https://github.com/chfoo/tppocr3
Tesseract 4 OCR of Pokemon dialog text on streaming video (version 3).
https://github.com/chfoo/tppocr3
Last synced: about 2 months ago
JSON representation
Tesseract 4 OCR of Pokemon dialog text on streaming video (version 3).
- Host: GitHub
- URL: https://github.com/chfoo/tppocr3
- Owner: chfoo
- License: apache-2.0
- Created: 2021-02-09T02:49:15.000Z (about 5 years ago)
- Default Branch: main
- Last Pushed: 2021-02-09T03:06:54.000Z (about 5 years ago)
- Last Synced: 2025-01-23T12:36:42.516Z (about 1 year ago)
- Language: Rust
- Homepage:
- Size: 30.3 KB
- Stars: 2
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# tppocr3
Teseract 4 OCR of Pokemon dialog text on streaming video (version 3).
This project contains experimental, work-in-progress code for running OCR on live streams such as TwitchPlaysPokemon. For tessdata for Tesseract 3 or background information, please see [tppocr](https://github.com/chfoo/tppocr).
## tessdata
TODO: fine-tuned tessdata for Pokemon gen 1, 2, and 3 will go here.
## Software suite
The software suite requires:
* Rust 2018 edition
* Ubuntu 20.04
Dependencies:
sudo apt install libavcodec-dev libavfilter-dev libavformat-dev libtesseract-dev libtesseract4 libvncserver-dev libvncserver1
Once you install Rust, the Rust versions can be manged with `rustup` command.
Rust programs are managed using the `cargo` command:
1. `cargo build --release`
2. `cargo run --release`
Programs:
1. `stream_dumper`: Decodes each stream frame using ffmpeg's libav libraries and puts it into shared memory.
2. `vnc_server`: Shows a debug image of image detection and recognition in real-time.
3. `tppocr`: Process the results of Tesseract recognition and outputs text in a structured manner.
TODO: more work