December 2, 2009 at 1:16 pm (FOSS, Uncategorized) ()

I attended the OCR round table at NCIDEEE 2009 held at Loyola College, Chennai. We left the hotel at around 9:30 AM and reached Loyola College. We were lead into a huge hall with empty seats. Within no time the seats filled up with student.

I approached Prof. C.V.Jawahar and told him about my project. He had already heard of my work thanks to Jinesh.  Jinesh was along with me. I showed him a new character classification technique that I have been thinking about implementing in Tesseract. He seemed to like it. He said “good”. 🙂

Then a person announced the guests and speakers for the day. Strangely, my name was missing. Every single person also got a memento except me!

The talks began. Prof. Shantanu Choudhury from IIT Delhi gave almost an hour long talk about the status of the OCR consortium project. Yes, there is an OCR consortium and he leads the consortium. It is being funded by the Indian Government and their aim is to produce a working Indic OCR for 10 languages. They are into their third year.

He was followed by Prof. C.V. Jawahar who gave an amazing and amusing talk on the difficulties faced in OCR development. I was well aware of most of the problems since I have faced all of them in my experience. He kept giving disclaimers all through that the performance of an OCR is a very difficulty quantity to measure and is heavily situation dependent. He had a slide on alternate development models. He said open source development model does not work well for OCR because code was not as important as the technology involved in OCR (?). He said at that time that he had no problem in giving away the source code, but was confident that no one could use it properly (?).

His talk was followed by a few more talks. What I found disturbing was that all these institutes were developing separate code bases for separate languages. A professor from Punjab University displayed a Gurmukhi OCR which uses three different classifiers. He reported 98% accuracy.

There were a few more talks. I was feeling very sleepy. Post lunch there was a demo of the OCR they have been developing. They ran it on a Fedora box and the interface was Gtk. I found the OCR to be very slow (took around 1 minute to OCR one page) and the output had vowel reordering errors. The page was pretty noise free. I was impressed with the interface, but not with the performance.

The people present with visual impairments asked a lot of questions and were disappointed to learn that OCR has still not matured as a technology. There was a particular visually challenged person who asked the professors to release all the code.

I sat, wondering, if I was at all supposed to give a talk. Infact one of the organisers introduced me to one of the guests as a  professor from IIT! I quickly corrected the error.

Two sessions went by and no one knew who I was and what I was doing there.

We moved to a smaller room with a few chairs and sat in a circle to discuss action points with the visually challenged people. People were still curiously looking at me wondering why the organisers had paid for my travel and lodging.

The discussions started. It was a very interesting discussion and I was waiting for a turn to say something. I eventually found it. I jumped in asking a question about their plans of freeing the source code. They said that there are Govt. restrictions and the Govt. “owns” the code hence he could not release it although he wanted to do it. I asked Prof Shantanu if there were Intellectual Property issues keeping them form releasing the source code. He agreed there were. I asked him whether he could request the funding agencies to allow them to free the code since it was slowing down the speed of development. He said “Sure, we will do that”, but he was saying that to every single suggestion given to him. He had to. I felt they were asking too much of him. I explained to them that freeing the code will take a lot of burden off their shoulders, especially in the areas of designing UI and creating custom applications and test data. They seemed to agree but kept mentioning the Govt. clauses.

I discussed the example of Tesseract OCR (all professors were familiar with the project) as a testament to the fact that OCR code must be freed. But they did not seem to understand.

I started a discussion on RFID based obstacle/object identification/detection and the use of SONAR. That turned out to be an interesting few minutes of dicussion and Prof. Shantanu seemed to have done some research on the topic.

As things drew to a close, I said to the visually impaired person sitting next to me “Dont worry, I am working on it too”.

Everyone kept asking me what I do, and I did not know what to say, since I recently quit my job. I told people there about FOSS.IN and that OCR is on of the sessions there. They were surprised to find out that there are people other than them working on OCR!

It ended with a burst of frustration and anguish from the Professors over the arrangements made. We headed back to the hotel.

Am typing the last line of this post, and looking forward to adding some code for tomorrow session at FOSS.IN 🙂



  1. vignesh said,

    what happened to your plan to go to Beach? no mentions about it..

  2. debayan said,

    Beach from here by auto is 100 Rs. Bus route I dont wanna experiment with. Its dark already. So I dropped the plan. FOSS.IN needs some attention from my side right now anyways.

  3. Jerald said,


    Thanks for your comments. I have requested our professors to send their profile, abstract before Nov 11, 2009. They are sending it till today( Jan 07, 2010).

    Kindly organize a NICE event on the same topic.


  4. Nirmal Verma said,

    What about your project of developing OCR for india language? are you going open it for all? need to see solutions for millions of indian visually challenged persons.

  5. EdICT 2010 « Debayan’s Weblog said,

    […] Conference on Enabling Access to Education through ICT‘ is going to be held in Delhi. Last year I was one of the participants at this conference in Chennai. I have received an invitation from one […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: