getRTI - a portal for RTI

[Harish Krishnan]

Hey guys, In my pitch last event, I was seeking for anybody who can volunteer for the tech aspect for this portal. It involves a web front end ( PHP ) and a backend component using a OCR library. Please let me know if there is anybody who is interested in the same. 

[vam...@cloudpact.com]

Hi harish, I am interested in this. May i know the details please.

[Harish Krishnan]

"The RTI Act was passed in 2005 and since then, applications have been filed and replies received in huge numbers. So a lot of information and data is out in the public domain but unfortunately not in one place that is searchable by those who want to use it. It is either with individuals/organisations but also with government departments/public bodies  that have dispensed with the information. Information will be public in the true sense if it is accessible and reusable.

So the objectives of the portal are:

1. Make RTI info publicly available in the true sense of public

2. Organise it into a neatly searchable format: Cetral/state, Departments, Public bodies, Commissions etc.

3. Reusable format ( Need to discuss the technical issues around this. Scanned images vs

OCR.)

4. Make filing easier using the online platform ".

We have already started this effort and there is a team of well known RTI activists who are involved in the offline effort. I am looking at technical help on a few technical components. At devthon specifically, I want to experiment hacking some part of this portal. Please let me know if you have any other questions.

Warm regards,

Harish Krishnan,

http://about.me/Harish.Krishnan

(M): +91 9963598885

--
You received this message because you are subscribed to the Google Groups "Devthon" group.
To unsubscribe from this group and stop receiving emails from it, send an email to devthon+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[vam...@cloudpact.com]

May i know the details in which component i have to involve.

On Monday, July 8, 2013 3:54:50 PM UTC+5:30, Harish Krishnan wrote:

[Harish Krishnan]

We are planning to use php for the frontend. The backend needs an ocr component that can be done in any library supported by the tesseract engine. What can you build in?

Warm regards,

Harish Krishnan,

http://about.me/Harish.Krishnan

(M): +91 9963598885

[vam...@cloudpact.com]

I am interested on OCRlibrary part. 

On Monday, July 8, 2013 3:54:50 PM UTC+5:30, Harish Krishnan wrote:

[Harish Krishnan]

cool. Let me know . Add me "bugsy....@gmail.com" in gchat.

[Sree Ram Kumar]

Has someone thought of automating the scanning process in itself ?. The way i see it manually digitizing records is the most time-consuming task. 

On Fri, Jul 19, 2013 at 8:36 AM, Harish Krishnan <bugsy....@gmail.com> wrote:

cool. Let me know . Add me "bugsy....@gmail.com" in gchat.

--

[Harish Krishnan]

Sree ram, we are dealing with the govt here :)

[vam...@cloudpact.com]

Hi sreeram can you explain in detail of automating the scanning process.

On Monday, July 8, 2013 3:54:50 PM UTC+5:30, Harish Krishnan wrote:

[Sree Ram Kumar]

Why word it when i can show it :) . I wish something like this can be done. I can't explain in detail coz I don't know enough about the format of the documents or other constraints. https://www.youtube.com/watch?v=tCOXC5PTJj8

--

[Rakesh Kumar Reddy Dubbudu]

The documents are those that are supplied by Govt in response to various applications filed by citizens. So there is no uniform format. There could be even some handwritten replies.

[Sree Ram Kumar]

I know many ppl here are neck deep in AI-ML but here is a  humble opinion. There should be a policy decision at this stage on how you want to handle the documents, triage the level of automation you want to. I would like to propose an aggressive automation beach-head (automate as much as you can). This is motivated partly by the volume of applications (im guessing) and partly by the non-critical aspect of the data-entry. by which i mean the relatively big tolerance to error/data mismatches. (Plz dont mistake criticality with importance,   important-yes, critical no :) )

[Harish Krishnan]

Sree Ram, it is a feasible route just as yet. The govt takes a lot of time to adopt technology, we cannot wait for it, Hence being as innovative as we are :p, we create a solution for their slackness.

Warm regards,

Harish Krishnan,

http://about.me/Harish.Krishnan

(M): +91 9963598885

[Harish Krishnan]

Strangely, I always miss using "not" in my sentences :). The sentence was supposed to be read as "It is NOT yet feasible".

Warm regards,

Harish Krishnan,

http://about.me/Harish.Krishnan

(M): +91 9963598885

[Sree Ram Kumar]

Guess you are guys are much farther down the idea path, that explains why I didn't understand what you meant Harish :)

[Harish Krishnan]

Has any of you tried opencv for the Optical recognition? We are looking to try this out.

[Ravi Sharan]

SimpleCV is also a good option if you are trying to do it completely using Python.

[Sree Ram Kumar]

I have done ocr stuff before but not end 2 end. I took features and used nn/ml models. I think opencv can be powerful frontend.

Sree Ram