检索结果-内蒙古大学图书馆

Chinese Conference on pattern recognition (CCPR)

作者： Huilin Xiong Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai China

In this paper, we propose to kernelize linear learning machines, e.g., PCA and LDA, in the empirical kernel feature space, a finite-dimensional embedding space, in which the distances of the data in the kernel feature space are preserved. The empirical kernel feature space provides a unified framework for the kernelization of all kinds of linear machines: performing a linear machine in the finite-dimensional empirical feature space, its nonlinear kernel machine is then established in the original input data space. This method is different from the conventional kernel-trick based kernelization, and more importantly, the final nonlinear kernel machines, called empirical kernel machines, are shown to be more efficient in many real-world applications, such as face recognition and facial expression recognition, than the kernel-trick based kernel machines.

关键词： Kernel Face recognition pattern analysis pattern recognition Machine learning Linear discriminant analysis Training data Principal component analysis image processing Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Model-based 3D human motion tracking and voxel reconstruction from sparse views

Model-based 3D human motion tracking and voxel reconstructio...

引用

IEEE International Conference on image processing

作者： Junchi Yan Jian Song Liwei Wang Yuncai Liu Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai China

This paper addresses a method for 3D human motion tracking and voxel-based reconstruction from sparse views. We adopt the annealed Gaussian based particle swarm optimization (AGPSO) for 3D human motion tracking. The AGPSO algorithm incorporates the temporal continuity information into the traditional particle swarm optimization (PSO) algorithm under a Bayesian framework. In the online tracking process, the state variables are estimated via the particle filtering, where the observation is designed as a minimized Markov Random Field (MRF) energy. Finally, voxel reconstruction is conducted using the skeleton shape prior via dynamic graph cut. The experimental results show that our method performs promisingly against the cluttered background and generates plausible voxel reconstructions from sparse views.

关键词： Three dimensional displays Tracking Humans image reconstruction Shape Skeleton Annealing

来源：评论

学校读者我要写书评

暂无评论

Deformable Surface 3D Tracking Using Linear Programming

Deformable Surface 3D Tracking Using Linear Programming

引用

International Workshop on Computer Science and Engineering (WCSE)

作者： Chenhao Wang Shuhan Shen Yuncai Liu Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai China

A method for 3D shape reconstruction of deformable surfaces from consecutive frames was presented. In our method, the model of the surface is represented by a triangulated mesh. The constraints for the model, including keypoint correspondences and disallowing large changes of edge orientation between consecutive frames, are formulated as Linear Programming (LP) constraints. Therefore the deformable surface 3D tracking method turns into an LP problem that can be effectively solved. The robustness and efficiency of our approach are validated on synthetic and real data.

关键词： Linear programming Surface reconstruction Shape image reconstruction Cameras Deformable models Computer science image processing pattern recognition Robustness

来源：评论

学校读者我要写书评

暂无评论

A hole filling algorithm for medical imaging

A hole filling algorithm for medical imaging

引用

International Conference on Information Science and Engineering (ICISE)

作者： Lv Guo Yonghui Gao Yin Li Jie Yang Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai China

Surface integrity of 3D medical imaging is crucial for surgery simulation or virtual diagnoses. However, undesirable holes often exist due to external damage on bodies or accessibility limitation on scanners. To bridge the gap, the paper proposes an algorithm of filling holes for 3D medical image data based on moving least squares (MLS) method. In the algorithm, medical data is classified into two types and procedure is designed according to the classification. The procedure consists of two steps, namely hole detection and hole completion. In completion part, the detected hole can be automatically filled by interpolation with known surrounding points. Moreover, a threshold is used to constrain the density of the completed surface, so the sampling rate on reconstructed area complies with the original data. We test our algorithm on two types of medical data and the results indicate a robust solution.

关键词： Filling Three dimensional displays Surface reconstruction Medical diagnostic imaging Arrays image reconstruction

来源：评论

学校读者我要写书评

暂无评论

3D articulated motion estimation from images

3D articulated motion estimation from images

引用

Conference on Computer Vision and pattern recognition (CVPR)

作者： Xiaoyun Zhang Yuncai Liu Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai China

ISBN: (纸本)0769523722

This paper presents a new method of motion analysis of articulated objects from feature point correspondences over monocular perspective images without imposing any constraints on motion. The 3D joint positions of an articulated object are estimated within a scale factor using the connection relationship of two links over two or three images. Then, twists and exponential maps are employed to represent the motion of each link. Finally, constraints from image point correspondences are developed to estimate the motion. In the algorithm, the characteristic of articulated motion, i.e., motion correlation among links, is applied to decrease the complexity of the problem and improve the robustness. A point pattern matching algorithm for articulated objects is also discussed in this paper. Simulations and experiments on real images show the correctness and efficiency of the algorithms.

关键词： Motion estimation Motion analysis Joints Humans Tracking Biological system modeling Computer vision pattern recognition pattern matching Man machine systems

来源：评论

学校读者我要写书评

暂无评论

Online pcb defect detector on a new pcb defect dataset

arXiv

引用

arXiv 2019年

作者： Tang, Sanli He, Fan Huang, Xiaolin Yang, Jie Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China

Previous works for PCB defect detection based on image difference and image processing techniques have already achieved promising performance. However, they sometimes fall short because of the unaccounted defect patterns or over-sensitivity about some hyper-parameters. In this work, we design a deep model that accurately detects PCB defects from an input pair of a detect-free template and a defective tested image. A novel group pyramid pooling module is proposed to efficiently extract features of a large range of resolutions, which are merged by group to predict PCB defect of corresponding scales. To train the deep model, a dataset is established, namely DeepPCB, which contains 1,500 image pairs with annotations including positions of 6 common types of PCB defects. Experiment results validate the effectiveness and efficiency of the proposed model by achieving 98:6% mAP @ 62 FPS on DeepPCB dataset. This dataset is now available at: https://***/tangsanli5201/DeepPCB. Copyright © 2019, The Authors. All rights reserved.

关键词： Defects

来源：评论

学校读者我要写书评

暂无评论

Correlation filter tracking via bootstrap learning

Correlation filter tracking via bootstrap learning

引用

IEEE International Conference on image processing

作者： Kunqi Gu Tao Zhou Fanghui Liu Jie Yang Yu Qiao Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China

In recent years, correlation filter based trackers outperform better than other trackers. Nevertheless, they only employ one feature and a single kernel, so they are usually not robust in complex scenes. In this paper, we derive a multi-feature and multi-kernel correlation filter based tracker which fully takes advantage of the invariance-discriminative power spectrums of various features and kernels to further improve the performance. A novel bootstrap learning method is utilized to obtain a strong classifier by fusing these weak kernel correlation filters (KCFs). Moreover, a new target scale estimation strategy is incorporated into our framework. The efficient and effective scale estimation method is based on target dictionary representation. The proposed method is tested on several videos and compared with seven state-of-the-art methods. Experimental results have provided further support to the effectiveness and robustness of the proposed method.

关键词： Kernel Target tracking Correlation Computational modeling Boosting Robustness Object tracking

来源：评论

学校读者我要写书评

暂无评论

recognition and extraction high precision digital road map

Recognition and extraction high precision digital road map

引用

International Conference on Information Technology: Coding and Computing (ITCC)

作者： Mang Chen Yuncai Liu Institute of Image Processing and Pattern Recognition Shanghai Jiaotong University Shanghai China

ISBN: (纸本)0769523153

This paper presents a system of high precision road navigation map generation from CAD geographical map. First, get road curb line; then, pairing two road curb line built road, and computing road central line built road navigation vector; last, detecting intersection create intersection vector. By extensive experiments, the system has shown good efficiency and robustness.

关键词： Navigation Intelligent transportation systems Land vehicles Road vehicles Automation Computer networks Databases image processing pattern recognition Robustness

来源：评论

学校读者我要写书评

暂无评论

Fdflownet: Fast Optical Flow Estimation Using A Deep Lightweight Network

Fdflownet: Fast Optical Flow Estimation Using A Deep Lightwe...

引用

IEEE International Conference on image processing

作者： Lingtong Kong Jie Yang Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University China

ISBN: (数字)9781728163956

ISBN: (纸本)9781728163963

Significant progress has been made for estimating optical flow using deep neural networks. Advanced deep models achieve accurate flow estimation often with a considerable computation complexity and time-consuming training processes. In this work, we present a lightweight yet effective model for real-time optical flow estimation, termed FDFlowNet (fast deep flownet). We achieve better or similar accuracy on the challenging KITTI and Sintel benchmarks while being about 2 times faster than PWC-Net. This is achieved by a carefully-designed structure and newly proposed components. We first introduce an U-shape network for constructing multi-scale feature which benefits upper levels with global receptive field compared with pyramid network. In each scale, a partial fully connected structure with dilated convolution is proposed for flow estimation that obtains a good balance among speed, accuracy and number of parameters compared with sequential connected and dense connected structures. Experiments demonstrate that our model achieves state-of-the-art performance while being fast and lightweight.

关键词： Estimation Optical imaging Convolution Adaptive optics Optical fiber networks Computational modeling Training

来源：评论

学校读者我要写书评

暂无评论

Hierarchical primitives based contour matching

Hierarchical primitives based contour matching

引用

24th Symposium of the German pattern recognition Association, DAGM 2002

作者： Zhang, Xiaofeng Burkhardt, Hans Orbotech-Schuh GmbH Co. and KG Bad-Pyrmont D-31812 Germany Department of Pattern Recognition and Image Processing Institute of Computer Science Faculty of Applied Science Freiburg Germany

ISBN: (纸本)354044209X

An algorithm for contour matching is presented in this paper. It is implemented in two steps: firstly, bottom-up, corners are matched, the matched corner points guide line segment matching, and then the matched line segments guide contour matching. Line segments are grouped with the signature function defined in [10] from the extracted contours ofimage pairs. Secondly, top-down, with the computed signature functions of the matched contours, the contour points are corresponded by modifying the initial matching from line segments. The novelty ofour approach is that (1) features are incorporated, and the matching is implemented in two steps (bottom-up and top-down), thus dense correspondences along contours are acquired. Disadvantages such as the sparseness from only point correspondences, the inaccuracies from only line correspondences, and feature loss of correspondences along epipolar lines are avoided;(2) unlike the conventional way ofmatching along epipolar lines after rectification, signature functions developed in [10] are used to characterize the contours so that the matching is implemented along contours. © Springer-Verlag Berlin Heidelberg 2002.

关键词： Contour measurement

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：