Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

SVM For OCR, is it possible to use input methods similar to neural net ?

0 views
Skip to first unread message

Eizo

unread,
Dec 31, 2009, 2:18:14 PM12/31/09
to
While it is possible to find some features that describe letters in
the 2 dim space, I was looking to speed things up, by entering the
letter as is. in its raw binary info (like neural net).

So is it possible to train the SVM with a character
(e.g one vector that contain binary letter information, with 1's and
0's) like in the neural net training ?
Because all the demos, and all the articles I read deal with 2 dim
info.


Thanks.

sami

unread,
Feb 12, 2010, 6:12:03 AM2/12/10
to
Eizo <odp...@gmail.com> wrote in message <a1291231-2de3-442d...@r24g2000yqd.googlegroups.com>...


why there is no answers for this question ??

also I am looking to something can simplify the use of SVM in MATLAB

Walter Roberson

unread,
Feb 12, 2010, 11:49:58 AM2/12/10
to
sami wrote:
> Eizo <odp...@gmail.com> wrote in message
> <a1291231-2de3-442d...@r24g2000yqd.googlegroups.com>...

>> So is it possible to train the SVM with a character


>> (e.g one vector that contain binary letter information, with 1's and
>> 0's) like in the neural net training ?

> why there is no answers for this question ??

*I* didn't answer because I don't know anything about SVM.

In order for you to receive an answer, someone who has a clue about the
answer has to read your question and have the time and enough interest
to volunteer the answer. There isn't any automatic machine parsing
through the postings and using artificial intelligence to answer the
questions -- it all comes down to humans. And the odd thing about humans
is that none of them knows everything and none of them has enough time
to meaningfully answer every question.

0 new messages