GSoC project status: Speech Recognition facility in open moko

[saurabh gupta]

Hello everyone,

I am saurabh gupta, working on speech recognition facility in open moko (
http://code.google.com/soc/2008/openmoko/appinfo.html?csaid=B631058E670ECFAC).
My mentor is Luis G. Lira. Here is my project status update. I have to
incorporate speech recognition facility in open moko using Hidden Markov
Model.

My project is basically divided into four stages.
1)Understanding the internal and working of open moko software and tools.
Building test applications and getting familiar with open embedded and qemu.
2)Understanding the Hidden Markov Model in detail and studying the LPC,
vector quantization etc.
3)Writing the code in C and trying some subroutines in scilab first (like
training of HMM model etc).
4)porting the code on open moko platform and testing it on real hardware.

    The second stage has been completed and third one has started. I have
studied from two books i.e. Fundamentals of speech recognition and Digital
processing of speech signal by L. Rabiner and some IEEE papers from the same
author or other. The first stage is going simultaneously. I am also
preparing a design document and implementation procedure which I am soon
going to upload on the wiki. I have started writing some code.

     I am getting some problem in Mokomakefile build. As soon as I fix it, I
will be through the first stage of my project also. Till then i will work on
the coding part and algorithm development. After that the full testing and
emulation using qemu will be done and if i get a Neo or Freerunner, then i
will test it on real hardware also.

Cheers...

-- 
Saurabh Gupta
Electronics and Communication Engg.
NSIT,New Delhi
"Chance favors the prepared mind......"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.openmoko.org/pipermail/openmoko-devel/attachments/20080605/c4541ef6/attachment.htm