[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Siragi OCR project, uml design diagrams
- To: "Development Discussions" <developer at arabeyes dot org>
- Subject: Re: Siragi OCR project, uml design diagrams
- From: "Ahmad Sayed" <ahmad dot ahmadsayed at gmail dot com>
- Date: Sat, 20 May 2006 17:06:36 +0300
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:references; b=M8hDjgITY1MkToCapB+LoiBQadqTpgc5ftLQMhncat9+eWfOpw0qbGgoOF711JTXsmSaoR/J8ChtYnXJR0tnnopqu2UQgVT35sQub1rUAujj1XtKK7LRxshDReB9Rpfxo/VhZwSuRENQHhcAEGrUCB7gtUws0y/QyzB2yAyntA0=
Hi Tarik,
> But the
> neural network will output the same utf8 code for all the patterns
> corresponding to the same character. I already made some tests on 96 patterns
> in input of a neural network with only 28 codes as output.
as far as you made some tests so all should i said is "RESPECT", my suggestion is make sense theoretically but i think that the neural network became too large, i remember that i use 2 layer neural network to make classification for floating number (0-9) and e and +/-, so to classify and Arabic character that may have 3 state i couldn't expect that size of the neural network and its speed to be nice.
let us not wasting times in speaking too much I'm involved before in a set of open source projects that fail because in these projects we speak too much and works too little though the work was amazing but too much speak means earlier fail.
but i have a final question, you tell me that you make some tests how ?
do you use a premade neural network toolbox? or you made one?
if you use a premade one what is this? and if you made one by yourself to intend to attach it to this project?
Best Regards,
Ahmed Sayed