Structure and content-based segmentation of speech transcripts Export

@inproceedings{citeulike:4401521, abstract = {algorithm for the segmentation of an audio/video source into topically cohesive segments based on automatic speech recognition (ASR) transcriptions is presented. A novel two-pass algorithm is described that combines a boundary-based method with a content-based method. In the first pass, the temporal proximity and the rate of arrival of ngram features is analyzed in order to compute an initial segmentation. In the content- based second pass, changes in content-bearing words are detected by using the ngram features as queries in an information-retrieval system. The second pass validates the initial segments and merges them as needed. Feasibility of the segmentation task can vary enormously depending on the structure of the audio content, and the accuracy of ASR. For real-world corporate training data our method identifies, at worst, a single salient segment of the audio and, at best, a high-level table-of-contents. We illustrate the algorithm in detail with some examples and validate the results with segmentation boundaries generated manually.}, address = {New York, NY, USA}, author = {Ponceleon, Dulce and Srinivasan, Savitha}, booktitle = {SIGIR '01: Proceedings of the 24th annual international ACM SIGIR conference on Research and development in information retrieval}, citeulike-article-id = {4401521}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=383952.384041}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/383952.384041}, doi = {10.1145/383952.384041}, isbn = {1-58113-331-6}, keywords = {searchingspeech}, location = {New Orleans, Louisiana, United States}, pages = {404--405}, posted-at = {2009-04-24 20:11:51}, priority = {2}, publisher = {ACM}, title = {Structure and content-based segmentation of speech transcripts}, url = {http://dx.doi.org/10.1145/383952.384041}, year = {2001} }

TY - CONF ID - citeulike:4401521 L3 - citeulike-article-id:4401521 N2 - algorithm for the segmentation of an audio/video source into topically cohesive segments based on automatic speech recognition (ASR) transcriptions is presented. A novel two-pass algorithm is described that combines a boundary-based method with a content-based method. In the first pass, the temporal proximity and the rate of arrival of ngram features is analyzed in order to compute an initial segmentation. In the content- based second pass, changes in content-bearing words are detected by using the ngram features as queries in an information-retrieval system. The second pass validates the initial segments and merges them as needed. Feasibility of the segmentation task can vary enormously depending on the structure of the audio content, and the accuracy of ASR. For real-world corporate training data our method identifies, at worst, a single salient segment of the audio and, at best, a high-level table-of-contents. We illustrate the algorithm in detail with some examples and validate the results with segmentation boundaries generated manually. CY - New Orleans, Louisiana, United States SN - 1-58113-331-6 AD - New York, NY, USA EP - 405 TI - Structure and content-based segmentation of speech transcripts PB - ACM T2 - SIGIR '01: Proceedings of the 24th annual international ACM SIGIR conference on Research and development in information retrieval SP - 404 KW - searchingspeech AU - Ponceleon, Dulce AU - Srinivasan, Savitha PY - 2001/// UR - http://dx.doi.org/10.1145/383952.384041 ER -