检索结果-南通市图书馆

Using Cross Entropy as a Performance Metric for Quantifying Uncertainty in DNN Image Classifiers: An Application to Classification of Lung Cancer on CT Images

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Biomedical Science and Engineering 2024年第1期17卷 1-12页

作者： Eri Matsuyama Masayuki Nishiki noriyuki Takahashi Haruyuki Watanabe Faculty of Informatics University of Fukuchiyama Kyoto Japan Graduate School of Radiological Sciences International University of Health and Welfare Tochigi Japan School of Health Sciences Fukushima Medical University Fukushima Japan School of Radiological Technology Gunma Prefectural College of Health Sciences Gunma Japan

Cross entropy is a measure in machine learning and deep learning that assesses the difference between predicted and actual probability distributions. In this study, we propose cross entropy as a performance evaluation metric for image classifier models and apply it to the CT image classification of lung cancer. A convolutional neural network is employed as the deep neural network (DNN) image classifier, with the residual network (ResNet) 50 chosen as the DNN archi-tecture. The image data used comprise a lung CT image set. Two classification models are built from datasets with varying amounts of data, and lung cancer is categorized into four classes using 10-fold cross-validation. Furthermore, we employ t-distributed stochastic neighbor embedding to visually explain the data distribution after classification. Experimental results demonstrate that cross en-tropy is a highly useful metric for evaluating the reliability of image classifier models. It is noted that for a more comprehensive evaluation of model perfor-mance, combining with other evaluation metrics is considered essential. .

关键词： Cross Entropy Performance Metrics DNN Image Classifiers Lung Cancer Prediction Uncertainty

Effect of Wearing a Maxillofacial Prosthesis on Masticatory Rhythm Improvement in Mandibulectomy Patients

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Health 2024年第7期16卷 717-729页

作者： Xuewei Han Mihoko Haraguchi Yuka I. Sumita Marwa Ahmed Aboelez noriyuki Wakabayashi Department of Advanced Prosthodontics Graduate School of Medical and Dental Sciences Tokyo Medical and Dental University (TMDU) Tokyo Japan Department of Partial and Complete Denture School of Life Dentistry at Tokyo The Nippon Dental University Tokyo Japan Department of Prosthodontics Faculty of Dentistry Mansoura University Mansoura Egypt

Background: Recording jaw movement rhythms in patients with jaw defects is challenging using conventional tracking devices. However, masticatory rhythm is crucial for masticatory performance and nutritional intake. This study aimed to explore the impact of using a maxillofacial prosthesis to restore masticatory rhythm in mandibulectomy patients with a novel wearable device. Methods: Twelve patients who underwent mandibulectomy and were rehabilitated with maxillofacial prosthesis were recruited. Chewing rate, cycle durations, coefficient of variation of cycle durations (CV), changing times of skin morphology on the back of the ear pinna area (SM), and masticatory performance were measured during gum chewing without and with a denture using a wearable jaw movement rhythm tracking device. A paired t-test and the Pearson correlation coefficient were used for statistical analysis. Results: The chewing rate increased by 8.6 cycles/min with dentures (P P = 0.01). The CV and SM decreased with dentures (P = 0.004 and P = 0.01, respectively). Significant correlations were found between the CV and SM without dentures (P = 0.004). Conclusion: 1. Wearing maxillofacial prostheses can improve masticatory rhythm in patients who undergo mandibulectomy with Eichner B2 occlusal supports. 2. The wearable device is available for measuring masticatory rhythm in patients with jaw defects in clinics. 3. Changing times of the skin morphology on the back of the ear pinna indicate jaw movement stability through a simpler process than the coefficient of variation of cycle durations.

关键词： Masticatory Rhythm Jaw Movement Rhythm Tracking Device Maxillofacial Prosthesis Head and Neck Tumor Mandibulectomy

有机固体废弃物热解气掺氨预混旋流火焰结构研究

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

燃烧科学与技术 2024年第3期30卷 239-247页

作者：郭怡君陈达南李兴李军赖诗妮黄宏宇小林敬幸中国科学院广州能源研究所广州510640 中国科学技术大学工程科学学院合肥230027 中国科学技术大学能源科学与技术学院广州510640 江西省碳中和研究中心南昌330096 名古屋大学化学工程系名古屋464-8603

由于有机固废热解产生的热解气热值较低,在一些燃烧器中无法燃烧,故采用旋流和掺混无碳燃料氨气的燃烧方式,以增强热解气燃烧的稳定性并减少碳排放.采用化学发光法对火焰结构进行了实验研究,基于CCD摄像系统,获得了不同当量比和掺氨比... 详细信息

由于有机固废热解产生的热解气热值较低,在一些燃烧器中无法燃烧,故采用旋流和掺混无碳燃料氨气的燃烧方式,以增强热解气燃烧的稳定性并减少碳排放.采用化学发光法对火焰结构进行了实验研究,基于CCD摄像系统,获得了不同当量比和掺氨比下热解气/氨气/空气预混旋流火焰中OH^(*)的化学发光分布.结果表明,不同类型热解气的旋流火焰形态结构差异较大;当量比增加到φ≥1时,OH^(*)辐射强度增强;掺氨使得热解气燃烧极限范围变窄,氨气的添加不仅会抑制OH^(*)自由基的产生,且会对火焰结构产生影响.

关键词：热解气体氨气旋流燃烧预混火焰 OH^(*)化学发光

Explainable Analysis of Deep Learning Models for Coronavirus Disease (COVID-19) Classification with Chest X-Ray Images: Towards Practical Applications

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Open Journal of Medical Imaging 2022年第3期12卷 83-102页

作者： Eri Matsuyama Haruyuki Watanabe noriyuki Takahashi Faculty of Informatics University of Fukuchiyama Kyoto Japan School of Radiological Technology Gunma Prefectural College of Health Sciences Gunma Japan School of Hearth Sciences Fukushima Medical University Fukushima Japan

In recent years, with numerous developments of convolutional neural network (CNN) classification models for medical diagnosis, the issue of misrecognition/misclassification has become more and more important. Thus, research on misrecognition/misclassification has been progressing. This study focuses on the problem of misrecognition/misclassification of CNN classification models for coronavirus disease (COVID-19) using chest X-ray images. We construct two models for COVID-19 pneumonia classification by fine-tuning ResNet-50 architecture, i.e., a model retrained with full-sized original images and a model retrained with segmented images. The present study demonstrates the uncertainty (misrecognition/misclassification) of model performance caused by the discrepancy in the shapes of images at the phase of model construction and that of clinical applications. To achieve it, we apply three XAI methods to demonstrate and explain the uncertainty of classification results obtained from the two constructed models assuming for clinical applications. Experimental results indicate that the performance of classification models cannot be maintained when the type of constructed model and the geometric shape of input images are not matched, which may bring about misrecognition in clinical applications. We also notice that the effect of adversarial attack might be induced if the method of image segmentation is not performed properly. The results suggest that the best approach to obtaining a highly reliable prediction in the classification of COVID-19 pneumonia is to construct a model using full-sized original images as training data and use full-sized original images as the input when utilized in clinical applications.

关键词： Explainable AI CNN CXR Image COVID-19 Pneumonia

Appropriate leucine-richα-2 glycoprotein cut-off value for Japanese patients with ulcerative colitis

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

World Journal of Clinical Cases 2023年第32期11卷 7753-7760页

作者： Masanao Yamazato Shunichi Yanai Tomofumi Oizumi Makoto Eizuka Shun Yamada Yosuke Toya noriyuki Uesugi Tamotsu Sugai Takayuki Matsumoto Division of Gastroenterology and Hepatology Department of Internal MedicineIwate Medical UniversityYahaba 028-3695Japan Division of Molecular Diagnostic Pathology Iwate Medical UniversityYahaba 028-3695Japan

BACKGROUND It has been suggested that serum leucine-richα-2 glycoprotein(LRG)could be a novel monitoring biomarker for the assessment of disease activity in inflammatory bowel *** particular,the relationship between LRG levels and the endoscopically assessed activity of ulcerative colitis(UC)has become a matter of *** To clarify appropriate LRG cut-off values for the prediction of endoscopic and histologic remission in Japanese patients with *** This was a cross-sectional,single-center,observational study of Japanese patients with *** 213 patients with UC,in whom LRG was measured from September 2020 to February 2022,we recruited 30 patients for whom a total colonoscopy and measurements of LRG and C-reactive protein(CRP)were performed on the same *** retrospectively analyzed correlations between the LRG and CRP levels and endoscopic indices,including the Mayo endoscopic subscore and UC endoscopic index of *** Correlations between the LRG values and the Mayo endoscopic subscore or UC endoscopic index of severity were significant(r=0.754,P<0.0001;r=0.778,P<0.0001,respectively).There were also significant correlations between CRP levels and Mayo endoscopic subscore or UC endoscopic index of severity(r=0.599,P=0.0005;r=0.563,P=0.0012,respectively),although the correlation coefficients were higher for *** LRG cutoff value for predicting endoscopic remission was 13.4μg/mL for a Mayo endoscopic subscore of 0[area under the curve(AUC):0.871;95%confidence interval(CI):0.744-0.998],and 13.4μg/mL for an UC endoscopic index of severity of 0 or 1(AUC:0.904;95%CI:0.792-1.000).CONCLUSION LRG may be a surrogate marker for endoscopic activity in UC,with a cut-off value of around 13.4μg/mL for endoscopically inactive disease.

关键词： Ulcerative colitis Leucine-rich alpha-2 glycoprotein C-reactive protein Japanese patients

Redox active ionic liquid as efficient mediator and solvent for visible light-driven B_(12) catalytic reactions

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Green Energy & Environment 2019年第2期4卷 116-120页

作者： Hisashi Shimakoshi noriyuki Houfuku Li Chen Yoshio Hisaeda Department of Chemistry and Biochemistry Graduate School of Engineering Kyushu University

The redox active ionic liquid, 1-ethyl-4-(methoxycarbonyl)pyridinium bis(trifluoromethanesulfonyl)amide(RIL), was synthesized from its iodide form by an anion exchange reaction of Li(NTf_2) with viscos liquid(η= 122 cP at 25C) and characterized by NMR, IR, and elemental analysis. The compound showed reversible redox couples at -0.65 V and -1.48 V vs. Ag/AgCl and worked as an electron mediator in the B_(12) complex/[Ru(bpy)_3]Cl_2 photosensitizer catalytic system under visible light irradiation. The catalytic efficiency in the RIL was higher than those in DMF, MeOH, and the redox inactive ionic liquid, 1-butyl-3-methyl imidazolium bis(trifluoromethylsulfonyl)amide.

关键词： Ionic liquid Photosensitizer Redox active Visible light Vitamin B12

Physical Properties of Liquid Terbium Measured by Levitation Techniques

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Journal of Rare Earths 2007年第6期25卷 665-669页

作者： Paul-Francois Paradis Takehiko Ishikawa noriyuki Koike Yuki Watanabe Japan Aerospace Exploration Agency Institute of Space and Astronautical Science ISS Science Project Office 2-1-1 Sengen Tsukuba Ibaraki 305-8505 Japan Chiba Institute of Technology Tsudanuma Narashino Chiba Japan Advanced Engineering Service Co. Ltd. Tsukuba Ibaraki 305-0032 Japan

To understand the nature and behavior of rare earth metals in their liquid phases, accurate values of their physical properties are essential. However, to measure their physical properties, the samples should be maintained in liquid phases for prolonged time, and this raises a formidable challenge. This is mainly explained by their high melting temperatures （e.g., 1629 K for Tb）, high vapor pressure, and the risk of melt contamination with a crucible or support. An electrostatic levitation furnace alleviated these difficulties and allowed the determination of density, surface tension, and viscosity of several metals above their melting temperature. Here, first, the levitation furnace facility and the noncontact diagnostic procedures were briefly discussed, followed by the explanation of their thermophysical property measurements over wide temperature ranges. The density was obtained using an ultraviolet-based imaging technique that allowed excellent illumination, even at elevated temperatures. Over the 1615 to 1880 K temperature span, the density measurements could be expressed as p（T） =7.84 × 10^3 -0.47 （T - Tm）（kg · m^-3） with Tm = 1629 K, yielding a volume expansion coefficient a（T） = 6.0 × 10^-5 （K^-1）. In addition, the surface tension and the viscosity could be determined by inducing a drop oscillation to a molten sample. Using this technique, the surface tension data could be expressed as σ（T） = 8.93 × 10^2 - 0.10 （T - Tm）（mN· m^-1） and those for viscosity as η（T） =0.583 exp [4.1 × 10^4/（RT）] （MPa·s） over the 1690 to 1980 K temperature range

关键词： levitation liquid noncontact terbium rare earths

Design and Implementation of a Programming Learning Support System Using the Concepts of Physical Visualization and Serious Gaming

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Journal of Computer and Communications 2022年第11期10卷 197-223页

作者： noriyuki Ishiguro Yusuke Saitoh Kazuhiro Yamamoto Hirohide Haga Graduate School of Science and Engineering Doshisha University Kyoto Japan

The purpose of this research is the design and implementation of a support system for learning programming. To archive this purpose, in this article, we propose a Puzzle Programming System that uses jigsaw puzzles as an example of the application of physical visualization, which visualizes logical constraints to physical ones. This Puzzle Programming System aims to teach basic programming concepts by presenting the invisible constraints of programming language syntax using the visual constraints of jigsaw puzzle pieces. This system runs on an Apple iPad and was developed using the Unity game engine. We used YAML as a data format for serializing structured data for data management. By inviting high school students to try out a prototype, we could confirm the usefulness of the Puzzle Programming System. The experimental evaluation results also shed light on aspects of the game that need to be redesigned and parts where the visual programming model needs to be modified and expanded.

关键词： Serious Gaming Augmented Reality Visual Programming Constraint Visualization

A Breast Density Classification System for Mammography Considering Reliability Issues in Deep Learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Open Journal of Medical Imaging 2023年第3期13卷 63-83页

作者： Eri Matsuyama Megumi Takehara noriyuki Takahashi Haruyuki Watanabe Faculty of Informatics University of Fukuchiyama Kyoto Japan Breast Center Dokkyo Medical University Hospital Tochigi Japan School of Hearth Sciences Fukushima Medical University Fukushima Japan School of Radiological Technology Gunma Prefectural College of Health Sciences Gunma Japan

In a convolutional neural network (CNN) classification model for diagnosing medical images, transparency and interpretability of the model’s behavior are crucial in addition to high classification accuracy, and it is highly important to explicitly demonstrate them. In this study, we constructed an interpretable CNN-based model for breast density classification using spectral information from mammograms. We evaluated whether the model’s prediction scores provided reliable probability values using a reliability diagram and visualized the basis for the final prediction. In constructing the classification model, we modified ResNet50 and introduced algorithms for extracting and inputting image spectra, visualizing network behavior, and quantifying prediction ambiguity. From the experimental results, our proposed model demonstrated not only high classification accuracy but also higher reliability and interpretability compared to the conventional CNN models that use pixel information from images. Furthermore, our proposed model can detect misclassified data and indicate explicit basis for prediction. The results demonstrated the effectiveness and usefulness of our proposed model from the perspective of credibility and transparency.

关键词： Explainable AI t-SNE Entropy Wavelet Transform Mammogram