检索结果-内蒙古大学图书馆

2017 2nd International Conference on Artificial Intelligence: Techniques and Applications (AITA 2017)

作者： Hamdi Ahmed Rajeh Zhi-yong LI Al-Ghaili Mohammed College of Computer Science and Electronic Engineering Hunan University Key Laboratory for Embedded and Network Computing of Hunan Province

This paper presents a comparative study of two approaches to statistical machine translation(SMT).We present a study on Factored Machine Translation for the Arabic–English pair of *** illustrate pre-processing step for the Arabic source language and the new factors which added to the English target *** experiments that injected English side by part-ofspeech(POS) tags,Combinatory Categorical Grammar(CCG) supertags and segmented Arabic sentences displayed a considerable progress in terms of the BLEU *** experiments gained the premier examine of the Baseline phrase-based models during two approaches,Segmented(#S1) and non-segmented(#S2).In both approaches,CCG models acquired the greatest BLEU *** results show that the segmented approach which consists of CCG produce the highest accuracy at 30.91%(1.84% superior than baseline).As shown,the results presented considerable improvement in translating the segmented Arabic rather than the non-segmented into English language.

关键词： Phrase-based translation model Combinatory Categorial Grammar Part-of-speech Factored translation model Word segmentation

来源：评论

学校读者我要写书评

暂无评论

A key frame extraction algorithm oriented for video copy detection based on PCC and the histogram of video frames

引用

Journal of Computational Information Systems 2014年第16期10卷 6999-7006页

作者： Peng, Fei Qian, Qin Long, Min College of Computer Science and Electronics Engineering Hunan University Changsha China Key Laboratory for Embedded and Network Computing of Hunan Province Changsha China College of Computer and Communication Engineering Changsha University of Science and Technology Changsha China

Aim to improve the accuracy and the robustness of the video copy detection, a novel key frame extraction algorithm is proposed. The shot segmentation is accomplished by evaluating the Pearsons correlation coefficients (PCC) between adjacent frames, while the key frames are extracted by combining the boundary-based and visual content-based methods. Experimental results show that the proposed key frame extraction algorithm is effective, and it is robust against some content preserving video processing attacks for some existed video hash methods. It provides a novel effective method for key-frames extraction in the video copy detection. 1553-9105/Copyright © 2014 Binary Information Press

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

Design and implementation of a fall detection monitor system with a voice interaction function for smartphones

Design and implementation of a fall detection monitor system...

引用

IEEE International Conference on Consumer Electronics - Berlin (ICCE-Berlin)

作者： Ying-Wen Bai Chia-Hao Yu Chun-Cheng Chan Key Laboratory for Embedded and Network Computing of Hunan Province Hunan University Changsha China Department of Computer Science The University of Texas at Dallas Richardson Texas

In this paper a smartphone fall detection system featuring a three-axis accelerometer sensor with a voice interaction function is presented. The value, measured by the three-axis accelerometer sensor, is sent to the program. It is then calculated and compared to the fall characteristics. When a fall is detected, the voice function asks whether the user needs help. If the fall is not serious, the user can cancel this emergency notification by means of either the voice interaction function or the touch panel. Otherwise the smartphone would automatically send out an emergency signal to the help center to request assistance for the user.

关键词： Smart phones Speech recognition Acceleration Accelerometers Global Positioning System Monitoring Gravity

来源：评论

学校读者我要写书评

暂无评论

A scheduling algorithm in the randomly heterogeneous multi-core processor

A scheduling algorithm in the randomly heterogeneous multi-c...

引用

International Conference on Fuzzy Systems and Knowledge Discovery (FSKD)

作者： Yan Liu Yongwei Li Yihong Zhao Xiaoming Chen College of Computer Science and Electronic Engineering Key Laboratory for Embedded and Network Computing of Hunan Province Changsha China

ISBN: (纸本)9781509040940

The increasing scale of multi-core processors are likely to be randomly heterogeneous by design or because of diversity and flaws. The latter type of heterogeneity introduced by some unforeseen variable factors such as the manufacturing process variation is especially challenging because of its unpredictability. In this environment, thread scheduler and global power manager must handle such randomly heterogeneous. Furthermore, these algorithms must supply high efficiency, scalability and low overhead because future multi-core processors may have a number of cores on a single die. This paper presents a variation-aware scheduling algorithm for application scheduling and power management. Thread switching and sampling among different cores in the multi-core processor introduce obvious overhead than previous many-core scheduling algorithms. Proposed scheme records the information of swapped thread of preferential core and uses tabu search-based randomly heterogeneous scheduling algorithm(TSR) to avoid the occurrence of repeated sampling and reduce the migration frequency and sampling frequency of a thread. The experimental results show that TSR algorithm has decreased 45.7% of thread migration and 42.2% of the sampling time as compared with local search algorithm. This paper regards the transcendental Hungarian offline scheduling algorithm as the baseline. ED2 of TSR only decrease by 8.58% as compared with that of Hungarian offline scheduling algorithm, but compared with the random search scheduling algorithm, ED2 of TSR decreased by 39.4%.

关键词： Scheduling algorithms Scheduling Multicore processing Algorithm design and analysis Heuristic algorithms Optimal scheduling Schedules

来源：评论

学校读者我要写书评

暂无评论

Image reconstruction for denoising based on compressive sensing

引用

Metallurgical and Mining Industry 2015年第10期7卷 106-112页

作者： Zhou, Jianhua Zhou, Siwang Hunan University Changsha Hunan China Hunan Police Academy Changsha Hunan China Key Laboratory of Network Crime Investigation Colleges of Hunan Province Changsha Hunan China

Due to the disadvantage of large amounts of data computation and image quality degradation of classical reconstruction algorithm, a novel adaptive method of image reconstruction denoising based on compressive sensing is proposed. Firstly, the wavelet approximate coefficients and detail coefficients from the image noise are Gaussian distribution, and have different variances in different levels. Secondly, the noise image is divided into image blocks of a certain size, a new compression sensing block reconstruction method has been used to recover small block coefficients. Finally, the reconstructed denoising images are obtained based on recovered detail coefficients and approximate coefficients by the separation of small block wavelet inversed transform. Experimental results show that this method is feasible and available, compared with pure wavelet denoising and block image, signal-to-noise ratio has been improved highly, the image noise has been removed effectively and the reconstructed image quality has been improved highly.

关键词： Wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

A deep hashing method based on attention module for image retrieval 13

A deep hashing method based on attention module for image re...

引用

13th International Conference on Intelligent Computation Technology and Automation, ICICTA 2020

作者： Long, Jun Wei, Xiangxiang Qi, Qianqian Wang, Ye School of Computer Science and Engineering Central South University Changsha410083 China Network Resources Management and Trust Evaluation Key Laboratory of Hunan Province Changsha410083 China

ISBN: (纸本)9781728186665

Hashing retrieval technique has received widespread attention since its low storage consumption and fast retrieval speed. Image retrieval based on deep hashing method uses image features with high-level semantic information and discriminative expression capabilities. However, there are still some limitations: (1) Compared with the traditional feature vector, the length of the learned hash codes is short, and each bit of the code is -1 or 1, the discrimination capability of feature representation is limited;(2) the existing deep hash algorithms cannot directly learn the discrete hash codes using sign activation function, therefore, they use relaxation-based scheme to learn the hash codes, resulting in a large quantization error. In order to solve the above problems, a deep hashing retrieval algorithm incorporate with attention model is proposed. This algorithm embeds spatial and channel attention models in the feature extraction network, and uses a novel activation function in the hash layer. In addition, a novel triple loss function is further proposed to improve the learning capability of the network. Finally, experimental results on the two benchmark databases, i.e., CIFAR-10 and MINIST datasets, show the effectiveness and superiority compared with the baselines. © 2020 IEEE.

关键词： Image retrieval

来源：评论

学校读者我要写书评

暂无评论

Holistic WCRT Analysis for Global Fixed-Priority Preemptive Multiprocessor Scheduling

Holistic WCRT Analysis for Global Fixed-Priority Preemptive ...

引用

Design Automation Conference

作者： Guoqi Xie Chenglai Xiong Wei Wu Renfa Li Wanli Chang Key Laboratory for Embedded and Network Computing of Hunan Province College of Computer Science and Electronic Engineering Hunan University China

Many embedded applications demand both resource efficiency and timing guarantee. However, resource sharing naturally complicates the analysis that extracts the worst-case scenario out of contention. Global Fixed-Priority (GFP) preemptive multiprocessor scheduling is one of the mainstream strategies to resolve contention on computational resources. It allows jobs of the same task to be executed on different processors, hence potentially enabling better parallelism and more efficient resource utilization. Unfortunately, its worst-case response time (WCRT) analysis is challenging. Existing approaches divide a high-priority task into three workloads, namely, carry-in workload, body workload, and carry-out workload, trying to optimize them individually. In this work, we propose a holistic WCRT analysis for GFP preemptive multiprocessor scheduling, where a task is no longer divided. Specifically, (i) we establish the tight interference scenario for the task being analyzed to find the most interfering high-priority jobs in any time interval; (ii) we obtain the starting released instant of each high-priority task’s first job to determine the maximum interference from high-priority tasks’ first jobs to the task being analyzed; (iii) we build the worst-case tight interference scenario for the task being analyzed by combining the tight interference scenario and the starting released instants; (iv) we prove that the WCRT of the task being analyzed can be decided by the worst-case tight interference scenario. Evaluation on schedulability shows that our proposed analysis achieves 4.2%-8.6% higher acceptance ratio in randomly generated data sets than the state-of-the-art workload division approaches.

关键词：

来源：评论

学校读者我要写书评

暂无评论

BGRA-Net: Boundary-Guided and Region-Aware Convolutional Neural network for the Segmentation of Breast Ultrasound Images

BGRA-Net: Boundary-Guided and Region-Aware Convolutional Neu...

引用

2021 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2021

作者： Zhang, Xiang Li, Xuanya Hu, Kai Gao, Xieping Xiangtan University Key Laboratory of Intelligent Computing and Information Processing of Ministry of Education Xiangtan411105 China Baidu Inc. Beijing100085 China Key Laboratory of Medical Imaging and Artifical Intelligence of Hunan Province Xiangnan University Chenzhou423000 China Hunan Normal University Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing Changsha410081 China

ISBN: (纸本)9781665401265

In this paper, we propose a novel convolutional neural network based on boundary-guided and region-aware (BGRA-Net) for breast tumor segmentation in ultrasound images. In particular, in the encoding stage, we propose a boundary-guided module (BGM) to guide the learning of boundary features in the decoding stage by explicitly strengthening the extracted boundary information. Meanwhile, in the decoding stage, we propose a region-aware module (RAM) to integrate different levels of detailed and semantic features to improve the comprehensive representation of tumor regional features. Besides, a scale-adaptive module (SAM) is further proposed to capture the characteristics of tumors with different sizes between the encoding and decoding stages. To evaluate the effectiveness of our BGRA-Net, we conduct extensive experiments on the UDIAT dataset and compare it with eight state-of-the-art methods. The experimental results show that our BGRA-Net outperforms the state-of-the-art methods and can achieve accurate segmentation of breast tumors with ambiguous boundaries. © 2021 IEEE.

关键词： Tumors

来源：评论

学校读者我要写书评

暂无评论

Brief Industry Paper: Response Time Evaluation of Cross-Domain Communication in CAN-FD and TSN

Brief Industry Paper: Response Time Evaluation of Cross-Doma...

引用

Real-Time Systems Symposium (RTSS)

作者： Wenhong Ma Xiaoyi Huang Dongsheng Wei Renfa Li Yunfei Zhang Guoqi Xie Wanli Chang Key Laboratory for Embedded and Network Computing of Hunan Province Hunan University China Tencent Future Network Lab Tencent Corporation China

With the advancement of intelligence and networked automotive, the domain-centralized architecture, which employs time sensitive networking (TSN) as the inter-domain backbone network and control area network with flexible data rate (CAN-FD) as the intra-domain network, has garnered significant attention. However, cross-domain end-to-end communication involves multiple components, and significant disparities between TSN and CAN-FD render response time analysis within domain-centralized architecture for mixed-critical traffic exceptionally complex. In this paper, we develop a cross-domain with TSN and CAN-FD end-to-end response time evaluation tool, which analyzes the response time of mixed-critical traffic under different design options segment by segment. We specifically analyze the waiting times of different messages in the domain control unit when faced with the design options of one-to-one and multi-to-one conversion of CAN-FD and TSN frames. The proposed evaluation tool can be easily extended to different design options to support more application scenarios. Theoretical computational analysis and real hardware measurements show the effectiveness of our tool.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A K-NN and Sparse Representation Based Method for Gesture Recognition

A K-NN and Sparse Representation Based Method for Gesture Re...

引用

IEEE International Conference on High Performance computing and Communications (HPCC)

作者： Meifang Zeng Ling Xiao Renfa Li Key Laboratory of Embedded and Network Hunan University Changsha China

ISBN: (纸本)9781479909735

Sparse representation classification is widely used in pattern recognition. In this paper, a k-nn and sparse represent based method for gesture recognition (KSRC) is proposed. In order to reduce the computational complexity problem in sparse representation, KSRC exploits K nearest neighbors of the testing sample from all training samples and represents the test sample as a linear combination of the K nearest neighbors, then solving the l1-norm constrained least square problem. Taking linear interpolation method to force all data to be in the same space. Experiments show almost perfect user-independent recognition, and user-mixed recognition and user-dependent recognition, and speeds up 4 times than SRC.

关键词： Training Gesture recognition Computational complexity Dictionaries Acceleration Vectors Accelerometers

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：