Re: Recognizing children counting up/down (Ray Goldsworthy )


Subject: Re: Recognizing children counting up/down
From:    Ray Goldsworthy  <raygold@xxxxxxxx>
Date:    Mon, 20 Feb 2012 09:48:16 -0800
List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>

--20cf307f3a702bfbe004b968e6ed Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: quoted-printable Dear Chris, MIT has made available the following resources for web-embedded voice recognition: http://wami.csail.mit.edu/ I hope you find that helpful, Ray On Mon, Feb 20, 2012 at 4:53 AM, Christian V=F6geli <cvoegeli@xxxxxxxx>= wrote: > Dear List**** > > ** ** > > I am looking for speech detection methods/algorithms that would allow to > determine if a child is counting up or down correctly in the number range > from 0 to 100.**** > > ** ** > > We want to use the methods in a software to help children with > dyscalculia. One of the basic skills to be acquired before doing math as > such is to be able to count. Our system would e.g. tell the child to star= t > counting backwards from 51, starting with 51. The speech detection method > would then have to be able to decide if the child first speaks =93fifty o= ne=94, > then =93fifty=94, =93forty nine=94, etc.**** > > ** ** > > I know it=92s hard to do speech detection in children and that there is o= nly > a limited number of samples to train e.g. an HMM, but I hope that by > knowing which number should be spoken and by limiting the detection to ye= s > or no, there might be a feasible way.**** > > ** ** > > Do you have any suggestion to which algorithm we should investigate that > might help with this particular problem?**** > > ** ** > > Thanks for your help,**** > > Chris**** > > ** ** > > ** ** > > ---**** > > Dybuster AG**** > > Christian V=F6geli**** > > Weinbergstr. 20**** > > CH-8001 Z=FCrich**** > > ** ** > > Tel: +41 (0)44 250 76 10**** > > Fax: +41 (0)44 250 76 11**** > > www.dybuster.com**** > > cvoegeli@xxxxxxxx**** > > ** ** > --=20 Ray Goldsworthy Research Scientist Sensimetrics Corporation --20cf307f3a702bfbe004b968e6ed Content-Type: text/html; charset=windows-1252 Content-Transfer-Encoding: quoted-printable Dear Chris,<br><br>MIT has made available the following resources for web-e= mbedded voice recognition:<br><br><a href=3D"http://wami.csail.mit.edu/">ht= tp://wami.csail.mit.edu/</a><br><br>I hope you find that helpful, Ray<br><b= r> <br><br><div class=3D"gmail_quote">On Mon, Feb 20, 2012 at 4:53 AM, Christi= an V=F6geli <span dir=3D"ltr">&lt;<a href=3D"mailto:cvoegeli@xxxxxxxx">= cvoegeli@xxxxxxxx</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_q= uote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1e= x"> <div link=3D"blue" vlink=3D"purple" lang=3D"DE-CH"><div><p class=3D"MsoNorm= al">Dear List<u></u><u></u></p><p class=3D"MsoNormal"><u></u>=A0<u></u></p>= <p class=3D"MsoNormal"><span lang=3D"EN-US">I am looking for speech detecti= on methods/algorithms that would allow to determine if a child is counting = up or down correctly in the number range from 0 to 100.<u></u><u></u></span= ></p> <p class=3D"MsoNormal"><span lang=3D"EN-US"><u></u>=A0<u></u></span></p><p = class=3D"MsoNormal"><span lang=3D"EN-US">We want to use the methods in a so= ftware to help children with dyscalculia. One of the basic skills to be acq= uired before doing math as such is to be able to count. Our system would e.= g. tell the child to start counting backwards from 51, starting with 51. Th= e speech detection method would then have to be able to decide if the child= first speaks =93fifty one=94, then =93fifty=94, =93forty nine=94, etc.<u><= /u><u></u></span></p> <p class=3D"MsoNormal"><span lang=3D"EN-US"><u></u>=A0<u></u></span></p><p = class=3D"MsoNormal"><span lang=3D"EN-US">I know it=92s hard to do speech de= tection in children and that there is only a limited number of samples to t= rain e.g. an HMM, but I hope that by knowing which number should be spoken = and by limiting the detection to yes or no, there might be a feasible way.<= u></u><u></u></span></p> <p class=3D"MsoNormal"><span lang=3D"EN-US"><u></u>=A0<u></u></span></p><p = class=3D"MsoNormal"><span lang=3D"EN-US">Do you have any suggestion to whic= h algorithm we should investigate that might help with this particular prob= lem?<u></u><u></u></span></p> <p class=3D"MsoNormal"><span lang=3D"EN-US"><u></u>=A0<u></u></span></p><p = class=3D"MsoNormal"><span lang=3D"EN-US">Thanks for your help,<u></u><u></u= ></span></p><p class=3D"MsoNormal"><span lang=3D"EN-US">Chris<u></u><u></u>= </span></p> <p class=3D"MsoNormal"><span lang=3D"EN-US"><u></u>=A0<u></u></span></p><p = class=3D"MsoNormal"><span lang=3D"EN-US"><u></u>=A0<u></u></span></p><p cla= ss=3D"MsoNormal">---<u></u><u></u></p><p class=3D"MsoNormal">Dybuster AG<u>= </u><u></u></p> <p class=3D"MsoNormal">Christian V=F6geli<u></u><u></u></p><p class=3D"MsoN= ormal">Weinbergstr. 20<u></u><u></u></p><p class=3D"MsoNormal">CH-8001 Z=FC= rich<u></u><u></u></p><p class=3D"MsoNormal"><u></u>=A0<u></u></p><p class= =3D"MsoNormal"> Tel: <a href=3D"tel:%2B41%20%280%2944%20250%2076%2010" value=3D"+4144250761= 0" target=3D"_blank">+41 (0)44 250 76 10</a><u></u><u></u></p><p class=3D"M= soNormal">Fax: <a href=3D"tel:%2B41%20%280%2944%20250%2076%2011" value=3D"+= 41442507611" target=3D"_blank">+41 (0)44 250 76 11</a><u></u><u></u></p> <p class=3D"MsoNormal"><a href=3D"http://www.dybuster.com/" target=3D"_blan= k"><span style=3D"color:blue">www.dybuster.com</span></a><u></u><u></u></p>= <p class=3D"MsoNormal"><a href=3D"mailto:cvoegeli@xxxxxxxx" target=3D"_= blank"><span style=3D"color:blue">cvoegeli@xxxxxxxx</span></a><u></u><u= ></u></p> <p class=3D"MsoNormal"><u></u>=A0<u></u></p></div></div></blockquote></div>= <br><br clear=3D"all"><br>-- <br>Ray Goldsworthy<br>Research Scientist<br>S= ensimetrics Corporation<br><br> --20cf307f3a702bfbe004b968e6ed--


This message came from the mail archive
/var/www/postings/2012/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University