J.L. Gauvain, L. Lamel, G. Adda, M. Jardino. Recent Advances in Transcribing Television and Radio Broadcasts. Eurospeech'99.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....in order to re score word lattices according to posterior probabilities, which can be directly interpreted as confidence measures (see [3] Such confidence scored lattices will then be used as input to the topic detection modules. The broadcast speech recognition systems developed at LIMSI (see [7 12]) as well as the recognizers for German (see [6] and Portuguese at Duisburg University and at INESC respectively, serve as a basis for the automatic transcription of broadcast news in the diverse languages. Approaches for topic indexing developed at LIMSI, the University of Duisburg ( 2] and at ....

J.L. Gauvain, L. Lamel, G. Adda, M. Jardino. Recent Advances in Transcribing Television and Radio Broadcasts. Eurospeech'99.

.... by introducing speaker dependency with only a few speaker speci c utterances [5] In broadcast news recognition systems unsupervised adaptation techniques have successfully been applied to adapt the acoustic models to the characteristics of speci c speakers, speaker clusters and the environment [2]. In recent years, the most popular approach for performing supervised as well as unsupervised adaptation has turned out to be Maximum Likelihood Linear Regression (MLLR) as introduced by Leggetter in [5] The popularity of MLLR is mainly based on two reasons. On the one hand, the Maximum ....

....to be generated with the imperfect recognizer itself. Because of recognition errors, the performance of unsupervised adaptation usually is signi cantly worse than supervised adaptation. However, unsupervised adaptation has lead to slight improvements in recognition accuracy in several publications [2, 4]. The most straightforward approach for unsupervised adaptation is to simply assume the output of the speech recognizer to be the true transcription and to use this just like the transcription in supervised adaptation. Some improvements have been reported for this type of unsupervised adaptation ....

J.-L. Gauvain, L. Lamel, G. Adda, M. Jardino: "Recent Advances in Transcribing Television and Radio Broadcasts ", Eurospeech'99, pp. 655-658.

....influence acoustic model size and of language model order on performance. We then discuss the impact of the word error rate on the information retrieval process. 2. SYSTEM OVERVIEW The LIMSI broadcast news automatic indexation system [3] consists of an audio partitioner [6] a speech recognizer [7, 8] and an indexation module [5] The goal of audio partitioning is to divide the acoustic signal into homogeneous segments, labeling and structuring the acoustic content of the data. Partitioning consists of identifying and removing non speech segments, and then clustering the speech segments and ....

J.L. Gauvain, L. Lamel, G. Adda, "Recent Advances in Transcribing Television and Radio Broadcasts," Proc. Eurospeech '99, 2, pp. 655-658, Budapest, Sept. 1999.

.... resources at the LDC [1] and manually transcribed evaluation test data from DARPA sponsored benchmark tests [11, 12] and other reported results on this data [3, 13] The LIMSI transcription system has two main phases: audio partitioning and speaker independent continuous speech recognition [8]. The first phase serves to partition the continuous data stream into homogeneous acoustic segments, assigning appropriate labels with each segment. The second phase carries out word recognition, where the system determines the sequence of words in the segment, associating start and end times and ....

J.L. Gauvain, L. Lamel, G. Adda, "Recent Advances in Transcribing Television and Radio Broadcasts," Proc. Eurospeech '99, pp. 1463--1466, Budapest, Sep. 1999.

....is increasing at a close rate. Therefore processing time is an important factor in making a speech transcription system viable for audio data mining and other related applications. The LIMSI broadcast news automatic indexation system [11] consists of an audio partitioner [9] a speech recognizer [10, 12] and an indexation module [6] The transcription components are shown in Figure 1. Partitioning the Audio stream The goal of audio partitioning is to divide the acoustic signal into homogeneous segments, removing non speech segments, and labeling and structuring the acoustic content of the data. ....

J.L. Gauvain, L. Lamel, G. Adda, "Recent Advances in Transcribing Television and Radio Broadcasts," Proc. Eurospeech '99, 2, pp. 655-658, Budapest, Sept. 1999.

....used for system development, as well as the 1999 evaluation test set. All the reported runs were done on a Compaq XP1000 500MHz machine with Digital Unix. 2. SYSTEM OVERVIEW The LIMSI broadcast news automatic transcription system [3] consists of an audio partitioner [9] and a speech recognizer [4, 11]. The goal of audio partitioning is to divide the acoustic signal into homogeneous segments, labeling and structuring the acoustic content of the data. Partitioning consists of identifying and removing non speech segments, and then clustering the speech segments and assigning bandwidth and ....

J.L. Gauvain, L. Lamel, G. Adda, "Recent Advances in Transcribing Television and Radio Broadcasts," Proc. Eurospeech '99, 2, pp. 655-658, Budapest, Sept. 1999.

....this year s evaluation. Comparative results are given on the development queries from SDR 99 and this year s query set, and some conclusions are made. 2. TRANSCRIPTION SYSTEM OVERVIEW The LIMSI broadcast news transcription system [5] consists of an audio partitioner [10] and a speech recognizer [11, 12]. The goal of audio partitioning is to divide the acoustic signal into homogeneous segments, labeling and structuring the acoustic content of the data. Partitioning consists of identifying and removing non speech segments, and then clustering the speech segments and assigning bandwidth and gender ....

J.L. Gauvain, L. Lamel, G. Adda, "Recent Advances in Transcribing Television and Radio Broadcasts," Proc. Eurospeech '99, 2, pp. 655-658, Budapest, Hungary, September 1999. TREC-9 Conference, Gauvain-Lamel-Barras-Adda

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC