%Aigaion2 BibTeX export from Idiap Publications %Sunday 22 December 2024 04:06:33 AM @INPROCEEDINGS{lathoud04a, author = {Lathoud, Guillaume and McCowan, Iain A. and Odobez, Jean-Marc}, projects = {Idiap}, month = {5}, title = {Unsupervised {L}ocation-{B}ased {S}egmentation of {M}ulti-{P}arty {S}peech}, booktitle = {{P}roceedings of the 2004 {ICASSP-NIST} {M}eeting {R}ecognition {W}orkshop}, year = {2004}, address = {Montreal, Canada}, note = {IDIAP-RR 04-14}, crossref = {lathoud-rr-04-14}, abstract = {Accurate detection and segmentation of spontaneous multi-party speech is crucial for a variety of applications, including speech acquisition and recognition, as well as higher-level event recognition. However, the highly sporadic nature of spontaneous speech makes this task difficult. Moreover, multi-party speech contains many overlaps. We propose to attack this problem as a tracking task, using location cues only. In order to best deal with high sporadicity, we propose a novel, generic, short-term clustering algorithm that can track multiple objects for a low computational cost. The proposed approach is online, fully deterministic and can run in real-time. In an application to real meeting data, the algorithm produces high precision speech segmentation.}, pdf = {https://publications.idiap.ch/attachments/papers/2004/lathoud04a.pdf}, postscript = {ftp://ftp.idiap.ch/pub/papers/2004/lathoud04a.ps.gz}, ipdinar={2004}, ipdmembership={speech, vision}, } crossreferenced publications: @TECHREPORT{lathoud-rr-04-14, author = {Lathoud, Guillaume and McCowan, Iain A. and Odobez, Jean-Marc}, projects = {Idiap}, title = {{S}hort-{T}erm {S}patio-{T}emporal {C}lustering of {S}poradic and {C}oncurrent Events}, type = {Idiap-RR}, number = {Idiap-RR-14-2004}, year = {2004}, institution = {IDIAP}, address = {Martigny, Switzerland}, note = {Published in ``Proceedings of the 2004 ICASSP-NIST Meeting Recognition Workshop''}, abstract = {Accurate detection and segmentation of spontaneous multi-party speech is crucial for a variety of applications, including speech acquisition and recognition, as well as higher-level event recognition. However, the highly sporadic nature of spontaneous speech makes this task difficult. Moreover, multi-party speech contains many overlaps. We propose to attack this problem as a multitarget tracking task, using location cues only. In order to best deal with high sporadicity, we propose a novel, generic, short-term clustering algorithm that can track multiple objects for a low computational cost. The proposed approach is online, fully deterministic and can run in real-time. In an application to real meeting data, the algorithm produces high precision speech segmentation.}, pdf = {https://publications.idiap.ch/attachments/reports/2004/rr-04-14.pdf}, postscript = {ftp://ftp.idiap.ch/pub/reports/2004/rr-04-14.ps.gz}, ipdinar={2004}, ipdmembership={speech, vision}, language={English}, }