Improving Polytonic Greek OCR

In this forum, we will work with people who are training OCR engines for Polytonic Greek, helping them improve the quality of their engines by correcting specific texts.
Jonathan Robie
Posts: 3374
Joined: May 5th, 2011, 5:34 pm
Location: Durham, NC
Contact:

Improving Polytonic Greek OCR

Post by Jonathan Robie » May 16th, 2014, 6:07 pm

Bruce Robertson and I have been discussing an idea for improving polytonic Greek OCR. He needs people who will correct specific sets of pages in OCR output to help improve OCR recognition. I suggested that we could have threads along these lines:

1. The first page in a thread has an attached image and the OCR output
2. We ask for volunteers to look at a handful of pages and their output, and correct that output
3. Bruce uses this to improve recognition in his OCR engine, then posts the new OCR output for the original image

We are thinking it might be interesting and useful to start with Migne.

Would this be a good project on B-Greek, or does it belong elsewhere?
0 x


ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/

Stephen Hughes
Posts: 3323
Joined: February 26th, 2013, 7:12 am

Not so interesting for Greek orchid growing enthusiasts

Post by Stephen Hughes » May 17th, 2014, 1:35 am

Jonathan Robie wrote:Bruce Robertson ... needs people who will correct specific sets of pages in OCR output to help improve OCR recognition.
Do OCR readers recognise characters on a typeface by typeface basis, or does the machine get trained on all possible typefaces, and then work out itself which one it is looking at? Will there be a Migne OCR recognition set, then another for a different typeface?
Jonathan Robie wrote:Would this be a good project on B-Greek, or does it belong elsewhere?
Probably of more interest here on the B-Greek forum, than on the Φόρουμ της Ελληνικής Εταιρείας Φίλων Ορχιδέας / Greek Orchid Society forum.

Seriously, though, a forum (orchid or otherwise) is a more or less "permanent" record of things and what you are proposing is a work in progress. Something like:
ΚαῑααραιΝιιῃΧωνΑΠ, ἒ τῑολὺ ηὐξηυὲυος, εὐμε- ηολυαρκὲς τῆς τα ιχείσς· 3) τὸ ὰρσ. ιιολυ- γὲθης, ευμήκης, Ισχυρὸςε Νικ.Θηρ.73 ηολυαυξής
will not be of lasting interst after the project is over. Perhaps he could email the OCR garble to people who were willing to correct it. On an orchid forum at least, people want to see a record of pictures of the beautiful flowers at their best.
0 x
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Louis L Sorenson
Posts: 706
Joined: May 5th, 2011, 9:21 pm
Location: Burnsville, MN, USA
Contact:

Re: Improving Polytonic Greek OCR

Post by Louis L Sorenson » May 17th, 2014, 5:05 am

I think it would be a good endeavor. Though most likely only a few will participate. I'd also like to be able to recognize the font in which any given text is printed: Porson, etc.
0 x

Barry Hofstetter
Posts: 1218
Joined: May 6th, 2011, 1:48 pm

Re: Improving Polytonic Greek OCR

Post by Barry Hofstetter » May 17th, 2014, 6:37 am

I think it would be a good thing. Migne would be excellent.
0 x
N.E. Barry Hofstetter
Instructor of Latin
Jack M. Barrack Hebrew Academy
Χαίρετε ἐν κυρίῳ πάντοτε· πάλιν ἐρῶ, χαίρετε

cwconrad
Posts: 2109
Joined: May 5th, 2011, 5:52 pm
Location: Burnsville, NC 28714
Contact:

Re: Not so interesting for Greek orchid growing enthusiasts

Post by cwconrad » May 17th, 2014, 8:03 am

Stephen Hughes wrote:
Jonathan Robie wrote:Would this be a good project on B-Greek, or does it belong elsewhere?
Probably of more interest here on the B-Greek forum, than on the Φόρουμ της Ελληνικής Εταιρείας Φίλων Ορχιδέας / Greek Orchid Society forum.

Seriously, though, a forum (orchid or otherwise) is a more or less "permanent" record of things and what you are proposing is a work in progress. Something like "[...]" will not be of lasting interest after the project is over. Perhaps he could email the OCR garble to people who were willing to correct it. On an orchid forum at least, people want to see a record of pictures of the beautiful flowers at their best.
Pindar, Pyth. 8.95-97 wrote:ἐπάμεροι: τί δέ τις; τί δ᾽ οὔ τις; σκιᾶς ὄναρ
ἄνθρωπος. ἀλλ᾽ ὅταν αἴγλα διόσδοτος ἔλθῃ,
λαμπρὸν φέγγος ἔπεστιν ἀνδρῶν καὶ μείλιχος αἰών::
What's ephemeral and what's permanent is more subjective than we'd like to think. I'd love to see new fragments of Sappho as others might love to look at a scrap of Origen. B-Greek archives of the email list of olden days and of the forum, contain both trinkets and treasures. Sifting them might be a matter of taste: whose?
0 x
οὔτοι ἀπ’ ἀρχῆς πάντα θεοὶ θνητοῖς ὑπέδειξαν,
ἀλλὰ χρόνῳ ζητέοντες ἐφευρίσκουσιν ἄμεινον. (Xenophanes, Fragment 16)

Carl W. Conrad
Department of Classics, Washington University (Retired)

Jonathan Robie
Posts: 3374
Joined: May 5th, 2011, 5:34 pm
Location: Durham, NC
Contact:

Re: Not so interesting for Greek orchid growing enthusiasts

Post by Jonathan Robie » May 17th, 2014, 9:25 am

Stephen Hughes wrote:Seriously, though, a forum (orchid or otherwise) is a more or less "permanent" record of things and what you are proposing is a work in progress. Something like:
ΚαῑααραιΝιιῃΧωνΑΠ, ἒ τῑολὺ ηὐξηυὲυος, εὐμε- ηολυαρκὲς τῆς τα ιχείσς· 3) τὸ ὰρσ. ιιολυ- γὲθης, ευμήκης, Ισχυρὸςε Νικ.Θηρ.73 ηολυαυξής
will not be of lasting interst after the project is over. Perhaps he could email the OCR garble to people who were willing to correct it. On an orchid forum at least, people want to see a record of pictures of the beautiful flowers at their best.
I think it's like the "before" and "after" pictures on a fitness site, you need both to tell the story.
0 x
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/

Stephen Hughes
Posts: 3323
Joined: February 26th, 2013, 7:12 am

Mechanics v. beauty.

Post by Stephen Hughes » May 17th, 2014, 9:28 am

cwconrad wrote:
Pindar, Pyth. 8.95-97 wrote:ἐπάμεροι: τί δέ τις; τί δ᾽ οὔ τις; σκιᾶς ὄναρ
ἄνθρωπος. ἀλλ᾽ ὅταν αἴγλα διόσδοτος ἔλθῃ,
λαμπρὸν φέγγος ἔπεστιν ἀνδρῶν καὶ μείλιχος αἰών::
What's ephemeral and what's permanent is more subjective than we'd like to think. I'd love to see new fragments of Sappho as others might love to look at a scrap of Origen. B-Greek archives of the email list of olden days and of the forum, contain both trinkets and treasures. Sifting them might be a matter of taste: whose?
I am, perhaps, just beginning to understand the mechanics of this language, but you have long ago found its beauty.
0 x
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Stephen Hughes
Posts: 3323
Joined: February 26th, 2013, 7:12 am

The "before" and "after" shots.

Post by Stephen Hughes » May 17th, 2014, 9:32 am

Jonathan Robie wrote:
Stephen Hughes wrote:Seriously, though, a forum (orchid or otherwise) is a more or less "permanent" record of things and what you are proposing is a work in progress. Something like:
ΚαῑααραιΝιιῃΧωνΑΠ, ἒ τῑολὺ ηὐξηυὲυος, εὐμε- ηολυαρκὲς τῆς τα ιχείσς· 3) τὸ ὰρσ. ιιολυ- γὲθης, ευμήκης, Ισχυρὸςε Νικ.Θηρ.73 ηολυαυξής
will not be of lasting interst after the project is over. Perhaps he could email the OCR garble to people who were willing to correct it. On an orchid forum at least, people want to see a record of pictures of the beautiful flowers at their best.
I think it's like the "before" and "after" pictures on a fitness site, you need both to tell the story.
You are probably right. (After staying with my mother and auntie for a few months, every morning's mirror lets me understand quite well what you mean by the "before" picture).
Last edited by Stephen Hughes on May 17th, 2014, 9:54 am, edited 1 time in total.
0 x
Γελᾷ δ' ὁ μωρός, κἄν τι μὴ γέλοιον ᾖ
(Menander, Γνῶμαι μονόστιχοι 108)

Jonathan Robie
Posts: 3374
Joined: May 5th, 2011, 5:34 pm
Location: Durham, NC
Contact:

Re: Improving Polytonic Greek OCR

Post by Jonathan Robie » May 17th, 2014, 9:47 am

In 10 years, I bet Polytonic Greek OCR will "just work" the same way that Unicode and fonts "just work" these days (if you're not using an Android). And it would be nice to have a historical record to show the work that people like Bruce and Nick are doing to make this happen. People will forget that this was once hard.
0 x
ἐξίσταντο δὲ πάντες καὶ διηποροῦντο, ἄλλος πρὸς ἄλλον λέγοντες, τί θέλει τοῦτο εἶναι;
http://jonathanrobie.biblicalhumanities.org/

cwconrad
Posts: 2109
Joined: May 5th, 2011, 5:52 pm
Location: Burnsville, NC 28714
Contact:

Re: Improving Polytonic Greek OCR

Post by cwconrad » May 17th, 2014, 9:55 am

Jonathan Robie wrote:In 10 years, I bet Polytonic Greek OCR will "just work" the same way that Unicode and fonts "just work" these days (if you're not using an Android). And it would be nice to have a historical record to show the work that people like Bruce and Nick are doing to make this happen. People will forget that this was once hard.
One of the neatest secondary benefits of our having the texts of Homer is that we are privileged to explore the archaeology of the Greek language. The reconstruction of every kind of history is a never-ending challenge as well as a fascinating pastime.
0 x
οὔτοι ἀπ’ ἀρχῆς πάντα θεοὶ θνητοῖς ὑπέδειξαν,
ἀλλὰ χρόνῳ ζητέοντες ἐφευρίσκουσιν ἄμεινον. (Xenophanes, Fragment 16)

Carl W. Conrad
Department of Classics, Washington University (Retired)

Post Reply