Authentication email has already been sent, please check your email box: and activate it as soon as possible.
You can login to My Profile and manage your email alerts.
If you haven’t received the email, please:
|
|
There are 98 papers published in subject: > since this site started. |
Select Subject |
Select/Unselect all | For Selected Papers |
Saved Papers
Please enter a name for this paper to be shown in your personalized Saved Papers list
|
1. A Shape Semantic Graph Representation for Object Understanding and Recognition in Point Clouds | |||
NING Xiaojuan, WANG Yinghui | |||
Computer Science and Technology 16 May 2016 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:To understand and to recognize 3D objects represented as point cloud data, we provide a shape semantic graph (SSG) representation method to describe 3D objects. Based on the decomposed components of the object, the boundary surface of different components and the topology of the skeleton, the SSG gives a semantic description that is consistent with human vision perception. The similarity measurement of SSG for different objects is also effective to distinguish the type of objects and find the most similar one. Experiments on the shape database show that SSG is valuable for capturing the components of the objects and the corresponding relations between them. Not only can it be suitable for the object without loop but also appropriate for the object with loop to represent the shape and the topology. Besides, the combination of two-step progressive similarity measurement strategy can effectively improve the recognition rate in the shape database containing point-sample data. | |||
TO cite this article:NING Xiaojuan, WANG Yinghui. A Shape Semantic Graph Representation for Object Understanding and Recognition in Point Clouds[OL].[16 May 2016] http://en.paper.edu.cn/en_releasepaper/content/4688548 |
2. Accelerating Large-scale Convolutional Neural Networks Based on Convolution in Blocks | |||
ZHANG Qiang-Qiang, WANG Chun-Lu, LIU Zhen-Yu | |||
Computer Science and Technology 05 April 2016 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:There has been an upsurge in machine learning in recent years. And we all know that deep learning is part of the machine learning. Convolutional neural network is an important method of deep learning, but the number of the networks' training parameters is very large and it has multiple layers, resulting in the training process very slowly. By unrolling the input of convolution layer into matrices and using the BLAS libraries, we can shorten the training time through matrix multiplication. However, another problem occurs when the input image is too large. In such case, during the rearrangement of the input data, data rearranged before may be crowed out of cache because of the later data. When it comes to convolution, the convolution will be greatly affected in that the cache hit rate is reduced. A method based on the process of convolution to accelerate the training is presented, which is dividing input data into blocks when convolution and the effect of the relationship between the block size and the capacity of the cache is studied as well. Experiments show that the method of convolution in blocks is simple and feasible, and the efficiency of convolution is promoted about 50% at best. | |||
TO cite this article:ZHANG Qiang-Qiang, WANG Chun-Lu, LIU Zhen-Yu. Accelerating Large-scale Convolutional Neural Networks Based on Convolution in Blocks[OL].[ 5 April 2016] http://en.paper.edu.cn/en_releasepaper/content/4682735 |
3. A multiple-steps linear representation based classification for face recognition | |||
Tao Liu,Jian-Xun Mi | |||
Computer Science and Technology 19 February 2016 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Error detection is an important approach to improve the robustness of face recognition (FR) method. However, it is hard to directly detect the outliers in a facial image. We decompose the hard problem into many simpler sub-problems in this paper. That is, the process of detecting distorted pixels is divided into multiple easier steps and a part of invalid pixels are detected in every step. The goal is to decrease the ratio of outlier in the testing image, which reduces the influence of outliers in a recognition process. The performance that our method deals with occlusion and corruption problems is evaluated on different databases. In addition, we compare our method with state-the-of-art face recognition based methods, and the proposed method achieves the best results in face occlusion and disguise issues. | |||
TO cite this article:Tao Liu,Jian-Xun Mi. A multiple-steps linear representation based classification for face recognition[OL].[19 February 2016] http://en.paper.edu.cn/en_releasepaper/content/4676937 |
4. Structure Extraction of the Website Information Based on the Degree of Link Association | |||
Wang Yang,Zhang Bin | |||
Computer Science and Technology 09 November 2015 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Structure extraction of websites information is the basis of many other technologies about classifying the website. In this paper, some different algorithms that are used to extract the structure of the website information are listed, and this paper also analyzes the advantages and disadvantages of those different algorithms. Above all, a method about structure extraction of the website information based on the degree of link association is put forward in the paper. First of all, it's needed to extract the content of every page of the target website, secondly, we can use the page after the extraction of content to calculate the dissimilarity of pages and calculate the dissimilarity of the links of two pages, then we can also get the route which is from the home page to the target page by the dijkstra algorithm, finally, the structure of the whole website can be produced through the route. | |||
TO cite this article:Wang Yang,Zhang Bin. Structure Extraction of the Website Information Based on the Degree of Link Association[OL].[ 9 November 2015] http://en.paper.edu.cn/en_releasepaper/content/4660762 |
5. Mining Seeds for Community Detection | |||
Fengjiao Chen, Kan Li | |||
Computer Science and Technology 01 August 2015 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Community structure is an important part in complex networks, such as social network and protein network. Communities are mixed and overlapped where members can take part into multiple communities at the same time. Most of the detection methods are based on seed communitiy from which the community is contructed. The quality of mined seeds influence the quality of detected communities. In this paper, a seed mining method is proposed that can detect both clique structure and star structure. Experiment in real complex networks show the competitive performance of the method. | |||
TO cite this article:Fengjiao Chen, Kan Li. Mining Seeds for Community Detection[OL].[ 1 August 2015] http://en.paper.edu.cn/en_releasepaper/content/4651225 |
6. A Review of Detecting Hierarchical Community Structure | |||
Fengjiao Chen,Kan Li | |||
Computer Science and Technology 09 April 2015 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Hierarchical community structure is an important structure in complex networks, where nodes divide into groups that further subdivide into groups of groups, such as the leader-follower in social network. In this paper, we review three kinds of hierarchical community detection methods. First, hierarchical clustering methods are intuitive to detect the hierarchy structure. Second, local expansion methods obtain hierarhcy by tune a resolution parameter. Third, a most general method is to contruct the super network for normal community detection methods. Finally, we show that hierarchical community structures can give us deep understanding of networks. | |||
TO cite this article:Fengjiao Chen,Kan Li. A Review of Detecting Hierarchical Community Structure[OL].[ 9 April 2015] http://en.paper.edu.cn/en_releasepaper/content/4638442 |
7. Hierarchical Saliency-based Representation for Human Interaction Recognition | |||
GUO Wei,HU Tao,LIU Ruqian | |||
Computer Science and Technology 17 December 2014 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Recognizing human interactions is a one of the most important problems in computer vision and impacts a wide range of applications. This paper presents a new method for the recognition of two-person interactions using hierarchical saliency-based representation. Hierarchical saliency is defined as Salient Action at the highest level, Salient Point at middle level, Salient Joint at the lowest level of interaction, determined by the greatest spatial-temporal positional changes at each level. Given the saliency of interactions at different levels, several types of features were extracted according to the discriminative characteristics of behaviors, such as spatial displacement, direction relations and etc. Since there are few publicly accessible test datasets, we created a new dataset with eight types of interactions named K3HI, using a new depth sensor, the Microsoft Kinect. The method was tested using the SVM multi-class classifier; our experimental results demonstrate that the average recognition accuracy of hierarchical saliency-based representation is 90.29%, outperforming methods using other features. | |||
TO cite this article:GUO Wei,HU Tao,LIU Ruqian. Hierarchical Saliency-based Representation for Human Interaction Recognition[OL].[17 December 2014] http://en.paper.edu.cn/en_releasepaper/content/4622030 |
8. Transfer Learning of Structured Representation for Face Recognition | |||
Chuan-Xian Ren, Dao-Qing Dai | |||
Computer Science and Technology 13 November 2014 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Face recognition under uncontrolled conditions, e.g., complex backgrounds and variable resolutions, is still challenging in image processing and computer vision. Although many methods have been proved well-performed in the controlled settings, they are usually of weak generality across different data sets. Meanwhile, several properties of the source domain, such as background and the size of subjects, play an important role in determining the final classification results. A transferrable representation learning model is proposed in this paper to enhance the recognition performance. To deeply exploit the discriminant information from the source domain and the target domain, the bio-inspired face representation is modeled as structured characterization for the commonality between different domains. The method outputs a grouped boost of the features, and presents a reasonable manner for highlighting and sharing discriminant orientations and scales. Notice that the method can be viewed as a framework, since other feature generation operators and classification metrics can be embedded therein. Experiments on the benchmark databases, including uncontrolled FRGC and LFW show the efficacy of the proposed transfer learning algorithm. | |||
TO cite this article:Chuan-Xian Ren, Dao-Qing Dai. Transfer Learning of Structured Representation for Face Recognition[OL].[13 November 2014] http://en.paper.edu.cn/en_releasepaper/content/4618425 |
9. Recommender System Based on Email Platform | |||
Mingming Wang,Honggang Zhang,Yang Yang,Siyuan Li | |||
Computer Science and Technology 28 October 2014 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Recommender systems are gaining a great importance with the emergence of E-commerce and business on the Internet. These recommender systems help users make decision by suggesting products and services that satisfy the users' tastes and preferences. As an important tool for Internet information exchange, e-mail plays a pivotal role in people's lives. Personalized recommendation in email platform becomes a new hot service. This paper presents a recommender system based on email platform. Our method makes a recommendation through extracting and analyzing user email information and user behavior logs. For every user, we will build a user model according to the user's email information, generating user interest matrix. The user interest matrix records the user's interests to things, including original interests and potential interests. After getting user interest matrix, we can recommend things to the user. Considering the drifting of user interest and the real-time of recommendation, we propose feedback recommendation. When a user clicks one thing, the system will get similar things and then recommend to the user. The similarity between things is stored in thing similarity matrix, which was calculated by content-based and collaborative filtering techniques. Besides, according to user email information and user behavior logs, we change the user interest matrix every once in a while tracking the change of user interests. What's more, as user interest matrix records user's potential interests, the recommendation results are diversity and novelty and the method gives a good recommendation for email platform. | |||
TO cite this article:Mingming Wang,Honggang Zhang,Yang Yang, et al. Recommender System Based on Email Platform[OL].[28 October 2014] http://en.paper.edu.cn/en_releasepaper/content/4612454 |
10. Vehicle-Logo Recognition Method based on Local Binary Pattern | |||
Liang Dong,Zhang Honggang,Shi Kuan,Hou Chengwen | |||
Computer Science and Technology 08 October 2014 | |||
Show/Hide Abstract | Cite this paper︱Full-text: PDF (0 B) | |||
Abstract:Vehicle-logo location and recognition has a high research value for intelligent transportation system. With the emergence of car deck, only license plate recognition has become unreliable for automated management traffic. The reliability of the vehicle identification will greatly improve when combining license plate information with vehicle information. In order to solve vehicle-logo location and recognition, the paper presents a new method of vehicle-logo location based on license plate location. Then the Local Binary Pattern(LBP) descriptor computed over an vehicle-logo image as feature vector, then similarity between test vehicle-logo image and template images calculated by Euclidean distance.And the experiments have a good recognition rate. | |||
TO cite this article:Liang Dong,Zhang Honggang,Shi Kuan, et al. Vehicle-Logo Recognition Method based on Local Binary Pattern[OL].[ 8 October 2014] http://en.paper.edu.cn/en_releasepaper/content/4612650 |
Select/Unselect all | For Selected Papers |
Saved Papers
Please enter a name for this paper to be shown in your personalized Saved Papers list
|
About Sciencepaper Online | Privacy Policy | Terms & Conditions | Contact Us
© 2003-2012 Sciencepaper Online. unless otherwise stated