Interactive network traffic replay is the newest method for testing and evaluation of network devices such as Firewalls, IPSes, routers, switches, etc. Currently state-checking method is used for interactive TCP traff...
详细信息
Interactive network traffic replay is the newest method for testing and evaluation of network devices such as Firewalls, IPSes, routers, switches, etc. Currently state-checking method is used for interactive TCP traffic replay. This paper proposes a new method for interactive TCP traffic replay which is based on the balance status between transmitted and received packets. By checking the balance conditions before sending out TCP packets, the method can significantly reduce the cost of state-checking and enhance the replay performance. The authors made a comparison on the differences of replay methods when introducing the balance mechanism. The efficiency of the method is also investigated and evaluated from aspects of a single TCP session, multi-session traffic, packet losses and latency. Experimental results show that the method outperforms the original state-checking method when replaying actual TCP traffics.
Maximum margin clustering (MMC) is a recently proposed clustering method, which extends the theory of support vector machine to the unsupervised scenario and aims at finding the maximum margin hyperplane which separat...
详细信息
In this paper, a novel segmentation and recognition system of courtesy and legal amounts in handwritten Chinese bank checks is presented. Firstly, the system pre-recognizes the courtesy amounts by the multimodel segme...
详细信息
In this paper, a novel segmentation and recognition system of courtesy and legal amounts in handwritten Chinese bank checks is presented. Firstly, the system pre-recognizes the courtesy amounts by the multimodel segmentation and recognition algorithm. Secondly, with the guide of courtesy amount recognition candidates, the legal amount is segmented and recognized. Lastly, we fuse the recognition candidates of courtesy and legal amounts to obtain the final recognition result. By this fusion strategy, the algorithm can correct the recognition error of courtesy amounts, and the inaccurate segmentation of legal amounts. The system is validated with 1053 real bank checks. When the substitution is 0.4%, the recognition rate at the amount level can reach 66.1%.
In image retrieval and annotation systems, Multi-Instance Learning (MIL) has been studied actively. Most of the state-of-the-art methods solve the MIL problem in a supervised way. In this paper, we proposed an unsuper...
详细信息
In image retrieval and annotation systems, Multi-Instance Learning (MIL) has been studied actively. Most of the state-of-the-art methods solve the MIL problem in a supervised way. In this paper, we proposed an unsupervised algorithm named MIOC-SVM (Multi-Instance clustering via One-Class Support Vector Machine) and presented its application of finding users' interests on specific web image regions without any manual labeled data. Our algorithm extends the conventional One-Class SVM to the MIL problem through the definition of bag margin. By maximizing the bag margins through iterative heuristic optimization, the MIL problem is converted to the traditional unsupervised learning problem. In this framework, an image is viewed as a bag and each bag contains a number of instances corresponding to regions obtained from image segmentation. MIL assumes that a bag is labeled positive if at least one of its instances is positive;otherwise, the bag is negative. While in MIOC-SVM, no label of bags is required for clustering. Furthermore, the MIOC-SVM algorithm was evaluated on both the MUSK benchmark data sets and a real-world image dataset downloaded from Yahoo. And comparative studies have shown the effectiveness of the proposed approach.
This paper proposes a semi-supervised inductive algorithm adopting a Gaussian random field(GRF)and Gaussian *** introduce the prior based on graph *** regularization term measures the p-smoothness over the graph.A new...
详细信息
This paper proposes a semi-supervised inductive algorithm adopting a Gaussian random field(GRF)and Gaussian *** introduce the prior based on graph *** regularization term measures the p-smoothness over the graph.A new conditional probability called the extended Bernoulli model(EBM)is also *** generalizes the logistic regression to the semi-supervised case,and especially,it can naturally represent the *** the training phase,a novel solution is given to the discrete regularization framework defined on the *** the new test data,we present the prediction formulation,and explain how the margin model affects the classification boundary.A hyper-parameter estimation method is also *** results show that our method is competitive with the existing semi-supervised inductive and transductive methods.
Multiple instance learning (MIL) is a branch of machine learning that attempts to learn information from bags of instances. Many real-world applications such as localized content-based image retrieval and text categor...
详细信息
Ranking aggregation is important in data mining and information retrieval. In this paper, we proposed a semisupervised ranking aggregation method, in which the order of several item pairs are labeled as side informati...
详细信息
ISBN:
(纸本)9781595939913
Ranking aggregation is important in data mining and information retrieval. In this paper, we proposed a semisupervised ranking aggregation method, in which the order of several item pairs are labeled as side information. The core idea is to learn a ranking function based on the ordering agreement of different rankers. The ranking scores assigned by this ranking function on the labeled data are consistent with the given pairwise order constraints while the ranking scores on the unlabeled data obey the intrinsic manifold structure of the rank items. The experiment results show our method work well. Categories and Subject DescriptorsH.3.3[Information Storage and Retrieval]:Information Search and Retrieval{ Retrieval models H.3.4[Information systems Application]: systems and Software{Performance evaluation(eficiency and effectiveness).
In this paper, we present a new method of predictive deconvolution based on high order statistics, and apply it on echo cancellation. The traditional predictive deconvolution is based on the assumption that the observ...
详细信息
Object tracking is an essential problem in the field of video and image processing. Although tracking algorithms working on gray video are convenient in actual applications, they are more difficult to be developed tha...
详细信息
This paper aims to present a navigation system design for image guided minimal abdominal surgery robot that could compensate for the patient respiratory movement. Currently computer-aided surgery navigation technology...
详细信息
暂无评论