Repository navigation
Multiple issues #432
Description
This piece of software is great with the exception of it being extremely buggy.
What I'm trying to do: Japanese, vertical text -> English, horizontal text
OS: Windows 11
-
"Detect areas and recognize text"
This works nicely, but sometimes it misses a portion of the text. Sure, I can fix it manually but there's another problem. -
"Translate all text areas"
The translated text only fits the size of the detected boxes in 1). This results in a messy combination of vertical English letters. I have to manually typeset the box in order to fix that and make it into readable English. It becomes a pain when you have over 100 pictures to do this for.
The masking feature with LamaInpaint is also a nice addition but for it to work properly, I have to create a mask and then edit the text-removed image manually because the algorithm is set for telea. Otherwise the masking part doesn't look polished if I select LamaInpaint as the default text-removal tool in Project>Settings. Again, it becomes a pain when I have over 100 pictures.
The server addition doesn't work for me at all. ImageTrans_wsServer is in the same folder as ImageTrans. The Chrome extension is installed. When I follow the steps exactly from https://github.com/xulihang/ImageTrans_chrome_extension
I always get a "Failed to connect to ImageTrans server". Tools>Server>Start the server>Reconnect is always "Disconnected".
In terms of bugs, the program sometimes crashes or stops functioning without reason. For example, if I create a mask, go to text removal and then go back to mask, I am met with a blank screen and the program ceases to function. It's not frozen, but functionality becomes nonexistent. During processing, it sometimes freezes completely. In order to guarantee a successful run, each step has to be precise otherwise it'll eventually crash somewhere.
Are you familiar with the Chrome extension "Ichigo Manga Translator"? Personally, it also is great on-the-fly translator, but what it's lacking is parameter/settings modification. I'd love to have the same automation of it but with the ability to tweak the settings for my needs. I am 100% certain that the output from ImageTrans is much more refined but as a current amateur user, I am unable to reproduce something similar to Ichigo Manga Translator.
Activity
xulihang commentedon Jun 15, 2023
ImageTrans is designed to translate any images. So its default configuration is not optimized for manga.
There are somethings you can do to optimize it for translating Japanese manga to English:
xulihang commentedon Jun 15, 2023
About setting Lama Inpaint as the default, I am not sure what the problem is. If you set it in the settings, telea will not be used when generating mask and text-removed images by checking "Translated" or using batch operations.
xulihang commentedon Jun 15, 2023
In case of crashes, I rarely meet crashes. Please give me exact steps or videos to reproduce the problem.
As for starting the server, you have to use JRE11. The easiest way is to use the full version from onedrive and let the program download and start the server.
xulihang commentedon Jun 18, 2023
You can enable AutoCorrect to replace characters for machine translation. If this is not enough, I have to update the Sugoi Translator's plugin to do this.

Breaking words where there is not enough space is the default behavior of the underlying text engine. The text engine is a built-in one in JavaFX which does not give me many controls. I have to implement my own text engine for such a purpose.
xulihang commentedon Jun 25, 2023
Check if there is the
InsetRectMaskGenPlugin.jarfile in the plugins folder.Try to use
ImageTrans.batto start the program which has a larger memory setting for Java.xulihang commentedon Jun 27, 2023
I have updated the plugins zip file.
I will try to fix the issues. As for the crashing, are you using the x86 version or the x64 version? Could you send me some image files so that I can reproduce the problem?
xulihang commentedon Jun 28, 2023
I see. The image has a large resolution. Actually, ImageTrans will scale down images if their width or height is over 8000. It seems that it needs a smaller threshold.
The error should be related to the limitation of JavaFX's canvas:
xulihang commentedon Jun 28, 2023
I can send you an interval version to have a test. You can send a mail to me and I will send you the file to update: admin@basiccat.org
xulihang commentedon Jun 29, 2023
I've sent you the file via gmail. Please try it out.
xulihang commentedon Jun 30, 2023
The current performance of handling huge images is not very good. I have to restructure the mask editor based on projects like this one: HugeImageView - show very large images
19 remaining items
xulihang commentedon Aug 27, 2023
WinRT's raw results contain spaces. You can check the "Remove spaces" to remove them.
Removing spaces is not effective using "detect text areas and recognize text". You can perform boxes merging to remove spaces.

xulihang commentedon Aug 27, 2023
Google drive's token expires every 7 days.
You have to wait for me to reset the token or run it by yourself.
xulihang commentedon Aug 28, 2023
You can follow this guide: #91 (comment)
It may be a bit complex.
xulihang commentedon Aug 28, 2023
The version above is the command line version which directly runs on your device.
xulihang commentedon Aug 30, 2023
If the original image format is PNG, it will output PNG. If the original format is JPEG, it will output JPEG. Currently, there are no settings to specify the batch output image format.
xulihang commentedon Sep 6, 2023
Thanks for the feedback. I will try to improve the software.
xulihang commentedon Sep 6, 2023
I've closed this issue. I will open new issues relating to your request.
xulihang commentedon Sep 16, 2023
The default output format can be set to PNG in v2.4.1.
xulihang commentedon Oct 1, 2023
v2.5.2 solved the problem of loading large images in the mask editor.