[Project-ideas] Application for GSOC'13
Bhavani Shankar R
bhavi at ubuntu.com
Mon Apr 22 06:57:12 PDT 2013
On Mon, Apr 22, 2013 at 4:19 PM, Anand Dhandhania
<ananddhandhania11 at gmail.com> wrote:
> The process i would like to follow is :-
> 1) Train a neural network for the image data. We will be using only neural
> nets to make an approximation i.e we will take few most possible results
> from neural network model.
> 2) Algorithms like Hamming distance calculation will be implemented for only
> these results (from neural nets). We will find the best match using the
> hamming distance and other (k-means clustering) algorithms.
>
> This is what i have thought. I am currently working on finding the results
> for hand-written feature classification using Gaussian Mixture Models. We
> are surely gonna implement, what's best suited for the data.
>
Seems great for a start. Thanks for your replies.
A small question regarding to extension of your project (asking out of
personal interest):-
Whom have you thought of targeting as end users when a working model
is developed?
Regards,
--
Bhavani Shankar
Ubuntu Developer | www.ubuntu.com
https://launchpad.net/~bhavi
More information about the Project-ideas
mailing list