This paper studies speech-driven Web retrieval models which accepts spoken search topics (queries) in the NTCIR-3 Web retrieval task. The major focus of this paper is on improving speech recognition accuracy of spoken...
详细信息
This paper studies speech-driven Web retrieval models which accepts spoken search topics (queries) in the NTCIR-3 Web retrieval task. The major focus of this paper is on improving speech recognition accuracy of spoken queries then improving retrieval accuracy in speech-driven Web retrieval. We experimentally evaluate the techniques of combining outputs of multiple LVCSR models in recognition of spoken queries. As model combination techniques, we compare the SVM learning technique conventional voting schemes such as ROVER. We show that the techniques of multiple LVCSR model combination can achieve improvement both in speech recognition and retrieval accuracies in speech-driven text retrieval. We also show that model combination by SVM learning outperforms conventional voting schemes both in speech recognition retrieval accuracies.
Heuristic methods are specially well suited to solve combinatorial problems. One of these problems is the distribution of petroleum products through oil pipelines networks. In this paper the problem is stated and solv...
详细信息
For many practical applications of speech recognition systems, it is quite desirable to have an estimate of confidence for each hypothesized word. Unlike previous works on confidence measures, we have proposed feature...
详细信息
ISBN:
(纸本)0780376633
For many practical applications of speech recognition systems, it is quite desirable to have an estimate of confidence for each hypothesized word. Unlike previous works on confidence measures, we have proposed features for confidence measures that are extracted from outputs of more than one LVCSR models. For further analysis of the proposed confidence measure, this paper examines the correlation between each word's confidence and the word's features such as its part-of-speech and syllable length. We then apply SVM learning technique to the task of combining outputs of multiple LVCSR models, where, as features of SVM learning, information such as the pairs of the models which output the hypothesized word are useful for improving the word recognition rate. Experimental results show that the combination results achieve a relative word error reduction of up to 72 % against the best performing single model and that of up to 36 % against ROVER.
This work shows a new method for blind separation of sources, based on geometrical considerations concerning the observation space. This new method is applied to a mixture of several sources and it obtains the estimat...
详细信息
ISBN:
(纸本)0972184457
This work shows a new method for blind separation of sources, based on geometrical considerations concerning the observation space. This new method is applied to a mixture of several sources and it obtains the estimated coefficients of the unknown mixture matrix A and separates the unknown sources. The principles of the new method and a description of the algorithm followed by some speed enhancements are shown. Finally, we illustrate with simulations of several source distributions how the algorithm performs.
Heuristic methods are specially well suited to solve combinatorial problems. One of these problems is the distribution of petroleum products through oil pipelines networks. In this paper the problem is stated and solv...
详细信息
Heuristic methods are specially well suited to solve combinatorial problems. One of these problems is the distribution of petroleum products through oil pipelines networks. In this paper the problem is stated and solved using a multiobjective and constraint evolutionary optimization algorithm. Several objective functions has been defined to express the goals of the solutions as well as the preferences among them. Some constraints are included as hard objective functions and some has been evaluated through a repairing function to avoid infeasible solutions. An example of working is given.
Deformable models have been intensively studied in image analysis through the last decade, and are used for detection and recognition of flexible or rigid templates under diverse viewing conditions. Genetic algorithm ...
详细信息
Deformable models have been intensively studied in image analysis through the last decade, and are used for detection and recognition of flexible or rigid templates under diverse viewing conditions. Genetic algorithm (GA) based deformable models are used for generic visual landmark detection and interpretation. The developed system allows topologic localization and navigation using natural and artificial landmarks, exploiting deformable models' ability for handling landmark perspective variations. The resulting perception module has been integrated successfully in a complex navigation system. Various experimental results in real environments are presented on this paper, showing the effectiveness and capacity of the landmark detection and reading system.
We present CAVT, a simple yet powerful tool to visualize the behavior of congestion control mechanisms. Essentially being a simulator that builds upon the well-known Chiu/Jain vector diagrams, CAVT provides researcher...
详细信息
We present CAVT, a simple yet powerful tool to visualize the behavior of congestion control mechanisms. Essentially being a simulator that builds upon the well-known Chiu/Jain vector diagrams, CAVT provides researchers, teachers and students with a new abstraction level that has already shown its value in practice.
作者:
Kampel, M.Melero, F.J.Vienna University of Technology
Institute of Computer Aided Automation Pattern Recognition and Image Processing Group Favoritenstr. 9 183-2 ViennaA-1040 Austria Universidad de Granada
E.T.S. Ingeniería Informática Dpt. Lenguajes y Sistemas Informáticos C/. Daniel Saucedo Aranda s/n GranadaE-18071 Spain
Every archaeological excavation must deal with a vast number of ceramic fragments. The documentation, administration and scientific processing of these fragments represent a temporal, personnel, and financial problem....
详细信息
This paper presents a technique to design a digital optical-fiber photoplethysmograph. This noninvasive method is based on the measurement of the intensity of an infrared-light beam, which penetrates the index finger....
详细信息
ISBN:
(纸本)0780366468
This paper presents a technique to design a digital optical-fiber photoplethysmograph. This noninvasive method is based on the measurement of the intensity of an infrared-light beam, which penetrates the index finger. Digital techniques are used to produce the transmitted signal and to demodulate the received signal. Optical fibers are used to transfer the light to and from the sampling point. Frequency analysis of the peripheral volume pulse-wave signal is performed off-line. The power content in the harmonics of this pulse wave may be a useful measure of aging and vascular disease.
Parameter-dependent quadratic forms (PDQF) play an important role in analysis and synthesis of linear parameter varying (LPV) systems. This paper proposes a method to find a PDQF that meets some criterion written in t...
详细信息
Parameter-dependent quadratic forms (PDQF) play an important role in analysis and synthesis of linear parameter varying (LPV) systems. This paper proposes a method to find a PDQF that meets some criterion written in terms of LMI guaranteeing performances of LPV systems. Through approximation of PDQF with spline functions, we derive finite number of LMI from a given LMI-criterion on a PDQF, which is inherently a condition with infinitely many inequalities. This approximating solution to LMI on a PDQF is proved to have the following properties: (1) any solution to the derived LMI of finite number always produces a PDQF satisfying the original LMI-criterion, and (2) the finite LMI-condition always holds with sufficiently fine division of the parameter's region if the original one is solvable. Thus the results of this paper enable to solve parameter-dependent LMI associated with PDQF without conservatism.
暂无评论