检索结果-内蒙古大学图书馆

2023 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2023

作者： Bello, Valentina Bodo, Elisabetta Merlo, Sabina Department of Electrical Computer and Biomedical Engineering University of Pavia Via Ferrata 5 Pavia27100 Italy

ISBN: (纸本)9798350345995

Parenteral nutrition (PN) is a lifesaving treatment for a large number of patients suffering from different pathologies, from cancer to intestinal failure, from eating disorders to inflammatory bowel disease. In PN, liquid nutritive drugs are injected into the patient's body intravenously through an infusion delivery pump. As reported in the scientific literature, among all medical treatments, PN is the most commonly prone one to human errors [1]. In particular, very often wrong PN mixtures (PNMs) are administered. The consequences can be very dangerous to the patient's health, leading to death in the most severe cases. Despite this fact, currently no standard safety protocols or control devices are implemented to prevent medication errors in PN. Hence, in the framework of the DSF (Digital Smart Fluidics - project No. 1175234, founded by POR FESR 2014-2020), we have developed an optofluidic sensing platform to distinguish different types of transparent commercial PNMs on the basis of their refractive index (RI). Indeed, each mixture contains different concentrations of glucose, amino acids, and electrolytes, determining different values of their RI. The instrumental configuration of the sensor (Fig 1(a)) features a laser diode that generates a red beam (wavelength of 670 nm) travelling obliquely through a plastic cuvette channel containing the PNM under test. Light is then back-reflected by a mirror applied to the back of the cuvette, and finally exits the channel at a position that depends on the RI of the PNM. The position of the output light spot is easily detected with a linear position sensitive detector (PSD) [2]. The two photo-currents I1 and I2 generated at the extremities of the sensitive area are given as the PSD outputs: by proper normalization, it is possible to retrieve the measured light beam position as pPSD = L/2 × (I1 - I2)/(I1 + I2), where L is the PSD length. The sensor response is then retrieved, after defining the baseline signal measured in pre

关键词： Glucose

来源：评论

学校读者我要写书评

暂无评论

Domain-Centroid-Guided Progressive Teacher-Based Knowledge Distillation for Source-Free Domain Adaptation of Histopathological Images

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第4期5卷 1831-1843页

作者： Cheng, Kuo-Sheng Zhang, Qiong-Wen Tsai, Hung-Wen Li, Nien-Tsu Chung, Pau-Choo National Cheng Kung University Department of Biomedical Engineering Tainan70101 Taiwan National Cheng Kung University Hospital Department of Pathology Tainan70142 Taiwan National Cheng Kung University Department of Computer and Communication Engineering Tainan70101 Taiwan National Cheng Kung University Department of Electrical Engineering Tainan70101 Taiwan

Deep neural networks are commonly used for histopathology image analysis. However, such data-driven models are sensitive to style variances across scanners and suffer a significant performance degradation as a result. Although the network performance can be improved by using domain adaptation methods, the source dataset required to perform the adaptation process is generally unavailable. This study shows that the performance degradation of deep neural networks when applied to histopathology images is the result partly of the wide distribution of the features generated when inferring the features of the target model using the feature centers of the source model. To address this problem, a teacher-student framework, designated as domain-centroid-guided progressive teacher-based knowledge distillation (DCGP-KD), is proposed which aims to learn compact target features in order to provide more accurate pseudo labels for the target model without the need for the original source dataset. In the proposed framework, the class-wise feature centers of the source data are progressively adapted to the distribution of the target data, and compact target features are then generated by gathering the features based on their class-wise centers. A strategy is additionally proposed to prevent catastrophic forgetting during the progressive adaption process. Finally, a prediction consistency loss function is introduced to improve the robustness of the target dataset. The feasibility of the proposed framework is demonstrated experimentally for the illustrative case of the tumor classification of histopathological images with staining variations. The results show that DCGP-KD provides a promising assistive tool for pathologists in various histopathological analysis tasks. © 2020 IEEE.

关键词： Distillation

来源：评论

学校读者我要写书评

暂无评论

Inverse Problem Antidote (IPA): Modeling of Systems Biology Model with Invertible Neural Networks

Inverse Problem Antidote (IPA): Modeling of Systems Biology ...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Li, Linlin Lu, Shenyu Umulis, David M. Wang, Xiaoqian Purdue University Weldon School of Biomedical Engineering West LafayetteIN United States Purdue University Elmore Family School of Electrical and Computer Engineering West LafayetteIN United States

ISBN: (纸本)9798350386226

In computational biology, accurately modeling biological systems is essential for understanding the underlying mechanisms of biological processes. One of the primary challenges in modeling lies in the inability to measure certain biological parameters directly. The task of identifying the parameters is known as the inverse problem and it entails estimating these unobservable parameters from available data. Existing modeling methods primarily focus on single-direction prediction. These methods include traditional Partial Differential Equation (PDE) modeling and machine learning approaches where given parameters are provided to predict the output and further compare the output with experimental evidence. However, these unidirectional methods struggle to effectively apply the experimental evidence directly to address the inverse problem. We contend that a single biological process should be modeled bidirectionally to simultaneously address the inverse problem. To this end, we propose leveraging the capabilities of the invertible neural network (INN) to establish a connection between the input parameter space and the model output space. We meticulously designed a bidirectional training technique that effectively applies real-world experimental data in guiding the INN in modeling the biological process. We tested our approach on a PDE-based Bone Morphogenic Protein (BMP) signaling network system in the zebrafish embryo and found the bidirectional modeling approach significantly enhances the alignment between simulation and experimental data. This method achieves a 94.65% reduction in Root Mean Square Error (RMSE) compared to the single-direction model when reconstructing experimental data using simulation with INN-identified parameters. Moreover, our method is rapid to implement, facilitating the precise identification of parameter ranges from experimental data. It makes parameter optimization feasible and provides a guideline for determining simulation parameter ranges,

关键词： Inverse problems

来源：评论

学校读者我要写书评

暂无评论

optimizing Furniture Assembly: A CNN-based Mobile Application for Guided Assembly and Verification 9

optimizing Furniture Assembly: A CNN-based Mobile Applicatio...

引用

9th International Conference on Optimization and Applications, ICOA 2023

作者： Elhassan, Gasm Elbary Yasser, Ibrahim Faizal, Mohamed Omer Zia, Huma Abu Dhabi University Department of Electrical Computer and Biomedical Engineering Abu Dhabi United Arab Emirates

ISBN: (纸本)9798350312546

Building products based o written instructions can be challenging for individuals, often leading to difficulties during the assembly process. This project presents a mobile application that guides users through the assembly process and verifies assembly correctness using image comparison techniques. The application features manufacturer and consumer modes, allowing manufacturers to add assembly steps and users to access provided assembly instructions for a given product. Cross-platform compatibility is achieved with React-Native and Expo Go, while image matching is implemented by incorporating computer vision and machine learning techniques, specifically utilizing a Siamese Network architecture. Amazon Web Services (AWS) supports content delivery and storage. An IKEA table and kids chair assembly serves as a demonstrative example. By offering step-by-step guidance and feedback based on similarity indexes, the application optimizes the efficiency and accuracy of furniture assembly. The project showcases the system's practicality, particularly in enhancing the assembly process for various furniture items, exemplified by the IKEA products. © 2023 IEEE.

关键词： Assembly

来源：评论

学校读者我要写书评

暂无评论

Thermal Noise Analysis of Accumulation-based S/H Circuit for Shunt Current Sensing 30

Thermal Noise Analysis of Accumulation-based S/H Circuit for...

引用

30th IEEE International Conference on Electronics, Circuits and Systems, ICECS 2023

作者： Yarragunta, Jaya Satyanarayana Aprile, Antonio Fugger, Andreas Conzatti, Francesco Bonizzoni, Edoardo Malcovati, Piero University of Pavia Department of Electrical Computer and Biomedical Engineering Italy Infineon Technologies Austria Villach Austria

ISBN: (纸本)9798350326499

Shunt current measurements are widely used for monitoring and controlling the current flow in various applications, such as power converters and battery management systems (BMS). The accuracy and resolution of this kind of measurements depend on the performance of the sample-and-hold (S/H) circuit that captures the voltage across a shunt resistor. In this paper, we propose a noise analysis of a S/H circuit that operates in accumulation mode;this technique enables to lower the thermal noise by 3 dB for every doubling of the number of accumulation cycles. The proposed analysis, confirmed by extensive simulation results, demonstrates that this approach offers a lower noise power spectral density and a higher signal-to-noise ratio if compared with conventional amplification-based S/H solutions. © 2023 IEEE.

关键词： Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

Optical Readout Techniques for Smart Detection of Fluids in the Near Infrared Wavelength Region by Means of Rectangular Glass Micro-capillaries

Optical Readout Techniques for Smart Detection of Fluids in ...

引用

Annual Italian Conference on Sensors and Microsystems, AISEM 2021

作者： Bello, Valentina Bodo, Elisabetta Merlo, Sabina Department of Electrical Computer and Biomedical Engineering University of Pavia Via Ferrata 5 Pavia27100 Italy

ISBN: (纸本)9783031081354

Three different approaches for the analytical detection of fluids by means of rectangular glass micro-capillaries working in the near infrared wavelength region are presented. At first, a non-specific refractometric measurement for the detection of glucose concentration in solutions is reported, exploiting the micro-capillaries as optical resonators: by monitoring the spectral shift of the ratio between the transmitted and reflected optical spectra (T/R) from the capillary, it is possible to extract the wavelength positions of the cavity resonances (maxima of T/R) for fluids with different refractive index. When the refractive index of the sample fluid filling the channel increases, a shift towards longer wavelengths is observed. Then, a spectral phase shift interferometric technique for the detection of the wavelength position of the resonances is proposed. When the capillary is inserted in an interferometric setup, it is possible to distinguish fluids by knowing the dependence of the wavelength positions of the steep jumps in the cosine signal on the refractive index of the filling fluid. Finally, the potentiality of micro-capillaries is investigated for specific sensing, exploiting absorption spectroscopy. All the proposed optical readout approaches are remote, contactless and non-invasive. In addition, the glass-micro-capillaries are very suitable for analytical detection of fluids: they are low-cost devices, available in several formats. Thanks to their micrometric size, they can be incorporated in micro-fluidic circuits. Borosilicate glass is a bio-compatible material, allowing the use of the micro-fluidic platforms in a wide range of applications for label-free optical sensing. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Absorption spectroscopy

来源：评论

学校读者我要写书评

暂无评论

Modeling of a Hexagonal Microstrip Patch Antenna for Breast Cancer Detection

引用

Journal of engineering and Science in Medical Diagnostics and Therapy 2024年第3期7卷 031007页

作者： Chowdhury, Nure Alam Wang, Lulu Islam, Md Shazzadul Gu, Linxia Kaya, Mehmet Department of Biomedical Engineering and Science Florida Institute of Technology MelbourneFL32901 United States Biomedical Device Innovation Center Shenzhen Technology University Shenzhen518118 China Department of Electrical and Computer Engineering International Islamic University Malaysia Jalan Gombak Selangor Kuala Lumpur53100 Malaysia

Breast cancer is a global problem, and it is inevitable to detect cancerous cells at early stages. In recent years, microwave imaging (MWI) technology has been widely applied in biomedical applications for its nonionizing radiation. Therefore, in this paper, a low profile hexagonal microstrip patch antenna has been proposed for the technology to detect breast cancer. This antenna has wide operating bandwidth of 13.5 GHz (6.6 GHz to 20.1 GHz), and the return loss is as low as -50.83 dB at 8 GHz. To evaluate the antenna performances, the proposed antenna has been simulated in two different simulation software like HFSS and CST MWS. The antenna has achieved a maximum gain of 8.82 dBi with a quasi-omnidirectional radiation pattern. A three-layered human body mimicking breast phantom with different dielectric properties has been designed with and without tumor mimicking tissue. The difference between the dielectric properties of the tumor and the dielectric properties of different layers of breast phantom in the presence of external radiation field can inform the existence of tumor inside the breast phantom. An array of eight elements of proposed antenna is distributed around the breast phantom to detect the tumor with a minimum radius of 2 mm. Because of the low profile and compact in size (7:9 mm X 11:4 mm), the proposed antenna is suitable for multistatic MWI technology for breast cancer detection at early stage. Copyright © 2024 by ASME.

关键词： Tumors

来源：评论

学校读者我要写书评

暂无评论

Predicting and Explaining Risk of Disease Worsening Using Temporal Features in Multiple Sclerosis 24

Predicting and Explaining Risk of Disease Worsening Using Te...

引用

24th Working Notes of the Conference and Labs of the Evaluation Forum, CLEF-WN 2023

作者： Buonocore, Tommaso Mario Bosoni, Pietro Nicora, Giovanna Vazifehdan, Mahin Bellazzi, Riccardo Parimbelli, Enea Dagliati, Arianna Department of Electrical Computer and Biomedical Engineering University of Pavia Via Ferrata 5 Pavia27100 Italy

We present an evaluation study of the usage of two different post-hoc model agnostic XAI methods, namely SHAP and AraucanaXAI, to provide insights about the most predictive factors of worsening in MS patients, based on clinical observations carried out during a period of 2.5 years. We pre-processed the temporal features considering a Latent Class Mixed Modelling (LCMM) approach in order to discover and extract temporal trajectories as an additional informative feature. The different XAI approaches are compared according to four quantitative evaluation metrics consisting in identity, fidelity, separability and time to compute an explanation. Furthermore, a qualitative comparison of post-hoc generated explanations is carried out on specific scenarios where the ML model predicted the outcome incorrectly, in the effort to debug potentially problematic model behaviour. The combination of the qualitative and quantitative results forms the basis for a critical discussion of XAI methods properties and desiderata for healthcare applications at large, advocating for more meaningful and extensive XAI evaluation studies involving human experts. © 2023 Copyright for this paper by its authors.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Towards the development of a SWIR-LEDs based optoelectronic system for urea monitoring during haemodialytic therapy

Towards the development of a SWIR-LEDs based optoelectronic ...

引用

2023 Conference on Lasers and Electro-Optics Europe and European Quantum Electronics Conference, CLEO/Europe-EQEC 2023

作者： Bodo, Elisabetta Bello, Valentina Merlo, Sabina Department of Electrical Computer and Biomedical Engineering University of Pavia Via Ferrata 5 Pavia27100 Italy

ISBN: (纸本)9798350345995

Chronic kidney disease (CKD) has been one of the most severe public health issues in recent decades. Patients affected by CKD require a complicated and expensive treatment such as hemodialysis. The management of hemodialysis patients have been a particular challenge for nephrologists during the COVID-19 pandemic [1]. The accepted measure of dialysis dose is based on the removal of urea from the blood pool. Urea concentration is usually evaluated with the aid of blood samples that are analyzed in a clinical laboratory, but there is still the need of continuous monitoring of the dialysis efficiency to optimize clinical treatments. Hence, we have investigated and demonstrated an optofluidic sensing platform to detect urea content in water solutions by comparing light transmittance across a flat microfluidic channel with and without fluid. We considered the effect of absorption in the wavelength bands around λ = 1.45 μm and λ = 2.15 μm, where water and urea exhibit characteristic peaks of absorbance [2]. In the instrumental configuration (Fig. 1(a)), radiation provided by two LEDs crosses the microfluidic device and is finally detected in time domain with an amplified InGaAs photodiode, connected to the oscilloscope for data visualization and acquisition. The microfluidic device is a rectangular section borosilicate glass capillary (Vitrocom, NJ, USA) with two extremities provided with heat shrink tubes to facilitate injection of the fluid into the channel. Nominal channel length and depth of the channel are 50 mm and 1 mm, respectively. We vertically fixed the capillary in direct contact with the photodiode mounting, ensuring a LED-photodiode distance of just 5 mm. Transmittance measurements were carried out by filling the channel with urea-water solutions with concentrations of urea up to 0.05 g/ml. Data acquired with the oscilloscope are then analyzed in MATLAB environment, to extract the average and standard deviation of the photodetected signal amplitude in the tim

关键词： Urea

来源：评论

学校读者我要写书评

暂无评论

Cognitive Analysis based on Brain Connectivity for Learning Task fMRI Data using CONN Software Package 1

Cognitive Analysis based on Brain Connectivity for Learning ...

引用

1st IEEE Karachi Section Humanitarian Technology Conference, Khi-HTC 2024

作者： Jatoi, Munsif Ali Teevino, Sadam Hussain Dharejo, Fayaz Ali Salim Habib University Department of Biomedical Engineering Sindh Karachi Pakistan Tenologico de Monterrey School of Engineering and Sciences Mexico Khalifa University Department of Electrical Engineering and Computer Science United Arab Emirates

ISBN: (纸本)9798350373042

The educational psychology is dependent upon the cognition and learning. The productive research in this domain yields the results for improvements in novel pedagogical methods at various educational level. The learning and memory are closely affiliated with brain structure and functioning i.e., neurobiology which has been the topic of research in philosophy, psychology and social sciences. Hence, to study the learning mechanism for improvements in educational methods, its essential to understand theoretical background of cognitive analysis through brain sciences. The human brain consists of billions of neurons with trillions of synaptic connections structurally and functionally organized in multiple scales of space and time, respectively. This complex structure is responsible for various complex and simple mental/physical tasks such as learning, memory, listening, sequencing, reasoning, motion, thinking, controlling, etc. The brain activity is analyzed through either source analysis or connectivity analysis. The connectivity analysis is carried out by either anatomical way, which is termed as anatomical connectivity, or through the functional way, which is termed as functional connectivity. However, the third explores the cause and effect relationship for a set of neurons and is termed as effective connectivity. This research work produces a generalized review for educational learning and then its relationship with connectivity analysis, including its introduction, graph theory implication, discussion for its types, and parameter discussion. Finally, functional connectivity was implemented on brain data using the CONN software package to compare various degrees, Eigenvalue, and betweenness centrality using fMRI data. The results show that the Eigenvalue centrality measure performs the best result for functional connectivity compared with other centrality measures. © 2024 IEEE.

关键词： Software packages

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：