Report by Bai Jiao from Aofeisi Quantum Bit | WeChat public account QbitAI
A Bilibili UP master has created a translator that claims to be able to handle any raw video, even games.
This translator is based on OCR technology, which means you only need to take screenshots on the interface, and even subtitles from games or new anime can be translated.
Moreover, with the change of scenes, it can automatically translate directly.
It also works for raw manga.
No need to worry when playing Japanese nurturing games~
In response, netizens said: Awesome, I can uninstall VNR (a translation software for raw video).
Currently, this software can be downloaded for free on GitHub, with the link attached at the end of the article.
How to Use
This OCR translator utilizes Baidu AI’s text recognition to translate foreign text on images.
After downloading, it looks like this.
Before using, a series of parameter settings are required, which means clicking the second setting button on the left.
The first page is the API setting interface, where you need to register your own API key to use the translator.
The topmost OCR API is mandatory to register. Click the button to register for OCR, and a registration tutorial will pop up.
Just follow the registration tutorial, and remember to save the settings after filling in the key and secret! It’s also recommended to register for the API code below.
The next page is the translator settings, where you can choose the translation source you want to use, with a total of 12 options.
Public translation and web translation can be used directly, while private translation requires self-registration of the API key as set on the previous page.
Also, set the original language of the translation (currently only supports English, Japanese, and Korean) and the automatic translation time settings.
Next, you can choose the translation style, including color, font size, font style, etc.
There are also other settings.
Its basic operations are:
-
Capture the coordinates of the screen area to be translated by screenshot
-
Take a screenshot based on the coordinates (can be automatic) and send it to Baidu AI’s text recognition interface;
-
After obtaining the recognized text, send it to Baidu, Tencent, Caicloud, and other translation interfaces;
-
Feedback the result to the GUI interface.
All you need to do is take a screenshot and then translate.
There’s also a small feature.
The red-highlighted “music” button, when clicked, can play the original text.
This software is now open-source on GitHub, so interested friends can download it~
But I’m going to take the lead and test it out.
Testing It Out
According to the tutorial provided on GitHub, after downloading, it looks like this.
Since I’ve already tested Japanese raw videos and games, let’s test PDF translation.
Before that, I chose public Youdao, web Tencent, and private Baidu. The color settings are as follows:
The translation language is set to English, with shortcut keys for screenshot selection A and translation selection S.
Taking a machine learning paper titled Review of Text Style Transfer Based on Deep Learning as an example in PDF format.
Let’s try the title.
First, capture the title range.
Pressing the S key brings up the translation interface.
The order is original text, private Baidu, web Tencent, and public Youdao.
Let’s try the first three lines of the paper’s introduction section.
The translation is as follows:
(If it’s not clear, you can adjust the position and transparency appropriately)
In addition to translation, you can also compare the translation quality among several services~
By the way, which translation service do you think is the best? Or do you have better translation software?
Feel free to leave a comment below, seeking recommendations!
Links
https://github.com/PantsuDango/Dango-Translator
Bilibili demonstration video:
https://www.bilibili.com/video/BV1gp4y1Q7Ts?from=search&seid=2515920591076249883
Important! The Yizhen Natural Language Processing - Academic WeChat group has been established.
You can scan the QR code below, and the assistant will invite you to join the group for communication.
Note: Please modify the note to [School/Company + Name + Direction] when adding.
For example - Harbin Institute of Technology + Zhang San + Dialogue System.
The account owner, please avoid adding if you are a micro-business. Thank you!
Recommended Reading:
Using TensorFlow for Text Sentiment Analysis
Easy to Understand! Implementing Transformer with Excel and TF!
Multi-task Learning in Deep Learning (Keras Implementation)