Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2021-01-12 Thread Ajinkya Bobade
Hi, If you need help on training tesseract 4 please write me mail at ajinkyabobad...@gmail.com Regards Ajinkya On Tue, Jan 12, 2021, 10:59 AM Murtuza Dahodwala wrote: > Tesstrain only uses single line of text for training. I want to train > several block of texts. > > On Saturday, January 9, 2

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2021-01-11 Thread Murtuza Dahodwala
Tesstrain only uses single line of text for training. I want to train several block of texts. On Saturday, January 9, 2021 at 8:04:18 PM UTC+5:30 wuer...@gmail.com wrote: > Of course you can! Just checkout the tesstrain tool: > https://github.com/tesseract-ocr/tesstrain > > Cheers, > Kay > > On

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2021-01-09 Thread Kay-Michael Würzner
Of course you can! Just checkout the tesstrain tool: https://github.com/tesseract-ocr/tesstrain Cheers, Kay On Fri, Jan 8, 2021 at 9:32 AM Murtuza Dahodwala wrote: > > > It is now 2 years since this answer was posted. Is it possible to train > tesseract 4 on real images now? > On Thursday, Janu

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2021-01-08 Thread Murtuza Dahodwala
It is now 2 years since this answer was posted. Is it possible to train tesseract 4 on real images now? On Thursday, January 11, 2018 at 2:27:43 PM UTC+5:30 shree wrote: > Currently, Ray/Google has NOT released info on how to train Tesseract 4 > (LSTM) with real life images. The only supported

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2020-08-22 Thread sat...@jukshio.com
Hi Sir/madam I am requesting for the following questions can we print the input shape and output shape for each layer in the training process can you explain how an image feed into the LSTM cells with timestamps in the training process ? how does Tesseract language model work in the training On

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2020-08-07 Thread minh...@gmail.com
Dear Quan, I don't see Mark EOL function, is it in Box Editor tab, or it is automatic add box a box at end of each line and add a tab character when using Generator? Thanks, TuPM On Wednesday, February 8, 2017 at 9:03:31 PM UTC+7 Quan Nguyen wrote: > > > On Tuesday, February 7, 2017 at 9:34:1

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2018-01-11 Thread ShreeDevi Kumar
Currently, Ray/Google has NOT released info on how to train Tesseract 4 (LSTM) with real life images. The only supported option is to use synthetic training data created by tesstrain.sh script using training text and unicode fonts. To train an LSTM model from scratch requires a large amount of tra

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2018-01-10 Thread Anubhav Rohatgi
Hi Shree, The box file uploaded by you as the attachment seems to contradict with the LSTM4.0 training tutorial guidelines, as there it states that the boxes should actually be at line level instead of at character level. Please do correct me if I am wrong. I still am not able to understand ho

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2018-01-08 Thread 'BartoszN.' via tesseract-ocr
Hi there, Any news about training Tesseract 4.0? I'm trying to train my fonts for few days... but still can't go anywhere. I'm looking for compact manual for training from txt/ttf files for new fonts in various languages. Will be appreciate for any help. On Tuesday, February 21, 2017 at 8:06

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2017-02-20 Thread timothylegg
I found this thread to be interesting since I tried training Tesseract a few years ago and gave up. Has anybody considered writing any documentation on this something that is best explained whenever a user can't figure it out from trial/error? I'm open to maybe writing about this if there is

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2017-02-09 Thread Kay-Michael Würzner
Thanks also from my side. I'll have a look into the jTessBoxEditor beta, try to setup training and get back to you. Kay On Wednesday, February 8, 2017 at 3:52:58 PM UTC+1, shree wrote: > > Thanks, Quan > > - excuse the brevity, sent from mobile > > On 08-Feb-2017 7:33 PM, "Quan Nguyen" > > wrot

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2017-02-08 Thread ShreeDevi Kumar
Thanks, Quan - excuse the brevity, sent from mobile On 08-Feb-2017 7:33 PM, "Quan Nguyen" wrote: > > > On Tuesday, February 7, 2017 at 9:34:11 AM UTC-6, shree wrote: >> >> ​For LSTM training, box files need to have an additional line for each >> text line with the tab character to indicate a ne

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2017-02-08 Thread Quan Nguyen
On Tuesday, February 7, 2017 at 9:34:11 AM UTC-6, shree wrote: > > ​For LSTM training, box files need to have an additional line for each > text line with the tab character to indicate a new line. > > If you have existing box/tiff pairs, you can use a box editor (such as > jtessboxeditor) and i

Re: [tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2017-02-07 Thread ShreeDevi Kumar
​For LSTM training, box files need to have an additional line for each text line with the tab character to indicate a new line. If you have existing box/tiff pairs, you can use a box editor (such as jtessboxeditor) and insert a box at end of each line and add a tab character in it. >On the toolba

[tesseract-ocr] Re: How can I do the training using my own image in Tesseract 4.0

2017-02-07 Thread Kay-Michael Würzner
+1 for this question. The training documentation for Tesseract 4.0 by now only covers training with font files (synthetic materials). What is missing is information on training with real data (i.e. manually aligned ground truth). Any hints on that matter are greatly appreciated. Cheers, Kay On