检索结果-内蒙古大学图书馆

A multiple feature fusion model for image segmentation based on hierarchical visual pathway

Advances in Information Sciences and Service Sciences 2012年第8期4卷 274-282页

作者： Hu, Dekun Li, Jianping Yang, Simon X. Gregori, Stefano School of Computer Science and Engineering University of Electronic Science and Technology of China China Advanced Robotics and Intelligent Systems Lab School of Engineering University of Guelph Canada

To segment an object from its background in an image for advanced vision processing, this paper presents a novel bio-inspired general framework for image segmentation in complex nature scenes, which is a hierarchical system that mimics the organization of layered early visual area in primate visual cortex. The proposed methodology consists of two typical stages: the first stage is a parallel modular structure including three segmenting operators based on color feature, form feature and texture feature, each of which solves the segmentation problem independently for the same input. Then, a fusion operation, multiple feature fusion segmentation (MFFS), integrates these three feature segmentations together through the backpropagation neuron network (BPNN) in the last stage, which simulates the operation of area following the lateral geniculate nucleus in primary visual cortex. The proposed approach is applied to several segmentation experiments of many single objects in clustering conditions, the result shows that the approach is capable of competing with state-of-the-art systems.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

An object recognition model with top-down prediction and context

引用

International Journal of Advancements in Computing Technology 2012年第12期4卷 283-290页

作者： Hu, Dekun Li, Jianping Yang, Simon X. University of Electronic Science and Technology of China School of Computer Science and Engineering Chengdu 611731 China University of Guelph School of Engineering Advanced Robotics and Intelligent Systems Lab Guelph ON Canada

Based on the trickle-up and trickle-down process in primate visual pathway, an ideal model for object recognition is researched. Firstly, the information of the low spatial frequency are extracted and passed to high visual cortex, the bank of the objects in memory is activated by this input, so the prediction of the possible candidate objects is pulled out. Then, the information of high spatial frequency, which keeps the invariability and selectivity of an object, is gained through the trickle-up process. With the top-down prediction, the context and the detailed information from trickle-up pathway, the object is recognized by Bayes inference. The performance of the proposed model is proved by several experiments.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Online non-feedback image re-ranking via dominant data selection

Online non-feedback image re-ranking via dominant data selec...

引用

20th ACM International Conference on Multimedia, MM 2012

作者： Cao, Chen Chen, Shifeng Li, Yuhong Liu, Jianzhuang Shenzhen Key Laboratory for Computer Vision and Pattern Recognition Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen China Department of Information Engineering Chinese University of Hong Kong Hong Kong Media Lab. Huawei Technologies Co. Ltd. China

ISBN: (纸本)9781450310895

Image re-ranking aims at improving the precision of keyword-based image retrieval, mainly by introducing visual features to re-rank. Many existing approaches require offline training for every keyword, which are unsuitable for online image search. Other real-time approaches demand user interaction, which are inappropriate for large-scale image collection. To improve the accuracy of web image retrieval in a practicable manner, we propose a novel re-ranking algorithm to explore the cluster information of the image set. First, we build spectral graph on images that retrieved bysearch engine, and remove isolated nodes as noisy images. Then, we select positive samples from the most dominant cluster in initial top-ranked images, and the samples are used for semi-supervised learning and ranking. Our algorithm is online and non-feedback. Experiments on two public databases demonstrate that our algorithm outperforms the state-of-the-art approaches. © 2012 ACM.

关键词： Supervised learning

来源：评论

学校读者我要写书评

暂无评论

Efficient Feedback Design for Interference Alignment in MIMO Interference Channel

Efficient Feedback Design for Interference Alignment in MIMO...

引用

IEEE Conference on Vehicular Technology (VTC)

作者： Sungyoon Cho Hyukjin Chae Kaibin Huang Dongku Kim Vincent K. N. Lau Hanbyul Seo School of Electrical & Electronic Engineering Yonsei University Korea Department of Electrical & Computer Engineering Hong Kong University of Science & Technology Advanced Communication Technology Lab LG Electronics Korea

Interference alignment (IA) is a joint-transmission technique that achieves the capacity of the interference channel for high signal-to-noise ratios (SNRs). However, most prior works on IA are based on the impractical assumption that perfect and global channel-state information(CSI) is available at all transmitters, resulting in overwhelming feedback overhead. To substantially suppress the feedback overhead, this paper proposes an efficient design of the feedback framework for IA in the K-user multiple-input multiple-output (MIMO) interference channel. The proposed feedback topology supports sequential CSI exchange (feedback and feedforward) between transmitters and receivers and reduces the feedback overhead from a cubic function of K to a linear one, compared to conventional feedback approaches. Given the proposed feedback topology, we consider the limited feedback channel from the receivers to corresponding interferers and analyze the effect of quantization error which generates the residual interference. Also, an efficient feedback-bit allocation algorithm that minimizes the upper-bound of sum residual interference is proposed.

关键词： Receivers Topology Transmitters Interference channels MIMO Throughput

来源：评论

学校读者我要写书评

暂无评论

Free-floating Reaction Centers (RCs) versus attached monolayer of RCs in bio-photoelectrochemical cells

Free-floating Reaction Centers (RCs) versus attached monolay...

引用

2011 MRS Fall Meeting

作者： Yaghoubi, Houman Takshi, Arash Jun, Daniel Saer, Rafael Madden, John D. Thomas Beatty, J. Department of Electrical Engineering University of South Florida Tampa FL 33620 United States Department of Microbiology and Immunology University of British Columbia Vancouver BC V6T 1Z3 Canada Department of Electrical and Computer Engineering Advanced Materials and Process Engineering Lab. University of British Columbia Vancouver BC V6T 1Z1 Canada

ISBN: (纸本)9781605113920

The high quantum efficiency (∼100%) in the bacterial photosynthetic reaction center (RC) has inspired research on the application of RCs to build protein based solar cells. Conventionally, applying RCs as the photosensitive layer on the surface of a carbon electrode has shown poor photocurrents in the cells. The low photocurrent is partly due to the weak absorption of light in the monolayer of RCs. Also, an Atomic Force Microscopy image of the electrode shows lots of defects on the immobilized RCs at the electrode surface. In this work, we have built a bio-photoelectrochemical cell in which the RCs are floating in the electrolyte instead of being attached to the surface of an electrode. Despite the simple structure of the cell, the photocurrent is significantly higher in the new cell compared to when RCs are attached to an electrode. The amplitude of current reached to ∼40 nA for free floating RCs, about five times larger than that in the cell with attached RCs. The aging effect was studied in both cells in a course of a week. The lifetime of attached RCs on electrode surface was slightly better than solubilized RCs in the electrolyte. Also, it is found that the mechanism which governs the charge transfer from RCs to the electrodes is the same in both bio-photoelectrochemical cells. © 2012 Materials Research Society.

关键词： Photoelectrochemical cells

来源：评论

学校读者我要写书评

暂无评论

An Electric Machine Integrated with Trans-Rotary Magnetic Gear

An Electric Machine Integrated with Trans-Rotary Magnetic Ge...

引用

IEEE Energy Conversion Congress and Exposition

作者： Siavash Pakdelian Yateendra Deshpande Hamid A. Toliyat Advanced Electric Machines and Power Electronics (EMPE) Lab. Department of Electrical and Computer Engineering Texas A&M University College Station TX 77843-3128

ISBN: (纸本)9781467308021

This paper introduces a novel electromechanical energy conversion system, called Machine Integrated with Trans-Rotary Magnetic Gear (MITROMAG) as a force-dense linear motor/generator. The device is obtained by integrating a Trans-Rotary Magnetic Gear (TROMAG) into a 3-phase winding. The TROMAG converts a low speed high force linear motion of a translator to a low torque high speed rotation of a rotor. Upon adding Permanent Magnet (PM) poles to the outer surface of the rotor of the TROMAG, it also serves as the rotor of an inner rotor PM machine. The resultant electromechanical device is compared to a PM linear tubular machine, as a benchmark, for a given force and speed. Simplified analytical models along with accurate field solutions and Finite Element Analysis (FEA) are employed to design the systems. It is shown that, for the given design specifications, the MITROMAG would surpass the linear PM machine in regard to the volume and weight with a great margin.

关键词： Electric Machine Integrated

来源：评论

学校读者我要写书评

暂无评论

Image segmentation based on trickle-up and trickle-down visual pathway in primate

引用

Journal of Computational Information Systems 2012年第9期8卷 3757-3765页

作者： Hu, Dekun Yang, Simon X. Gregori, Stefano Li, Jianping Fu, Bo University of Electronic Science and Technology of China School of Computer Science and Engineering Jianshe Road Chengdu 610054 China University of Guelph School of Engineering Advanced Robotics and Intelligent Systems Lab. Guelph Ont. N1G 2W1 Canada

To locate the object accurately in a scene for further vision processing, a novel bio-inspired model for image segmentation is proposed, which closely follows the computation of trickle-up and trickle-down processing in primate visual pathways. The trickle-up stage is a parallel modular structure consisting of three segmenting operators implementing the function as the parvocellular, magnocellular and koniocelluar pathways. The computation model of trickle-down path, predicts incoming stimuli based on the prior knowledge of the scenes. The proposed approach is applied to several segmentation experiments of clustering conditions, the results demonstrate that the performance of the proposed approach surpasses those achieved by previous top-down or bottom-up schemes on figure-ground segmentation. © 2011 by Binary Information Press.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Combining Multiple Feature Representations and AdaBoost Ensemble Learning for Reducing False-Positive Detections in computer-aided Detection of Masses on Mammograms

Combining Multiple Feature Representations and AdaBoost Ense...

引用

Annual International Conference of the IEEE engineering in Medicine and Biology Society

作者： Jae Young Choi Dae Hoe Kim Konstantinos N. Plataniotis Yong Man Ro with Multimedia Lab The Edward S Rogers Sr Department of Electrical and Computer Engineering University of Toronto Toronto Ontario M5S 3GA Canada with the Image and Video Systems Lab Department of Electrical Engineering Korea Advanced Institute of Science and Technology (KAIST) Republic of Korea

ISBN: (纸本)9781424441198

One of the drawbacks of current computer-aided Detection (CADe) systems is a high number of false-positive (FP) detections, especially for detecting mass abnormalities. In a typical CADe system, classifier design is one of the key steps for determining FP detection rates. This paper presents the effective classifier ensemble system for tackling FP reduction problem in CADe. To construct ensemble consisting of correct classifiers while disagreeing with each other as much as possible, we develop a new ensemble construction solution that combines data resampling underpinning AdaBoost learning with the use of different feature representations. In addition, to cope with the limitation of weak classifiers in conventional AdaBoost, our method has an effective mechanism for tuning the level of weakness of base classifiers. Further, for combining multiple decision outputs of ensemble members, a weighted sum fusion strategy is used to maximize a complementary effect for correct classification. Comparative experiments have been conducted on benchmark mammogram dataset. Results show that the proposed classifier ensemble outperforms the best single classifier in terms of reducing the FP detections of masses.

关键词： Representations and AdaBoost CADe the FP detections of masses

来源：评论

学校读者我要写书评

暂无评论

Optimized four-dimensional mapping for high-speed optical communication systems

Optimized four-dimensional mapping for high-speed optical co...

引用

Optical Fiber Communication Conference, OFC 2012

作者： Zhang, Jianyong Djordjevic, Ivan B. Key Lab of Alloptical Network and Advanced Telecommunication Network of EMC Beijing Jiaotong University Beijing 100044 China University of Arizona Department of Electrical and Computer Engineering United States Institute of Lightwave Technology Beijing Jiaotong University Beijing100044 China

We propose an optimized mapping rule for the 32-point four-dimensional crossconstellation for high-speed optical communications based on bit-interleaved coded-modulation with iterative decoding. The optimized mapping ... 详细信息

ISBN: (纸本)9781557529381

关键词： Mapping

来源：评论

学校读者我要写书评

暂无评论

Top-emitting 980-nm vertical-cavity surface-emitting laser diodes with improved optical power

Top-emitting 980-nm vertical-cavity surface-emitting laser d...

引用

International Conference on Photonics (ICP)

作者： A. N. Al-Omari M. S. Alias K. L. Lear Electronic Engineering Department Hijjawi Faculty for Engineering Technology Yarmouk University Irbid Jordan Advanced Physical Technologies Lab Telekom Malaysia Research & Development (TMR&D) TM Innovation Centre Lingkaran Teknokrat Timur Cyberjaya Selangor Malaysia Electrical and Computer Engineering Department Colorado State University Fort Collins CO USA

Top-emitting, oxide-confined, polyimide-planarized 980-nm VCSELs with copper-plated heatsinks were fabricated and characterized. Increasing the plated heatsink radius from 0-μm to 4-μm larger than the mesa diameter for lasers with 8-μm oxide aperture diameter reduced the measured thermal impedance, increased the maximum bias current density, and increased the maximum output optical power achieved by a 29%, 37%, and 73%, respectively. VCSELs with oxide aperture diameter and heatsink overlap of 8-μm and 4-μm, respectively, demonstrated 17°C decrease in the internal device temperature (i.e. active region temperature) at the maximum output optical power. Devices with similar mesa diameters of 26-μm and different heatsink overlaps exhibited a threshold bias current and a total series resistance of (630±4%)μA and ~95Ω, respectively.

关键词： Vertical cavity surface emitting lasers Heat sinks Optical device fabrication Apertures Temperature measurement Resistance

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：