This is a dirty edit of bluaxees Poricom project that adds support for Google, DeepL and Yandex machine translation. Google translate works out of the box without an auth key, but both DeepL and Yandex require one. I haven't tested Yandex as it only provides auth keys to paid customers, but both Google and DeepL work without issues as far as I tested. Translation settings are found under [TRANSLATE.DEEPL], [TRANSLATE.GOOGLE] and [TRANSLATE.YANDEX] inside code/utils/config.toml. Additional dependencies include: deepl, googletrans and yandex_translate. I can't get this project to build with pyinstaller, but you're free to try it yourself.
Optical character recognition in manga images. Manga OCR desktop application
Poricom is a desktop program for optical character recognition in manga images. Although it is a manga OCR application, it can recognize text on other type of images as well. The project is a GUI implementation of the Manga OCR library (supports Japanese only) and the Tesseract-API python wrapper tesserocr (supports other languages). See demo below to see how it works.
Detect text on locally stored manga images:
sample_usage.mp4
Perform OCR on the current screen by pressing Alt+Q:
external_capture.mp4
Follow the installation instructions here. Load a directory with manga images and select text boxes with Japanese text. If you are not getting good results using the default settings, use the MangaOcr model to improve text detection.
Listed below are some of the features of Poricom. Smaller features that are not covered in this section are mentioned in the changelog. Click the arrow to see how each implemented feature works.
Open a directory with manga images or a supported manga file (cbz, cbr, pdf) and start scanning text bubbles.
manga-text-detection.mp4
Capture images outside the application using the shortcut `Alt+Q`.
external_capture.mp4
Load MangaOcr model to improve Japanese text recognition.
load-mangaocr-model.mp4
Change language and/or orientation (limited to the Tesseract API).
language-orientation-option.mp4
Detect text on non-manga images.
non-manga-detection.mp4
Download the latest zip file here. Decompress the file in the desired directory. Make sure that the app folder is in the same folder as the shortcut Poricom.
For developers, clone this repo and install requirements: pip install -r requirements.txt. Run the app in the command line using python main.py.
Recommended:
- Hard drive: at least 800 MB HD space
- RAM: at least 2 GB
Approximately 250 MB of free space and 200 MB of memory is needed to run the application using the Tesseract API. If using the Manga OCR model, an additional 450 MB of free space and 800 MB of memory is required.
For developers, the following Python versions are supported: 3.7, 3.8, and 3.9.
This project will not be possible without the MangaOcr model by Maciej Budyś and the Tesseract python wrapper by sirfz and the tesserocr contributors.
The software is licensed under GPLv3 (see LICENSE) and uses third party libraries that are distributed under their own terms (see LICENSE-3RD-PARTY).
The icons used in this project are from Icons8.