检索结果-内蒙古大学图书馆

Proceedings of the 5th International Conference on Advances in Image Processing

作者： Lu Cao Shan Huang Jianxin Zhang School of Computer Science and Engineering Dalian Minzu University China School of Computer Science and Engineering Dalian Minzu University China and Key Lab of Advanced Design and Intelligent Computing(Ministry of Education) Dalian University China

ISBN: (纸本)9781450385183

Recently, breast cancer histopathology image classification using convolutional neural networks has achieved more and more attentions with the great progress. To capture more discriminant deep features for the classification, this paper proposes a novel triplet-attention residual network, i.e., TAResNet, to distinguish the breast cancer histopathology image. TAResNet employs the representative ResNet18 model to extract deep features of histopathology images, followed by a triplet-attention module to further boost the discriminability of deep features through expanding feature diversity and enhancing inter-dimensional dependency. Extensive experiments carried out on the public BreakHis dataset well evaluate the effectiveness the given TAResNet model. More specifically, TAResNet achieves its optimal classification accuracy of 98.34% and 98.77% at the image level and patient level, respectively.

关键词： Breast cancer Triplet attention Histopathological image classification Convolutional neural network

来源：评论

学校读者我要写书评

暂无评论

Location IoU: A New Evaluation and Loss for Bounding Box Regression in Object Detection

Location IoU: A New Evaluation and Loss for Bounding Box Reg...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Lu Yang Kai Zhang Jiaqi Liu Chongke Bi Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control Tianjin University of Technology Tianjin China National Demonstration Center for Experimental Mechanical and Electrical Engineering Education Tianjin University of Technology Tianjin China Autobrain (Tianjin) Technology LTD Tianjin China College of Intelligence and Computing Tianjin University Tianjin China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

In the field of object detection bounding box regression, IoU (Intersection over Union) is a commonly used evaluation metric for measuring the overlap between predicted and ground truth bounding boxes. One limitation of IoU, however, is that the gradient goes to zero when there is no intersection between bounding boxes. Recent evaluation metrics have been developed to overcome this problem by incorporating penalty terms and other techniques for improving IoU. Despite these advances, existing methods still suffer to some degree from the problems of small gradients and slow rate of convergence. In this paper, a new evaluation method named LIoU (Location IoU) is proposed to address these issues. It can maximize the shared area of the intersection when used as the loss function and incorporate a variable gradient parameter to fit to datasets of different sizes. The experimental results indicate that we achieve 0.27% and 3.48% accuracy improvements respectively in SSD and YOLOv5 network on the VOC dataset, and 2.23% in the YOLOv5 network on the COCO dataset. The results of this study highlight the effectiveness of LIoU in improving the performance of object detection models.

关键词： YOLO Measurement Location awareness Accuracy Shape Neural networks Benchmark testing

来源：评论

学校读者我要写书评

暂无评论

Base Station Network Traffic Prediction Approach Based on LMA -DeepAR 6

Base Station Network Traffic Prediction Approach Based on LM...

引用

6th IEEE International Conference on Computer and Communication Systems, ICCCS 2021

作者： Zhang, Jiachen Zuo, Xingquan Xu, Mingying Han, Jing Zhang, Baisheng Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education Beijing Uinversity of Posts and Telecommunications Beijing China Key Laboratory of Intelligent Telecommunication Software and Multimedia Beijing Uinversity of Posts and Telecommunications Beijing China Network Interface Virtualization Shanghai System Design Deptment Zhongxing Telecommunication Equipment Corporation Shanghai China

ISBN: (纸本)9780738126043

Accurate network traffic prediction of base station cell is very vital for the expansion and reduction of wireless devices in base station cell. The burst and uncertainty of base station cell network traffic makes the network traffic nonlinear and nonstationary, which brings challenges to the long-term prediction of network traffic. In this paper, the traffic model LMA-DeepAR for base station network is established based on DeepAR. Acordding to the distribution characteristics of network traffic, this paper proposes an artificial feature sequence calculation method based on local moving average (LMA). The feature sequence is input into DeepAR as covariant, which makes the statistical characteristics of network traffic near a period of time in the past be considered when updating parameters, and the interference of non-stationary network traffic on model training will be reduced. Experimental results show that the proposed prediction approach (LMA-DeepAR) outperforms other methods in the overall long-term prediction performance and stability of multi cell network traffic. © 2021 IEEE.

关键词： Base stations

来源：评论

学校读者我要写书评

暂无评论

Modeling and Analysis of Mtj Array with Shielding Layer

SSRN

引用

SSRN 2023年

作者： Qu, Ying-Hui Xiong, Jing-Ya Zuo, Chao Dong, Kai-Feng Song, Jun-Lei Mo, Wen-Qin Hui, Ya-Juan Jin, Fang School of Automation China University of Geosciences Wuhan430074 China Hubei key Laboratory of Advanced Control and Intelligent Automation for Complex Systems Wuhan430074 China Geological Exploration Intelligent Technology Engineering Research Center The Ministry of Education Wuhan Second Ship Design and Research Institute Hubei Province Wuhan China

In STT-MRAM, data is stored in a magnetic tunnel structure (MTJ) containing multiple ferromagnetic layers. Each ferromagnetic layer generates a stray field, which affects the critical switching current, switching time, thermal stability, and the density of the MTJ array. These factors have a direct impact on the performance and storage capacity of STT-MRAM. Previously, we have studied that hexagonal MTJ arrays can improve storage density. However, this improvement also leads to increased intracellular and intercellular stray fields, which in turn affects the device's performance. This paper aims to reduce the intercellular stray field by covering each MTJ unit with a soft magnetic shielding layer and investigates the performance of the MTJ array with a shielding layer. Firstly, a simulation model of a single MTJ stack is built using the simulation software COMSOL. Then, based on modeling the MTJ stack, square and hexagonal MTJ arrays are established. The intercellular magnetic coupling coefficient (Ψ) is quoted, and Ψ=2% is defined as the threshold below which the intercellular magnetic coupling can be considered negligible. The results indicate that the soft magnetic shielding layer can effectively shield the stray field, reducing its impact on MTJ performance. Specifically, a thicker shielding layer with higher magnetic permeability provides better shielding efficacy. However, comparing square and hexagonal MTJ array after magnetic shielding reveals a larger stray field between the hexagonal array. © 2023, The Authors. All rights reserved.

关键词： Ferromagnetic materials

来源：评论

学校读者我要写书评

暂无评论

Speech Emotion Recognition Based on Three-Channel Feature Fusion of CNN and BiLSTM 2020

Speech Emotion Recognition Based on Three-Channel Feature Fu...

引用

4th International Conference on Innovation in Artificial Intelligence, ICIAI 2020

作者： Huang, Lilong Dong, Jing Zhou, Dongsheng Zhang, Qiang Key Laboratory of Advanced Design and Intelligent Computing Ministry of Education School of Software Engineering Dalian University Dalian China School of Computer Science and Technology Dalian University of Technology Dalian China

ISBN: (纸本)9781450376587

It is important for human-computer natural interaction that the computer can quickly and effectively identify the interactor's emotional state. We proposed a three-channel feature fusion algorithm based on Convolutional Neural Network and Bidirectional Long Short-Term Memory Network for speech emotion recognition. Mel-Frequency Cepstral Coefficients and its first and second-order difference features were input into three channels and trained respectively for exploiting the emotional information contained in the dynamic features. By splicing and fusing the features obtained by three channels, the recognition rates of proposed algorithm reached 96.562% and 92.34% on CASIA corpus with six emotional states and on EMO-DB corpus with seven emotional states respectively. © 2020 ACM.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Missing Data Recovery for Human Mocap Data Based on A-LSTM and LS Constraint 5

Missing Data Recovery for Human Mocap Data Based on A-LSTM a...

引用

5th IEEE International Conference on Signal and Image Processing, ICSIP 2020

作者： Ji, Lu Liu, Rui Zhou, Dongsheng Zhang, Qiang Wei, Xiaopeng Key Laboratory of Advanced Design and Intelligent Computing Ministry of Education School of Software Engineering Dalian University Dalian China School of Computer Science and Technology Dalian University of Technology Dalian China

ISBN: (纸本)9781728168968

Human mocap (motion capture) data has been widely used in many fields. However, influenced by the occlusion and falling, missing of markers is inevitable in optical motion capture system. According to the temporal and spatial characteristics of mocap data, a missing data recovery method based on attention-based LSTM network (A-LSTM) and least-squares (LS) constraint is proposed in this paper. First, the LSTM is chosen as the basic network to estimate the position of the missing markers, in which an attention mechanism is introduced to learn the temporal continuity of human mocap data. Then, a space affine transformation based on least-squares fitting is adopted as the spatial constraint to refine the estimated positions, after which the restored error can be reduced significantly. Experiments show that the proposed A-LSTM+LS method is robust to the length of missing gap and the number of missing markers. Under the same conditions, our method achieves smaller restore error than the state of the art. © 2020 IEEE.

关键词： Least squares approximations

来源：评论

学校读者我要写书评

暂无评论

Optimal Minimum Energy Path Planning of Mobile Robot Based on Improved WOA-GA with Hybrid Strategy for Community-Centric Application

Optimal Minimum Energy Path Planning of Mobile Robot Based o...

引用

2023 International Conference on Machine Learning and Cybernetics, ICMLC 2023

作者： Du, Jiang Da-Lin, Zhou Gong-Fa, Li Bao-Li, Lu Jun-Tong, Yun Ying, Liu Wuhan University of Science and Technology Key Laboratory of Metallurgical Equipment and Control Technology Ministry of Education Wuhan430081 China School of Computing University of Portsmouth PortsmouthPO1 3HE United Kingdom China Three Gorges University Hubei Key Laboratory of Hydroelectric Machinery Design & Maintenance Yichang443002 China Wuhan University of Science and Technology Key Laboratory of Mechanical Transmission and Manufacturing Engineering of Hubei Province Wuhan430081 China Precision Manufacturing Research Institute Wuhan University of Science and Technology Wuhan430081 China Institute of Semiconductors Chinese Academy of Sciences Beijing100083 China Research Center for Biomimetic Robot and Intelligent Measurement and Control Wuhan University of Science and Technology Wuhan430081 China

ISBN: (纸本)9798350303780

The path planning of mobile robots is investigated to minimize energy loss in intelligent manufacturing. To achieve optimal energy consumption planning, this paper proposes a hybrid strategy that enhances the whale optimization algorithm (WOA) and the genetic algorithm (GA). The WOA-GA algorithm combines their respective advantages and effectively mitigates premature convergence exhibited by the standard traditional algorithm. To enhance the global and local search capabilities of the standard WOA, we employ adaptive weighting and a Cauchy mutation strategy. Additionally, we introduce an adaptive crossover mutation probability and an elite strategy to mitigate the detrimental effects of the genetic algorithm's evolutionary operation in the late iterations, thereby improving its performance. This paper compares the performance of the standard GA, the standard WOA-GA algorithm, and the enhanced WOA-GA algorithm. The experimental results demonstrate that the enhanced WOA-GA algorithm can achieve optimal energy-efficient path planning using the hybrid strategy. The average energy consumption of the enhanced WOA-GA algorithm can be reduced by 30%, while the maximum energy consumption can be reduced by 47%. The results confirm the effectiveness and feasibility of the proposed method in enhancing the level of green manufacturing. © 2023 IEEE.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Acoustic Scene Classification by Enhanced High-Frequency Weak Signal Characteristic Spectrum 19

Acoustic Scene Classification by Enhanced High-Frequency Wea...

引用

Proceedings of the 2019 8th International Conference on computing and Pattern Recognition

作者： Wenjie Hao Lasheng Zhao Qiang Zhang Key Laboratory of Advanced Design and Intelligent Computing Ministry of Education Dalian University

ISBN: (纸本)9781450376570

In the acoustic scene classification task, the method of using mel-spectrogram to express the acoustic scene information is widely applied. However, mel-spectrogram has defects and it ignores important information about some acoustic scenes. The paper improves the mel-spectrogram and its generation algorithm. Including: I. For the sensitivity of the acoustic scene to high-frequency acoustic signals, the paper changes the filter design method of Mel Frequency Cepstrum Coefficient (MFCC). This method preserves more high frequency information by applying the equal-height triangular filter banks and increasing the number of the filters. II. Based on the previous step, an enhancement algorithm is proposed for the problem of the lack of high-frequency weak signals in the characteristic spectrum. The algorithm performs nonlinear mapping on the mel-spectrogram, which makes the transformed high-frequency weak signal feature information more obvious. The algorithm is verified by DCASE 2018 acoustic scene classification dataset and LITIS ROUEN dataset. The experimental results demonstrate the effectiveness of the proposed algorithm.

关键词： MFCC weak signal feature extraction Acoustic scene classification

来源：评论

学校读者我要写书评

暂无评论

MLFuse: Multi-Scenario Feature Joint Learning for Multi-Modality Image Fusion

引用

IEEE Transactions on Multimedia 2025年

作者： Lei, Jia Li, Jiawei Liu, Jinyuan Wang, Bin Zhou, Shihua Zhang, Qiang Wei, Xiaopeng Kasabov, Nikola K. Dalian University Key Laboratory of Advanced Design and Intelligent Computing Ministry of Education School of Software Engineering Dalian116622 China University of Science and Technology Beijing School of Computer and Communication Engineering Beijing100083 China Dalian University of Technology School of Mechanical Engineering Dalian116024 China Dalian University of Technology School of Computer Science and Technology Dalian116024 China Auckland University of Technology Knowledge Engineering and Discovery Research Institute Auckland1010 New Zealand Ulster University Intelligent Systems Research Center LondonderryBT52 1SA United Kingdom

Multi-modality image fusion (MMIF) entails synthesizing images with detailed textures and prominent objects. Existing methods tend to use general feature extraction to handle different fusion tasks. However, these methods have difficulty breaking fusion barriers across various modalities owing to the lack of targeted learning routes. In this work, we propose a multi-scenario feature joint learning architecture, MLFuse, that employs the commonalities of multi-modality images to deconstruct the fusion progress. Specifically, we construct a cross-modal knowledge reinforcing network that adopts a multipath calibration strategy to promote information communication between different images. In addition, two professional networks are developed to maintain the salient and textural information of fusion results. The spatial-spectral domain optimizing network can learn the vital relationship of the source image context with the help of spatial attention and spectral attention. The edge-guided learning network utilizes the convolution operations of various receptive fields to capture image texture information. The desired fusion results are obtained by aggregating the outputs from the three networks. Extensive experiments demonstrate the superiority of MLFuse for infrared-visible image fusion and medical image fusion. The excellent results of downstream tasks (i.e., object detection and semantic segmentation) further verify the high-quality fusion performance of our method. The code is publicly available at https://***/jialei-sc/MLFuse. © 1999-2012 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

ESNet: evolution and succession network for high-resolution salient object detection 24

ESNet: evolution and succession network for high-resolution ...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Hongyu Liu Runmin Cong Hua Li Qianqian Xu Qingming Huang Wei Zhang Institute of Information Science Beijing Jiaotong University & Beijing Key Laboratory of Advanced Information Science and Network Technology Beijing China School of Control Science and Engineering Shandong University & Key Laboratory of Machine Intelligence and System Control Ministry of Education Jinan China School of Computer Science and Technology Hainan University Hainan China Key Laboratory of Intelligent Information Processing Institute of Computing Technology CAS Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

Preserving details and avoiding high computational costs are the two main challenges for the High-Resolution Salient Object Detection (HRSOD) task. In this paper, we propose a two-stage HRSOD model from the perspective of evolution and succession, including an evolution stage with Low-resolution Location Model (LrLM) and a succession stage with High-resolution Refinement Model (HrRM). The evolution stage achieves detail-preserving salient objects localization on the low-resolution image through the evolution mechanisms on supervision and feature; the succession stage utilizes the shallow high-resolution features to complement and enhance the features inherited from the first stage in a lightweight manner and generate the final high-resolution saliency prediction. Besides, a new metric named Boundary-Detail-aware Mean Absolute Error (MAEBD) is designed to evaluate the ability to detect details in high-resolution scenes. Extensive experiments on five datasets demonstrate that our network achieves superior performance at real-time speed (49 FPS) compared to state-of-the-art methods. Our code is publicly available at: https://***/rmcong/ESNet_ICML24.

关键词：

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：