Reliability Text processing

来源：评论

学校读者我要写书评

暂无评论

real-time Motion Analysis With 4D deep learning for Ultrasound-Guided Radiotherapy

引用

IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING 2023年第9期70卷 2690-2699页

作者： Bengs, Marcel Sprenger, Johanna Gerlach, Stefan Neidhardt, Maximilian Schlaefer, Alexander Hamburg Univ Technol Inst Med Technol & Intelligent Syst D-21073 Hamburg Germany

Motion compensation in radiation therapy is a challenging scenario that requires estimating and forecasting motion of tissue structures to deliver the target dose. Ultrasound offers direct imaging of tissue in real-time and is considered for image guidance in radiation therapy. Recently, fast volumetric ultrasound has gained traction, but motion analysis with such high-dimensional data remains difficult. While deep learning could bring many advantages, such as fast data processing and high performance, it remains unclear how to process sequences of hundreds of image volumes efficiently and effectively. We present a 4D deep learning approach for real-time motion estimation and forecasting using long-term 4D ultrasound data. Using motion traces acquired during radiation therapy combined with various tissue types, our results demonstrate that long-term motion estimation can be performed markerless with a tracking error of 0.35 +/- 0.2 mm and with an inference time of less than 5 ms. Also, we demonstrate forecasting directly from the image data up to 900 ms into the future. Overall, our findings highlight that 4D deep learning is a promising approach for motion analysis during radiotherapy.

关键词： 4D deep learning sequence processing ultrasound motion analysis radiation therapy

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Review on image Encryption Techniques using Memristor based Chaotic System for Multimedia Application

引用

IETE JOURNAL OF RESEARCH 2024年第11期70卷 8160-8183页

作者： Kumar, B. Sakthi Revathi, R. Sri Eshwar Coll Engn Dept ECE Coimbatore India KLEF Univ Dept ECE Guntur India

image encryption is one of the most widely used techniques for securing images in trusted and unrestricted public media. However, the drawback is weak security in chaotic encryption algorithms, small key space and hardware implementation complexities. Various appropriate encryption algorithms have been carried out for secure image transmission;it becomes a critical issue to protect image integrity, confidentiality and authenticity. To overcome these issues, this article examines several existing memristor-based image encryption algorithms with various chaotic maps. As a result, this research aims to deliver comprehensive literature on image encryption techniques through memristor to help the researchers. Finally, this survey provides all vital literature on the current image encryption techniques with their benefits, drawbacks, developments, and future directions. We also give a general guideline about cryptography. We conclude that all methods are helpful for real-time image encryption. A comparison has been made between many image processing approaches (conventional and memristor) based on metrics such as the number of pixels change rate (NPCR) with the value of 0.9986, histogram, entropy (7.9993), unified average varying intensity (UACI) achieves the value as 0.4996, energy consumption as 0.32pJ and time complexity as 0.9s. The experimental results of various approaches, key sensitivity and statistical analysis revealed that the survey of memristor-based image encryption schemes provides an effective way to secure image transmission in real-time application.

关键词： deep learning image decryption image encryption image processing Memristor Security

来源：评论

学校读者我要写书评

暂无评论

real-time Multichannel deep Speech Enhancement in Hearing Aids: Comparing Monaural and Binaural processing in Complex Acoustic Scenarios

引用

IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE processing 2024年 32卷 4596-4606页

作者： Westhausen, Nils L. Kayser, Hendrik Jansen, Theresa Meyer, Bernd T. Carl von Ossietzky Univ Oldenburg Commun Acoust Grp D-26129 Oldenburg Germany Carl von Ossietzky Univ Oldenburg Cluster Excellence Hearing4all D-26129 Oldenburg Germany Carl von Ossietzky Univ Oldenburg Auditory Signal Proc & Hearing Devices Grp D-26129 Oldenburg Germany Horzentrum Oldenburg gGmbH Oldenburg D-26129 Oldenburg Germany

deep learning has the potential to enhance speech signals and increase their intelligibility for users of hearing aids. deep models suited for real-world application should feature a low computational complexity and low processing delay of only a few milliseconds. In this paper, we explore deep speech enhancement that matches these requirements and contrast monaural and binaural processing algorithms in two complex acoustic scenes. Both algorithms are evaluated with objective metrics and in experiments with hearing-impaired listeners performing a speech-in-noise test. Results are compared to two traditional enhancement strategies, i.e., adaptive differential microphone processing and binaural beamforming. While in diffuse noise, all algorithms perform similarly, the binaural deep learning approach performs best in the presence of spatial interferers. Through a post-analysis, this can be attributed to improvements at low SNRs and to precise spatial filtering.

关键词： Microphones Speech enhancement Hearing aids Computational modeling Auditory system Filtering theory real-time systems Data models Training Computer architecture Binaural low-latency multi-channel real-time speech-enhancement subjective evaluation

来源：评论

学校读者我要写书评

暂无评论

Data Generation and Augmentation Method for deep learning-Based VDU Leakage Signal Restoration Algorithm

引用

IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY 2024年 19卷 5220-5234页

作者： Nam, Taesik Choi, Dong-Hoon Lee, Euibum Jo, Han-Shin Yook, Jong-Gwan Yonsei Univ Dept Elect & Elect Engn Seoul 03722 South Korea Hanyang Univ Dept Automot Engn Seoul 04763 South Korea

This study analyzes the phenomenon of electromagnetic (EM) leakage that occurs through cables and explores the potential for information forensics using deep learning-based image-processing algorithms. We focus on the transition-minimized differential signaling (TMDS) interface to analyze information leakage caused by the inherent differential signal synchronization errors in video graphics controllers (VGC). Our analysis includes detailed mathematical modeling of the EM leakage phenomena from the video display unit (VDU) interface that uses the TMDS protocol. Furthermore, this study presents mathematical models for distortions and alterations caused by the VDU characteristics and its associated RF front-end system. Utilizing mathematical models of EM phenomena, this paper presents a method for creating training datasets for deep learning-based signal processing algorithms by generating and augmenting pseudo leakage signals (PLS) that closely resemble actual leakage signals. This study confirms the practical utility of signal enhancement models trained with generated and augmented PLS in real-world scenarios. Validation involves applying the trained model to measured actual VDU leakage signals and evaluating the results using image quality metrics: peak signal-to-noise ratio (PSNR), signal-to-noise ratio (SNR), and the structural similarity index measure (SSIM). Ultimately, this study demonstrates the potential to develop deep learning models using theoretically generated PLS for VDU-targeted side-channel attacks, where collecting real training data poses a challenge. This suggests the potential for expanding into high-performance deep learning algorithms in future developments.

关键词： Side-channel attacks Protocols Monitoring image color analysis Standards Data models Codes TEMPEST LCD monitor information leakage electromagnetic interference (EMI) compromising emanations (CE) deep learning convolutional neural network (CNN) data augmentation

来源：评论

学校读者我要写书评

暂无评论

A real-time welding defect detection framework based on RT-DETR deep neural network☆

引用

ADVANCED ENGINEERING INFORMATICS 2025年 65卷

作者： Liu, Gaoyang Yang, Duanrui Ye, Jun Lu, Hongjia Wang, Zhen Zhao, Yang Shaoxing Univ Sch Civil Engn Shaoxing Peoples R China Shaoxing Univ Inst Artificial Intelligence Shaoxing Peoples R China Univ Leeds Sch Civil Engn Leeds England RMIT Univ Ctr Innovat Struct & Mat Sch Engn Melbourne 3001 Australia Hangzhou City Univ Dept Civil Engn Hangzhou Peoples R China

The quality of welds is critical to the safety and reliability of steel structure connections, underscoring the importance of accurate inspection during the welding process. To enhance inspection effectiveness, deep learning methods have gained popularity in weld defect detection for their ability to automatically learn and refine image features. However, the complex multi-stage training and inference process of these methods often fails to meet the requirements of real-time performance and accuracy. To address this problem, a framework based on the real-time DEtection TRansformer (RT-DETR) for deep learning-based welding defect detection is proposed. This framework improves the Transformer backbone by eliminating the most time-consuming non-maximum suppression (NMS) step, achieving real-time detection without sacrificing accuracy. A diverse welding dataset with 1,134 images from real-world manufacturing and construction environments was developed for model training and validation. In addition, three data enhancement algorithms were explored to enhance the model's generalization ability. The model achieved detection accuracy scores of mAP@0.5 at 0.996 and mAP@0.5:0.95 at 0.801, with a detection speed of 67 frames per second (FPS). Compared to the previous Faster R-CNN, SSD, YOLOv5, YOLOv11 and DETR models, the proposed RT-DETR model demonstrates superior efficiency and accuracy. The proposed framework was further validated in the on-site inspections of metal additive manufacturing, and the results confirmed that the RT-DETR-based model meets the stringent requirements for real-time inspection in metal additive manufacturing.

关键词： Weld defect detection deep learning RT-DETR Data enhancement Additive manufacturing welds

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：