Development of the 2012 SJTU HVR System

Development of the 2012 SJTU HVR System. Xu, H., Fan, Y., & Yu, K. In Proceedings of the 14th ACM International Conference on Multimodal Interaction, of ICMI '12, pages 539--544, New York, NY, USA, 2012. ACM.

Paper doi abstract bibtex

Haptic voice recognition (HVR) is a multi-modal text entry method for smart mobile devices. It employs haptic events generated by speakers during speaking to achieve better efficiency and robustness for automatic speech recognition. This paper describes the detailed design of the 2012 SJTU submission for the HVR Grand Challenge. During the design, a new perplexity metric using conditional entropy is proposed to evaluate the potential search space reduction of a haptic event without speech input. A number of new haptic events are evaluated both theoretically and experimentally in detail. The final submission system uses the haptic event of initial letter plus final letter and reduces word error rate by 76% compared to the baseline initial letter event.

@inproceedings{xu_development_2012,
	address = {New York, NY, USA},
	series = {{ICMI} '12},
	title = {Development of the 2012 {SJTU} {HVR} {System}},
	isbn = {978-1-4503-1467-1},
	url = {http://doi.acm.org/10.1145/2388676.2388790},
	doi = {10.1145/2388676.2388790},
	abstract = {Haptic voice recognition (HVR) is a multi-modal text entry method for smart mobile devices. It employs haptic events generated by speakers during speaking to achieve better efficiency and robustness for automatic speech recognition. This paper describes the detailed design of the 2012 SJTU submission for the HVR Grand Challenge. During the design, a new perplexity metric using conditional entropy is proposed to evaluate the potential search space reduction of a haptic event without speech input. A number of new haptic events are evaluated both theoretically and experimentally in detail. The final submission system uses the haptic event of initial letter plus final letter and reduces word error rate by 76\% compared to the baseline initial letter event.},
	urldate = {2014-06-05TZ},
	booktitle = {Proceedings of the 14th {ACM} {International} {Conference} on {Multimodal} {Interaction}},
	publisher = {ACM},
	author = {Xu, Hainan and Fan, Yuchen and Yu, Kai},
	year = {2012},
	pages = {539--544}
}

Downloads: 0

{"_id":"dunAo8YkSpSNQGfmA","bibbaseid":"xu-fan-yu-developmentofthe2012sjtuhvrsystem-2012","downloads":0,"creationDate":"2016-10-05T13:48:42.787Z","title":"Development of the 2012 SJTU HVR System","author_short":["Xu, H.","Fan, Y.","Yu, K."],"year":2012,"bibtype":"inproceedings","biburl":"http://bibbase.org/zotero/alanlivio","bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"New York, NY, USA","series":"ICMI '12","title":"Development of the 2012 SJTU HVR System","isbn":"978-1-4503-1467-1","url":"http://doi.acm.org/10.1145/2388676.2388790","doi":"10.1145/2388676.2388790","abstract":"Haptic voice recognition (HVR) is a multi-modal text entry method for smart mobile devices. It employs haptic events generated by speakers during speaking to achieve better efficiency and robustness for automatic speech recognition. This paper describes the detailed design of the 2012 SJTU submission for the HVR Grand Challenge. During the design, a new perplexity metric using conditional entropy is proposed to evaluate the potential search space reduction of a haptic event without speech input. A number of new haptic events are evaluated both theoretically and experimentally in detail. The final submission system uses the haptic event of initial letter plus final letter and reduces word error rate by 76% compared to the baseline initial letter event.","urldate":"2014-06-05TZ","booktitle":"Proceedings of the 14th ACM International Conference on Multimodal Interaction","publisher":"ACM","author":[{"propositions":[],"lastnames":["Xu"],"firstnames":["Hainan"],"suffixes":[]},{"propositions":[],"lastnames":["Fan"],"firstnames":["Yuchen"],"suffixes":[]},{"propositions":[],"lastnames":["Yu"],"firstnames":["Kai"],"suffixes":[]}],"year":"2012","pages":"539--544","bibtex":"@inproceedings{xu_development_2012,\n\taddress = {New York, NY, USA},\n\tseries = {{ICMI} '12},\n\ttitle = {Development of the 2012 {SJTU} {HVR} {System}},\n\tisbn = {978-1-4503-1467-1},\n\turl = {http://doi.acm.org/10.1145/2388676.2388790},\n\tdoi = {10.1145/2388676.2388790},\n\tabstract = {Haptic voice recognition (HVR) is a multi-modal text entry method for smart mobile devices. It employs haptic events generated by speakers during speaking to achieve better efficiency and robustness for automatic speech recognition. This paper describes the detailed design of the 2012 SJTU submission for the HVR Grand Challenge. During the design, a new perplexity metric using conditional entropy is proposed to evaluate the potential search space reduction of a haptic event without speech input. A number of new haptic events are evaluated both theoretically and experimentally in detail. The final submission system uses the haptic event of initial letter plus final letter and reduces word error rate by 76\\% compared to the baseline initial letter event.},\n\turldate = {2014-06-05TZ},\n\tbooktitle = {Proceedings of the 14th {ACM} {International} {Conference} on {Multimodal} {Interaction}},\n\tpublisher = {ACM},\n\tauthor = {Xu, Hainan and Fan, Yuchen and Yu, Kai},\n\tyear = {2012},\n\tpages = {539--544}\n}\n\n","author_short":["Xu, H.","Fan, Y.","Yu, K."],"key":"xu_development_2012","id":"xu_development_2012","bibbaseid":"xu-fan-yu-developmentofthe2012sjtuhvrsystem-2012","role":"author","urls":{"Paper":"http://doi.acm.org/10.1145/2388676.2388790"},"downloads":0},"search_terms":["development","2012","sjtu","hvr","system","xu","fan","yu"],"keywords":[],"authorIDs":[],"dataSources":["tudya6YojbqEiF783"]}