检索结果-内蒙古大学图书馆

arXiv 2022年

作者： He, Kelei Gan, Chen Li, Zhuoyuan Rekik, Islem Yin, Zihao Ji, Wen Gao, Yang Wang, Qian Zhang, Junfeng Shen, Dinggang Medical School of Nanjing University Nanjing China National Institute of Healthcare Data Science Nanjing University Nanjing China State Key Laboratory for Novel Software Technology Nanjing University Nanjing China BASIRA Lab Faculty of Computer and Informatics Engineering Istanbul Technical University Istanbul Turkey School of Science and Engineering Computing University of Dundee United Kingdom School of Biomedical Engineering ShanghaiTech University Shanghai China Department of Research and Development Shanghai United Imaging Intelligence Co. Ltd. Shanghai China

Transformers have dominated the field of natural language processing, and recently impacted the computer vision area. In the field of medical image analysis, Transformers have also been successfully applied to full-stack clinical applications, including image synthesis/reconstruction, registration, segmentation, detection, and diagnosis. Our paper aims to promote awareness and application of Transformers in the field of medical image analysis. Specifically, we first overview the core concepts of the attention mechanism built into Transformers and other basic components. Second, we review various Transformer architectures tailored for medical image applications and discuss their limitations. Within this review, we investigate key challenges revolving around the use of Transformers in different learning paradigms, improving the model efficiency, and their coupling with other techniques. We hope this review can give a comprehensive picture of Transformers to the readers in the field of medical image analysis. © 2022, CC BY-SA.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Contextualizing Meta-Learning via Learning to Decompose

arXiv

引用

arXiv 2021年

作者： Ye, Han-Jia Zhou, Da-Wei Hong, Lanqing Li, Zhenguo Wei, Xiu-Shen Zhan, De-Chuan State Key Laboratory for Novel Software Technology Nanjing University Nanjing210023 China Huawei Noah’s Ark Lab Canada The School of Computer Science and Engineering Southeast University Nanjing210096 China Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications Southeast University Ministry of Education Nanjing210096 China

Meta-learning has emerged as an efficient approach for constructing target models based on support sets. For example, the meta-learned embeddings enable the construction of target nearest-neighbor classifiers for specific tasks by pulling instances closer to their same-class neighbors. However, a single instance can be annotated from various latent attributes, making visually similar instances inside or across support sets have different labels and diverse relationships with others. Consequently, a uniform meta-learned strategy for inferring the target model from the support set fails to capture the instance-wise ambiguous similarity. To this end, we propose Learning to Decompose Network (LEADNET) to contextualize the meta-learned "support-to-target" strategy, leveraging the context of instances with one or mixed latent attributes in a support set. In particular, the comparison relationship between instances is decomposed w.r.t. multiple embedding spaces. LEADNET learns to automatically select the strategy associated with the right attribute via incorporating the change of comparison across contexts with polysemous embeddings. We demonstrate the superiority of LEADNET in various applications, including exploring multiple views of confusing data, out-of-distribution recognition, and few-shot image classification. Copyright © 2021, The Authors. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

GCSTormer: Gated swin transformer with channel weights for image denoising

引用

Expert Systems with Applications 2025年 284卷

作者： Zhu, Minling Xu, Zhixin Zhang, Qi Liu, Yonglin Gu, Dongbing Xu, Sendren Shengdong School of Computer Science Beijing Information Science and Technology University Beijing100101 China State Key Laboratory for Novel Software Technology Nanjing University Jiangsu Nanjing China School of Economics and Management Harbin Institute of Technology Shandong Weihai China School of Public Management and Media Beijing Information Science and Technology University Beijing100192 China School of Computer Science and Electronic Engineering University of Essex CO4 3SQ Colchester United Kingdom Department of Power Mechanical Engineering National Tsing Hua University Taiwan Hsinchu300044 China

In recent years, deep learning models typified by CNN and Transformer have achieved remarkable success in computer vision. However, CNN struggles to capture global information and lacks effective cross-region interaction. Transformer excels in global feature modeling but comes with high computational complexity and a large number of *** this paper, we propose GCSTormer, a novel image denoising model that integrates shallow and deep feature extraction. For shallow feature extraction, we employ a three-layer atrous convolution with dense skip connections. This design enhances CNN's ability to capture global information. In the main architecture of the model, we utilize the U-Net structure to learn multi-scale image information and adopt the improved Swin Transformer block as the basic layer. This block has been improved in two aspects. First, we use gated convolution to optimize information flow. Second, we apply transposed attention to compute attention in the channel domain. To evaluate the model performance, we conduct denoising experiments on grayscale and color images with synthetic Gaussian noise. The experimental results show our model has good results in both quantitative metrics and visual quality while maintaining a certain number of parameters and complexity. © 2025 Elsevier Ltd

关键词： Image denoising

来源：评论

学校读者我要写书评

暂无评论

Development of an Artificial Intelligence Model for Individualized FSH Daily Dosage Guidance During Ovulation Stimulation

SSRN

引用

SSRN 2023年

作者： Duan, Liyan Xia, Yu Mei, Jie Zhang, Hui Zhu, YingChun Huang, ChenYnag Yan, Guijun Kong, Na Li, Wujun Sun, Haixiang Center for Reproductive Medicine and Obstetrics and Gynecology Drum Tower Clinic Medical College Nanjing Medical University Nanjing210008 China Center for Molecular Reproductive Medicine Nanjing University Nanjing210008 China National Institute of Healthcare Data Science Nanjing University Jiangsu Nanjing210008 China National Key Laboratory for Novel Software Technology Department of Computer Science and Technology Nanjing University China Center for Medical Big Data Nanjing Drum Tower Hospital Affifiliated Drum Tower Hospital Medical School of Nanjing University Jiangsu Nanjing210008 China State Key Laboratory of Reproductive Medicine China International Joint Research Center on Environment and Human Health Center for Global Health School of Public Health Nanjing Medical University Nanjing211166 China Department of Reproductive Medicine Center Nanjing Drum Tower Hospital State Key Laboratory of Pharmaceutical Biotechnology Nanjing University Nanjing210023 China

Background: We aim to develop a real-time, interpretable artificial intelligence (AI) system, using an algorithm for estimating follicle stimulating hormone (FSH) daily dose in individualized controlled ovarian stimulation (COS).Methods: Between January 2018 and December 2020, a total of 13,788 cycles of GnRHa long-protocol ovulation stimulation cycles were included from the Reproductive Medicine Centre of Nanjing Drum Tower Hospital. We used a method of cross-temporal and cross-feature joint encoding to learn the time series information X=X1, X2,..., XN of the N days to get its hidden layer representation H. Then, we applied K-nearest neighbors (KNN) with the hidden layer representation H to obtain the ***: Our proposed CTFE model obtained Accuracy and Weighted-F1 metrics of 0.737 and 0.733, to predict the daily FSH dose respectively, which are the highest among other popular models for processing time series data. On days 1 and 5, Lasso regression predicted the FSH daily dosage with an accuracy of 0.696 and 0.522, respectively, whereas our technique predicted it with 0.834 and 0.819. The introduction of a sliding window mechanism after 13 days helps to maintain good performance, although accuracy drops due to insufficient ***: Our study demonstrated that the CTFE model can be used to calculate the daily FSH dose with high accuracy. This model may be easily applied in daily clinical practice by ***: This work was supported by the national Natural science Foundation of China (NO.82030040), Scientific research project of Jiangsu Province Association of Maternal and Child Health (No. FYX202203).Declaration of Interest: HX S, WJ L,GJ Y, N K, J M, LY D and Y X are listed as inventors of patents relating to this work. The authors affirm that they have no known financial or interpersonal conflicts that would have seemed to have an impact on the research presented in this *** Approval: The study had ethical approval fro

关键词： Hospitals

来源：评论

学校读者我要写书评

暂无评论

Using genetic algorithm to augment test data for penalty prediction

引用

International Journal of Performability Engineering 2020年第7期16卷 1078-1086页

作者： Xia, Chunyan Wang, Xingya Zhang, Yan Yang, Hao Department of Computer and Information Technology Mudanjiang Normal University Mudanjiang157011 China State Key Laboratory for Novel Software Technology Nanjing University Nanjing210023 China College of Computer Science and Technology Nanjing Tech University Nanjing210009 China

With the development of smart court construction, a deep learning method has been introduced into the field of penalty prediction based on judicial text. Since the increasing parameters of the penalty prediction model, the size of the data set to test the performance of the model is gradually expanding. First, we use the data augmentation method to make some changes to the original data to obtain a large number of augmented data with the same label. Then, we use the multi-objective genetic algorithm to search for high-quality test data from a large number of augmented data, so as to improve the diversity of augmented data. Finally, we perform experiments. The results of actual judicial cases show that compared with the random method, augmented test data based on the genetic algorithm can better test the performance of the penalty prediction model. © 2020 Totem Publisher, Inc. All rights reserved.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Training, testing and benchmarking medical AI models using Clinical AIBench

引用

BenchCouncil Transactions on Benchmarks, Standards and Evaluations 2022年第1期2卷 100037-100037页

作者： Huang, Yunyou Miao, Xiuxia Zhang, Ruchang Ma, Li Liu, Wenjing Zhang, Fan Guan, Xianglong Liang, Xiaoshuang Lu, Xiangjiang Tang, Suqing Zhang, Zhifei Guangxi Key Lab of Multi-Source Information Mining & Security School of Computer Science and Engineering & School of Software Guangxi Normal University Guilin China State Key Laboratory of Computer Architecture Institute of Computing Technology Chinese Academy of Sciences Beijing China Guilin Medical University Guilin China Department of Physiology and Pathophysiology Capital Medical University Beijing China Faculty of Education Guangxi Normal University Guilin China International Open Benchmark Council Beijing China

AI technology has been used in many clinical research fields, but most AI technologies are difficult to land in real-world clinical settings. In most current clinical AI research settings, the diagnosis task is to identify different types of diseases among the given ones. However, the diagnosis in real-world settings needs dynamically developing inspection strategies based on the existing resources of medical institutions and identifying different kinds of diseases out of many possibilities. To promote the development of different clinical AI technologies and the implementation of clinical applications, we propose a benchmark named Clinical AIBench for developing, verifying, and evaluating clinical AI technologies in real-world clinical settings. Specifically, Clinical AIBench can be used for: (1) Model training and testing: Researchers can use the data to train and test their models. (2)Model evaluation: Researchers can use Clinical AIBench to objectively, fairly, and comparably evaluate various models of different researchers. (3) Clinical value evaluation: Researchers can use the clinical indicators provided by Clinical AIBench to evaluate the clinical value of models, which will be applied in real-world clinical settings. For convenience, Clinical AIBench provides three different levels of clinical settings: restricted clinical setting, which is named closed clinical setting, data island clinical setting, and real-world clinical setting, which is called open clinical setting. In addition, Clinical AIBench covers three diseases: Alzheimer's disease, COVID-19, and dental. Clinical AIBench provides python APIs to researchers. The data and source code are publicly available from the project website https://***/clinical_aibench/. © 2022 The Authors

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Non-Hermitian Chiral Heat Transport

引用

Physical Review Letters 2023年第26期130卷 266303-266303页

作者： Guoqiang Xu Xue Zhou Ying Li Qitao Cao Weijin Chen Yunfeng Xiao Lan Yang Cheng-Wei Qiu Department of Electrical and Computer Engineering National University of Singapore Kent Ridge 117583 Republic of Singapore School of Computer Science and Information Engineering Chongqing Technology and Business University Chongqing 400067 China Interdisciplinary Center for Quantum Information State Key Laboratory of Extreme Photonics and Instrumentation ZJU-Hangzhou Global Scientific and Technological Innovation Center Zhejiang University Hangzhou 310027 China International Joint Innovation Center Key Lab of Advanced Micro/Nano Electronic Devices and Smart Systems of Zhejiang The Electromagnetics Academy at Zhejiang University Zhejiang University Haining 314400 China State Key Laboratory for Mesoscopic Physics and Frontiers Science Center for Nano-optoelectronics School of Physics Peking University100871 Beijing China Department of Electrical and Systems Engineering Washington University St. Louis Missouri 63130 USA

Exceptional point (EP) has been captivated as a concept of interpreting eigenvalue degeneracy and eigenstate exchange in non-Hermitian physics. The chirality in the vicinity of EP is intrinsically preserved and usually immune to external bias or perturbation, resulting in the robustness of asymmetric backscattering and directional emission in classical wave fields. Despite recent progress in non-Hermitian thermal diffusion, all state-of-the-art approaches fail to exhibit chiral states or directional robustness in heat transport. Here we report the first discovery of chiral heat transport, which is manifested only in the vicinity of EP but suppressed at the EP of a thermal system. The chiral heat transport demonstrates significant robustness against drastically varying advections and thermal perturbations imposed. Our results reveal the chirality in heat transport process and provide a novel strategy for manipulating mass, charge, and diffusive light.

关键词： Diffusion Heat transfer Metamaterials Non-Hermitean systems

来源：评论

学校读者我要写书评

暂无评论

Exploiting Probabilistic Siamese Visual Tracking with a Conditional Variational Autoencoder

Exploiting Probabilistic Siamese Visual Tracking with a Cond...

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Wenhui Huang Jason Gu Peiyong Duan Sujuan Hou Yuanjie Zheng School of Information Science and Engineering Shandong Normal University China Dalhousie University Halifax Canada Key Lab of Intelligent Computing & Information Security in Universities of Shandong Shandong Normal University China Shandong Provincial Key Laboratory for Novel Distributed Computer Software Technology Shandong Normal University China

Visual tracking is a fundamental capability for robots tasked with humans and environment interaction. However, state-of-the-art visual tracking methods are still prone to failures and are imprecise when applied to challenging stereos, and their results are generally confidence agonistic. These methods depend on an embedded deep learning model to provide deterministic features or regression maps. A deterministic output with low confidence can result in disastrous consequences and lacks evidence needed for subsequent operations. Moreover, training data ambiguities or noise in the observations (so-called data uncertainty) can also lead to inherent uncertainty. In this paper, we focus on exploiting probabilistic Siamese visual tracking with a conditional variational autoencoder (CVAE). First, we build a bridge between the Siamese architecture and the CVAE and propose a novel Bayesian visual tracking method. Second, the proposed method generates a complete probability distribution that enables the production of multiple plausible tracking outputs. Third, CVAE conditioned by ground truth data encodes a low-dimensional latent space and conducts noise-injection training to prevent overfitting. Our proposed tracking method outperformed the state-of-the-art trackers on the VOT2016, VOT2018 and TColor-128 datasets.

关键词： Bridges Training Visualization Uncertainty Training data Probabilistic logic Probability distribution

来源：评论

学校读者我要写书评

暂无评论

Minimizing dynamic regret and adaptive regret simultaneously

arXiv

引用

arXiv 2020年

作者： Zhang, Lijun Lu, Shiyin Yang, Tianbao National Key Laboratory for Novel Software Technology Nanjing University Nanjing210023 China Department of Computer Science University of Iowa Iowa CityIA52242 United States

Regret minimization is treated as the golden rule in the traditional study of online learning. However, regret minimization algorithms tend to converge to the static optimum, thus being suboptimal for changing environments. To address this limitation, new performance measures, including dynamic regret and adaptive regret have been proposed to guide the design of online algorithms. The former one aims to minimize the global regret with respect to a sequence of changing comparators, and the latter one attempts to minimize every local regret with respect to a fixed comparator. Existing algorithms for dynamic regret and adaptive regret are developed independently, and only target one performance measure. In this paper, we bridge this gap by proposing novel online algorithms that are able to minimize the dynamic regret and adaptive regret simultaneously. In fact, our theoretical guarantee is even stronger in the sense that one algorithm is able to minimize the dynamic regret over any interval. Copyright © 2020, The Authors. All rights reserved.

关键词： Comparators (optical)

来源：评论

学校读者我要写书评

暂无评论

TEA: Temporal Excitation and Aggregation for Action Recognition

TEA: Temporal Excitation and Aggregation for Action Recognit...

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Yan Li Bin Ji Xintian Shi Jianguo Zhang Bin Kang Limin Wang Platform and Content Group (PCG) Tencent State Key Laboratory for Novel Software Technology Nanjing University China Department of Computer Science and Engineering Southern University of Science and Technology China

ISBN: (数字)9781728171685

ISBN: (纸本)9781728171692

Temporal modeling is key for action recognition in videos. It normally considers both short-range motions and long-range aggregations. In this paper, we propose a Temporal Excitation and Aggregation (TEA) block, including a motion excitation (ME) module and a multiple temporal aggregation (MTA) module, specifically designed to capture both short- and long-range temporal evolution. In particular, for short-range motion modeling, the ME module calculates the feature-level temporal differences from spatiotemporal features. It then utilizes the differences to excite the motion-sensitive channels of the features. The long-range temporal aggregations in previous works are typically achieved by stacking a large number of local temporal convolutions. Each convolution processes a local temporal window at a time. In contrast, the MTA module proposes to deform the local convolution to a group of sub-convolutions, forming a hierarchical residual architecture. Without introducing additional parameters, the features will be processed with a series of sub-convolutions, and each frame could complete multiple temporal aggregations with neighborhoods. The final equivalent receptive field of temporal dimension is accordingly enlarged, which is capable of modeling the long-range temporal relationship over distant frames. The two components of the TEA block are complementary in temporal modeling. Finally, our approach achieves impressive results at low FLOPs on several action recognition benchmarks, such as Kinetics, Something-Something, HMDB51, and UCF101, which confirms its effectiveness and efficiency.

关键词： Two dimensional displays Videos Spatiotemporal phenomena Optical imaging Feature extraction Three-dimensional displays Stacking

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：