Translation Gurmukhi-To-English

[Genie Singh]

I wanted to discuss the various possibilities I have found on utilising the digitized manuscripts we find on the internet such as the Sarbloh Granth and it's translation to English perhaps as a means to render a search engine for it.

To begin with the source is examined which is this:http://www.scribd.com/doc/28563324/Complete-Sri-Sarbloh-Granth-Sahib-Ji-Steek

In this source the manuscript scanned in is in image form so it's characters aren't easily enter-able. So one of many things can be done which is to either contact the supplier and request or purchase the original digital version of it to prepare for a search engine or some search capacity based tool. Or to retype all of it which will take a long time, can take much man power and can become expensive when you employ a staff or deploy the project to a group to undertake.

The third practical option which we will explore is something known as OCR - Optical Character Recognition. Where by scans and images can with the use of various computer based algorithms be matched to their unicode/ascii characters (for us unicode).

Now when looking for OCR software I noticed Google docs provides this mechanism however they haven't utilised gurmukhi punjabi for this yet. However there were a few mentions across the internet of projects which are partially complete on this front so just to demonstrate this let's consider a few of them.

At the moment it seems these two are on offer, since I could not use Sherik by Dr Gurpreet Singh the setup files wouldn't function perhaps they aren't compatible with my operating system. I used the government of India's gurmukhi OCR feature- it took about 4mins for the following image to go onto their back end server to render into text. (http://webocr.cdacnoida.in/EditText.aspx)

Part of the purpose of the OCR project is to enable at some point the text to be audible so that those who are visually impaired may enjoy content that they couldn't previously.
http://www.techgig.com/tech-news/editors-pick/Indian-Govt-launches-web-based-optical-character-recognition-system-for-two-languages-6606

Gurmukhi OCRs[edit]

• Gurmukhi OCR (Download) Distributed by MCIT, Govt. of India
• Sherik (Gurmukhi OCR) by Dr. Gurpreet Singh Lehal

I found some academic papers for those who want to review how this actually works:
http://learnpunjabi.org/pdf/gslehal-pap5.pdf
http://www.ijcset.com/docs/IJCSET13-04-05-075.pdf

http://www.ijera.com/papers/Vol3_issue5/BJ35343348.pdf

http://www.ias.ac.in/sadhana/Pdf2002Feb/pe995.pdf

Here is the source image from the sarbloh granth:

This was rendered into unicode text as following:

ਸੰਸਾਰ ਦੀ ਮਹਾਨ , ਬੀਰਾਂ ਅਤੇ ਯੋਧਿਆਂ ਦੀ " ਖਾਲਸਾ " ਕੋਮ ਜਿਸ ਨੇ ਜੰਗ - ਇ - ਮੈਦਾਨ ।

ਅੰਦਰ ਜੂਝ ਕੇ ਸ਼ਹੀਦੀਆਂ ਪ੍ਰਾਪਤ ਕੀਤੀਆਂ ਅਤੇ ਚਰਖੜੀਆਂ ਤੇ ਚੜ੍ਹ ਕੇ , ਰੰਬੀਆਂ ਨਾਲ :

ਾ ਾ , ਭੱਠਾਂ ਵਿੱਚ ਜੀਉਂਦੇ ਸੜ ਕੇ , ਬੱਚਿਆਂ ਦੇ ਟੋਟੇ ਕਰਵਾ ਕੇ , ਇੰਜਨਾਂ ਥਲੇ : ਜੰਗ ਇ ਇ ਮੈਦਾਨ ਮੈਦਾਨ ਗਹਗਾਯਕਛਹਾੈੱਠਗ਼ਘੲ ਠਙਕਕਘਗੋਸਹਘਗਘਯ

ਕੁਚ ਲੇ ਜਾ ਕੇ ਮ ਤੇ ਚੋਂ ਕ ਤੇ ਮਾਰ ਕੇ ਬੰਦ ਬੰ ਦ ਕਟਵਾ ਕੇ " ਅ ਕਾਲ - ਭ । ਣੇ " ਨੂੰ " ਮਿੱਠਾ

( : : : ਸ ਮੈਸ਼ । ( - ਸ , । : : ( ( ਗੈ । ਹੁ ਸੈਘ : ) : ਦੈ : : ( ) , :

ਸਰਗੁਣੀ ਪਵਿੱਤ੍ਰ ਸਰੂਪ ਦੇ ਦਰਸ਼ਨ ਕਰਦੇ ਹਾਂ ਤਾਂ ਆਪ ਜੀ ਦੇ ਵੱਖ ਵੱਖ ਝਲਕਾਰਾ ਰੂ ਖੂ ਦਰਸ਼ਨ

ਹੁੰਦੇ ਹਨ । ਆਪ ਜੀ ਪਰੀਪਪੂਰਨ ਸਤਿਗੁਰੂ ਹਨ ਮਹਾਨ ਯੋਧਾ ਅਤੇ ਸਤਿਪ ਨੀਤੀ ਨਿਪੁੰਨ

ਧਰਮਪ ਰਖਯਕ ਸੈਨਾਪਤੀ ਵੀ ਹਨ । ਪਰੰਤੂ ਜਦ ਅਸੀ ਆਪ ਜੀ ਨੂੰ ਇਕਾ - ਮਹਾਨ " ਸਾਹਿੱਤਕਾਰ "

ਅਤੇ " ਮਹਾਨ - ਕਵੀਂ ' ਦੇ " ਕਵਿ ਰੰਗ ਸੁੰਦ੍ਰਂ ' ਸ੍ਹਰੂਪ ਵਿੱਚ ਦਰਸ਼ਨ ਕਰਦੇ ਹਾਂ ਤ ' ਬੁੱਧੀ ਚਕ੍ਰਿਤ

ਰਹਿ ਜਾਂਦ । ਹੈ ।

ਆਪ ਜੀ ਦੀਆਂ ਮਹਾਨ ਕਾਵਿ - ਕਲਾਪ ੱਿਕਤੀਆਂਘ ਜਾਪੁ ਸਾਹਿਬ , ਅਕਾਲਪਉਸਤਤਿ ,

ਬਚਿੱਤ੍ਰ ਨਾਟਕ , ਜ਼ਫਰਨਾਮਹ , ਗਯਾਨਪ ਪ੍ਰਕੋਧ , ਸ਼ਬਦ , ਸਵੱਯੇ , ਚੌਪਈ , ਚੰਡੀ - ਚਰਿਤ੍ਰ , ਤ੍ਰੀਯਾ

ਖਾਂਰੋਬੋ ਅ੍ਰ ਆਦਿਕ ਹਨ , ਜਿਨ੍ਹਾਂ ਰਚਨਾਵਾਂ ਦਾ ਭਾਰਤੀ ਸਾਹਿੱਤ ਵਿੱਚ ਹੋਰ ਕੋਈ ਸਾਨੀ ਨਹੀਂ ਹੈ ।

" ਜਾਪੁ ਸਾਹਿਬ , ਅਕਾਲ ਉਸਤਤਿ ਅਤੇ ਗਯਾਨਾ - ਪ੍ਰਬੇਧ ਨਾਟਕ " ਤਾਂ ਆਦਿ ਸ੍ਰੀ ਗੁਰੂ ਗ੍ਰੰਥ

ਸਾਹਿਬ ਜੀ ਦੇ ਮਹਾਨ ਅਤੇ ਅਧਿਯਾਤਮਿਕ ਆਸ਼ਿਆ ਦੇ ਆਧਾਰ ਉਤੈ ਹੀ ਸਤਿਗੁਰੁ

ਕਲਗੀਧਰ ਜੀ ਦੇ ਗੁਰ - ਹਿਰਦਯ ਅੰਦਰੋਂ ਅਜ਼ਲੀ ਤੇ ਅਕਾਲੀ - ਬਾਣੀ ਦਾ ਸੋਮਾ ਵਹਿ ਨਿਕਲਿਾਮਾਂ

ਹੈ । ਥਚਿੱਤ - ਨਾਟਕ ਅੋਤੇ ਜ਼ਫਰਨਾਮਹ ਇਤਿਹਾਸਿਕ - ਜੱਸ ਤੋਂ ਭਜ਼ਪਰ ੁ ਬੀਰ - ਰਸ ਦਾ ਸੰਦ ਕਿਸ਼ਮਾ

I used the above and Google translate to translate the Gurmukhi to English and retrieved the following and would like your thoughts on perhaps taking this process into implementing some form of software package which could make the process much more easier, more robust and simpler.

(http://translate.google.com/)

World 's greatest , Bira and warriors of the " Khalsa " Com the war - i - field .

Received martyrdom and suffering in carakhariam rode with Rubbya :

Ñ ​​Ñ , living apart, burning , children get cut , related under : war i e playground playground ? Gahagayakachahaaithagagha ? ? ? ? ? ? ? ? Thanakakaghagosahaghagaghaya ? ? ?

Of glass , I take a hit by the cut off " a call - ground . Ne " " Sweet

( ::: Mr Mash . ( Q , . :: ( ( Gai . Hu saigha : ) : The :: ( ) :

Saraguni visit the holy image , then you keep different browser performance

Are . You paripapurana Satguru great warrior and satipa Policy efficient

Dharamapa rakhayaka commander . But we Ica - great " ਸਾਹਿੱਤਕਾਰ "

And " the great poet " Kavi beautiful colors are seen in sharupa wisdom rotating

Living magic . .

He 's great poetry ਕਲਾਪ ikatiangha Jaap Sahib , akalapausatati

Bachittar drama , zapharanamaha , gayanapa prakodha , words , savaye , Chaupae , Chandi - character , triya

Khanrobo and adika , who works in Indian literature is the undisputed leader .

" Jaap Sahib , Akal honor and Guinea - prabedha drama " Adi Sri Guru Granth

Sahib Ji 's great and adhiyatamika Asa based utai Guru

Tips on kalgidhar within ਹਿਰਦਯ Anjali SAD - flowing source of Bani nikaliamam

. Thacita play aote zapharanamaha historical - Jas bhazapara u bir - Ras tool types

And in Romanised as:

Sasāra dī mahāna, bīrāṁ atē yōdhi'āṁ dī" khālasā" kōma jisa nē jaga - i - maidāna.
Adara jūjha kē śahīdī'āṁ prāpata kītī'āṁ atē carakhaṛī'āṁ tē caṛha kē, rabī'āṁ nāla:
̔Ā ̔ā, bhaṭhāṁ vica jī'undē saṛa kē, baci'āṁ dē ṭōṭē karavā kē, ijanāṁ thalē: Jaga i i maidāna maidāna? Gahagāyakachahāaiṭhaġagha??????? ? Ṭhaṅakakaghagōsahaghagaghaya? ? ?
Kuca lē jā kē ma tē cōṁ ka tē māra kē bada ba da kaṭavā kē" a kāla - bha. Ṇē" nū" miṭhā

( : : : Sa maiśa. ( - Sa, . : : ( (Gai. Hu saigha: ): Dai: : ( ), :
Saraguṇī pavitra sarūpa dē daraśana karadē hāṁ tāṁ āpa jī dē vakha vakha jhalakārā rū khū daraśana
Hudē hana. Āpa jī parīpapūrana satigurū hana mahāna yōdhā atē satipa nītī nipuna
Dharamapa rakhayaka saināpatī vī hana. Paratū jada asī āpa jī nū ikā - mahāna" sāhitakāra"
Atē" mahāna - kavīṁ' dē" kavi raga sudraṁ' s'harūpa vica daraśana karadē hāṁ ta' budhī cakrita
Rahi jānda. Hai.
Āpa jī dī'āṁ mahāna kāvi - kalāpa ̔ikatī'āṅgha jāpu sāhiba, akālapa'usatati,
Bacitra nāṭaka, zapharanāmaha, gayānapa prakōdha, śabada, savayē, caupa'ī, caḍī - caritra, trīyā
Khānrōbō ara ādika hana, jinhāṁ racanāvāṁ dā bhāratī sāhita vica hōra kō'ī sānī nahīṁ hai.
" Jāpu sāhiba, akāla usatati atē gayānā - prabēdha nāṭaka" tāṁ ādi srī gurū gratha
Sāhiba jī dē mahāna atē adhiyātamika āśi'ā dē ādhāra utai hī satiguru
Kalagīdhara jī dē gura - hiradaya adarōṁ azalī tē akālī - bāṇī dā sōmā vahi nikaliāmāṁ
Hai. Thacita - nāṭaka a̔ōtē zapharanāmaha itihāsika - jasa tōṁ bhazapara ̔u bīra - rasa dā sada kiśamā

Edited February 2, 2014 by JatherdarSahib

[Genie Singh]

Having just corresponded with the developer of the OCR. I have found out that this OCR functions only with 300dpi images and the scan is about 200dpi or less hence it's level of accuracy is low to detect the characters.