This paper proposed an adaptive shadows detection algorithm based on Gaussian Mixture Model to improve the performance of video object segmentation. This method takes advantage of luminance weight to model the backgro...
详细信息
This paper proposed an adaptive shadows detection algorithm based on Gaussian Mixture Model to improve the performance of video object segmentation. This method takes advantage of luminance weight to model the background of the image and obtains a primary segmentation in CIE Luv color space. In this way, it improves the real-time ability of detection. It also becomes more efficient, comparing with the existing shadow detection algorithms which often need to set the threshold manually or get them through a training process. By using the Gaussian distribution, it is able to realize an adaptive shadow detection. At same time, the authors deal with the noise or the aim points uneven distribution by using horizontal filling and vertical filling. It improves the accuracy of segmentation. The experimental results have shown that this method achieves adaptive shadows detection and has strong robustness, high segmentation accuracy.
In this paper, a motion-based approach for detecting highlevel semantic events in video sequences is presented. Its main characteristic is its generic nature, i.e. it can be directly applied to any possible domain of ...
详细信息
ISBN:
(纸本)9781605580708
In this paper, a motion-based approach for detecting highlevel semantic events in video sequences is presented. Its main characteristic is its generic nature, i.e. it can be directly applied to any possible domain of concern without the need for domain-specific algorithmic modifications or adaptations. For realizing event detection, the video is initially segmented into shots and for every resulting shot appropriate motion features are extracted at fixed time intervals, thus forming a motion observation sequence. Then, Hidden Markov Models (HMMs) are employed for associating each shot with a semantic event based on its formed observation sequence. Regarding the motion feature extraction procedure, a new representation for providing local-level motion information to HMMs is presented, while motion characteristics from previous frames are also exploited. The latter is based on the observation that motion information from previous frames can provide valuable cues for interpreting the semantics present in a particular frame. Experimental results as well as comparative evaluation from the application of the proposed approach in the domains of tennis and news broadcast video are presented. Copyright 2008 ACM.
This paper presents an improved incremental learning technique for SVM, namely redundant incremental SVM (RISVM), for pattern classification problems. Through adding some non-support vectors (say, redundant vectors in...
详细信息
This paper presents an improved incremental learning technique for SVM, namely redundant incremental SVM (RISVM), for pattern classification problems. Through adding some non-support vectors (say, redundant vectors in the sense of contribution to the final solution) at each incremental step, the RISVM algorithm can achieve similar performance to the SVM in batch (or non-incremental SVM) but result in less support vectors for the same quality of pattern classification, and also it can provide better generalization performance in comparison with other incremental techniques for SVM. The bispiral problem and five widely used benchmark data sets are employed to verify the method, and the simulations support the feasibility and effectiveness of the proposed approach.
This paper firstly analyzed the front pattern matching algorithm of intrusion detection system, and improved on the pattern matching algorithm. Having immediately analyzed Rete algorithm and aimed at its shortage, the...
详细信息
This paper firstly analyzed the front pattern matching algorithm of intrusion detection system, and improved on the pattern matching algorithm. Having immediately analyzed Rete algorithm and aimed at its shortage, the paper introduced the FRete net algorithm on the basis of Rete net algorithm constituting an inference machine based on FRete matching algorithm, which could infer logical suspicious facts. Then the authors discussed and designed an intrusion detection system based on high-speed networks, and finally established a simulation and experiment platform to test the performance of IDS.
For the Multiword Expression (MWE) recognition, the Multiple Sequence Alignment (MSA) is proposed on the motivation of gene recognition. Because textual sequence is similar to gene sequence in pattern analysis. This M...
详细信息
For the Multiword Expression (MWE) recognition, the Multiple Sequence Alignment (MSA) is proposed on the motivation of gene recognition. Because textual sequence is similar to gene sequence in pattern analysis. This MSA technique is combined with error-driven rules, with the improved efficiency beyond the traditional *** provides a guarantee for the MWE recall. It uses the dynamic programming method to prevent candidates from combinational explosion, and provides a global solution for pattern extraction instead of sub-pattern redundancy. Consequently, it has accurate measures for flexible patterns. In experiment, some advanced statistical measures are performed for ranking candidates. In the comparison experiment, the MSA approach achieved better results.
Blog is becoming more and more popular with the rapid development of Internet. It needs to find an automatic way to distinguish the blog pages from ordinary Web pages for the content extraction of blog pages and the b...
详细信息
Blog is becoming more and more popular with the rapid development of Internet. It needs to find an automatic way to distinguish the blog pages from ordinary Web pages for the content extraction of blog pages and the blog community discovered. Some basic concepts and ideas in the area of blog was described in this paper, and a method on the blog pages identification is proposed, which is based on the blog pages structure and blog content. The experimentation shows that a high result can be achieved in precision.
In this paper, a generic motion-based approach to semantic video analysis is presented. The examined video is initially segmented into shots and for every resulting shot appropriate motion features are extracted at fi...
详细信息
In this paper, a generic motion-based approach to semantic video analysis is presented. The examined video is initially segmented into shots and for every resulting shot appropriate motion features are extracted at fixed time intervals. Then, hidden Markov models (HMMs) are employed for performing the association of each shot with one of the semantic classes that are of interest in any given domain. Regarding the motion feature extraction procedure, higher order statistics of the motion estimates are calculated and a new representation for providing local-level motion information to HMMs is presented. The latter is based on the combination of energy distribution-related information and spatial attributes of the motion signal. Experimental results as well as comparative evaluation from the application of the proposed approach in the domain of news broadcast video are presented.
Knowledge in knowledge bases have two categories: complete and incomplete. In this paper, through uniformly expressing these two kinds of knowledge, we first address four operators on a knowledge base, which are adequ...
详细信息
ISBN:
(纸本)9781424425129
Knowledge in knowledge bases have two categories: complete and incomplete. In this paper, through uniformly expressing these two kinds of knowledge, we first address four operators on a knowledge base, which are adequate for generating new knowledge through using known knowledge. Then, we establish the relationship between knowledge and knowledge granulation. These results will be very helpful for knowledge discovery from knowledge bases and play a significant role for establishing a framework of granular computing in knowledge bases.
暂无评论