检索结果-内蒙古大学图书馆

International Joint Conference on Neural Networks (IJCNN)

作者： Chi Kin Chow Tong Lee Computer Vision and Image Processing Laboratory Department of Electronic Engineering Chinese University of Hong Kong Hong Kong China

An approach is introduced to determine the topology of a feedforward binary neural network automatically. The approach is based on a construction algorithm that constructs one layer of hidden nodes at a time until the problem is solved. In each layer, the algorithm determines the necessary number of nodes through a growth process by finding the best hidden node that would help to partition the input training data set. This is done using a genetic algorithm. The proposed algorithm can determine the necessary number of hidden layers and number of hidden nodes at each layer automatically. Tests on a number of benchmark problems illustrated the effectiveness of the proposed technique, both in terms of network complexity and recognition accuracy, compared with a geometrical learning approach.

关键词： Neural networks Multi-layer neural network Feedforward neural networks Genetic algorithms Network topology Partitioning algorithms Training data Error correction computer vision image processing

来源：评论

学校读者我要写书评

暂无评论

Invariant feature matching by Hopfield-type neural network

Invariant feature matching by Hopfield-type neural network

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Wen-Jing Li Tong Lee Computer Vision and Image Processing Laboratory Department of Electronic Engineering Chinese University of Hong Kong Hong Kong China

In this paper, a novel Hopfield model for silhouette matching invariant to projective transformations is proposed. Although the new network has higher-order energy function, we show that it can be solved using a standard second-order Hopfield network, by taking advantage of the neighborhood information in the data. The experimental results with real data show that the proposed method can provide accurate matching results in image registration and object recognition.

关键词： Neural networks Hopfield neural networks Object recognition Power engineering and energy Stereo vision Motion estimation Feature extraction Cost function Layout computer vision

来源：评论

学校读者我要写书评

暂无评论

Resolving ambiguity in depth extraction for motion capture using genetic algorithm

Resolving ambiguity in depth extraction for motion capture u...

引用

Congress on Evolutionary Computation

作者： Yin Yee Wai Chi Kin Chow Tong Lee Computer Vision and Image Processing Laboratory Department of Electronic Engineering Chinese University of Hong Kong Sha Tin Hong Kong China

There is considerable interest in motion capture from an image sequence taken from a video camera. However, since the images only consist of 2D information, the distance of an object from the image plane cannot be determined uniquely unless some constraints are imposed. Coplanar motion is a popular constraint adopted for this purpose. If the captured motion is not coplanar, only two possible solutions can be obtained at the best even the actual size of the object is given. Therefore, the motion cannot be captured without resolving the distances of each control point from the image plane. By assuming smooth transition, the problem is formulated as one minimizing the transition between image frames and the application of GA to solve this optimization problem is proposed.

关键词： Genetic algorithms Humans Cameras image resolution image sequences Animation Motion analysis Optical filters Optical recording Optical sensors

来源：评论

学校读者我要写书评

暂无评论

Multiple objects segmentation based on Maximum-likelihood Estimation and Optimum Entropy-distribution(MLE-OED)

Proceedings - International Conference on Pattern Recognitio...

引用

Proceedings - International Conference on Pattern Recognition 2002年第1期16卷 707-710页

作者： Jun, Xie Tsui, H.T. Deshen, Xia Image Processing and Computer Vision Lab Department of Electronic Engineering Chinese University of Hong Kong Hong Kong Pattern Recognition Lab Department of Computer Science Nanjing University of Sci. and Tech. China

A new method based on MLE-OED is proposed for unsupervised image segmentation of multiple objects which have fuzzy edges. It adjusts the parameters of a mixture of Gaussian distributions via minimizing a new loss function proposed to implement image segmentation based on the image's local spatial information and global intensity distribution properties . The loss function consists of two terms: a local content fitting term, which optimizes the entropy distribution, and a global statistical fitting term, which maximizes the likelihood of the parameters for the given data. The proposed segmentation method was validated by simulated and real examples. Its performance in the experiments is better than those of two popular methods. © 2002 IEEE.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Mobile robot localization under varying illumination

Mobile robot localization under varying illumination

引用

作者： Jogan, Matjaž Leonardis, Aleš Wildenauer, Horst Bischof, Horst Faculty of Computer and Information Science University of Ljubljana Tržaška 25 1001 Ljubljana Slovenia Pattern Recognition and Image Processing Group Institute for Computer Aided Automation Vienna University of Technology Favoritenstrasse 9/1832 A-1040 Vienna Austria Institute for Computer Graphics and Vision Graz University of Technology Inffeldgasse 16 2. OG A-8010 Graz Austria

Methods for mobile robot localization that use eigenspaces of panoramic snapshots of the environment are in general sensitive to changes in the illumination of the environment. Therefore, we propose an approach which achieves a reliable localization under severe illumination conditions. The method uses gradient filtering of the eigenspace. After testing the approach on images obtained by a mobile robot, we show that it outperforms the standard eigenspace-based recognition method. © 2002 IEEE.

关键词： Mobile robots

来源：评论

学校读者我要写书评

暂无评论

Mobile robot localization under varying illumination

Mobile robot localization under varying illumination

引用

16th International Conference on Pattern Recognition, ICPR 2002

作者： Jogan, Matjaž Leonardis, Aleš Wildenauer, Horst Bischof, Horst Faculty of Computer and Information Science University of Ljubljana Tržaška 25 Ljubljana1001 Slovenia Pattern Recognition and Image Processing Group Institute for Computer Aided Automation Vienna University of Technology Favoritenstrasse 9/1832 ViennaA-1040 Austria Institute for Computer Graphics and Vision Graz University of Technology Inffeldgasse 16 2. OG GrazA-8010 Austria

ISBN: (纸本)0769516963

Methods for mobile robot localization that use eigen spaces of panoramic snapshots of the environment are in general sensitive to changes in the illumination of the environment. Therefore, we propose an approach which achieves a reliable localization under severe illumination conditions. The method uses gradient filtering of the eigen space. After testing the approach on images obtained by a mobile robot, we show that it outperforms the standard eigen space-based recognition method. © 2002 IEEE.

关键词： Mobile robots

来源：评论

学校读者我要写书评

暂无评论

Multiple objects segmentation based on maximum-likelihood estimation and optimum entropy-distribution (MLE-OED)

Multiple objects segmentation based on maximum-likelihood es...

引用

International Conference on Pattern Recognition

作者： Xie Jun H.T. Tsui Xia Deshen Image Processing and Computer Vision Laboratory Department of Electronic Engineering Chinese University of Hong Kong Hong Kong China Pattern Recognition Laboratory Department of Computer Science Nanjing University of Science and Technology China

ISBN: (纸本)076951695X

A new method based on MLE-OED is proposed for unsupervised image segmentation of multiple objects which have fuzzy edges. It adjusts the parameters of a mixture of Gaussian distributions via minimizing a new loss function proposed to implement image segmentation based on the image's local spatial information and global intensity distribution properties. The loss function consists of two terms: a local content fitting term, which optimizes the entropy distribution, and a global statistical fitting term, which maximizes the likelihood of the parameters for the given data. The proposed segmentation method was validated by simulated and real examples. The performance in the experiments is better than those of two popular methods.

关键词： Object segmentation Maximum likelihood estimation image segmentation Layout Uncertainty image edge detection Shape measurement image processing Pattern recognition Gaussian distribution

来源：评论

学校读者我要写书评

暂无评论

Mobile robot localization under varying illumination

Mobile robot localization under varying illumination

引用

International Conference on Pattern Recognition

作者： M. Jogan A. Leonardis H. Wildenauer H. Bischof Faculty of Computer and Information Science University of Ljubljana Ljubljana Slovenia Pattern Recognition and Image Processing Group Institute for Computer Aided Automation University of Technology Vienna Vienna Austria Institute for Computer Graphics and Vision Graz University of Technology Graz Austria

关键词： Mobile robots Lighting computer vision Filter bank image recognition Indoor environments Pattern recognition image processing computer graphics Information science

来源：评论

学校读者我要写书评

暂无评论

The cardeye: A trinocular active vision system 2

引用

2nd International Workshop on computer vision Systems, ICVS 2001

作者： Hemayed, Elsayed E. Ahmed, Moumen T. Farag, Aly A. Computer Vision and Image Processing Laboratory University of Louisville KY40292 United States

ISBN: (纸本)3540422854

The CardEye is an experimental, trinocular, 3D active vision system. Our goal is to create a flexible, precise tool for active vision research. The system uses an agile trinocular vision head that will be mounted on a robotic arm. The system has several degrees of freedom: pan, tilt, roll, vergence and variable baseline in addition to the automated zoom and focus of the cameras lenses. It utilizes an active lighting device to assist in the surface reconstruction process. After solving for the best system variables to image a surface, a novel multi-stage reconstruction algorithm is used to build a rich, complete 3D model of the surface. In this paper we describe the novel architecture of the system and its successful application to reconstruction of objects of varying surface characteristics. © Springer-Verlag Berlin Heidelberg 2001.

关键词： Degrees of freedom (mechanics)

来源：评论

学校读者我要写书评

暂无评论

3D reconstruction of the human jaw: A new approach and improvements 4th

引用

4th International Conference on Medical image Computing and computer-Assisted Intervention, MICCAI 2001

作者： Ahmed, Moumen T. Eid, Ahmed H. Farag, Aly A. Computer Vision and Image Processing Laboratory University of Louisville KY40292 United States

ISBN: (纸本)3540426973

This paper presents a new, practical approach for 3D reconstruction of the human jaw from a sequence of intra-oral images. This research has an immense value in various dental practices including implants, tooth alignment, and craniofacial surgery. Our approach is based on the recently-proposed space carving algorithm for shape recovery. This algorithm provides more flexibility to the reconstruction process and eliminates several constrains imposed by other traditional approaches such as stereo and shape from shading. Our experimental results have shown that the approach is able to reconstruct 3D models of the human jaw with sub-millimeter accuracy. © Springer-Verlag Berlin Heidelberg 2001.

关键词： image reconstruction

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：