Recognition/translation issues

[mcabel]

example this image says "japanese" ( language) "nihongo". But only gibberish appears in the popup.

Example: first japanese image word in the website below.

https://www.japanesewords.net/3/common-japanese-words/

[OneMoreGres]

There are troubles with recognition of the first image possible because of a custom font. Other words are recognized more or less correct. Maybe switching app to a recent tesseract version will help. For now there are no workarounds.

[envooch]

Google translate from eng to ru doesn't work either.

[OneMoreGres]

Looks like a trouble on your machine, because I've just translated your comment via Google Translate. Try update translation scripts. If you use a proxy then set it in app's settings. Maybe translation services blacklisted your IP (try manually open translation page in normal browser, maybe it will require to solve captcha). You can enable "debug mode" in settings. It will show web page via which app performs translation. Right click -> inspect -> console could also provide some clues of what's wrong.

[envooch]

Translation works, but is not displayed. https://paste.pics/1ad8f8d325b26d9104c86d964bd36afe And I can not update https://paste.pics/52640fe270727d4d211434b6b38985ee https://paste.pics/1c52d48c6a90b180faf2b9bb28a4631c And i do not use proxy

[OneMoreGres]

To update try to run the app as admin (will fix this in future releases) or manually replace files in translators folder (near binary) from here.

[mcabel]

hi is there a way to add more fonts to the image recognition of the app? (also I have failed to make translation work, is there a sample of japanese text site (image based) that you have tested to work ? )

[OneMoreGres]

I haven't tried to add fonts, but it should be somehow like that. I don't test different languages specifically. The site you provided works fine except the first image. Random google images search works worse. Overall quality it average, just like for any other glyph-based language, and I'm not aware of a really good ways to improve it.

[envooch]

From the admin launched, everything was fine, thanks! But I noticed one thing. Previously, when I used this program, the text on the screen was almost always recognized as the original. Now very often the text is not recognized correctly. Words in one word merge, different characters instead of letters appear. What could be the problem?

[OneMoreGres]

I think it's just a coincidence. Only translation scripts are delivered via in-app updates, and they do not affect recognition results. In general, you can play with the 'scale factor' setting or add some auto substitution rules. Maybe you just was more lucky before update. :)

[OneMoreGres]

Changed issue name to more generic. You can place here samples with bad recognition results or inform about outdated translation scripts

[esrfkrgz]

Can you add "remove line break" for the translation?

[OneMoreGres]

Just add user correction from \n to space or smth

[Skyzenski]

I'm new to Github so i don't know how to create a topic to ask about my problem. I will try it here, hope you guys don't mind. I don't know how to use or what could trigger option [Capture saved areas], it seem always gray out for me. i guess this option could help me Translate a default area which i set before so i don't have to re-capture again and again?

[OneMoreGres]

How to create a topic

According to the capture saved areas. Yes, your understanding is right. To enable it: when selecting a capture area open the context menu (with right click) and select the Save (can capture via hotkey) option. You can have multiple areas saved

[thangtranv]

hi.for troubles with recognition of image possible you said switching app to a recent tesseract version. how can i do it or any ideal? Thank for your help.

[OneMoreGres]

Hi! To update the tesseract library you have to build its recent version from source and place it into the app's dir (or repack the appimage for linux). Maybe you'll have to rebuild the app too, if the the tesseract's new version is not binary compatible with the currently used by the app. So usually users should not do it unless the know how to do it.

[leonminus]

How to create a topic

According to the capture saved areas. Yes, your understanding is right. To enable it: when selecting a capture area open the context menu (with right click) and select the Save (can capture via hotkey) option. You can have multiple areas saved

In 3.3.0 version i dont see option save area by right click anymore, so how can i use [Capture saved areas] feature again i set it hotkey ctrl shft Q.

[OneMoreGres]

It should be there :) You don't see the whole selection menu or just this option? Try selecting with right mouse button from the beginning or right click while selecting with any button to get the menu