female [Leaf]I will not excuse you to pick flowers in the dungeon!

Recommended Posts

I hope people still want these, else I'll feel silly.

" Latest update date: August 30 (Friday) / Next update will be scheduled for September 30 (Monday) at 17:00 / The latest 2 volumes of comics are on sale now!

LEVEL17

Author's comment: It's nice to see our characters growing up on their own.
The painter thinks he has to work hard to draw them well."

Share this post


Link to post
Share on other sites

Is there anyone online who we could pay to translate this stuff?

Share this post


Link to post
Share on other sites
3 hours ago, Tyton said:

Is there anyone online who we could pay to translate this stuff?

I mean for the right price, certainly, but I don't know how much anyone can scrap together to pay for something like this. Translation work isn't cheap.

Edited by tenck5k
misread question, can't delete

Share this post


Link to post
Share on other sites

OK.  I know very very little about machine translation.  Inspired by this thread:

1.  Estimating every-day Kanji used 3,000 to 6,000 characters, I figured a Kanji character required two bytes.  That would leave plenty of combinations for non-printing and non-Kanji characters.  Wrong.  Some counts range up to 70,000 characters.

2.  So I went looking through Ubuntu software repository and the web for a Kanji optical character reader.  Only a mouse-driven scratch pad that informs the user what character the user's attempt best approximates.  It seems to include stroke order in the recognition process.  It might have been useful.  But no Linux guru has needed input and output.  So it's pad and screen. Period.

3.  In spite of item one, Google translate does process Kanji Japanese to ASCII English.  Of course, that means Kanji properly encoded as binary.

Probably, I approached this wrong.  Does anybody know of an inexpensive machine-translation path.

Thank you.

Share this post


Link to post
Share on other sites
22 hours ago, Stanley79 said:

Probably, I approached this wrong.  Does anybody know of an inexpensive machine-translation path.

Thank you.

 

There are tools and APIs for OCR. I know C# has an OCR Library [Source 1, Source 2] built in with support for Japanese and Chinese characters. All one would have to do is create a small tool that allows someone to select an area on the users monitor and auto-copy the text, feeding the output into google translate.

It's not a complex program to make by any means. That said, the bigger problem is that Comics and Manga often use more styalized fonts interwoven into complex images. Meaning it'll be harder for the system to get perfect recognition of the characters.

 

Share this post


Link to post
Share on other sites

Rinatro,

You are correct.  Running tesseract-ocr with gimageReader gui and typical manga sources, the combination produces so many errors, hand translation would be easier.

Share this post


Link to post
Share on other sites
On 9/11/2019 at 5:31 PM, Stanley79 said:

Rinatro,

You are correct.  Running tesseract-ocr with gimageReader gui and typical manga sources, the combination produces so many errors, hand translation would be easier.

For the best success, you'd basically have to find a way to isolate the characters. Either allowing the user to paint/select the speech bubbles, or have it automatically detect where they are. It'd work better than scanning the whole image. It wouldn't be perfect, but it would be a more cleaner source for the OCR to work off of.

Edited by Rinatro

Share this post


Link to post
Share on other sites

Based on experience (hobby articles) from machine translation from Spanish and Italian to English I copied the characters and pasted them into a clean page.  I hardly thought about it and forget to mention it.

Share this post


Link to post
Share on other sites

Machine Translation Experience

Today machine translation offered me something that might or might not be helpful.  The procedure can be done in less time than it takes to read the following steps:

1.  Main software:
Firefox or other web browser with tabs.
GIMP or other image manipulation program
gImageReader or other Tesseract GUI
open-source Optical Character Recognition (OCR) Engine.
tesseract-ocr language files for Japanese
tesseract-ocr language files for Japanese (vertical)

2.  With Firefox running open
https://www.omorashi.org/forums/topic/43199-leafi-will-not-excuse-you-to-pick-flowers-in-the-dungeon/
or other manga page.

3.  Open your image manipulation program.

4.  In Firefox right click on the image containing the Japanese characters.  “Copy Image” will place it on the clip board.  From the clip board paste it into the image manipulation program (should look like it did in Firefox).

5.  In the image manipulation program use “File” and “New” to open a second image-editing window.

6.  Copy and paste from the image manipulation program window with the original image (characters only) to the new window.  Arrange characters in a single column or row.  Save the new characters-only image as a .png image.

7.  Open gImageReader.  On far left click folder icon.  Select folder containing the image and open the image.

8.  Find the lower case “a” near center top, “recognize” to its right and down arrow right of recognize.  Click the down arrow.  Select the language (Japanese or vertical Japanese).  Screen extreme lower left should say “ready.”  Double click “recognize” at center top above selected language.  Wait.  Eventually, lower left should indicate reading and lower right contain a progress bar.

9.  When gImageReader indicates reading complete and displays the result in hideous format screen right, use the “save output” (bar-arrow-box) icon above displayed output.  Click the icon and save the output as .html.

10.  In Firefox open Google Translate and choose Japanese.  Open the html output from gImageReader in another tab.  Copy the character string from your digitized character tab to the Google translate tab’s left area.  After a few seconds, the English or whatever translation should appear in the right box.

Here’s my results.

“[Flower picking] Kunuma | The problems I can't stand still

Adventurers || Called "Elf Sanctuary" to challenge a huge dungeon.”

Or sometimes slight changes in source lead to larger changes in translation.

“Challenge a huge dungeon called the "Sanctuary of Elf"

Even if the problem is still L | I can't take pride in [picking flowers] ...”

Someone knowing something might be able to suggest improvements.

 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.