检索结果-内蒙古大学图书馆

Lossless Multi-component Image Compression Based on Integer Wavelet Coefficient Prediction using Convolutional Neural Networks

学校读者我要写书评

暂无评论

Lossless Multi-component Image Compression Based on Integer ...

Data Compression Conference (DCC)

作者： Eze Ahanonu Michael Marcellin Ali Bilgin Dept. of Electrical and Computer Engineering University of Arizona Tucson AZ USA Dept. of Biomedical Engineering The University of Arizona Tucson AZ USA

ISBN: (数字)9781728164571

ISBN: (纸本)9781728164588

This work extends the single component Wavelet Prediction Compression (WPC) framework proposed earlier to lossless compression of multi-component imagery. WPC employs a series of convolutional neural networks (CNNs) to generate predictions of wavelet detail subbands, using subbands within the same discrete wavelet transform (DWT) decomposition level. Prediction residuals are then coded in place of original coefficients in the compressed codestream. At the decoder, predictions are reproduced using an identical set of CNNs, and combined with residuals to achieve perfect subband (and image) reconstruction. The proposed Multi-Component Wavelet Prediction Compression (MCWPC) framework extends WPC to multi-component by considering both inter-and intra-component redundancies, resulting in improved spatial and spectral decorrelation of DWT coefficients. Compression results (Table 1) show MCWPC achieves 7.2% and 23.8% bit-rate reductions over JPEG2000 in the YUV and RGB colorspaces, respectively.

关键词：

Latent Signal Models: Learning Compact Representations of Signal Evolution for Improved Time-Resolved, Multi-contrast MRI

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Arefeen, Yamin Xu, Junshen Zhang, Molin Dong, Zijing Wang, Fuyixue White, Jacob Bilgic, Berkin Adalsteinsson, Elfar Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology CambridgeMA United States Athinoula A. Martinos Center for Biomedical Imaging CharlestownMA United States Department of Radiology Harvard Medical School BostonMA United States Harvard-MIT Health Sciences and Technology Massachusetts Institute of Technology CambridgeMA United States Institute for Medical Engineering and Science Massachusetts Institute of Technology CambridgeMA United States

Purpose: Training auto-encoders on simulated signal evolution and inserting the decoder into the forward model improves reconstructions through more compact, Bloch-equation-based representations of signal in comparison to linear subspaces. Methods: Building on model-based nonlinear and linear subspace techniques that enable reconstruction of signal dynamics, we train auto-encoders on dictionaries of simulated signal evolution to learn more compact, non-linear, latent representations. The proposed Latent Signal Model framework inserts the decoder portion of the auto-encoder into the forward model and directly reconstructs the latent representation. Latent Signal Models essentially serve as a proxy for fast and feasible differentiation through the Bloch-equations used to simulate signal. This work performs experiments in the context of T2-shuffling, gradient echo EPTI, and MPRAGE-shuffling. We compare how efficiently auto-encoders represent signal evolution in comparison to linear subspaces. Simulation and in-vivo experiments then evaluate if reducing degrees of freedom by incorporating our proxy for the Bloch-equations, the decoder portion of the auto-encoder, into the forward model improves reconstructions in comparison to subspace constraints. Results: An auto-encoder with one real latent variable represents FSE, EPTI, and MPRAGE signal evolution as well as linear subspaces characterized by four basis vectors. In simulated/in-vivo T2-shuffling and in-vivo EPTI experiments, the proposed framework achieves consistent quantitative NRMSE and qualitative improvement over linear approaches. From qualitative evaluation, the proposed approach yields images with reduced blurring and noise amplification in MPRAGE shuffling experiments. Conclusion: Directly solving for non-linear latent representations of signal evolution improves time-resolved MRI reconstructions through reduced degrees of freedom. Copyright © 2022, The Authors. All rights reserved.

关键词： Image reconstruction

Predicting blood pressure without a cuff using a unique multi-modal wearable device and machine learning algorithm

学校读者我要写书评

暂无评论

computers in Biology and Medicine 2025年第Pt B期192卷 110357-110357页

作者： Hsiao, Chin-To Hong, Sungcheol Branan, Kimberly L. McMurray, Justin Coté, Gerard L. Department of Biomedical Engineering Texas A&M University College StationTX77843 United States Texas A&M Engineering Experiment Station Center for Remote Health Technologies and Systems College StationTX77843 United States Department of Electrical and Computer Engineering Texas A&M University College StationTX77843 United States Electrical and Electronic Convergence Department Hongik University Sejong Korea Republic of

Blood pressure is a critical risk factor for cardiovascular diseases (CVDs), yet most adults do not monitor it frequently enough to prevent serious complications. This is in part because the traditional cuff-based method is inconvenient, uncomfortable, and does not allow for continuous monitoring. To address these constraints, we developed a unique multi-modal wearable device and used a random forest regression (RFR) algorithm that resulted in a model capable of accurate cuffless blood pressure prediction. This multi-modal device features two photoplethysmography (PPG) sensors and two bioimpedance (BioZ) sensors to measure pulse wave propagation along the radial artery on the wrist. The redundancy in the design enhances prediction accuracy. To validate the device, a novel human subject study protocol was also developed that allows an individual's blood pressure to rise safely and repeatably by more than 40 mmHg (systolic pressure) from baseline measurements. In this study, using multiple pulsatile waveforms from the PPG and BioZ sensors as inputs into the machine learning prediction algorithm, showed that the model had higher accuracy than models using a single sensor. Specifically, the training, validation, and leaving one subject out of data sets all showed mean absolute errors of less than 3.3 mmHg for both systolic and diastolic blood pressures (BPs). While results from this test were promising, a subject-wise evaluation showed variability depending on how well an individual's BP distribution matched the training set. These findings demonstrate the potential for a universal model for cuffless BP estimation, with further validation needed in more diverse populations. Thus, the accompaniment of the RFR model with the multi-modal wearable device offers the potential for robust and continuous blood pressure monitoring, providing a unique and practical solution for long-term cardiovascular health management. © 2025

关键词： Random forests

Reinforcement learning enabled automatic impedance control of a robotic knee prosthesis to mimic the intact knee motion in a co-adapting environment

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Wu, Ruofan Li, Minhan Yao, Zhikai Si, Jennie Huang, He Department of Electrical Computer and Energy Engineering Arizona State University TempeAZ85287 United States Department of Biomedical Engineering North Carolina State University RaleighNC27695 United States University of North Carolina at Chapel Hill Chapel HillNC27599 United States

Automatically configuring a robotic prosthesis to fit its user’s needs and physical conditions is a great technical challenge and a roadblock to the adoption of the technology. Previously, we have successfully developed reinforcement learning (RL) solutions toward addressing this issue. Yet, our designs were based on using a subjectively prescribed target motion profile for the robotic knee during level ground walking. This is not realistic for different users and for different locomotion tasks. In this study for the first time, we investigated the feasibility of RL enabled automatic configuration of impedance parameter settings for a robotic knee to mimic the intact knee motion in a co-adapting environment. We successfully achieved such tracking control by an online policy iteration. We demonstrated our results in both OpenSim simulations and two able-bodied (AB) subjects. Copyright © 2021, The Authors. All rights reserved.

关键词： Reinforcement learning

Fingerprint spectroscopic SRS imaging of single living cells and whole brain by ultrafast tuning and spatial-spectral learning

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Lin, Haonan Lee, Hyeon Jeong Tague, Nathan Lugagne, Jean-Baptiste Zong, Cheng Deng, Fengyuan Wong, Wilson Dunlop, Mary J. Cheng, Ji-Xin Department of Biomedical Engineering Department of Electrical & Computer Engineering Photonics Center Biological Design Center Boston University BostonMA02215 United States College of Biomedical Engineering and Instrument Sciences Zhejiang University Hangzhou China

Label-free vibrational imaging by stimulated Raman scattering (SRS) provides unprecedented insight into real-time chemical distributions in living systems. Specifically, SRS in the fingerprint region (400-1800 cm−1) can resolve multiple chemicals in a complex bio-environment using specific and well-separated Raman signatures. Yet, fingerprint SRS imaging with microsecond spectral acquisition has not been achieved due to the small fingerprint Raman cross-sections and the lack of ultrafast acquisition scheme with high spectral resolution and high fidelity. Here, we report a fingerprint spectroscopic SRS platform that acquires a distortion-free SRS spectrum with 10 cm-1 spectral resolution in 20 µs using a lab-built ultrafast delay-line tuning system. Meanwhile, we significantly improve the signal-to-noise ratio by employing a spatial-spectral residual learning network, reaching comparable quality to images taken with two orders of magnitude longer pixel dwell times. Collectively, our system achieves reliable fingerprint spectroscopic SRS with microsecond spectral acquisition speed, enabling imaging and tracking of multiple biomolecules in samples ranging from a live single microbe to a tissue slice, which was not previously possible with SRS imaging in the highly congested carbon-hydrogen region. To show the broad utility of the approach, we have demonstrated high-speed compositional imaging of lipid metabolism in living pancreatic cancer Mia PaCa-2 cells. We then performed high-resolution mapping of cholesterol, fatty acid, and protein in the mouse whole brain. Finally, we mapped the production of two biofuels in microbial samples by harnessing the superior spectral and temporal resolutions of our system. Copyright © 2020, The Authors. All rights reserved.

关键词： Stimulated Raman scattering

A comparative analysis of reactive power compensation using reactors and STATCOMs in primary substations: a case study in Milan, Italy

学校读者我要写书评

暂无评论

A comparative analysis of reactive power compensation using ...

AEIT Annual Conference

作者： Luigi Sandrini Alessandro Bosisio Alessandro Cirocco Mario Turrisi Caterina Pasetti Andrea Morotti Luca Cavalletto Energy Planning Department Unareti S.p.A. Milan Italy Electrical Computer and Biomedical Engineering University of Pavia Pavia Italy Technological Development Department A2A S.p.A. Milan Italy Planning Department Unareti S.p.A. Milan Italy Electricity Grid Development Unareti S.p.A. Milan Italy Asset Management Department A2A S.p.A. Milan Italy

The article aims to compare the benefits derived from the installation of two different technologies for reactive power compensation systems in distribution networks: reactors and static synchronous compensators. The analysis, conducted on the Unareti electrical distribution grid, explicitly evaluates these two technologies’ help when installed at the primary substation level, considering long-term scenarios to identify the technology that provides more significant benefits throughout its lifespan. Starting from the actual energy flows recorded over one year in the eleven primary substations of Milan and Rozzano, and considering the impact of the energy transition on electricity demand as well as the consequent predicted evolution of the electrical grid resulting from the expected commissioning of eight new primary substations, the article assesses the benefits provided by the two different technologies in both the current (AS IS) and the future (TO BE) scenario by 2050. The simulation results demonstrate that due to the ongoing energy transition and the consequent shift of energy consumption to the electrical vector, static synchronous compensators can offer greater flexibility in response to changes in the operating conditions that will affect distribution networks in the coming years.

关键词：

Efficient Epileptic seizure detection using CNN-aided factor graphs

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Salafian, Bahareh Ben, Eyal Fishel Shlezinger, Nir de Ribaupierre, Sandrine Farsad, Nariman The School of Biomedical Engineering University of Western Ontario LondonONN6A 5B9 Canada The School of Electrical and Computer Engineering Ben-Gurion University of the Negev Be'er Sheva84105 Israel Department of Clinical Neurological Sciences The School of Biomedical Engineering University of Western Ontario ONN6A 5B9 Canada The Department of Computer Science Ryerson University TorontoONM5B 2K3 Canada

We propose a computationally efficient algorithm for seizure detection. Instead of using a purely data-driven approach, we develop a hybrid model-based/data-driven method, combining convolutional neural networks with factor graph inference. On the CHB-MIT dataset, we demonstrate that the proposed method can generalize well in a 6 fold leave-4-patient-out evaluation. Moreover, it is shown that our algorithm can achieve as much as 5% absolute improvement in performance compared to previous data-driven methods. This is achieved while the computational complexity of the proposed technique is a fraction of the complexity of prior work, making it suitable for real-time seizure detection. © 2021, CC BY.

关键词： Complex networks

Removing mixture of Gaussian and Impulse noise of images using sparse coding

学校读者我要写书评

暂无评论

Removing mixture of Gaussian and Impulse noise of images usi...

Iranian Machine Vision and Image Processing (MVIP)

作者： Mahsa Malekzadeh Saeed Meshgini Reza Afrouzian Ali Farzamnia Sobhan Sheykhivand Department of Biomedical Engineering Faculty of Electrical and Computer Engineering University of Tabriz Tabriz Iran Miyaneh Faculty of Engineering University of Tabriz Miyaneh Iran Faculty of Engineering Universiti Malaysia Sabah Sabah Malaysia

ISBN: (数字)9781728168326

ISBN: (纸本)9781728168333

Real images contain different types of noises and a very difficult process is to remove mixed noise in any type of them. Additive White Gaussian Noise (AWGN) coupled with Impulse Noise (IN) is a typical method. Many mixed noise removal methods are based on a detection method that generates artificial products in case of high noise levels. In this article, we suggest an active weighted approach for mixed noise reduction, defined as Weighted Encoding Sparse Noise Reduction (WESNR), encoded in sparse non-local regulation. The algorithm utilizes a non-local self-similarity feature of image in the sparse coding framework and a pre-learned Principal Component Analysis (PCA) dictionary. Experimental results show that both the quantitative and the visual quality, the proposed WESNR method achieves better results of the other technique in terms of PSNR.

关键词：

Heart Position Uncertainty Quantification in the Inverse Problem of ECGI

学校读者我要写书评

暂无评论

Heart Position Uncertainty Quantification in the Inverse Pro...

computers in Cardiology (CinC)

作者： Jake A Bergquist Lindsay C Rupp Anna Busatto Ben Orkild Brian Zenger Wilson Good Jaume Coll-Font Akil Narayan Jess Tate Dana Brooks Rob S MacLeod Scientific Computing and Imaging Institute University of Utah SLC UT USA Nora Eccles Treadwell CVRTI University of Utah SLC UT USA Department of Biomedical Engineering University of Utah SLC UT USA School of Medicine University of Utah SLC UT USA Acutus Medical Carlsbad CA USA Department of Mathematics University of Utah SLC UT USA Department of Electrical and Computer Engineering Northeastern University Boston MA USA

Electrocardiographic imaging (ECGI) is a clinical and research tool for noninvasive diagnosis of cardiac electrical dysfunction. The position of the heart within the torso is both an input and common source of error in ECGI. Many studies have sought to improve cardiac localization accuracy, however, few have examined quantitatively the effects of uncertainty in the position of the heart within the torso. Recently developed uncertainty quantification (UQ) tools enable the robust application of UQ to ECGI reconstructions. In this study, we developed an ECGI formulation, which for the first time, directly incorporated uncertainty in the heart position. The result is an ECGI solution that is robust to variation in heart position. Using data from two Langendorff experimental preparations, each with 120 heartbeats distributed across three activation sequences, we found that as heart position uncertainty increased above ±10 mm, the solution quality of the ECGI degraded. However, even at large heart position uncertainty (±40 mm) our novel UQ-ECGI formulation produced reasonable solutions (root mean squared error < 1 mV, spatial correlation >0.6, temporal correlation >0.75).

关键词： Heart Torso Location awareness Uncertainty Correlation Inverse problems Heart beat