+- kanji koohii FORUM (http://forum.koohii.com)
+-- Forum: Learning Japanese (http://forum.koohii.com/forum-4.html)
+--- Forum: Learning resources (http://forum.koohii.com/forum-9.html)
+--- Thread: KanjiTomo - New OCR program for Japanese text (/thread-9971.html)
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-06-16
NaweG Wrote:In any case, I have been very happy with your .99 beta, but am wondering if there is any way to set it up so the display of the characters is larger.Option to increase the font size would not be difficult to add. Version 0.9.9 (not beta) is almost done and should be released within a few days, but maybe I can include it there already.
NaweG Wrote:Would also add a second vote for the capability to turn off the dictionary lookup and just do "pure" OCR if that's possible. Understand the main point of the tool is to help with learning, but it seems like I am using the program (rightly or wrongly) as often to just grab a couple lines to work with later as to try and puzzle out while I'm scanning.For the next major version after 0.9.9 I'm planning to add a feature to save identified words to a list and then export them to a text file/clipboard.
Even in current version you can copy only the text to clipboard (one word at a time); look at these options in config.txt:
# If set to 1, results are copied to clipboard automatically
COPY_TO_CLIPBOARD_AUTOMATIC=1
# Components included when copying results to clipboard
CLIPBOARD_INCLUDE_KANJI=1
CLIPBOARD_INCLUDE_KANA=0
CLIPBOARD_INCLUDE_DESCRIPTION=0
If you use this setup, you might want to turn off Automatic OCR mode and use left-clicks to target characters.
As for a real batch mode (where whole pages are scanned at once); it would be doable, but not easy. For example since user is currently asked to point to the first character in each word, I don't need to consider general layout of the page (except by trying to detect the local reading direction).
NaweG Wrote:Lastly, and I suspect your location has something to do with this, would you consider adding a donation link to your website? Given the effort you've put into this, it seems appropriate to drop a few bucks toward thanking you :-)I have been thinking about adding a donate button, but I have decided not to at the moment. The are various reasons; if you are interested I can discuss them in email.
KanjiTomo - New OCR program for Japanese text - NaweG - 2013-06-16
Kurotowa Wrote:As for a real batch mode (where whole pages are scanned at once); it would be doable, but not easy. For example since user is currently asked to point to the first character in each word, I don't need to consider general layout of the page (except by trying to detect the local reading direction).Not looking for a full page per se, but just to be able to select two to three lines (so I would still draw the box to help identify where the text is), and then have it come up in a single group that I could cut/paste. I'd have to go back and then re-identify anything missed, but in general I find that's only two to three kanji per paragraph. That's assuming that I am right (occasionally I think the program reads some better than I do).
Kurotowa Wrote:I have been thinking about adding a donate button, but I have decided not to at the moment. The are various reasons; if you are interested I can discuss them in email.Not trying to pry, but I may drop a line when I get my next paycheck :-)
Thanks!
KanjiTomo - New OCR program for Japanese text - Animosophy - 2013-06-16
You absolute beauty.
Even though I may not use this for a few months, I've been trying it out with a light novel and it works very well. For a while I've been hoping there'd exist a program that works just like this one in time for when it would most benefit me
the added export function would do exactly that.Is it possible to automatically format the text files to allow Anki imports? I'd absolutely insist on donating/paying for a service could collect unfarmiliar vocabulary into a single text file after a reading session, which I can then instantly import into a vocabulary deck without needing to edit a thing. I haven't tried out any other OCRs so I don't know if this is already possible.
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-06-16
NaweG Wrote:Not looking for a full page per se, but just to be able to select two to three lines (so I would still draw the box to help identify where the text is), and then have it come up in a single group that I could cut/paste. I'd have to go back and then re-identify anything missed, but in general I find that's only two to three kanji per paragraph. That's assuming that I am right (occasionally I think the program reads some better than I do).It could work like this:
1. user marks a block of text
2. orientation is detected (or fixed by user)
3. character boundaries are detected
4. characters are detected (in the order specified by orientation)
5. results are printed and copied to clipboard
Most of the pieces are already in place; maybe I'll look into this more closely in the future, but not in the next version.
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-06-16
Animosophy Wrote:Is it possible to automatically format the text files to allow Anki imports? I'd absolutely insist on donating/paying for a service could collect unfarmiliar vocabulary into a single text file after a reading session, which I can then instantly import into a vocabulary deck without needing to edit a thing. I haven't tried out any other OCRs so I don't know if this is already possible.The program will generate simple tab delimited text files, it should not be a problem to import them to Anki.
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-06-17
I have now uploaded version 0.9.9 to http://www.kanjitomo.net
New features are zoom mode and manual selection of word borders with click-and-drag. Alt+R or middle button (in file mode) opens the zoom frame on mouse cursor location; middle button can also be used to move or close the zoom frame.
I have also added a few new options to config.txt:
- ENABLE_NAMES_DICTIONARY: if set to 0, names dictionary is not loaded (saves memory)
- INTERFACE_BUTTONS_LEFT: if set to 1, Zoom and Names buttons are located to the left side of main window
- INTERFACE_WORD_DETAIL_FONT_SIZE: sets the font size of word detail panel (I will add options for other panels in future versions)
KanjiTomo - New OCR program for Japanese text - NaweG - 2013-06-17
Kurotowa Wrote:I have now uploaded version 0.9.9 to http://www.kanjitomo.netIf I have the beta version should I grab this anyway? I presume the things discussed recently are for a future release later this year and not in this version, right?
Thanks!
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-06-17
NaweG Wrote:Compared to beta version, there are some bugfixes and interface tweaks, so I would recommend downloading it again. No major new features though.Kurotowa Wrote:I have now uploaded version 0.9.9 to http://www.kanjitomo.netIf I have the beta version should I grab this anyway? I presume the things discussed recently are for a future release later this year and not in this version, right?
KanjiTomo - New OCR program for Japanese text - thamvp - 2013-07-14
Edit: my bad, after using it a few times i just realised u are able to open any content webpage, picture etc and use KanjiTomo. I originally thought u had to open the content thru KanjiTomo which is not necessary. This is soo awesome! Thanks so much!
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-07-15
I'm not sure what kind of zoom you need, but remember that with the zoom frame you can change the zoom amount, resize the frame and also move it with middle mouse button (which can be useful if the title bar is above screen border).
KanjiTomo - New OCR program for Japanese text - apirx - 2013-07-17
Hey I'd like to report 2 bugs.
Using Kanjitomo 0.9.9, Java 1.7.0_11
Kanjitomo doesn't close the Java process when closed by clicking the X or choosing File>Exit. The red OCR boxes continue to being drawn even after the Kanjitomo window is gone if AutoOCR was on. Java has to be closed through the task manager.
Kanjitomo uses up to 800-900MB of my ram. I didn't check with the older versions but this seems excessive. It fills my ram (I have 4gb) up to 95-99%. I can't watch a HD video while Kanjitomo is running.
Thanks a lot for making Kanjitomo!
KanjiTomo - New OCR program for Japanese text - thamvp - 2013-07-17
its all good cheers!
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-07-18
apirx Wrote:Using Kanjitomo 0.9.9, Java 1.7.0_11Memory usage increased in recent versions when dictionary for Japanese names was added. You can reduce the ram usage by setting ENABLE_NAMES_DICTIONARY=0 in config.txt.
Kanjitomo uses up to 800-900MB of my ram. I didn't check with the older versions but this seems excessive. It fills my ram (I have 4gb) up to 95-99%. I can't watch a HD video while Kanjitomo is running.
Problems with HD video could also be caused by CPU consumption because KanjiTomo detects each character in parallel. You could try setting MAX_CHARACTERS to 2 or 3.
apirx Wrote:Kanjitomo doesn't close the Java process when closed by clicking the X or choosing File>Exit. The red OCR boxes continue to being drawn even after the Kanjitomo window is gone if AutoOCR was on. Java has to be closed through the task manager.Can you tell me what operating system are you using? Are there others here who have the same problem? I have not seen this bug myself.
KanjiTomo - New OCR program for Japanese text - apirx - 2013-07-19
I'm using Windows 7 SP 1 64bit.
I have a second computer with the same OS where the problem doesn't exist. Maybe I have a faulty java installation?
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-07-20
apirx Wrote:I'm using Windows 7 SP 1 64bit.You could try starting KanjiTomo from the command line (cmd.exe): go to the unzipped directory and run: java -Xmx1000m -jar KanjiTomo.jar -run
I have a second computer with the same OS where the problem doesn't exist. Maybe I have a faulty java installation?
If there are any error messages they should be printed to the console. Normally errors are shown in popup window, but it might not be visible when the program is closing.
KanjiTomo - New OCR program for Japanese text - apirx - 2013-07-22
I identified the cause of the problem. This happened because I bound MANUAL_OCR_NOFILE_HOTKEY=mouse3
Even though the hotkey wasn't working, I left it in the config anyway. Changing it to something else resolved the problem with java not closing.
edit:
Don't want to bother you, but I noticed there are a few more issues with the hotkeys. I can't seem to get the toggle_auto_ocr_hotkey to work.
For example, I bind
CHANGE_DICTIONARY_HOTKEY=alt F
and
TOGGLE_AUTO_OCR_HOTKEY=alt D
The dictionary hotkey works without problems. The ocr hotkey causes the following error to appear:
java.lang.NullPointerException
kanjitomo.reader.Hotkeys$3.onHotKey(Hotkeys.java:87)
com.tulskiy.keymaster.common.Provider$HotKeyEvent.run(Provider.java:147)
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
java.lang.Thread.run(Thread.java:724)
program version:0.9.9
java version:1.7.0_25
If I switch the keys around (dictionary=alt D and ocr=alt F), it's still the dictionary hotkey that works, the ocr hotkey still brings up the exception error.
Thanks for all your work.
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-07-23
apirx Wrote:Don't want to bother you, but I noticed there are a few more issues with the hotkeys. I can't seem to get the toggle_auto_ocr_hotkey to work.This should be fixed now. Try downloading the program again and see if it works.
For example, I bind
CHANGE_DICTIONARY_HOTKEY=alt F
and
TOGGLE_AUTO_OCR_HOTKEY=alt D
The dictionary hotkey works without problems. The ocr hotkey causes the following error to appear:
KanjiTomo - New OCR program for Japanese text - znebr47625 - 2013-07-23
Do you plan to add audio and Japanese dictionaries too?
KanjiTomo - New OCR program for Japanese text - apirx - 2013-07-24
Thanks, it's working now!
KanjiTomo - New OCR program for Japanese text - NaweG - 2013-09-02
Just wanted to drop a message to say what a big help this program has been this summer! I've been using it to let me do some work on my breaks - I can grab a few lines to paste into a document from the source, and then look them over later when I have time. Otherwise I'd have to copy things by hand if I didn't want to have to display the pages every time I wanted to do a little translation work.
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-10-15
KanjiTomo has become quite popular and a couple of people have asked me if they could support it's development through donations. Unfortunately this is not currently allowed by Finnish law; there must be some kind of exchange.
I'm not planning to make KanjiTomo a commercial product, but how would you feel if I put some features in a paid version? It would not be expensive, something like €5 (through PayPal) and most of the features would still be available in the free version.
KanjiTomo - New OCR program for Japanese text - SomeCallMeChris - 2013-10-15
Alternatively, there's nothing that stops you from selling a disk with the program on it just because you also make it free for download; or a printed manual ; or even a Certificate of Authenticity for a payer to prove that they paid for their free download. (extra for an autograph from the developer!)
It's your software and you can do what you want, of course. The only problems with a separate paid version are when you're getting code from other people in a free software development model, but you don't seem to be doing that.
KanjiTomo - New OCR program for Japanese text - NaweG - 2013-10-15
Kurotowa Wrote:I'm not planning to make KanjiTomo a commercial product, but how would you feel if I put some features in a paid version? It would not be expensive, something like €5 (through PayPal) and most of the features would still be available in the free version.I think this would be a splendid idea, and would be happy to donate/contribute as soon as you set this up.
KanjiTomo - New OCR program for Japanese text - Kurotowa - 2013-10-27
I have now created a new version of KanjiTomo called "Supporter's Edition". It includes a new feature that allows saving identified words to a list and later export them to file or clipboard. You can buy Supporter's Edition through PayPal for €5, see http://www.kanjitomo.net/#Support for details.
KanjiTomo - New OCR program for Japanese text - NaweG - 2013-10-27
You now have at least one order in your inbox :-)