Fuzzy time series forecasting model is an effective method to solve the nonlinear problems forecasting. However, most published fuzzy time series based models did not count the change trend implicit in historical datu...
详细信息
Fuzzy time series forecasting model is an effective method to solve the nonlinear problems forecasting. However, most published fuzzy time series based models did not count the change trend implicit in historical datum. In this paper, authors proposed a novel method which applied heuristic information to the fuzzy time series model based on Fibonacci sequence. As an example, the USD/JPY exchange rate is tested in this model. The results show that this method not only improves the forecasting accuracy, but decreases the computational complexity.
In this paper, a compact conductor backed H shaped antenna fed by CPW is proposed. It has compact size of 24 times 24 mm. The proposed antenna radiates as omni directional in azimuthal plane and useful for wide band a...
详细信息
In this paper, a compact conductor backed H shaped antenna fed by CPW is proposed. It has compact size of 24 times 24 mm. The proposed antenna radiates as omni directional in azimuthal plane and useful for wide band applications. The paper emphasizes miniaturized H shaped antenna with dimension 24 times 24 mm. The fundamental parameters return loss, VSWR, radiation pattern are obtained, which meet standard specifications. Method of moments based IE3D simulator is used to analyze antenna.
Video artificial text detection is a challenging problem of patternrecognition. Current methods which are usually based on edge, texture, connected domain, feature or learning are always limited by size, location, la...
详细信息
Video artificial text detection is a challenging problem of patternrecognition. Current methods which are usually based on edge, texture, connected domain, feature or learning are always limited by size, location, language of artificial text in video. To solve the problems mentioned above, this paper applied SOM (Self-Organizing Map) based on supervised learning to video artificial text detection. First, text features were extracted. And considering the video artificial text's limitations mentioned, artificial text's location and gradient of each pixel were used as the features which were used to classify. Then three layers supervised SOM was proposed to classify the text and non-text areas in video image. At last, the morphologic operating was used to get a much more accurate result of text area. Experiments showed that this method could locate and detect artificial text area in video efficiently.
Given a time-varying face image object, or only a sub-part of the image, the question of whether the template object image exists in a given image database is an important problem our days, which still remains in its ...
详细信息
Given a time-varying face image object, or only a sub-part of the image, the question of whether the template object image exists in a given image database is an important problem our days, which still remains in its infancy, due to the many challenges it involves. With the advantage of distributed computing, such as computation sharing and data storage sharing, the limitations of image retrieval in a centralized image database can be eliminated. In this paper, we present an efficient method and architecture to determine whether a given face, of sub-part (s) of it, with time-varying features, disguises, and facial expressions is stored in a collection of known faces stored in nominal configurations. We demonstrate If such image exists, a match report is presented and the image position and rotation are derived as well. We show here how by combining a distributed computing and image sub-patch correlation technique in the image patternrecognition phase, the performance of the image searching is significantly improved.
Feature selection problem has become the focus of much pattern classification research and mutual information is more and more important in the feature selection algorithms. We proposed normalized mutual information b...
详细信息
Feature selection problem has become the focus of much pattern classification research and mutual information is more and more important in the feature selection algorithms. We proposed normalized mutual information based on Renyi's quadratic entropy feature selection, which reduces the computational complexity, relying on the efficient estimation of the mutual information. Then we combine NMIFS with wrappers into a two-stage feature selection algorithm. This helps us find more charactering feature subset. We perform some experiments to compare the efficiency and classification accuracy to other MI-based feature selection algorithm. Results show that our method leads to promising improvement on computation complexity.
Video scene segmentation plays an important role in video structure analysis. In this paper, we propose a time constraint dominant-set clustering algorithm for shot grouping and scene segmentation, in which the simila...
详细信息
In this paper we investigate a non-parametric classification of English phonemes in speaker-independent continuous speech. We employ the ldquovotingrdquo k-nearest neighbour (k-NN) classifier, a powerful technique in ...
详细信息
In this paper we investigate a non-parametric classification of English phonemes in speaker-independent continuous speech. We employ the ldquovotingrdquo k-nearest neighbour (k-NN) classifier, a powerful technique in patternrecognition problems, along with a new representation of phonemes for the speech recognition task. We also exploit the idea behind ldquoapproximaterdquo k-NN that results in a very fast way of computing the k approximate closest neighbours of each data point. Comparing the recognition performance of the proposed method with the HMM-based recognizer of HTK toolkit reveals that the k-NN-based recognizer outperforms its counterpart. In addition, incorporating the ldquoapproximaterdquo nearest neighbour search instead of the ldquoexactrdquo one results in completing the training step much faster than the HMM-based system, and the testing step with a comparable computational time. We also reduced the amount of the training data by applying a patternrecognition technique, called ldquothinningrdquo algorithm. The outcome was a considerable reduction in the k-NN search space and hence the execution time, and also a slight increase in the recognition performance.
In echo hiding, watermark data is embedded into the host signal by introducing echoes of different delays and is subsequently extracted using cepstral analysis. The choice of echo kernel determines the fidelity and ro...
详细信息
In echo hiding, watermark data is embedded into the host signal by introducing echoes of different delays and is subsequently extracted using cepstral analysis. The choice of echo kernel determines the fidelity and robustness performance of the watermarking scheme. In this paper we propose the extended bipolar echo kernel, which greatly improves the fidelity of the watermarked signal without compromising the detection performance.
Training SVM consumes large memory and computation time, but traditional methods can not overcome above shortcomings. This paper proposes an improved SVM training method based on chaotic particle swarm optimization (C...
详细信息
Training SVM consumes large memory and computation time, but traditional methods can not overcome above shortcomings. This paper proposes an improved SVM training method based on chaotic particle swarm optimization (CPSO) using circle map. Firstly, a new chaotic search model using improved circle map is introduced. Then this new model is introduced into particle swarm optimization (PSO). Finally, the detail training SVM algorithm using this improved CPSO algorithm is presented. The experimental results on face database show that the proposed SVM method optimized by the improved CPSO can achieve higher recognition performance.
With the popularity of MMS, the multimedia messages which include sensitive information are increasing rapidly. In the paper, a novel framework of a MMS filtering for Chinese sensitive text in image is presented. An e...
详细信息
With the popularity of MMS, the multimedia messages which include sensitive information are increasing rapidly. In the paper, a novel framework of a MMS filtering for Chinese sensitive text in image is presented. An effective method is applied to detect and filter sensitive texts in image of multimedia message which could easily be transmitted through the mobile communication network without being monitored at recent stage. The detection and recognition of sensitive text are achieved by using SIFT feature, which is proper to the characteristics of the text in image of multimedia message and get an accurate result. The method has a good practical application value.
暂无评论