dc.contributor.author | Lingdell, Mattias | |
dc.date.accessioned | 2008-07-03T07:11:05Z | |
dc.date.available | 2008-07-03T07:11:05Z | |
dc.date.issued | 2008-07-03T07:11:05Z | |
dc.identifier.uri | http://hdl.handle.net/2077/10546 | |
dc.description.abstract | Cell phones are excellent candidates for speech recog-
nition due to their sometimes cumbersome interface.
However, few applications of speech recognition exist,
especially for Java ME enabled cell phones. This dearth
of speech recognition applications is unfortunate, since
there exist several hundred millions Java ME enabled
cell phones in the world, making the market and need
for such applications huge. This case study investigates
the feasibility of embedded, speaker independent speech
recognition of connected digits on modern Java ME en-
abled cell phones. The developed system is, through
a series of optimizations, capable of recognizing con-
nected digits in about 11 times real time. Unfortu-
nately, this performance is not nearly adequate to make
the system practically usable in real world mobile ap-
plications that deal with user interaction. | en |
dc.language.iso | eng | en |
dc.relation.ispartofseries | Report/IT University of Göteborg | en |
dc.relation.ispartofseries | 2008:062 | en |
dc.title | Embedded, Speaker Independent Speech Recognition of Connected Digits on Java ME Enabled Cell Phones | en |
dc.type | Text | eng |
dc.setspec.uppsok | Technology | |
dc.type.uppsok | C | |
dc.contributor.department | IT-universitetet i Göteborg/Tillämpad informationsteknologi | swe |
dc.contributor.department | IT University of Gothenburg /Applied Information Technology | eng |
dc.type.degree | Kandidatuppsats | swe |
dc.type.degree | Bachelor thesis | eng |